Empowerment Zone and Enterprise Community Program

Improvements Occurred in Communities, but the Effect of the Program Is Unclear Gao ID: GAO-06-727 September 22, 2006

The Empowerment Zone/Enterprise Community (EZ/EC) program is one of the most recent large-scale federal effort intended to revitalize impoverished urban and rural communities. There have been three rounds of EZs and two rounds of ECs, all of which are scheduled to end no later than December 2009. The Community Renewal Tax Relief Act of 2000 mandated that GAO audit and report in 2004, 2007, and 2010 on the EZ/EC program and its effect on poverty, unemployment, and economic growth. This report, which focuses on the first round of the program starting in 1994, discusses program implementation; program oversight; data available on the use of program tax benefits; and the program's effect on poverty, unemployment, and economic growth. In conducting this work, GAO made site visits to all Round I EZs, conducted an e-mail survey of 60 Round I ECs, and used several statistical methods to analyze program effects.

Round I Empowerment Zones (EZ) and Enterprise Communities (EC) implemented a variety of activities using $1 billion in federal grant funding from the Department of Health and Human Services (HHS), and as of March 2006, the designated communities had expended all but 15 percent of this funding. Most of the activities that the grant recipients put in place were community development projects, such as projects supporting education and housing. Other activities included economic opportunity initiatives such as job training and loan programs. Although all EZs and ECs also reported using the program grants to leverage funds from other sources, reliable data on the extent of leveraging were not available. According to federal standards, agencies should oversee the use of public resources and ensure that ongoing monitoring occurs. However, none of the federal agencies that were responsible for program oversight--including HHS and the departments of Housing and Urban Development (HUD) and Agriculture (USDA)--collected data on the amount of program grant funds used to implement specific program activities. This lack of data limited both federal oversight and GAO's ability to assess the effect of the program. Moreover, because HHS did not provide the states and designated communities with clear guidance on how to monitor the program grant funds, the extent of monitoring varied across the sites. In addition, detailed Internal Revenue Service (IRS) data on the use of EZ/EC program tax benefits were not available. Previously, GAO cited similar challenges in assessing the use of tax benefits in other federal programs and stated that information on tax expenditures should be collected to ensure that these expenditures are achieving their intended purpose. Although GAO recommended in 2004 that HUD, USDA, and IRS work together to identify the data needed to assess the EZ/EC tax benefits and the cost effectiveness of collecting the information, the three agencies did not reach agreement on an approach. Without adequate data on the use of program grant funds or tax benefits, neither the responsible federal agencies nor GAO could determine whether the EZ/EC funds had been spent effectively or that the tax benefits had in fact been used as intended. Using the data that were available, GAO attempted to analyze changes in several indicators--poverty and unemployment rates and two measures of economic growth. Although improvements in poverty, unemployment, and economic growth had occurred in the EZs and ECs, our econometric analysis of the eight urban EZs could not tie these changes definitively to the EZ designation.



GAO-06-727, Empowerment Zone and Enterprise Community Program: Improvements Occurred in Communities, but the Effect of the Program Is Unclear This is the accessible text file for GAO report number GAO-06-727 entitled 'Empowerment Zone and Enterprise Community Program: Improvements Occurred in Communities, but the Effect of the Program is Unclear' which was released on September 25, 2006. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. Report to Congressional Committees: September 2006: Empowerment Zone And Enterprise Community Program: Improvements Occurred in Communities, but the Effect of the Program Is Unclear: GAO-06-727: GAO Highlights: Highlights of GAO-06-727, a report to congressional committees Why GAO Did This Study: The EZ/EC program is one of the most recent large-scale federal efforts intended to revitalize impoverished urban and rural communities. There have been three rounds of EZs and two rounds of ECs, all of which are scheduled to end no later than December 2009. The Community Renewal Tax Relief Act of 2000 mandated that GAO audit and report in 2004, 2007, and 2010 on the EZ/EC program and its effect on poverty, unemployment, and economic growth. This report, which focuses on the first round of the program starting in 1994, discusses program implementation; program oversight; data available on the use of program tax benefits; and the program‘s effect on poverty, unemployment, and economic growth. In conducting this work, GAO made site visits to all Round I EZs, conducted an e-mail survey of 60 Round I ECs, and used several statistical methods to analyze program effects. What GAO Found: Round I Empowerment Zones (EZ) and Enterprise Communities (EC) implemented a variety of activities using $1 billion in federal grant funding from the Department of Health and Human Services (HHS), and as of March 2006, the designated communities had expended all but 15 percent of this funding. Most of the activities that the grant recipients put in place were community development projects, such as projects supporting education and housing. Other activities included economic opportunity initiatives such as job training and loan programs. Although all EZs and ECs also reported using the program grants to leverage funds from other sources, reliable data on the extent of leveraging were not available. According to federal standards, agencies should oversee the use of public resources and ensure that ongoing monitoring occurs. However, none of the federal agencies that were responsible for program oversight”including HHS and the departments of Housing and Urban Development (HUD) and Agriculture (USDA)”collected data on the amount of program grant funds used to implement specific program activities. This lack of data limited both federal oversight and GAO‘s ability to assess the effect of the program. Moreover, because HHS did not provide the states and designated communities with clear guidance on how to monitor the program grant funds, the extent of monitoring varied across the sites. In addition, detailed Internal Revenue Service (IRS) data on the use of EZ/EC program tax benefits were not available. Previously, GAO cited similar challenges in assessing the use of tax benefits in other federal programs and stated that information on tax expenditures should be collected to ensure that these expenditures are achieving their intended purpose. Although GAO recommended in 2004 that HUD, USDA, and IRS work together to identify the data needed to assess the EZ/EC tax benefits and the cost effectiveness of collecting the information, the three agencies did not reach agreement on an approach. Without adequate data on the use of program grant funds or tax benefits, neither the responsible federal agencies nor GAO could determine whether the EZ/EC funds had been spent effectively or that the tax benefits had in fact been used as intended. Using the data that were available, GAO attempted to analyze changes in several indicators”poverty and unemployment rates and two measures of economic growth. Although improvements in poverty, unemployment, and economic growth had occurred in the EZs and ECs, our econometric analysis of the eight urban EZs could not tie these changes definitively to the EZ designation. What GAO Recommends: While not making recommendations, GAO makes observations that should be considered if these or similar programs are authorized in the future. HHS, HUD, and USDA provided comments. In particular, HUD disagreed with the observation that there was a lack of data to perform program oversight. [Hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-727. To view the full product, including the scope and methodology, click on the link above. To view the survey results, click on the following link: [Hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-734SP. For more information, contact William B. Shear at (202) 512-8678 or ShearW@gao.gov. [End of Section] Contents: Letter: Results in Brief: Background: Round I EZs and ECs Have Used Their Grant Funds to Implement a Wide Range of Program Activities: Oversight Was Hindered by Limited Program Data and Variation in Monitoring: Lack of Detailed Tax Data Made It Difficult to Assess the Use of Program Tax Benefits: In Aggregate, EZs and ECs Showed Some Improvements, but Our Analysis Did Not Definitively Link These Changes to the Program: Observations: Agency Comments and Our Evaluation: Appendixes: Appendix I: Objectives, Scope, and Methodology: Methodology for Site Visits: Methodology for Survey of EC Officials: Methodology for Qualitative Analysis of Site Visit and EC Survey Data: Methodology for Review of Program Oversight: Methodology for Survey of EZ Businesses: Methodology for Assessing the Effect of the Program on Poverty, Unemployment, and Economic Growth: Appendix II: Methodology for and Results of Our Econometric Models: Description of Our Models: Results of Our Models for Poverty: Results of Our Models for Unemployment: Results of Our Models for Economic Growth: Other Variables Tested for Use in Our Econometric Models: Appendix III: List of Communities Designated in Round I of the EZ/EC Program: Appendix IV: Description of the Empowerment Zones and Enterprise Communities We Visited: Atlanta Empowerment Zone: Baltimore Empowerment Zone: Chicago Empowerment Zone: Detroit Empowerment Zone: New York Empowerment Zone: Philadelphia-Camden Empowerment Zone: Cleveland Empowerment Zone: Los Angeles Empowerment Zone: Kentucky Highlands Empowerment Zone: Mid-Delta Mississippi Empowerment Zone: Rio Grande Valley, Texas Empowerment Zone: Providence, Rhode Island Enterprise Community: Fayette-Haywood, Tennessee Enterprise Community: Appendix V: Comments from the Department of Health and Human Services: Appendix VI: Comments from the Department of Housing and Urban Development: Appendix VII: Comments from the U.S. Department of Agriculture: Appendix VIII: GAO Contact and Staff Acknowledgments: Tables: Table 1: Round I EZ/EC Program Criteria and Benefits: Table 2: National Poverty, Unemployment, Economic Growth Data for 1990 to 2004: Table 3: Total EZ/EC Grant Funding Remaining as of March 31, 2006: Table 4: Number of Stakeholders Interviewed for EZ and EC Site Visits, by Type: Table 5: Coding of Data Reliability of HUD and USDA Performance Systems: Table 6: Confidence Intervals for Average Household Income and Average Housing Value in Constant 2004 Dollars: Table 7: Factors Selected for Choosing Comparison Tracts: Table 8: Estimates of the Association between the EZ Program and the Change in Poverty Rate, 1990-2000: Table 9: Estimates of the Association between the EZ Program and the Change in Unemployment Rate, 1990-2000: Table 10: Estimates of the Association between the EZ Program and Economic Growth, Measured by the Change in the Number of Businesses, from 1995-1999: Table 11: Estimates of the Association between the EZ Program and Economic Growth, Measured by the Change in the Number of Jobs, 1995- 1999: Table 12: Alternative Variables Considered in Our Analyses: Table 13: Changes in Selected Census Variables Observed in the Atlanta EZ and Its Comparison Area: Table 14: Changes in Selected Economic Growth Variables Observed in the Atlanta EZ and Its Comparison Area: Table 15: Changes in Selected Census Variables Observed in the Baltimore EZ and Its Comparison Area: Table 16: Changes in Selected Economic Growth Variables Observed in the Baltimore EZ and Its Comparison Area: Table 17: Changes in Selected Census Variables Observed in the Chicago EZ and Its Comparison Area: Table 18: Changes in Selected Economic Growth Variables Observed in the Chicago EZ and Its Comparison Area: Table 19: Changes in Selected Census Variables Observed in the Detroit EZ and Its Comparison Area: Table 20: Changes in Selected Economic Growth Variables Observed in the Detroit EZ and Its Comparison Area: Table 21: Changes in Selected Census Variables Observed in the New York EZ, the Bronx and Upper Manhattan (UM) Portions, and the EZ Comparison Area (Comp.): Table 22: Changes in Selected Economic Growth Variables Observed in the New York EZ, the Bronx and Upper Manhattan (UM) Portions, and the EZ Comparison Area (Comp.): Table 23: Changes in Selected Census Variables Observed in the Philadelphia-Camden EZ, the Camden (Cam.) and Philadelphia (Phila.) Portions, and the EZ Comparison Area (Comp.): Table 24: Changes in Selected Economic Growth Variables Observed in the Philadelphia-Camden EZ, the Camden (Cam.) and Philadelphia (Phila.) Portions, and the EZ Comparison Area (Comp.): Table 25: Changes in Selected Census Variables Observed in the Cleveland EZ and Its Comparison Area: Table 26: Changes in Selected Economic Growth Variables Observed in the Cleveland EZ and Its Comparison Area: Table 27: Changes in Selected Census Variables Observed in the Los Angeles EZ and Its Comparison Area: Table 28: Changes in Selected Economic Growth Variables Observed in the Los Angeles EZ and Its Comparison Area: Table 29: Changes in Selected Census Variables Observed in the Kentucky Highlands EZ: Table 30: Changes in Selected Economic Growth Variables Observed in the Kentucky Highlands EZ: Table 31: Changes in Selected Census Variables Observed in the Mid- Delta EZ: Table 32: Changes in Selected Economic Growth Variables Observed in the Mid-Delta EZ: Table 33: Changes in Selected Census Variables Observed in the Rio Grande Valley EZ: Table 34: Changes in Selected Economic Growth Variables Observed in the Rio Grande Valley EZ: Table 35: Changes in Selected Census Variables Observed in the Providence EC: Table 36: Changes in Selected Economic Growth Variables Observed in the Providence EC: Table 37: Changes in Selected Census Variables Observed in the Fayette- Haywood EC: Table 38: Changes in Selected Economic Growth Variables Observed in the Fayette-Haywood EC: Figures: Figure 1: Oversight Responsibilities in Round I of the EZ/EC Program: Figure 2: Remaining Grant Funds by EZ as of March 31, 2006: Figure 3: Distribution of EZ and EC Activities by Key Program Principle: Figure 4: Types of Activities Implemented by Urban and Rural EZs and ECs, by Percent of Total Activities: Figure 5: Local Government Involvement in Decision Making in the Urban EZs: Figure 6: Changes in Poverty, Unemployment, and Two Measures of Economic Growth Observed in Round I EZs: Figure 7: Number and Percentage of EZs and ECs Experiencing a Decrease in Poverty from 1990 to 2000: Figure 8: Comparison of Decreases in Poverty in Urban and Rural Designated Areas and Comparison Areas from 1990 to 2000: Figure 9: Number and Percentage of EZs and ECs that Experienced a Decrease in Unemployment from 1990 to 2000: Figure 10: Comparison of Decreases in Unemployment in Urban and Rural Designated Areas and Comparison Areas from 1990 to 2000: Figure 11: Number and Percentage of EZs and ECs That Experienced an Increase in One or Both Measures of Economic Growth between 1995 and 2004: Figure 12: Comparison of Changes in the Number of Businesses and the Number of Jobs in Urban and Rural Designated Areas and Comparison Areas between 1995 and 2004: Figure 13: Map of the Atlanta EZ and Its Comparison Area: Figure 14: Activities Implemented by the Atlanta EZ: Figure 15: Map of the Baltimore EZ and Its Comparison Area: Figure 16: Activities Implemented by the Baltimore EZ: Figure 17: Map of the Chicago EZ and Its Comparison Area: Figure 18: Activities Implemented by the Chicago EZ: Figure 19: Map of the Detroit EZ and Its Comparison Area: Figure 20: Activities Implemented by the Detroit EZ: Figure 21: Map of the New York EZ and Its Comparison Area: Figure 22: Activities Implemented by the Upper Manhattan portion of the New York EZ: Figure 23: Activities Implemented by the Bronx portion of the New York EZ: Figure 24: Map of the Philadelphia-Camden EZ and Its Comparison Area: Figure 25: Activities Implemented by the Philadelphia Portion of the Philadelphia-Camden EZ: Figure 26: Activities Implemented by the Camden Portion of the Philadelphia-Camden EZ: Figure 27: Map of the Cleveland EZ and Its Comparison Area: Figure 28: Activity Implemented by the Cleveland EZ: Figure 29: Map of the Los Angeles EZ and Its Comparison Area: Figure 30: Activity Implemented by the Los Angeles EZ: Figure 31: Map of the Kentucky Highlands EZ: Figure 32: Activities Implemented by the Kentucky Highlands EZ: Figure 33: Map of the Mid-Delta EZ: Figure 34: Activities Implemented by the Mid-Delta EZ: Figure 35: Map of the Rio Grande Valley EZ: Figure 36: Activities Implemented by the Rio Grande Valley EZ: Figure 37: Map of the Providence EC: Figure 38: Activities Implemented by the Providence EC: Figure 39: Map of the Fayette-Haywood EC: Figure 40: Activities Implemented by the Fayette-Haywood EC: Abbreviations: EC: Enterprise Community: EZ: Empowerment Zone: HHS: Department of Health and Human Services: HUD: Department of Housing and Urban Development: IRS: Internal Revenue Service: USDA: U.S. Department of Agriculture: September 22, 2006: The Honorable Charles E. Grassley: Chairman: The Honorable Max Baucus: Ranking Minority Member: Committee on Finance: United States Senate: The Honorable William M. Thomas: Chairman: The Honorable Charles B. Rangel: Ranking Minority Member: Committee on Ways and Means: House of Representatives: The Empowerment Zone and Enterprise Community (EZ/EC) program is one of the most recent in a series of large-scale federal efforts intended to address one of the nation's most persistent challenges--the revitalization of impoverished urban and rural communities. When it was enacted in 1993, the EZ/EC program provided grants to public and private entities for social services and community redevelopment and tax benefits to local businesses to attract or retain jobs and businesses in distressed communities. The program differs from earlier initiatives with similar goals in that it emphasizes the role of local communities in identifying solutions and the use of public-private partnerships to attract the investment necessary for sustainable economic and community development. To date, Congress has authorized three rounds of EZs and two rounds of ECs. Communities designated under Round I of the program shared a total of $1 billion in federal grant funding and also received tax and other benefits. The EZs received the bulk of this funding--$720 million in total--as well as more extensive tax benefits than the ECs. Communities designated in the two subsequent rounds of the program received a smaller amount of federal funding and more tax benefits. All three rounds of the EZ/EC program are scheduled to end no later than December 31, 2009. The Community Renewal Tax Relief Act of 2000 mandated that we audit and report in 2004, 2007, and 2010 on the EZ/EC program and a later initiative, the Renewal Community program, and their effect on poverty, unemployment, and economic growth.[Footnote 1] This report, the second of the mandated series, focuses on the first round of communities designated as EZs and ECs in 1994. It (1) describes how the designated communities implemented Round I of the EZ/EC program; (2) evaluates the extent of federal, state, and local oversight of the program; (3) examines the extent to which data are available to assess the use of program tax benefits; and (4) analyzes the effects that the Round I EZs and ECs had on poverty, unemployment, and economic growth in their communities. To address our first three objectives, we made site visits to all 11 Round I EZs and 2 of the 95 ECs--1 urban and 1 rural--to interview stakeholders and review documentation.[Footnote 2] To gather information from the ECs, we administered an e-mail survey to officials from the 60 Round I ECs that were still in operation as of June 2005 and did not receive a subsequent designation.[Footnote 3] We chose to exclude the 34 ECs that received subsequent designations, because we did not want their responses to be influenced by those programs. Because the states distributed the federal funding to the communities, we conducted telephone interviews with state officials in the 13 states containing the EZs and ECs that we visited. In addition, we interviewed officials from the federal agencies with primary responsibility for the program--the Department of Health and Human Services (HHS), the Department of Housing and Urban Development (HUD), the Internal Revenue Service (IRS), and the U.S. Department of Agriculture (USDA). We also analyzed fiscal and program data from the agencies and assessed the reliability of these data.[Footnote 4] To address our fourth objective, that is, the effect of the program on poverty, unemployment, and economic growth, we used several methods. First, we calculated the changes in the poverty and unemployment rates from 1990 to 2000 and measures of economic growth from 1995 to 2004 in the designated EZs and ECs and in comparison areas selected for their similarities to the designated communities.[Footnote 5] Then, we used econometric models to assess the effects of the program. Finally, we used testimonial information gathered during our site visits and our survey results to help put these changes in context. We conducted our work between November 2004 and July 2006 in accordance with generally accepted government auditing standards. Appendix I lists the communities we visited. Appendixes I and II provide details on our methodology, and appendix III shows a list of communities designated in Round I of the EZ/EC program. Appendix IV provides details on each of the sites we visited. Results in Brief: Round I EZs and ECs used most of the $1 billion in program grant funds to implement a wide range of activities designed to help revitalize the designated communities. As of March 31, 2006, 20 percent of the $720 million that EZs received and 2 percent of the $280 million that ECs received remained unspent, and some designees had received extensions of the original 10-year grant period that was set to expire in 2004. In general, EZs and ECs undertook more community development activities in areas such as education, housing, and infrastructure than they did economic opportunity activities such as job training and assistance to businesses. Although stakeholders from all EZs and ECs reported using the program grants to leverage funds from other sources and some said that they had required subgrantees to obtain other funds as a condition of receiving EZ/EC funds, reliable data on the extent of leveraging were not available. EZ and EC designees also reported other accomplishments and challenges and utilized a variety of governance structures to implement these activities. Data were not collected on program benefits for specific activities, limiting the ability of federal agencies to oversee the program, and the monitoring performed at the state and local levels varied. According to our Standards for Internal Control in the Federal Government, federal agencies should oversee the use of public resources and ensure that ongoing monitoring occurs.[Footnote 6] However, the three agencies responsible for overseeing the program--HHS, HUD, and USDA--did not collect data on how program funds were used. For instance, HHS data show that EZs and ECs have used most of the EZ/EC grant funds but do not show the specific activities or types of activities for which the funds were used. And, although the performance reporting systems maintained by HUD and USDA do contain some information on activities that were carried out, they do not contain information on how much of the EZ/EC funds actually were used for specific activities or types of activities.[Footnote 7] Further, HHS did not provide the states, EZs, and ECs with clear guidance on how to monitor the program grant funds, so the types and extent of monitoring performed by state and local participants varied. To some degree, the lack of reporting requirements may be an outcome of the program's design, which was intended to give communities flexibility in using program funds and relied on multiple agencies for oversight. But the result has been that little information is available on the amount of funds spent on specific activities, hindering the agencies' efforts to oversee the program. Similarly, only limited data are available on the use of EZ/EC tax benefits, which were estimated to be much more substantial than the amount of program grant funds. We have stated that information on tax expenditures should be collected to ensure that these expenditures are achieving their intended purpose.[Footnote 8] In 2004, we reported that IRS collected data on some but not all of the program tax benefits and that the data could not be linked to the individual communities.[Footnote 9] We also recommended that HUD, USDA, and IRS work together to identify the data needed to measure the use of EZ/EC tax benefits and the cost-effectiveness of collecting the information, but the three agencies did not reach agreement on a cost-effective approach.[Footnote 10] During our work for this report, officials from some EZs and ECs told us that some local businesses were using the tax benefits. However, these testimonial data were neither sufficient to allow us to determine the actual amount of tax benefits used by EZs and ECs nationwide nor to assess the extent to which the program tax benefits contributed to the achievement of program goals. Although improvements in poverty, unemployment, and economic growth had occurred in the EZs and ECs, our econometric analysis of the eight urban EZs could not tie these changes definitively to the EZ designation.[Footnote 11] As mentioned in our previous report, measuring the effect of initiatives such as the EZ/EC program is difficult for a number of reasons, such as data limitations and the difficulty of determining what would have happened in the absence of the program.[Footnote 12] Given these limitations, the effects of the EZ/EC program remain unclear. In some cases, communities did see decreases in poverty and unemployment and increases in economic growth. However, when we used econometric analyses to separate the effect of the program from other nonprogram factors we found that the comparison tracts we selected showed changes that were similar to those in the urban EZs. Further, EZ stakeholders and EC survey respondents said that program-related factors had influenced changes in their communities but noted that other unrelated factors had also had an effect. For example, stakeholders who observed a decrease in poverty in their communities believed that this change had resulted in part from EZ/EC activities, but they also noted that the population in their communities had changed, with original EZ/EC residents moving out of the area and individuals with higher incomes moving in. Ultimately, the evaluation techniques we developed were limited by the absence of data on the use of program grants and tax benefits. While all three rounds of the EZ/EC program are scheduled to end no later than December 31, 2009, we observe two limitations that should be considered if these or similar programs are authorized in the future. These include (1) oversight limitations that occurred because data were not collected on how program grant funds were used for specific activities and (2) the limited ability to evaluate the effect of the program due to the lack of data on the use of program grant funds, the extent of leveraging, and the extent to which program tax benefits were used. Given the magnitude of federal grant funds and tax benefits provided for the program, more should be done to better understand the extent to which these federal expenditures are having the desired effect. We provided a draft of this report to HHS, HUD, IRS, and USDA. We received comments from HHS, HUD, and USDA. HHS commented that a statement made in our report--that the agency did not provide guidance detailing the steps state and local authorities should take to monitor the program--unfairly represented the relationship between HHS and the other federal agencies that administered the EZ/EC program. However, we note in our report that the program's design may have led to a lack of clarity in oversight since no single federal agency had sole oversight responsibility. Nonetheless, we believe that, in accordance with federal standards, each of the federal agencies that administered the program bore at least some responsibility for ensuring that public resources were being used effectively and that program goals were being met. HUD disagreed with our observation that there was a lack of data on the use of program grant funds, the amount of funds leveraged, and the use of tax benefits. However, although we found evidence that activities were carried out with program funds, information contained in HUD's performance reporting system on the amounts of funds used and the amounts leveraged was not reliable. Both HUD and USDA provided suggestions for future evaluations of similar programs. The agencies' comments are discussed later in the report and are reproduced in appendixes V through VII. HHS, HUD and USDA also provided technical comments that we incorporated into the report where appropriate. Background: The concept behind the EZ/EC program originated in Great Britain in 1978 with the inception of the Enterprise Zone program. The main objective of the Enterprise Zone program was to foster an attractive business environment in specific areas where economic growth was lacking. In the United States, some states began to administer similar state Enterprise Zones in the 1980s. In 1993, the federal government established the federal EZ/EC program to help reduce unemployment and revitalize economically distressed areas. The authorizing legislation established the eligibility requirements and the package of grants and tax benefits for the EZ/EC program (table 1). Multiagency teams from HHS, HUD, USDA, and other federal agencies reviewed the applications in Round I, and HUD and USDA issued designations based on the effectiveness of communities' strategic plans, assurances that the plans would be implemented, and geographic diversity.[Footnote 13] In Round I, HUD designated a total of 8 urban EZs and 65 urban ECs, and USDA designated 3 rural EZs and 30 rural ECs.[Footnote 14] Table 1: Round I EZ/EC Program Criteria and Benefits: Eligibility criteria; To be considered for the program, communities were required to select census tracts that; * had above-average poverty according to 1990 Census data;; * had unemployment rates of at least the national average according to 1990 Census data;; * met certain 1990 population and area criteria; and; * exhibited other conditions of distress, such as high crime, deteriorating infrastructure, or population decline. In addition, they were required to submit a strategic plan that addressed the four key principles of the program: * economic opportunity,; * sustainable community development,; * community-based partnerships, and; * strategic vision for change. EZ program benefits; Round I EZs received Title XX Social Services Block Grants (EZ/EC grants); * Six urban EZs each received $100 million.[A]; * Three rural EZs each received $40 million. Businesses located in EZs initially received three tax benefits: * a tax credit for wages paid to employees who both live and work in an EZ,; * an increased expensing deduction for depreciable property, and; * tax- exempt bonds that could be used to issue loans to qualified businesses for financing certain property; By 2002, businesses in EZs also became eligible for two additional tax benefits related to the treatment of gains on the sale of EZ assets and stock. EC program benefits; 95 Round I ECs each received $2.95 million in EZ/ EC grants; Businesses located in ECs were eligible for one program tax benefit, the tax-exempt bond financing. Source: GAO. [A] This does not include two additional urban communities--Cleveland and Los Angeles--that initially received Supplemental EZ designations and received full Round I EZ status in 1998, because they did not receive EZ/EC grant funds. [End of table] HHS provided Round I EZs and ECs with a total of $1 billion in EZ/EC grant funds. EZs and ECs were allowed to use the EZ/EC grants for a broader range of activities than was generally allowed with those types of HHS funds. For instance, EZs and ECs could use funding for "traditional" activities, such as skills training programs for disadvantaged youth or drug and alcohol treatment programs, as well as for additional activities, such as the purchase of land or facilities related to an eligible program or the capitalization of a revolving loan fund. EZs and ECs were also permitted to use grant funds to cover some administrative costs and to change their goals and activities over time, with approval from HUD or USDA. In addition, HUD and USDA expected EZs and ECs to use the EZ/EC grant to leverage additional investment. Businesses operating in EZs and ECs were eligible for a substantial amount of program tax benefits. In 1993, the Joint Committee on Taxation estimated that the tax benefits available to businesses in Round I communities would result in a $2.5 billion reduction in tax revenues between 1994 and 1998. In 2000, the committee estimated that the combination of EZ/EC program tax benefits and the Renewal Community tax benefits would reduce tax revenues by a total of $10.9 billion between 2001 and 2010.[Footnote 15] The tax benefits for ECs expired in 2004, and the tax benefits for all EZs and Renewal Communities are currently set to expire at the end of 2009. Four federal agencies are responsible for administering the program in Round I. Oversight responsibilities for Round I were divided among three agencies, with HHS providing fiscal oversight and HUD and USDA providing program oversight (fig. 1). HHS issued grants to the states, which served as pass-through entities--that is, they distributed funds to individual EZs and ECs. According to their regulations, HUD and USDA are required to evaluate the progress each EZ and EC made on its strategic plan based on information gathered on site visits and on information reported to them by the designated communities. In addition, IRS is responsible for administering the program tax benefits. Figure 1: Oversight Responsibilities in Round I of the EZ/EC Program: [See PDF for image] Source: GAO analysis. [End of figure] In assessing the extent of EZ/EC program improvements, it is useful to understand the overall national trends in poverty, unemployment, and economic growth. National trends in these indicators have varied since Round I of the program was established. As shown in table 2, the national poverty and unemployment rates showed improvements (i.e., declines) in 2000 compared with 1990, but both were somewhat higher in 2004. In 1990, Round I EZs and ECs had poverty and unemployment rates that exceeded these national averages, as was required for program eligibility. Table 2: National Poverty, Unemployment, Economic Growth Data for 1990 to 2004: Indicator: Poverty; 1990: 13.5%; 1995: 13.8%; 2000: 11.3%; 2003: 12.5%; 2004: 12.7%. Indicator: Unemployment; 1990: 5.6%; 1995: 5.6%; 2000: 4.0%; 2003: 6.0%; 2004: 5.5%. Indicator: Number of businesses; 1990: 6.1 million; 1995: 6.6 million; 2000: 7.1 million; 2003: 7.3 million; 2004: [A]. Indicator: Number of jobs; 1990: 93.4 million; 1995: 100.3 million; 2000: 114.1 million; 2003: 113.4 million; 2004: [A]. Sources: Census Bureau and Bureau of Labor Statistics. [A] Data were not yet available for 2004. [End of table] In terms of economic growth, the table shows that the number of businesses increased gradually between 1990 and 2003, and the number of jobs increased from 1990 to 2000 but fell slightly between 2000 and 2003. Round I EZs and ECs Have Used Their Grant Funds to Implement a Wide Range of Program Activities: EZs and ECs used most of the program grant funds to implement a wide range of activities to carry out their respective revitalization strategies. In total, as of March 31, 2006, EZs and ECs had used all but 15 percent of the available grants. EZs and ECs implemented a variety of activities, but, in general, focused more on community development than economic opportunity. In addition, all designated communities reported leveraging additional resources, though a lack of reliable data prevented us from determining how much. Several designees also noted other accomplishments, such as increasing local coordination and capacity. The governance structures that Round I EZs and ECs established to implement these activities varied and included organizations to manage the day-to-day operations of the EZs, boards, and advisory committees. Most EZ/EC Grant Funds Have Been Expended, but Many EZs and Some ECs Received Grant Extensions: As of March 31, 2006, Round I EZs and ECs had spent all but 15 percent of the program grant funds they received. HHS data show that 20 percent of the program grant funds provided to EZs and 2 percent of the funds provided to ECs were unspent (table 3). In addition, HUD data show that the Cleveland and Los Angeles EZs, which originally received Supplemental EZ designations, had used significant portions of the Economic Development Initiative grants and Section 108 Loan Guarantees that came with their designations.[Footnote 16] Specifically, each of them had spent slightly more than 70 percent of their grants; Cleveland had used 72 percent of its loan guarantees, but Los Angeles had used less--about 33 percent. Table 3: Total EZ/EC Grant Funding Remaining as of March 31, 2006: EZs; Total funding: $720 million; Amount remaining: $146.6 million; Percent remaining: 20%. ECs; Total funding: $280 million; Amount remaining: $4.5 million; Percent remaining: 2%. Total; Total funding: $1 billion; Amount remaining: $151 million; Percent remaining: 15%. Source: GAO analysis of HHS data. [End of table] Most of the remaining $151 million in EZ/EC grants consists of the funds of four urban EZs: Atlanta, New York, Philadelphia-Camden, and Chicago, with Atlanta and New York accounting for the majority of the unspent funds (fig. 2). When the Atlanta EZ received a Renewal Community designation from HUD in 2002, the EZ designation was terminated, but HHS allowed the city of Atlanta to continue spending its remaining EZ grant funds through December 2009. The city of Atlanta elected to administer its remaining EZ grants in conjunction with its Renewal Community initiative, and prepared a strategic plan to address administration of both the remaining HHS funds and the HUD-designated Renewal Community. The Atlanta Renewal Community officials told us that they did not use the EZ funds for about 4 years after receiving the designation because of the time required for start-up but added that they planned to begin utilizing the funds soon. The New York EZ received matching funds from both the state and city governments, for a total of $300 million. New York EZ officials stated that they used equal parts of funding from these three sources for each activity, potentially explaining why they have drawn down funds at a slower rate than other EZs. Figure 2: Remaining Grant Funds by EZ as of March 31, 2006: [See PDF for image] Source: GAO analysis of HHS data. Note: Two urban EZs--Philadelphia-Camden and New York--implemented the program through two separate entities that split the $100 million grant. These separate entities are represented above for Philadelphia- Camden, but separate data for the New York EZ were not available from HHS. The Cleveland and Los Angeles EZs did not receive EZ grant funds. [End of figure] Although the grant period for Round I EZs and ECs was originally scheduled to end December 21, 2004, several EZs and some ECs received extensions from HHS to continue drawing down their remaining funds. The recipients had to demonstrate a legitimate need to complete project activities outlined in their strategic plans. Eight of the 11 EZs (6 urban, 2 rural) and 17 of the 95 ECs (11 urban and 6 rural) received extensions of their grants until December 31, 2009. In addition, 1 urban EZ and 9 ECs (6 urban and 3 rural) received extensions for a shorter time frame, such as 2005, 2006, or 2007. EZs and ECs Implemented a Wide Variety of Activities, Most Related to Community Development: The designated communities were encouraged to implement both community and economic development activities as part of their revitalization strategies. The EZ/EC program was designed to be tailored to address local needs, and the type of grant funds most EZs and ECs received from HHS allowed them to implement a wide range of activities. Overall, both EZs and ECs used the program grants to implement a larger number of community development activities--such as education, health care, and infrastructure--than economic opportunity activities--such as workforce development and providing assistance to businesses (fig. 3).[Footnote 17] Figure 3: Distribution of EZ and EC Activities by Key Program Principle: [See PDF for image] Source: GAO analysis of HUD and USDA data. Note: This figure shows the percent of the total number of activities implemented, not the funds devoted to those activity types. The Cleveland and Los Angeles EZs are not included in this graphic because they did not receive EZ grant funds. The numbers do not always add up to 100 due to rounding. [End of figure] The activities most often implemented by urban EZs and ECs were workforce development, human services, education, and assistance to businesses, which accounted for more than 50 percent of the activities in urban EZs and 60 percent of the activities in urban ECs (fig. 4). For example, the Baltimore EZ implemented a customized training program that provided EZ residents with individualized training and a stipend during the training period. In the Bronx portion of the New York EZ, stakeholders explained that they had funded an organization that trained women to become child care providers, a program that not only provided job skills and employment opportunities but also improved the availability of child care in the area. In addition, the Atlanta EZ and the Camden portion of the Philadelphia-Camden EZ implemented educational programs for EZ youth, such as after-school or summer programs. Also, stakeholders from the Upper Manhattan portion of the New York EZ mentioned contributing financial assistance to the business development of the Harlem USA project, a 275,000-square-foot retail development located in the EZ. Moreover, stakeholders from the Providence EC said they provided grants to a nonprofit that offered job training to youth and business development programs, such as "business incubators" that offered office space and technical assistance to new small businesses. Figure 4: Types of Activities Implemented by Urban and Rural EZs and ECs, by Percent of Total Activities: [See PDF for image] Source: GAO analysis of HUD and USDA data. Note: This figure shows the percent of the total number of activities implemented, not the funds devoted to those activity types. The data reporting systems for urban and rural designees used slightly different categories of activities. The Cleveland and Los Angeles EZs are not included in this graphic because they did not receive EZ grant funds. [End of figure] Rural EZs and ECs implemented many of the same types of activities as urban designees, such as business development and job training, but often included activities related to health care and public infrastructure. For example, stakeholders from the Kentucky Highlands and Mid-Delta Mississippi EZs said that they had attracted businesses to the areas using EZ loans, grants, or tax benefits, and stakeholders from the Rio Grande Valley EZ reported funding job training for EZ residents. In addition, stakeholders from Kentucky Highlands said the EZ purchased ambulances for an area that previously did not have those services. All three rural EZs reported using the EZ/EC grant to improve the water or sewerage infrastructure in their EZs, which some said was needed to foster additional economic development. Finally, stakeholders from the Fayette-Haywood EC reported having implemented several activities related to health care, such as recruiting doctors and providing funding to reopen a clinic that had been closed for several years. For more information on the types of activities implemented by the individual communities we visited, see appendix IV. EZs and ECs Used Program Grants to Leverage Additional Funds, but Reliable Data on the Extent of Leveraging Are Not Available: HUD and USDA also expected designees to use their grants to leverage additional investment. Stakeholders from all EZs and ECs we visited and all EC survey respondents reported having used their EZ/EC grants to leverage other resources, including both monetary and in-kind donations. EZs and ECs developed different policies that may have affected the extent to which they leveraged funds. For example, the Mid- Delta EZ required that direct grant recipients obtain at least 65 percent of their funding from other sources. Some other communities, such as the Atlanta EZ, did not have similar requirements for subgrantees, although in some cases subgrantees did leverage funds on their own initiative. EC survey respondents reported using the EZ/EC grants to leverage additional resources for capital improvements, social services, and funding for businesses, among other things. Some EC survey respondents also mentioned that the designation had helped them to leverage funds to implement additional programs or to expand EC programs. All EZs and ECs that provided us with a definition of leveraging said that they included all non-EZ/EC grant funds that were used in EZ/EC- funded programs. But only two of the four EZs that used the program tax- exempt bond included the amount of the bonds in their total leveraged funds. In addition, some EZs reported as leveraged funds other investments made in the EZ area, aside from those directly funded with the EZ/EC grant funds, although other designated communities did not. For example, the Baltimore EZ included all business investments made subsequent to infrastructure improvements the EZ made to an industrial park. USDA encouraged rural EZs and ECs to report all investment in the EZ as leveraged funds, not only those projects that received EZ/EC funds. For example, at USDA's instruction, the Fayette-Haywood EC included funding from other USDA programs operating in the EC, even when EC funds were not involved. However, not all rural sites used this broad definition of leveraging. Similarly, at one HUD official's instruction, the Cleveland EZ included as leveraged funds other investments made within the EZ, such as city Community Development Block Grant funds invested in the area.[Footnote 18] However, there was no written guidance telling the Cleveland EZ to include other investments, and it no longer includes these other investments as leveraged funds in performance reports.[Footnote 19] Although communities reported using the EZ/EC grants to leverage additional resources, we could not verify the actual amounts. HUD's and USDA's performance reporting systems include information on the amount of funds leveraged for each activity, but for the sample of activities we reviewed, either supporting documentation showed an amount conflicting with the reported amount or documentation could not be found.[Footnote 20] In addition, the definition of "leveraged" varied across sites, as the federal agencies did not provide EZs and ECs with a consistent definition of what leveraged funds should include. As a result, designated communities included different types of funds in the amounts they reported as leveraged. Designees Reported Other Accomplishments: In addition to the activities that were implemented, EZ and EC stakeholders with whom we spoke mentioned other accomplishments that were not as easy to quantify and report in the performance systems. For example, one of the aims of the EZ/EC program was to increase collaboration among local governments, nonprofits, community members, and the business community. Stakeholders from several sites we visited commented on how the designation facilitated increased collaboration among different groups of people and organizations. For instance, several stakeholders from the Rio Grande Valley EZ noted the value of having different communities and people work together, something that had not happened prior to the EZ/EC program. Several EC survey respondents also mentioned the importance of collaboration and partnerships in carrying out the EC program. Stakeholders from some sites we visited mentioned that the EZ/EC program had helped to empower local residents by giving them a better understanding of how government worked. In addition, stakeholders from some EZs said that the EZ/EC program had helped to build the capacity of local organizations. In Cleveland, local stakeholders said that the funding provided by the EZ had helped increase the organizational capacity of four local community development corporations and that participation in the governance of the EZ helped to foster communication between the groups. Designees Reported Implementation Challenges: EZ stakeholders also mentioned some issues that had made implementing the EZ/EC program more challenging. Stakeholders from some EZs noted that an initial lack of experience or expertise on the part of EZ officials had made it difficult to implement the program. In addition, stakeholders from the Camden portion of the Philadelphia-Camden EZ and the Rio Grande Valley EZ said that local subgrantee organizations generally had a low level of organizational capacity, which sometimes made it difficult to choose qualified applicants to implement EZ programs. Stakeholders from several sites also said that it was difficult to manage the expectations of both the EZ community and of residents and businesses that were not located in the zones and were not eligible for EZ/EC program benefits, especially when the individuals and businesses were located just across the street from the designated area. EZs and ECs Established a Variety of Governance Structures and Encouraged Community Participation: In addition to choosing the activities that their EZs or ECs implemented, designated communities were permitted to determine the structure they would use to govern and operate the program. Generally, these structures included an EZ/EC management entity--either a nonprofit organization or an entity that was part of the local government. Two urban EZs--New York and Philadelphia-Camden--became two separate entities that were managed by different types of organizations that split the $100 million EZ grant. In the Philadelphia-Camden EZ, for example, the Philadelphia portion was run by the city of Philadelphia and the Camden portion by a nonprofit organization. All designees had at least one board, and, in some cases, EZs included community advisory groups or separate "subzone" boards, which represented specific areas of the EZ in their governance structures. All three rural EZ boards made decisions about EZ activities without the direct involvement of local government entities. However, the extent of government involvement in urban EZ boards varied, regardless of whether the EZ was managed by a nonprofit or local government organization (fig. 5). For example, in two EZs, Cleveland and Chicago, local government had extensive control of the program, but in other EZs, such as Detroit, the board of the nonprofit organization that managed the EZ shared partial decision-making authority with the mayor and city council. Other EZs were operated with minimal local government involvement, with the boards determining which activities to implement, allocating resources, and deciding which entities would implement the programs. Appendix IV provides more details on the governance structures of the EZs we visited. Figure 5: Local Government Involvement in Decision Making in the Urban EZs: [See PDF for image] Source: GAO analysis of EZ documentation and interview data. [A] The Los Angeles EZ was operated by a for profit organization--the Los Angeles Community Development Bank--until 2002 when it filed for bankruptcy. Since then, a Los Angeles city department has continued its operations; however, the mayor and city council are not directly involved. [End of figure] Another program expectation was to encourage community participation within the designated communities. Regardless of the type of governance structure they used, EZs and ECs involved community participants in the planning and carrying out of program activities. According to stakeholders from all the EZs and the ECs we visited, residents were involved in meetings such as "visioning sessions" and town hall gatherings during the strategic planning process. Community groups, such as local colleges and universities, development corporations, and businesses, were also involved prior to designation. In addition, 56 out of 58 ECs responding to our survey reported that EC residents attended listening sessions, generated ideas for activities, or helped to establish priorities. Respondents also indicated that a variety of other groups participated in the strategic planning process for the ECs, including local government officials and representatives from community-based organizations. After designation, stakeholders from the EZs and ECs we visited said that residents often served on boards, and some stakeholders noted they relied on the boards to capture a wide range of viewpoints. Most EZs and ECs we visited also included as participants business representatives, officials from nonprofits, and clergy, among others. Some EZs and ECs also included residents from specific neighborhoods within the designated area or individuals with special expertise, such as in the areas of health care and housing. Oversight Was Hindered by Limited Program Data and Variation in Monitoring: According to our federal standards, federal agencies should oversee the use of public resources and ensure that ongoing monitoring occurs.[Footnote 21] However, HHS, HUD, and USDA did not collect data on how program funds were spent. In addition, HHS did not provide the states, EZs, and ECs with clear guidance on how to monitor the program grant funds, and the types and extent of monitoring performed by state and local participants varied. The lack of reporting requirements may be related to the program's design, which was intended to give communities flexibility in using program funds and relied on multiple agencies for oversight. However, these limitations have hindered the agencies' efforts to determine whether the public resources are being used effectively and program goals are met. Federal Agencies Are Required to Oversee the Use of Public Funds and Provide Ongoing Monitoring: According to federal standards established in the Standards for Internal Control in the Federal Government, program managers need both program and fiscal data to determine whether public resources are being used effectively and program goals are being met.[Footnote 22] In the case of the EZ/EC program, fiscal data would include not only the aggregate amount of program grant funding designated communities spent, but also data on the amount of funds spent on specific types of activities. Program data would include descriptions of the activities implemented and program outputs, such as the number of individuals trained in a job training program. The standards also state that federal agencies should ensure that ongoing monitoring occurs in the course of normal operations. For instance, the federal agencies should provide guidelines on what monitoring should occur, including whether on-site reviews or reporting are required. For the EZ/EC program, HHS regulations require states, EZs, and ECs to maintain fiscal control of program funds and accounting procedures sufficient to enable them to prepare reports and ensure the funds were not used in violation of the applicable statute. The Federal Agencies' Oversight Efforts Had Shortcomings in Data Collection: None of the federal agencies collected data showing how program funds had been spent. As we have noted, the EZ/EC grants were special Social Services Block Grants that gave recipients expanded flexibility in using the funds. The regulations for most grants of this type require states to report on, among other things, the amount of funding spent on each type of activity. However, because HHS did not require this level of reporting for the EZ/EC program, the agency's data show how much of each grant was used but not how much was spent on specific activities or types of activities. Further, HHS's data sometimes do not show how much of the grant a specific EC used, since states could aggregate drawdowns for multiple communities. For example, there are five urban ECs in Texas, but the data reported to HHS show only the aggregate amount of funds these ECs used, not the amount used by each. Similarly, although HUD's and USDA's reporting systems contained some information on the amount of EZ/EC grants budgeted for specific activities, the systems did not account for the amounts actually spent on those activities. Moreover, we found that the data on the amount of EZ/EC grant funding were often not reliable, as some EZs and ECs reported budgeted amounts and others reported actual amounts spent. Further, in our assessments of the reliability of these data, we found documentation showing that the designated communities had undertaken certain activities with program funding, but we were often unable to find documentation of the actual amounts allocated or expended.[Footnote 23] Program Monitoring by State and Local Participants Varied: Although HHS regulations require states, EZs, and ECs to maintain fiscal control of program grant funds, the agency also did not provide guidance detailing the steps state and local authorities should take to monitor the program. In the absence of clear guidance, the type and level of monitoring conducted at the state and local levels varied. For example, some state and EZ/EC officials applied guidelines from other programs, such as the Community Development Block Grant program, or developed their own policies. Officials from almost all states we interviewed said they reviewed audits of the EZs and ECs and were required to submit aggregate data to HHS, and most had performed site visits at least once during the program. State officials also said they reviewed requests to draw down grant funds, approving expenditures if the requests met the goals outlined in the strategic plans. However, most states did not maintain records showing the types of activities designated communities undertook. Some states said that they had taken corrective actions, such as withholding payments when designated communities had not properly reported how funds were used. However, only a few states also completed program monitoring activities, such as reviewing whether a project took place or benefited EZ or EC residents, in conjunction with their fiscal reviews. Most of the EZs and ECs we visited conducted on-site monitoring of subgrantees and reviewed their financial and performance data, and some communities required annual audits of their subgrantees. For example, the Rio Grande Valley EZ assigned a program staff member to monitor each subgrantee activity and required annual audits. In contrast, the Fayette-Haywood EC did not perform any site visits and relied on other funding organizations to monitor subgrantees. Some instances of misuse of program funds did occur during the EZ/EC program. For example, officials at the Mid-Delta EZ reported two cases of embezzlement by EZ personnel.According to an EZ official, in one case that was discovered through an independent audit, an individual was prosecuted for embezzling $28,000 in 1996 (only $1,800 was recouped). The second case of embezzlement of $31,000 by two EZ staff, discovered when the staff turned themselves in, is currently under joint State of Mississippi and FBI investigation as part of a larger investigation of misuse of EZ funds starting as early as 1996. In addition, three audits by the state of Georgia found that almost all the administrative funds designated for the Atlanta EZ ($4 million) had been used in the first 3 ½ years of the program, including approximately $44,000 used for questionable costs related to personnel and travel expenditures. To address this issue, the Atlanta EZ repaid some of the costs in question, provided additional documentation, and instituted better recordkeeping procedures. The city of Atlanta also initiated a restructuring of the EZ and fired the majority of EZ staff. Limitations in EZ/EC Oversight May Have Resulted from the Program Design: As discussed earlier, the EZ/EC program was designed to give the designated communities increased flexibility in deciding how to use program funds and used states as pass-through entities for providing funds. Part of the philosophy behind the program was to relieve states and localities of the burden of excessive reporting requirements. Furthermore, no single federal agency had sole responsibility for oversight of Round I of the EZ/EC program, although federal standards require that agencies provide adequate oversight over public resources. In the beginning, the agencies made some efforts to share information, but these efforts were not maintained. For example, HUD officials said that they had received fiscal data from HHS and reconciled that information with their program data on the activities implemented in the early years of the program.[Footnote 24] According to HUD, the agency made additional attempts to obtain data from HHS but only recently received a report. An HHS official said the agency no longer regularly shared detailed data with HUD and USDA, which the official said was likely due to a lack of program staff. These limitations do not necessarily apply to Rounds II and III of the EZ/EC program. For example, both fiscal and program oversight of the urban and rural EZs and ECs were provided directly through HUD and USDA in Round II because the program funding came directly through HUD and USDA appropriations. Officials from both agencies explained that information on the activity for which funds were used was linked to each drawdown of program funds. In addition, a HUD official said they had issued improved monitoring guidance in Round II, since designees receive funds directly from HUD. However, a USDA official said that they provided similar monitoring guidance to designees in Rounds I, II, and III. Because this report focuses on Round I of the program, we did not determine the effectiveness of the oversight of future rounds of the program. Lack of Detailed Tax Data Made It Difficult to Assess the Use of Program Tax Benefits: A lack of detailed tax data limited our ability to assess the extent to which businesses in the EZs and ECs used program tax benefits. We have previously reported that information on tax expenditures should be collected to ensure that these expenditures are achieving their intended purpose.[Footnote 25] IRS collects data on the use of some of the program tax benefits, but not all of them, and none of the data can be linked to the individual communities where the benefits were claimed. We also recommended that HUD, USDA, and IRS work together to identify the data needed to measure the use of EZ/EC tax benefits and the cost-effectiveness of collecting the information, but the three agencies did not reach agreement on a cost-effective approach.[Footnote 26] Officials from some EZs and ECs reported that some of the tax benefits were being used, but this information was not sufficient to allow us to determine the actual extent of usage. IRS Data on the Use of Program Tax Benefits Are Limited: Previously, we have noted that information on tax expenditures should be collected in order to evaluate their effectiveness as a means of accomplishing federal objectives and to ensure that they are achieving their intended purpose.[Footnote 27] Inadequate or missing data can impede such studies, especially given the difficulties in quantifying the benefits of tax expenditures. Nevertheless, we have stated that the nation's current and projected fiscal imbalance serves to reinforce the importance of engaging in such evaluations. However, as described in our 2004 report, the IRS collects limited data on the EZ/EC tax benefits. It does not collect data on benefits used in individual designated sites and for some benefits it does not have any data.[Footnote 28] For example, the IRS collects some information on EZ businesses' use of tax credits for employing EZ residents. However, the data cannot be separated to show how much was claimed in individual EZs. In addition, IRS does not have data on the use of the increased expensing deduction for depreciable property, because taxpayers do not report this benefit as a separate line item on their returns. The lack of data on the use of program tax benefits is consistent with findings of other reports we prepared citing data challenges in other similar community and economic development programs, such as the Liberty Zone program.[Footnote 29] Our 2004 report recommended that HUD, IRS, and USDA collaborate to identify a cost-effective means of collecting the data needed to assess the use of the tax benefits.[Footnote 30] In response, HUD, IRS, and USDA identified two methods for collecting the information--through a national survey or by modifying the tax forms. However, the three agencies did not reach agreement on a cost-effective method for collecting additional data. Given the lack of information at the federal level, we, some EZs, and other researchers have tried to assess the use of EZ/EC tax benefits by surveying businesses.[Footnote 31] However, these surveys have had low response rates and a high number of undeliverable surveys, suggesting that the results might not be representative. Reasons associated with the low response rates were cited in previous reports, including the difficulty of locating someone at the businesses who knew whether the tax benefit had been claimed and issues associated with multiple business locations.[Footnote 32] In addition, some EZ officials said that businesses were not willing to share their tax information. Further, a high rate of small business closures was determined to be a contributing factor to the high number of undeliverable surveys. We initiated a survey of businesses as a part of the audit work for this engagement, but discontinued the survey due to a low response rate.[Footnote 33] In the absence of other data, we relied on testimonial information to assess how often the EZ tax benefits were used and who used them. Although stakeholders from all EZs told us that they did not have any data on the extent to which EZ businesses had used program tax benefits, they provided us with some information that was consistent with the findings of past studies.[Footnote 34] For example, during our site visits, EZ stakeholders told us that they believed large businesses, which tend to use tax professionals who know and understand the benefits, were more likely to use the tax benefits than small businesses. They also noted that small businesses were less likely to make enough in profits to take advantage of the tax benefits.[Footnote 35] The stakeholders stated further that the credit for employing EZ residents was the most frequently used of the three original tax benefits. A few EZ officials commented that retail businesses were more likely to use the employment credit and manufacturing businesses were more likely to use the increased expensing deduction. Stakeholders from only 4 of the 11 EZs and 2 of the 58 ECs that responded to our EC survey told us that the tax-exempt bond benefit had been used in their communities. EZ stakeholders and EC survey respondents cited a variety of reasons that the tax-exempt bond financing had not been more widely used. For instance, some said that the bonds were not used because of the availability of the Industrial Development Revenue Bond, which EZ stakeholders explained had fewer restrictions and could be issued for larger amounts.[Footnote 36] In addition, some EZ stakeholders and one EC survey respondent said that it was difficult to find a large pool of qualified EZ residents to satisfy the employment requirement for the bond, which required that at least 35 percent of the workforce be EZ residents. Some EZ stakeholders also told us that the legal fees for an EZ bond were higher than for other types of bonds because the restrictions made the EZ bond more complex. For this reason, stakeholders explained, the cost of issuing the EZ bond was high relative to the bond cap, particularly early in the program.[Footnote 37] Finally, some EC survey respondents noted other reasons for not using the bond, such as the complicated nature of the bond or a lack of interested businesses or viable projects. IRS Officials Reported that They Have Data Sufficient to Enforce the Tax Code, but This Information Is Insufficient for Assessing the Extent of Usage: IRS officials said that the limited data the agency collected did not affect its ability to enforce compliance with the tax code. They told us that IRS's role is to administer tax laws and said that collecting more comprehensive data on the use of program tax benefits would not help the agency to achieve this objective. Further, they said that they allocate their resources based on the potential effect of abuse on federal revenue and noted that these tax benefits are not considered high risk, since the amount claimed is small compared with revenues collected from other tax provisions or the amount of potential losses from abusive tax schemes. Furthermore, both IRS officials and our previous reports have suggested that IRS generally does not collect information on the frequency of use or types of businesses claiming tax benefits unless legislatively mandated to do so.[Footnote 38] Although the total program tax benefits were estimated to be much larger than the federal grant funding--over $2.5 billion compared with the $1 billion in EZ/EC grants--we do not, as we have noted, know the actual amount of tax benefits claimed by Round I EZs and ECs nationwide or the amounts used in individual communities.[Footnote 39] As a result, we could not assess differences in the rates of usage among the designated communities. Although we understand IRS's concerns, the lack of data is likely to become increasingly problematic in light of the fact that future rounds of the EZ/EC program and the Renewal Community program rely heavily on tax benefits to achieve revitalization goals. It may also be a concern with the Gulf Opportunity Zone Act, which provides tax benefits in counties and parishes affected by the 2005 Gulf Coast hurricanes.[Footnote 40] In Aggregate, EZs and ECs Showed Some Improvements, but Our Analysis Did Not Definitively Link These Changes to the Program: Although EZs and ECs showed some improvements in poverty, unemployment, and economic growth, we did not find a definitive connection between these changes and the EZ/EC program. As mentioned in our previous report, measuring the effect of initiatives such as the EZ/EC program is difficult for a number of reasons, such as data limitations and the difficulty of determining what would have happened in the absence of the program.[Footnote 41] In some cases, communities saw decreases in poverty and unemployment and increases in economic growth. But, we could not conclusively determine whether these changes were a response to the EZ/EC program or to other economic conditions. EZ stakeholders and EC survey respondents said that program-related factors had influenced changes in their communities but that other unrelated factors also had an effect. Although the overall effects of the EZ/EC program remain unclear, having data on the use of program grants and tax benefits would have allowed for a richer assessment of the program. A Number of Challenges Affected Our Efforts to Measure the Effects of the EZ/EC Program: We attempted to assess the effects of the program on four indicators: poverty, unemployment, and two measures of economic growth--the number of businesses and the number of jobs.[Footnote 42] Although we used several quantitative and qualitative methods, including an econometric analysis to try to isolate the EZ/EC program's effect, we could not differentiate between the effects of the program and other factors. Among the challenges we encountered were the following: * A lack of adequate data on the use of program benefits. As mentioned earlier, data on the use of EZ/EC grant funds and tax benefits were very limited. * Limited demographic data. We used poverty and unemployment data from the 1990 and 2000 censuses, but these dates do not correspond well to the program dates, as communities were designated in 1994 and in some cases are still operating. * Demonstrating what would have happened in the absence of the program. For example, we attempted to identify comparison areas that did not receive EZ or EC designations and that reflected similar community characteristics of EZs and ECs.[Footnote 43] However, the designated communities sometimes had the highest poverty levels in the area, making it difficult to find exact matches among nearby census tracts. * Accounting for the spillover effects of the program to other areas, the effects of similar public and private programs, and the effects of regional and local economic trends. * Accounting for bias in the choice of program areas. For example, if program officials tended to pick census tracts that were already experiencing gentrification prior to 1994, we may be overstating the effect of the EZ designation.[Footnote 44] Conversely, if officials tended to choose census tracts that were experiencing economic declines prior to 1994, such as areas affected by the loss of major employers, we may be understating the program's impact. Several program-specific factors also limited our ability to assess the effects of the program. First, the program was designed to be tailored to the local sites, and each community was given broad latitude to determine its own needs and the program activities it thought would address those needs. Thus, each designee may or may not have selected program activities that directly related to the three factors--poverty, unemployment, and economic growth--mandated for our evaluation. Second, the time frame of actual program implementation may have varied among the designees. For instance, some EZ stakeholders mentioned that their programs took 2 or 3 years to get started, while others were able to begin drawing down funds in the first year. Third, the nature of the EZ/EC program, which focuses on changes in geographic areas rather than on individuals, makes it difficult to determine how the program affected residents who lived in an EZ/EC in 1994 but later moved. Stakeholders from most of the EZs and ECs we visited said that residents were moving out of the designated areas, often after finding a job. If true, this phenomenon may have masked some of the program's effects on poverty and unemployment, since these individuals would not be captured in the 2000 data. In Some Cases, EZs and ECs Showed Improvements in Poverty, Unemployment, and Economic Growth: Some EZs and ECs saw improvements in poverty, unemployment, and economic growth. Four of the 11 EZs--Cleveland, Detroit, Philadelphia- Camden, and Kentucky Highlands--showed improvements in both poverty and unemployment between 1990 and 2000 and at least one measure of economic growth between 1995 and 2004 (fig. 6). Some ECs also experienced similar improvements. For example, 25 out of 95 ECs saw positive changes in poverty and unemployment and at least one measure of economic growth.[Footnote 45] None of the EZs and ECs experienced negative changes in all three indicators, but many experienced negative changes in at least one. For instance, the Atlanta EZ experienced negative changes in unemployment and both measures of economic growth. However, the extent of these changes varied, particularly in our two measures of economic growth. For those EZs that saw improvements in the number of jobs, the increases ranged from a low of 2.6 percent in the Philadelphia-Camden EZ to a high of 67.8 percent in the Kentucky Highlands EZ. Of those EZs that saw decreases in the number of businesses, the amount varied from 2.7 percent in the Detroit EZ to 20.8 percent in the Atlanta EZ. Figure 6: Changes in Poverty, Unemployment, and Two Measures of Economic Growth Observed in Round I EZs: [See PDF for image] Source: GAO analysis of Census and Claritas data. Note: The changes in poverty and unemployment rates are based on the difference between 1990 and 2000 Census data, and the changes in the number of businesses and jobs are based on the difference between 1995 and 2004 data from a private data vendor, Claritas. All poverty and unemployment estimates had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the change in the number of businesses and jobs, we did not consider a change of plus or minus one percent or less as being significant. [End of figure] Most EZs and ECs Saw Some Decrease in the Poverty Rate, but These Changes Could Not Be Tied Definitively to the EZ/EC Program: In most of the 11 EZs and 95 ECs, both urban and rural, poverty rates fell between 1990 and 2000 (fig. 7).[Footnote 46] Most communities experienced statistically significant decreases in the poverty rate that ranged from 2.6 to 14.6 percent. Specifically, our analysis showed the following: * Almost all urban EZs experienced significant decreases ranging from a low of 4.1 percentage points in the New York EZ to 10.9 percentage points in the Detroit EZ. * All three rural EZs showed significant decreases--7.3 percentage points in the Rio Grande Valley EZ, 10.1 percentage points in the Kentucky Highlands EZ, and 10.7 percentage points the Mid-Delta EZ. * 44 out of the 65 urban ECs also saw significant decreases in poverty, with declines ranging from 2.6 percentage points in the Boston, Massachusetts EC to 14.6 percentage points in the Minneapolis, Minnesota EC. * Most rural ECs saw significant decreases, ranging from 3.4 percentage points in the Imperial County, California EC to 12.2 percentage points in the Eastern Arkansas EC. Figure 7: Number and Percentage of EZs and ECs Experiencing a Decrease in Poverty from 1990 to 2000: [See PDF for image] Source: GAO analysis of Census data. Note: All poverty estimates had 95 percent confidence intervals of plus or minus 5 percentage points or less. [End of figure] We also compared changes in poverty in designated areas and comparison areas and across urban and rural communities for both EZs and ECs. Our analysis showed the following: * When combining urban and rural areas, the poverty rate in the designated areas fell more than in the comparison areas--5.4 percentage points overall, compared with 3.9 percentage points in the comparison areas (fig. 8). * Rural designees experienced a larger significant decrease in poverty than urban designees--7.2 and 5 percentage points, respectively. * Urban and rural EZs experienced greater decreases in poverty than both their comparison areas and the ECs. Figure 8: Comparison of Decreases in Poverty in Urban and Rural Designated Areas and Comparison Areas from 1990 to 2000: [See PDF for image] Source: GAO analysis of Census data. Note: There are 1,557 census tracts in the designated areas and 1,504 in the comparison areas. All poverty estimates had 95 percent confidence intervals of plus or minus 5 percentage points or less. [End of figure] Because we could not separate the program's effects from other factors in these analyses, we developed an econometric model for the eight urban EZs and their comparison areas that considered a variety of factors related to the poverty rate.[Footnote 47] Among the nonprogram factors we considered were high school dropouts, the presence of households headed by females, and vacant housing units as reported in the 1990 Census. Our models indicated that the poverty rate in the comparison areas fell slightly more than in the EZs themselves (app. II). This result did not demonstrate that the declines in poverty in the EZs were directly associated with the EZ program. Finally, we conducted interviews of EZ stakeholders and surveyed EC officials to determine their views of the effects of the EZ/EC program on their communities. Their responses were consistent with the inconclusive results of our other analyses: in general, they believed that both the EZ/EC program and additional factors had affected the prevalence of poverty in their communities.[Footnote 48] Some EZ and EC stakeholders said that the EZ/EC designation and program activities had addressed poverty by bringing in jobs and helping to stabilize the area. For instance, stakeholders from several EZs, including the Chicago, Mid-Delta, and Kentucky Highlands EZs, mentioned the role of the EZ in job creation. In addition, stakeholders from other EZs, such as Detroit and Rio Grande Valley, mentioned the role of EZ programs that were related to housing. EC survey respondents commented that the EC designation gave them the opportunity to focus on initiatives that could improve poverty in the area, such as job creation, infrastructure and physical improvements, and housing. However, EZ and EC stakeholders also mentioned external factors that may have affected the changes in poverty, such as changes in the local population when original residents moved away and gentrification. In addition, stakeholders from three EZs mentioned the positive effects of changes to welfare policy during the EZ/EC program.[Footnote 49] In ECs where our data showed that the poverty rate fell, some EC survey respondents also mentioned an increase in the availability of social services as a contributing factor. At EZs where stakeholders had mixed opinions on the changes in poverty, some cited a loss of industry or shifts in the national economy. Of the three EC survey respondents in areas where poverty either remained the same or increased, respondents mentioned the decrease in the number of jobs, increase in housing and utility costs, and the out-migration of residents with middle or high incomes. Decreases in the Unemployment Rate in Some Communities Also Could Not Be Definitively Tied to the EZ/EC Program: As we did for the poverty rate, we analyzed changes in the unemployment rate in EZs and ECs, using the same quantitative and qualitative methods. We found an overall decline in unemployment across communities; but, once again we could not tie the decrease definitively to the program's presence. Further, fewer than half of the individual EZs and ECs experienced a decrease in unemployment (fig. 9), with declines ranging from 1.5 to 11.7 percentage points, and a number saw significant increases--up to 6.5 percentage points.[Footnote 50] Many communities did not experience a significant change. Specifically, our analysis showed the following: * Four of the eight urban EZs saw unemployment fall, with rates declining from 2.9 percentage points in the Philadelphia-Camden EZ to 10 percentage points in the Cleveland EZ. Two of the EZs saw unemployment rise--2 percentage points in New York and 6 percentage points in Atlanta--and two did not see a statistically significant change. * Changes in the unemployment rates of the rural EZs were also mixed. For example, unemployment in the Kentucky Highlands EZ fell 2 percentage points, but it rose 3.1 percentage points in the Mid-Delta EZ and did not change significantly in the Rio Grande Valley EZ. * Twenty-seven, or fewer than half, of the 65 urban ECs saw significant decreases from 1.5 percentage points (San Diego, California) to 8.7 percentage points (Flint, Michigan). Eleven saw a significant increase of between 2.1 percentage points (Rochester, New York) and 6.5 percentage points (Charlotte, North Carolina), while 27 did not experience a significant change. * Almost half of the rural ECs saw significant decreases, with declines ranging from 2.7 percentage points (Fayette-Haywood, Tennessee) to 11.7 percentage points (Lake County, Michigan). The unemployment rate remained about the same in 12 rural ECs, but 4 showed increases of between 2.8 and 3.5 percentage points (Williamsburg-Lake City, South Carolina and Central Savannah River Area, Georgia, respectively). Figure 9: Number and Percentage of EZs and ECs that Experienced a Decrease in Unemployment from 1990 to 2000: [See PDF for image] Source: GAO analysis of Census data. Note: All unemployment estimates had 95 percent confidence intervals of plus or minus 5 percentage points or less. [End of figure] Our analysis also looked at changes in unemployment across urban and rural communities and compared changes in designated areas and comparison areas for both EZs and ECs. The analysis showed the following results: * The designated areas saw a statistically significant decrease in unemployment of 1.4 percentage points, compared with a decrease of just under 1 percentage point in the comparison areas (fig. 10). * In general, rural designees saw unemployment fall more than urban designees, although these differences were not as marked as those we identified in our analysis of the changes in poverty. * Urban EZs and ECs saw a greater decrease in unemployment than their comparison areas, where the rates did not show a statistically significant change. * Unemployment in rural EZs and their comparison areas remained about the same, while rural ECs and their comparison areas both experienced a significant decrease of about 2 percentage points. Figure 10: Comparison of Decreases in Unemployment in Urban and Rural Designated Areas and Comparison Areas from 1990 to 2000: [See PDF for image] Source: GAO analysis of Census data. Note: Areas for which there was no statistically significant change are not shown. There are 1,557 census tracts in the designated areas and 1,504 in the comparison areas. All unemployment estimates had 95 percent confidence intervals of plus or minus 5 percentage points or less. [End of figure] Although our analyses of changes again showed that EZs experienced a larger decrease in unemployment than the comparison areas, these analyses did not separate the effect of the program from other factors. We again used an econometric model for the eight urban EZs that considered other factors, such as average household income and the presence of individuals with a high school diploma as reported in the 1990 Census. This analysis showed that the EZs experienced a decrease that was slightly greater than in the comparison areas, but the difference was not statistically significant (app. II). We also looked at the observations of EZ stakeholders that we interviewed and the responses to our EC survey. Once again, these observations generally saw both program and external factors as affecting the changes in unemployment.[Footnote 51] Some EZ stakeholders cited EZ programs--such as providing financial assistance to EZ businesses, fostering job creation, and offering job training--as helping to reduce unemployment. For example, the Upper Manhattan and Bronx portions of the New York EZ and the Chicago EZ required subgrantees and borrowers to create a certain number of jobs based on the size of the EZ grant or loan received. Similarly, EC survey respondents also mentioned the EC's involvement in creating jobs, attracting new businesses, and offering loans and technical assistance to businesses, along with a variety of social service programs designed to support employment. EZ stakeholders and EC survey respondents also noted additional factors that may have been associated with changes in unemployment. For example, some EZs cited the availability of social services not sponsored by the EZ as factors that influenced unemployment--for instance, daycare, transportation, and adult education or job placement programs. Some EZ stakeholders also suggested that changes in the national economy and in welfare policy had helped to reduce unemployment. Many survey respondents in ECs where unemployment fell reported that the decreases could be attributed to activities that may or may not have been part of the EC program, including adult educational services, higher skill levels among area residents, and social services such as childcare, programs for the homeless, and substance abuse treatment. Stakeholders from EZs where unemployment did not change or rose explained that EZ residents faced barriers to employment such as a lack of education or job skills, drug dependency, and criminal histories. Our Measures Showed that Some Economic Growth Occurred, but Results from Our Econometric Model Were Not Conclusive: A number of indicators can be used to measure economic growth, including data on the change in the number of local businesses, sales volumes, or home values. Our poverty and unemployment analyses used specific variables available in Census data, but to measure economic growth, we chose two measures--the number of businesses and the number of jobs.[Footnote 52] Overall, our analysis showed that most EZs and ECs experienced an increase in at least one measure of economic growth between 1995 and 2004 (fig. 11). Specifically: * Two of the eight urban EZs experienced significant increases in the number of both businesses and jobs, and three more experienced significant increases in one measure. The increases in businesses ranged from 4.2 percent in the Philadelphia-Camden EZ to 23.6 percent in the New York EZ. The increases in jobs ranged from 2.6 percent in the Philadelphia-Camden EZ to 30.5 percent in the Detroit EZ. However, some urban EZs experienced decreases in the number of businesses or jobs, some of which were large. Five experienced decreases in the number of businesses, ranging from 2.7 percent in the Detroit EZ to 20.8 percent in the Atlanta EZ, and four experienced decreases in the number of jobs, from 5.2 percent in the Los Angeles EZ to 22.3 percent in the Atlanta EZ. * All three rural EZs experienced increases in both businesses and jobs, with businesses increasing between 15.6 percent in the Mid-Delta EZ and 33 percent in the Kentucky Highlands EZ, and jobs rising between 5 and 67.8 percent in the same two EZs, respectively. * Fourteen of the 64 urban ECs experienced an increase in both economic growth measures, and an additional 24 saw an increase in one of the measures.[Footnote 53] However, 26 urban ECs saw a decrease in both measures. * Like rural EZs, the majority of the rural ECs experienced an increase in both measures of economic growth. Figure 11: Number and Percentage of EZs and ECs That Experienced an Increase in One or Both Measures of Economic Growth between 1995 and 2004: [See PDF for image] Source: GAO analysis of Claritas data. Note: We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Data were not available for the Miami/Dade County, Florida EC. [End of figure] Like the analyses of poverty and unemployment, our analysis of the changes in economic growth compared urban and rural designees, designated and comparison areas, and EZs and ECs (fig. 12). * In aggregate, both designated and comparison areas saw little change in the number of businesses, and both experienced an increase in the number of jobs of about 7 percent. * Overall, urban designees saw a decrease in the number of businesses, while rural designees saw a substantial increase. Both urban and rural designees saw an increase in the number of jobs, but the aggregate increase in rural areas was much greater (23.6 percent) than in urban areas (5.7 percent). Urban and rural comparison areas generally experienced changes similar to the designated areas. * Urban EZs experienced a decrease in the number of businesses, while the number in comparison areas remained about the same. But urban EZs saw an increase in the number of jobs, while their comparison areas saw a decrease. * Rural EZs fared better than their comparison areas in both measures of economic growth. Figure 12: Comparison of Changes in the Number of Businesses and the Number of Jobs in Urban and Rural Designated Areas and Comparison Areas between 1995 and 2004: [See PDF for image] Source: GAO analysis of Claritas data. Note: There are 1,557 census tracts in the designated areas and 1,504 in the comparison areas. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. These analyses do not include data for the Miami/Dade County, Florida EC. [End of figure] As explained earlier, our descriptive analyses could not isolate the effects of the EZ/EC program from other factors affecting the designated and comparison areas. We conducted an econometric analysis that incorporated other factors, such as the percentage of vacant housing units and population density as reported in the 1990 Census. However, the results of our models explained little of the relative changes in the number of businesses or jobs in the urban EZs with respect to their comparison areas (app. II). Because our proxy measures--the number of businesses and jobs--were not the only indicators representative of economic growth, we tested our models using different measures, such as the number of home mortgage originations, but found similar results. As a result, we could not determine with a reasonable degree of confidence the role that the EZs might have played in the changes in economic growth that we observed. We also reviewed the perceptions of EZ stakeholders interviewed and respondents to our survey of ECs on economic growth in their communities.[Footnote 54] These observations cited several aspects of the program that contributed to economic growth, including loan programs and other benefits that aided small businesses, infrastructure improvements, and tax benefits, especially when the tax benefits were combined with other federal, state, and local benefits. Additionally, several stakeholders mentioned that their EZ or EC had acted as a catalyst for other local development. EZ stakeholders also noted several external factors that affected the change in economic growth, such as the increase of jobs in businesses located within the EZ or EC, the role of other state and local initiatives in attracting businesses, and trends in the national economy. In ECs where our data showed an increase in the number of businesses or jobs, some survey respondents reported that the result was due to an increase in technical assistance for area businesses, such as entrepreneurial training programs, and others reported that financial assistance to businesses contributed to the growth, both of which may or may not have been EC programs. EZ stakeholders also mentioned challenges facing their communities, including the lack of infrastructure and residents with incomes that were not high enough to support local businesses. In ECs where our data showed a decrease in the number of businesses or jobs, survey respondents pointed to a decrease in the number of area businesses and downsizing of existing businesses as contributing factors. Additional Program Data Could Facilitate Evaluations of the Effects of the EZ/EC and Similar Programs: Our efforts to analyze the effects of Round I designation on poverty, unemployment, and economic growth were limited by the absence of data on the use of program grant funds, the amount of funds leveraged, and the use of tax benefits. Without these data, we could not account for the amount of funds EZs used to carry out specific activities, the extent to which they leveraged other resources, or how extensively businesses used the tax benefits. As a result, we could not assess differences in program implementation. In addition, as we reported in 2004, we could not evaluate the effectiveness of the tax benefits, although later rounds of the EZ/EC program have relied heavily on them.[Footnote 55] While we recognize, and discussed in our prior report on the EZ/EC program, the difficulties inherent in evaluating economic development programs, having more specific data would facilitate evaluations of this and similar programs.[Footnote 56] For example, the precision of our econometric models might have been improved by combining data on how program funds were used--such as the amounts used for assisting businesses--and the use of program tax benefits with other data we obtained, such as data on businesses and area jobs. Also, additional data would have allowed us to do in-depth evaluations of the extent to which various tax benefits were being used within each community, the size and type of businesses utilizing them, and the potential competitive advantages of using these benefits. Our previous reports have recommended that information on outlay programs and tax expenditures be collected to evaluate the most effective methods for accomplishing federal objectives.[Footnote 57] Observations: The EZ/EC program, one of the most recent large-scale federal programs aimed at revitalizing distressed urban and rural communities, resulted in a variety of activities intended to improve social and economic conditions in the nation's high-poverty communities. As of March 31, 2006, all but 15 percent of the $1 billion in program grant funds provided to Round I communities had been expended, and the program was reaching its end. All three rounds of the EZ/EC program are scheduled to end no later than December 31, 2009. However, given our findings from this evaluation of Round I EZs and ECs, the following observations should be considered if these or similar programs are authorized in the future. Based on our review, we found that oversight for Round I of the program was limited because the three agencies--HHS, HUD, and USDA--did not collect data on how program funds were used, and HHS did not provide state and local entities with guidance sufficient to ensure monitoring of the program. These limitations may be related in part to the design of the program, which offered increased flexibility in the use of funds and relied on multiple agencies for oversight. However, limited data and variation in monitoring hindered federal oversight efforts. In addition, the lack of data on the use of program grant funds, the extent of leveraging, and extent to which program tax benefits were used also limited our ability and the ability of others to evaluate the effects of the program. The lack of data on the use of tax benefits is of particular concern, since the estimated amount of the tax benefits was far greater than the amount of grant funds dedicated to the program. In response to the recommendation in our 2004 report, HUD, IRS, and USDA discussed options for collecting additional data on program tax benefits and determined two methods for collecting the information--through a national survey or the modification of tax forms. The three agencies, however, did not reach agreement on a cost- effective method for collecting the additional data. In our and others' prior attempts to obtain this information using surveys, survey response rates were low and thus did not produce reliable information on the use of program tax benefits. We acknowledge that the collection of additional tax data by IRS would introduce additional costs to both IRS and taxpayers. Nonetheless, a lack of data on tax benefits is significant given that subsequent rounds of the EZ/EC program and the Renewal Community program rely almost exclusively on tax benefits, and other federal economic development programs, such as the recent Gulf Opportunity Zone initiative, involve substantial amounts of tax benefits. Furthermore, the nation's current and projected fiscal imbalance serves to reinforce the importance of understanding the benefits of such tax expenditures. If Congress authorizes similar programs that rely heavily on tax benefits in the future, it would be prudent for federal agencies responsible for administering the program to collect information necessary for determining whether the tax benefits are effective in achieving program goals. Agency Comments and Our Evaluation: We provided a draft of this report for review and comment to HHS, HUD, IRS, and USDA. We received comments from HHS, HUD, and USDA. In general, the agencies provided comments related to the oversight of the program, the availability of data, and the methodology used to carry out the work. Their written comments appear in appendixes V through VII, respectively, and our responses to HUD's more detailed comments also appear in appendix VI. HHS, HUD, and USDA also provided technical comments, which we have incorporated into the report where appropriate. HHS commented that a statement made in our report--that the agency did not provide guidance detailing the steps state and local authorities should take to monitor the program--unfairly represented the relationship between HHS and the other federal agencies that administered the EZ/EC program. Specifically, HHS emphasized its responsibility for fiscal as opposed to programmatic oversight of the program. We note in our report that program design may have led to a lack of clarity in oversight, as no single federal agency had sole oversight responsibility. While this lack of clarity in oversight may be related in part to the design of the program, which offered increased flexibility in the use of funds and relied on multiple agencies for oversight, limited data and variation in monitoring hindered federal oversight efforts. Moreover, we believe that, in accordance with federal standards, each of the federal agencies that administered the program bore at least some responsibility for ensuring that public resources were being used effectively and that program goals were being met. HUD disagreed with GAO's observation that there was a lack of data on the use of program grant funds, the amount of funds leveraged, and the use of the tax benefits. HUD indicated that we could obtain data on the use of program funds and the amount of funds leveraged from its performance reporting system. As we discussed in our report, we used information from HUD's reporting system to report on the types of activities that designated communities implemented. We also noted that HUD maintained some information on the amount of EZ/EC grants budgeted for specific activities. Although we found evidence that activities were carried out with program funds, information contained in the performance reporting system on the amounts of funds used and the amount leveraged was not reliable. For example, we found evidence that communities had undertaken certain activities with program funding, but we were often unable to find documentation of the actual amounts allocated or expended. HUD also indicated that it did not agree that data on the use of the tax benefits were lacking. However, HUD indicated that the agency itself had attempted to gather such data by collaborating with IRS in identifying ways to collect data on tax benefits, by developing a methodology to administer a survey to businesses, and by compiling anecdotal evidence of the use of program tax benefits. We continue to believe that the lack of data on program tax benefits limits the ability of the agencies to administer and evaluate the EZ/EC program. Further, the lack of such data is likely to become increasingly problematic in light of the fact that future rounds of the EZ/EC program and the Renewal Community program rely heavily on tax benefits to achieve revitalization goals. HUD concurred that limitations in the oversight of the EZ/EC program may have resulted from the design of the program as no single federal agency had sole responsibility for oversight. HUD also recommended that we make clear that more oversight was not allowed in Round I and we include a statement that it met agency requirements to undertake periodic performance reviews and described some of its efforts to monitor the program according to applicable regulations. We do not believe that more oversight was not allowed. For example, early in the program HUD and HHS made some efforts to share information. Specifically, HUD officials said that they had received fiscal data from HHS and reconciled that information with their program data on the activities implemented, but these efforts to share information were not maintained. Further, as we previously stated, while we recognize that program design may have led to a lack of clarity in oversight, we believe that in accordance with federal standards, each of the federal agencies that administered the program bore at least some responsibility for ensuring that public resources were being used effectively and that program goals were being met. HUD also described changes it had made to ensure better oversight of program funds for Round II. We acknowledge HUD's efforts to improve oversight of the program and, as discussed in our report, the oversight limitations that we identified in Round I of the program may not apply to later rounds. HUD provided several comments related to the methodology we used to carry out our work. For example, HUD suggested that we measure the successes of the Round I program in meeting the four key principles of the program, which the designated communities were required to include in their strategic plans. Additionally, HUD commented that the indices we used to assess the effects of the EZ/EC program--poverty, unemployment and economic growth--were used in the application process for the program but were not intended to be used as performance measures. While we appreciate HUD's suggestions on our methodology, our congressional mandate was to determine the effect of the EZ/EC program on poverty, unemployment and economic growth. In designing our methodology, we conducted extensive research on evaluations that had been conducted on the EZ/EC program, including HUD's 2001 Interim Assessment, and spoke with several experts in the urban studies field. USDA stated that data and analyses on the effectiveness of programs such as EZ/EC were useful and offered areas to consider for future evaluations of economic development programs involving rural areas. For example, USDA mentioned issues involved in collecting data on rural areas, such as the limited availability of economic and demographic data for small rural populations, and discussed USDA's efforts for developing a methodology that focuses on economic impacts using county- level economic data. USDA also said it is especially important in rural areas to have a clear and adequately funded data collection process for program evaluations. In addition, USDA noted that evaluations of the EZ/EC program could go beyond the indicators of poverty, unemployment and economic growth to include measures on economic development capacity and collaboration. We agree that collecting data for rural areas is a challenge and appreciate USDA's effort to develop a methodology that focuses on economic impacts using county-level economic data and captures the short-term Gross Domestic Product changes in the impacted rural counties. Further, we appreciate USDA's suggestion that additional measures be considered in future evaluations of economic development programs and that a broader perspective on program results might be useful. USDA also commented that its performance reporting system was intended to be used as a management tool for both USDA and the individual EZs and ECs. According to USDA, the system was not designed to be an accounting tool but has been useful for providing a picture of each designated community's achievements. As we discussed in our report, we used information from USDA's reporting system to report on the types of activities that designated communities implemented and also noted that USDA maintained some information on the amounts of EZ/EC grants budgeted for specific activities. Moreover, while we recognize the system was not intended to be used as an accounting tool, we found that the data on the amounts of the EZ/EC grant funding were not reliable. For example, in our assessment of the reliability of data contained in USDA's performance reporting system, we were often unable to find documentation of the actual amounts allocated or expended for specific activities. USDA further commented that it had encouraged designated communities to report all investment that contributed to the EZ or EC in accomplishing its strategic plan as leveraged funds. We recognize USDA's efforts to encourage leveraging in the designated communities and to report such information in its performance reporting system. Our report notes that stakeholders from all EZs and ECs we visited and EC survey respondents reported having used their EZ/EC grants to leverage other resources. However, we were unable to evaluate the amounts of funds leveraged because the data contained in USDA's performance reporting system were not reliable. For example, USDA's performance reporting system included information on the amounts of funds leveraged for each activity, but for the sample of activities we reviewed, either supporting documentation showed an amount conflicting with the reported amount or documentation could not be found. Moreover, as we discuss in our report, the definition of leveraging used among the designated communities was inconsistent. We are sending copies of this report to interested Members of Congress, the Secretary of Health and Human Services, the Secretary of Housing and Urban Development, the Secretary of Treasury, the Commissioner of the Internal Revenue Service, and the Secretary of Agriculture. We will make copies of this report available to others upon request. In addition, this report will be available at no charge on the GAO Web site at [Hyperlink, http://www.gao.gov]. Please contact me at (202) 512-8678 or ShearW@gao.gov if you or your staff have any questions about this report. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. Key contributors to this report are listed in appendix VIII. Signed by: William B. Shear: Director, Financial Markets and Community Investment: [End of section] Appendix I: Objectives, Scope, and Methodology: The objectives of this study were to (1) describe how Round I of the Empowerment Zone and Enterprise Community (EZ/EC) program was implemented by the designated communities; (2) evaluate the extent of federal, state, and local oversight of the program; (3) examine the extent to which data are available to assess the use of program tax benefits; and (4) analyze the effects the Round I EZs and ECs had on poverty, unemployment, and economic growth in their communities. To address each of our objectives, we completed site visits to all Round I EZs and two Round I ECs and administered a survey to all ECs that did not receive subsequent designations, such as a Round II EZ designation. At each site, we asked uniform questions on implementation, oversight, tax benefits, and changes observed in the EZ and ECs. We also surveyed 60 ECs that were in operation as of June 2005 and did not receive later designations and asked about similar topics. We performed a qualitative analysis to identify common themes from our interview data and open- ended survey responses. To address our second objective, we also interviewed federal and state program participants, reviewed oversight guidance and documentation, and verified a sample of reported performance data by tracing it to EZ and EC records. To address our third objective, we attempted to administer a survey of EZ businesses, but discontinued it due to a low response rate. To address our fourth objective, we obtained demographic and socioeconomic data from the 1990 and 2000 decennial censuses and business data for 1995, 1999, and 2004 from a private data vendor, Claritas. We used 1990 Census data to select areas similar to the EZ and EC areas for purposes of comparison. We then calculated the percent changes in poverty, unemployment, and economic growth observed in the EZs and ECs and their comparison areas. In addition, for the eight urban EZs, we used an econometric model to estimate the effect of the program, by controlling for certain factors, such as average household income, in the EZs and their comparison areas. Finally, we used information gathered from our qualitative analysis to provide context for the changes observed in the EZs and ECs. Methodology for Site Visits: To answer our objectives, we completed site visits to all 11 EZs and 2 of the 95 ECs, one urban and one rural.[Footnote 58] These EZs and ECs were located in: * Atlanta, Georgia (EZ): * Baltimore, Maryland (EZ): * Chicago, Illinois (EZ): * Cleveland, Ohio (EZ): * Detroit, Michigan (EZ): * Los Angeles, California (EZ): * New York, New York (EZ): * Philadelphia, Pennsylvania and Camden, New Jersey (EZ): * rural Kentucky (Kentucky Highlands EZ): * rural Mississippi (Mid-Delta EZ): * rural Texas (Rio Grande Valley EZ): * Providence, Rhode Island (EC): * rural Tennessee (Fayette-Haywood EC): We interviewed stakeholders from each site on the implementation, governance, oversight, and tax benefits of the EZ or EC and asked about the changes the stakeholders had observed in their communities. Using a standardized interview guide, we interviewed some combination of the following program stakeholders at each location: EZ/EC officials, board members (including some EZ/EC residents), representatives of subgrantee organizations, and Chamber of Commerce representatives or individuals able to provide the perspective of the business community (table 4).[Footnote 59] We identified participants to interview at each site by soliciting opinions from EZ/EC officials and the current board chair. For each site, we reviewed strategic plans, organizational charts, and documentation on oversight procedures. In addition, we toured the EZ/EC to see some of activities implemented. Table 4: Number of Stakeholders Interviewed for EZ and EC Site Visits, by Type: EZ/EC: Urban: Atlanta EZ; EZ/EC officials: Urban: 2; Board members: Urban: 2; Representatives from subgrantee organizations: Urban: 6; Representatives from the Chamber of Commerce or other business perspective: Urban: 2; Officials from the state pass-through entities: Urban: 4; Other representatives[A]: Urban: 10. EZ/EC: Urban: Baltimore EZ; EZ/EC officials: Urban: 6; Board members: Urban: 4; Representatives from subgrantee organizations: Urban: 3; Representatives from the Chamber of Commerce or other business perspective: Urban: 2; Officials from the state pass-through entities: Urban: 1; Other representatives[A]: Urban: 5. EZ/EC: Urban: Chicago EZ; EZ/EC officials: Urban: 6; Board members: Urban: 4; Representatives from subgrantee organizations: Urban: 4; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 3; Other representatives[A]: Urban: 2. EZ/EC: Urban: Cleveland EZ; EZ/EC officials: Urban: 5; Board members: Urban: 4; Representatives from subgrantee organizations: Urban: 3; Representatives from the Chamber of Commerce or other business perspective: Urban: 0; Officials from the state pass-through entities: Urban: 4; Other representatives[A]: Urban: 8. EZ/EC: Urban: Detroit EZ; EZ/EC officials: Urban: 7; Board members: Urban: 4; Representatives from subgrantee organizations: Urban: 3; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 2; Other representatives[A]: Urban: 9. EZ/EC: Urban: Los Angeles EZ; EZ/EC officials: Urban: 1; Board members: Urban: 2; Representatives from subgrantee organizations: Urban: 2; Representatives from the Chamber of Commerce or other business perspective: Urban: 0; Officials from the state pass-through entities: Urban: 0; Other representatives[A]: Urban: 12. EZ/EC: Urban: New York EZ; EZ/EC officials: Urban: [Empty]; Board members: Urban: [Empty]; Representatives from subgrantee organizations: Urban: [Empty]; Representatives from the Chamber of Commerce or other business perspective: Urban: [Empty]; Officials from the state pass-through entities: Urban: [Empty]; Other representatives[A]: Urban: [Empty]. EZ/EC: Urban: Upper Manhattan portion; EZ/EC officials: Urban: 3; Board members: Urban: 3; Representatives from subgrantee organizations: Urban: 2; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 2; Other representatives[A]: Urban: 2. EZ/EC: Urban: Bronx portion; EZ/EC officials: Urban: 2; Board members: Urban: 4; Representatives from subgrantee organizations: Urban: 3; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 2; Other representatives[A]: Urban: 1. EZ/EC: Urban: Philadelphia/Camden EZ; EZ/EC officials: Urban: [Empty]; Board members: Urban: [Empty]; Representatives from subgrantee organizations: Urban: [Empty]; Representatives from the Chamber of Commerce or other business perspective: Urban: [Empty]; Officials from the state pass-through entities: Urban: [Empty]; Other representatives[A]: Urban: [Empty]. EZ/EC: Urban: Philadelphia portion; EZ/EC officials: Urban: 5; Board members: Urban: 2; Representatives from subgrantee organizations: Urban: 2; Representatives from the Chamber of Commerce or other business perspective: Urban: 0; Officials from the state pass-through entities: Urban: 8; Other representatives[A]: Urban: 3. EZ/EC: Urban: Camden portion; EZ/EC officials: Urban: 1; Board members: Urban: 3; Representatives from subgrantee organizations: Urban: 2; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 2; Other representatives[A]: Urban: 2. EZ/EC: Urban: Providence EC; EZ/EC officials: Urban: 2; Board members: Urban: 3; Representatives from subgrantee organizations: Urban: 3; Representatives from the Chamber of Commerce or other business perspective: Urban: 0; Officials from the state pass-through entities: Urban: 2; Other representatives[A]: Urban: 1. EZ/EC: Rural: Kentucky Highlands EZ; EZ/EC officials: Urban: 4; Board members: Urban: 6[B]; Representatives from subgrantee organizations: Urban: 1; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 4; Other representatives[A]: Urban: 4. EZ/EC: Rural: Mid-Delta Mississippi EZ; EZ/EC officials: Urban: 4; Board members: Urban: 1; Representatives from subgrantee organizations: Urban: 4; Representatives from the Chamber of Commerce or other business perspective: Urban: 2; Officials from the state pass-through entities: Urban: 2; Other representatives[A]: Urban: 2. EZ/EC: Rural: Rio Grande Valley EZ; EZ/EC officials: Urban: 4; Board members: Urban: 3; Representatives from subgrantee organizations: Urban: 3; Representatives from the Chamber of Commerce or other business perspective: Urban: 1; Officials from the state pass-through entities: Urban: 3; Other representatives[A]: Urban: 1. EZ/EC: Rural: Fayette Haywood EC; EZ/EC officials: Urban: 1; Board members: Urban: 2; Representatives from subgrantee organizations: Urban: 2; Representatives from the Chamber of Commerce or other business perspective: Urban: 0; Officials from the state pass-through entities: Urban: 4; Other representatives[A]: Urban: 4. Source: GAO. [A] Includes local governmental officials, business owners, active community members, and other representatives. [B] Includes directors of subzones. [End of table] Methodology for Survey of EC Officials: To gather similar information from the ECs, we administered an e-mail survey to officials from the 60 Round I ECs that were still in operation as of June 2005 and did not receive a subsequent designation. We chose to exclude the 34 ECs that received subsequent designations, because we did not want their responses to be influenced by those programs. A version of the survey showing aggregated responses can be viewed at [Hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-06-734SP]. We developed survey questions from existing program literature and interview data collected from Department of Housing and Urban Development (HUD) and U.S. Department of Agriculture (USDA) headquarters officials as well as our site visits to Round I EZs and ECs. The questionnaire items covered the implementation of the program, the types of governance structures used, usage of the program tax- exempt bond, and stakeholders' views of factors that influenced the changes they observed in poverty, unemployment, and economic growth in their ECs. We created two versions of the questionnaire, one for urban ECs and another for rural ECs, in order to tailor items to urban or rural sites. Department of Health and Human Services (HHS), HUD, and USDA officials reviewed the survey for content, and we conducted pretests at four urban and two rural ECs.[Footnote 60] Since the survey was administered by e-mail, a usability pretest was conducted at one urban EC (Akron, Ohio) to observe the respondent answering the questionnaire as it would appear when opened and displayed on their computer screen. In administering the survey, we took the following steps to increase the response rate. To identify survey participants, we obtained contact information for the Round I ECs that did not receive a subsequent designation from HUD and USDA in April 2005.[Footnote 61] We then sent a notification e-mail to inform the ECs of the survey, to identify the correct point of contact, and to ensure the e-mail account was active. Those who did not respond to the first e-mail received follow up e- mails and telephone calls. The questionnaire was e-mailed on August 25, 2005 to 27 rural ECs and 33 urban ECs, and participants were given the option to respond via e-mail, fax, or post mail. Between September and December 2005, multiple follow up e-mails and calls were made to increase the response rate. When the survey closed on December 20, 2005, all of the rural ECs and 31 of the 33 urban ECs had completed it. The overall response rate was high at 97 percent, with the response rates for the rural ECs at 100 percent and urban ECs at 94 percent. We did not attempt to verify the respondents' answers against an independent source of information. However, we used two techniques to verify the reliability of questionnaire items. First, we used in-depth interviewing techniques to evaluate the answers of pretest participants, and interviewers judged that all the respondents' answers to the questions were based on reliable information. Second, for the items that asked about changes to poverty, unemployment, and economic growth in the EC, we asked respondents to provide a source of data for their response. Responses to those questions that did not include a data source were excluded from our analysis of those items. The practical difficulties of conducting any survey may introduce certain types of errors, commonly referred to as nonsampling errors. For example, differences in how a particular question is interpreted, the sources of information available to respondents, or the types of people who do not respond can introduce unwanted variability into the survey results. We sought to minimize these errors by taking the following steps: conducting pretests, making follow-up contacts with participants to increase response rates, performing statistical analyses to identify logical inconsistencies, and having a second independent analyst review the statistical analyses. Returned surveys were reviewed for consistency before the data were entered into an electronic database. All keypunched or inputted data were 100-percent verified--that is, the data were electronically entered twice. Further, a random sample of the surveys was verified for completeness and accuracy. We used statistical software to analyze responses to close- ended questions and performed a qualitative analysis on open-ended questions to identify common themes. Methodology for Qualitative Analysis of Site Visit and EC Survey Data: To summarize the information collected at our site visits, we conducted a qualitative analysis of interview data. The goal of the analysis was to create a summary that would produce an overall "story" or brief description of the program as implemented in each site. In this process, we reviewed data from over 200 interviews to identify information pertaining to the following six broad topics: * strategic planning and census tract selection; * goals, implemented activities, leveraging activities, and sustainability; * governance structure and process; * program oversight; * perceptions of the use of tax benefits; and: * perceptions of poverty, unemployment, economic growth, and other changes within the zone. Based on initial reviews of the interview data, we produced general outlines for each topic. For example, a description of the governance structure and process included identifying the type of governance structure used, roles within the structure, opportunities for community involvement, the process for decision making, and successes and challenges related to governance. One reviewer was assigned to each of the six topics for an individual site. The reviewer examined all interviews completed at an individual site and created a topical summary based on interview data. Each summary was verified by (1) presenting the summaries to the group of six interview reviewers to ensure accuracy, clarity, and completeness and (2) having a second reviewer trace the summaries back to source documents. We also performed a qualitative analysis of the open-ended responses in the EC survey to determine reasons why the tax-exempt bond was not more widely used; why poverty, unemployment, and economic growth may have remained the same over the designation; and what role the EC played in changes in poverty, unemployment and economic growth, as well as obtaining general comments about the program. Responses to these questions were first reviewed by an analyst to identify common categories within the responses and then independently verified by a second analyst. Methodology for Review of Program Oversight: We interviewed and obtained documentation from federal, state, and local program participants regarding program oversight. We interviewed officials from the federal agencies involved with the program and obtained and analyzed fiscal and program data from the agencies. In addition, since the states were the pass-through entities for grant funds provided to the EZs and ECs--that is, they distributed federal funding to the communities--we conducted telephone interviews with state officials and obtained relevant documents in the 13 states containing EZs and ECs we visited. Finally, we interviewed EZ and EC officials on their oversight of subgrantees as well as the oversight they received from federal and state entities. We did not perform financial audits of the EZs and ECs. To determine the reliability of data in HUD and USDA Internet-based performance reporting systems, we randomly selected activities at each EZ and EC we visited and conducted a file review to determine the accuracy of: the data.[Footnote 62] In the files, we searched related documentation for the amounts reported in the system for certain categories, including EZ/EC grant funding, leveraged funds, and program outputs. We also determined whether, at a minimum, documentation existed to support that the activity was implemented. We then assigned each item we verified a code (table 5). Finally, we averaged the information for each site by category and calculated the average score for each urban and rural community. Table 5: Coding of Data Reliability of HUD and USDA Performance Systems: Code: 2; Description: Items with strong documentation, meaning that exact documentation existed or could be easily inferred with the provided documentation. Code: 1; Description: Items with weak documentation, meaning that some evidence existed, but numbers did not match. Code: 0; Description: Items for which no documentation existed. Source: GAO. V We found sufficient documentation that most EZ/EC activities contained in the Internet-based reporting systems had occurred, with average codes of 2.0 for urban areas and 1.9 for rural areas.[Footnote 63] We found that data on EZ/EC grant funding, leveraged funds, and program outputs were not sufficiently reliable for our purposes because only weak or no documentation could be found at most sites. Methodology for Survey of EZ Businesses: To assess the use of program tax benefits, we attempted to administer a survey to EZ businesses; however, we discontinued the survey due to a very low response rate. Based on past post-mailed and phone- administered surveys of EZ businesses, we knew that this would be a challenging population to survey. In fact, surveys we and Abt Associates conducted in 1998 obtained response rates of only 42 and 35 percent, respectively.[Footnote 64] In addition, both surveys had a relatively high number of undeliverable surveys. In anticipation of these issues, we attempted to administer a concise, high-level survey via mail to a stratified random sample (n=517) of EZ businesses.[Footnote 65] We implemented a sampling procedure using the 2004 Claritas Business Facts dataset that stratified businesses located in the EZ by three strata: urban small businesses (less than 50 employees), urban large businesses (50 or more employees), and rural businesses. The survey was targeted to private businesses rather than public and nonprofit businesses, since these for-profit businesses were the ones eligible for the tax benefits.[Footnote 66] Public and nonprofit businesses were excluded from the sample by the primary industry code identifier included in the Claritas data. A few of these types of businesses that were not initially excluded based on their industry code were later removed from the sample because the respondents said that they were not eligible for the tax benefits. We developed our survey after reviewing surveys used in previous studies, interviewing business owners, and conducting pretests with EZ businesses. The questionnaire was brief--containing 21 closed-ended items and 1 optional open-ended item--and took most pretest respondents approximately five minutes to complete. When we conducted pretests with 10 businesses from Baltimore, Philadelphia, and rural Kentucky, all pretest participants found the survey to be easy to complete and said that it did not ask for sensitive information. These business owners, however, often lacked complete information about their company's tax filings and were not always able to answer all of the survey questions. Several indicated that they would be unlikely to complete the survey because the topic was not relevant to them. We administered the survey according to standard survey data collection practices. We sent a letter notifying the 517 businesses of our survey about a week prior to the survey mailing, mailed a copy of the survey, and followed that mailing with a reminder postcard. We received a total of 63 responses after our initial mailing, a response rate of 12 percent. Our mailings to 104 businesses (20 percent) could not be delivered and were returned because of incorrect addresses or contact information. Methodology for Assessing the Effect of the Program on Poverty, Unemployment, and Economic Growth: To determine the effect of the EZ/EC program on changes in poverty, unemployment, and economic growth, we used a variety of quantitative methods that examined changes in the designated program areas and areas we identified as comparison areas. In addition, we incorporated interview data in our qualitative analysis to provide context for the changes observed. We calculated percent changes of demographic, socioeconomic, and business data between two points in time for the all Round I EZs and ECs.[Footnote 67] However, we used only urban EZs in our econometric analysis because of data limitations in rural areas and the amount of funds awarded to ECs. Description of Data Sources: To assess the changes in poverty and unemployment, we used census tract- level data on poverty rates and unemployment rates from the 1990 and 2000 decennial censuses. To determine changes in economic growth in EZ and ECs, we defined economic growth in terms of the number of private businesses created and the total number of jobs in the areas.[Footnote 68] We obtained year-end data on these variables for years 1995, 1999, and 2004 from the Business-Facts Database maintained by Claritas, a private data processing company. We explored several public and private data sources that contained the number of businesses and jobs at the census tract level and selected Claritas because it (1) maintained archival data, (2) provided data with a high level of reliability at the census tract level, and (3) used techniques to ensure the representation of small businesses. We also explored a variety of other data options to enhance our analysis, but were ultimately not able to use them. For example, we tried to acquire data throughout the period of the program, such as state unemployment data, local building permit and crime data, and data on students receiving free or reduced lunches. However, we were not able to use these data because they were not captured consistently across sites, not available at the census tract level, or not sufficiently reliable for our purposes. The decennial census data used are from the census long form that is administered to a sample of respondents. Because census data used in this analysis are estimated based on a probability sample, each estimate is based on just one of a large number of samples that could have been drawn. Since each sample could have produced different estimates, we express our confidence in the precision of our particular sample's results as a 95 percent confidence interval. For example, the estimated percent change in the poverty rate of EZs is a decrease of 6.1 percent, and the 95 percent confidence interval for this estimate ranges from 4.9 to 7.2 percent. This is the interval that would contain the actual population value for 95 percent of the samples that could have been drawn. As a result, we are 95 percent confident that each of the confidence intervals in this report will include the true values in the study population. All Census variables based on percentages, such as poverty rate and unemployment rate, have 95 percent confidence intervals of plus or minus 5 percentage points or less. The confidence intervals for average household income and average owner-occupied housing value are shown in table 6. Table 6: Confidence Intervals for Average Household Income and Average Housing Value in Constant 2004 Dollars[A]: Average household income: Atlanta EZ; 95 percent confidence interval: 1990 estimate: $18,343; 95 percent confidence interval: From: $17,466; 95 percent confidence interval: To: $19,220; 95 percent confidence interval: 2000 estimate: $28,552; 95 percent confidence interval: From: $27,205; 95 percent confidence interval: To: $29,899; 95 percent confidence interval: Percent change: 55.66; 95 percent confidence interval: From: 55.4; 95 percent confidence interval: To: 55.91. Average household income: Atlanta EZ: Comparison; 95 percent confidence interval: 1990 estimate: 30,567; 95 percent confidence interval: From: 29,741; 95 percent confidence interval: To: 31,393; 95 percent confidence interval: 2000 estimate: 39,500; 95 percent confidence interval: From: 38,328; 95 percent confidence interval: To: 40,672; 95 percent confidence interval: Percent change: 29.23; 95 percent confidence interval: From: 28.99; 95 percent confidence interval: To: 29.46. Average household income: Baltimore EZ; 95 percent confidence interval: 1990 estimate: 28,185; 95 percent confidence interval: From: 27,207; 95 percent confidence interval: To: 29,164; 95 percent confidence interval: 2000 estimate: 35,059; 95 percent confidence interval: From: 33,566; 95 percent confidence interval: To: 36,551; 95 percent confidence interval: Percent change: 24.39; 95 percent confidence interval: From: 24.1; 95 percent confidence interval: To: 24.67. Average household income: Baltimore EZ; Comparison; 95 percent confidence interval: 1990 estimate: 27,931; 95 percent confidence interval: From: 27,316; 95 percent confidence interval: To: 28,546; 95 percent confidence interval: 2000 estimate: 31,367; 95 percent confidence interval: From: 30,511; 95 percent confidence interval: To: 32,223; 95 percent confidence interval: Percent change: 12.3; 95 percent confidence interval: From: 12.05; 95 percent confidence interval: To: 12.56. Average household income: Chicago EZ; 95 percent confidence interval: 1990 estimate: 23,097; 95 percent confidence interval: From: 22,636; 95 percent confidence interval: To: 23,559; 95 percent confidence interval: 2000 estimate: 34,718; 95 percent confidence interval: From: 33,868; 95 percent confidence interval: To: 35,567; 95 percent confidence interval: Percent change: 50.31; 95 percent confidence interval: From: 50.13; 95 percent confidence interval: To: 50.49. Average household income: Chicago EZ: Comparison; 95 percent confidence interval: 1990 estimate: 28,431; 95 percent confidence interval: From: 28,030; 95 percent confidence interval: To: 28,832; 95 percent confidence interval: 2000 estimate: 39,985; 95 percent confidence interval: From: 39,367; 95 percent confidence interval: To: 40,604; 95 percent confidence interval: Percent change: 40.64; 95 percent confidence interval: From: 40.48; 95 percent confidence interval: To: 40.8. Average household income: Detroit EZ; 95 percent confidence interval: 1990 estimate: 22,644; 95 percent confidence interval: From: 22,034; 95 percent confidence interval: To: 23,253; 95 percent confidence interval: 2000 estimate: 33,751; 95 percent confidence interval: From: 32,660; 95 percent confidence interval: To: 34,842; 95 percent confidence interval: Percent change: 49.05; 95 percent confidence interval: From: 48.84; 95 percent confidence interval: To: 49.26. Average household income: Detroit EZ: Comparison; 95 percent confidence interval: 1990 estimate: 25,609; 95 percent confidence interval: From: 25,197; 95 percent confidence interval: To: 26,021; 95 percent confidence interval: 2000 estimate: 36,200; 95 percent confidence interval: From: 35,523; 95 percent confidence interval: To: 36,877; 95 percent confidence interval: Percent change: 41.36; 95 percent confidence interval: From: 41.19; 95 percent confidence interval: To: 41.52. Average household income: New York EZ; 95 percent confidence interval: 1990 estimate: 26,518; 95 percent confidence interval: From: 25,981; 95 percent confidence interval: To: 27,054; 95 percent confidence interval: 2000 estimate: 33,557; 95 percent confidence interval: From: 32,833; 95 percent confidence interval: To: 34,280; 95 percent confidence interval: Percent change: 26.54; 95 percent confidence interval: From: 26.34; 95 percent confidence interval: To: 26.75. Average household income: New York: Comparison; 95 percent confidence interval: 1990 estimate: 26,993; 95 percent confidence interval: From: 26,714; 95 percent confidence interval: To: 27,272; 95 percent confidence interval: 2000 estimate: 31,247; 95 percent confidence interval: From: 30,872; 95 percent confidence interval: To: 31,622; 95 percent confidence interval: Percent change: 15.76; 95 percent confidence interval: From: 15.59; 95 percent confidence interval: To: 15.93. Average household income: Upper Manhattan; 95 percent confidence interval: 1990 estimate: 26,559; 95 percent confidence interval: From: 25,971; 95 percent confidence interval: To: 27,147; 95 percent confidence interval: 2000 estimate: 34,041; 95 percent confidence interval: From: 33,239; 95 percent confidence interval: To: 34,844; 95 percent confidence interval: Percent change: 28.17; 95 percent confidence interval: From: 27.96; 95 percent confidence interval: To: 28.39. Average household income: Bronx; 95 percent confidence interval: 1990 estimate: 26,294; 95 percent confidence interval: From: 24,983; 95 percent confidence interval: To: 27,606; 95 percent confidence interval: 2000 estimate: 30,842; 95 percent confidence interval: From: 29,238; 95 percent confidence interval: To: 32,446; 95 percent confidence interval: Percent change: 17.29; 95 percent confidence interval: From: Average household income: 16.95; 95 percent confidence interval: To: 17.64. Average household income: Philadelphia-Camden EZ; 95 percent confidence interval: 1990 estimate: 23,188; 95 percent confidence interval: From: 22,259; 95 percent confidence interval: To: 24,117; 95 percent confidence interval: 2000 estimate: 28,562; 95 percent confidence interval: From: 27,197; 95 percent confidence interval: To: 29,927; 95 percent confidence interval: Percent change: 23.17; 95 percent confidence interval: From: 22.87; 95 percent confidence interval: To: 23.48. Average household income: Philadelphia-Camden EZ: Comparison; 95 percent confidence interval: 1990 estimate: 27,292; 95 percent confidence interval: From: 26,031; 95 percent confidence interval: To: 28,553; 95 percent confidence interval: 2000 estimate: 31,318; 95 percent confidence interval: From: 29,718; 95 percent confidence interval: To: 32,918; 95 percent confidence interval: Percent change: 14.75; 95 percent confidence interval: From: 14.4; 95 percent confidence interval: To: 15.1. Average household income: Philadelphia; 95 percent confidence interval: 1990 estimate: 22,269; 95 percent confidence interval: From: 21,262; 95 percent confidence interval: To: 23,276; 95 percent confidence interval: 2000 estimate: 27,851; 95 percent confidence interval: From: 26,309; 95 percent confidence interval: To: 29,392; 95 percent confidence interval: Percent change: 25.07; 95 percent confidence interval: From: 24.74; 95 percent confidence interval: To: 25.39. Average household income: Camden; 95 percent confidence interval: 1990 estimate: 26,742; 95 percent confidence interval: From: 24,465; 95 percent confidence interval: To: 29,018; 95 percent confidence interval: 2000 estimate: 31,158; 95 percent confidence interval: From: 28,228; 95 percent confidence interval: To: 34,088; 95 percent confidence interval: Percent change: 16.52; 95 percent confidence interval: From: 16.05; 95 percent confidence interval: To: 16.98. Average household income: Cleveland EZ; 95 percent confidence interval: 1990 estimate: 20,535; 95 percent confidence interval: From: 19,730; 95 percent confidence interval: To: 21,340; 95 percent confidence interval: 2000 estimate: 28,781; 95 percent confidence interval: From: 27,524; 95 percent confidence interval: To: 30,038; 95 percent confidence interval: Percent change: 40.16; 95 percent confidence interval: From: 39.9; 95 percent confidence interval: To: 40.42. Average household income: Cleveland EZ: Comparison; 95 percent confidence interval: 1990 estimate: 24,688; 95 percent confidence interval: From: 24,171; 95 percent confidence interval: To: 25,206; 95 percent confidence interval: 2000 estimate: 30,311; 95 percent confidence interval: From: 29,607; 95 percent confidence interval: To: 31,016; 95 percent confidence interval: Percent change: 22.78; 95 percent confidence interval: From: 22.56; 95 percent confidence interval: To: 23. Average household income: Los Angeles EZ; 95 percent confidence interval: 1990 estimate: 28,801; 95 percent confidence interval: From: 28,191; 95 percent confidence interval: To: 29,412; 95 percent confidence interval: 2000 estimate: 32,631; 95 percent confidence interval: From: 31,857; 95 percent confidence interval: To: 33,405; 95 percent confidence interval: Percent change: 13.3; 95 percent confidence interval: From: 13.06; 95 percent confidence interval: To: 13.54. Average household income: Los Angeles EZ: Comparison; 95 percent confidence interval: 1990 estimate: 34,087; 95 percent confidence interval: From: 33,478; 95 percent confidence interval: To: 34,696; 95 percent confidence interval: 2000 estimate: 37,843; 95 percent confidence interval: From: 37,058; 95 percent confidence interval: To: 38,628; 95 percent confidence interval: Percent change: 11.02; 95 percent confidence interval: From: 10.79; 95 percent confidence interval: To: 11.25. Average household income: Kentucky Highlands EZ; 95 percent confidence interval: 1990 estimate: 23,304; 95 percent confidence interval: From: 22,043; 95 percent confidence interval: To: 24,565; 95 percent confidence interval: 2000 estimate: 31,064; 95 percent confidence interval: From: 29,520; 95 percent confidence interval: To: 32,608; 95 percent confidence interval: Percent change: 33.3; 95 percent confidence interval: From: 32.99; 95 percent confidence interval: To: 33.61. Average household income: Mid-Delta EZ; 95 percent confidence interval: 1990 estimate: 25,872; 95 percent confidence interval: From: 24,321; 95 percent confidence interval: To: 27,424; 95 percent confidence interval: 2000 estimate: 35,559; 95 percent confidence interval: From: 33,392; 95 percent confidence interval: To: 37,726; 95 percent confidence interval: Percent change: 37.44; 95 percent confidence interval: From: 37.12; 95 percent confidence interval: To: 37.76. Average household income: Rio Grande Valley EZ; 95 percent confidence interval: 1990 estimate: 25,093; 95 percent confidence interval: From: 23,626; 95 percent confidence interval: To: 26,560; 95 percent confidence interval: 2000 estimate: 32,763; 95 percent confidence interval: From: 30,920; 95 percent confidence interval: To: 34,606; 95 percent confidence interval: Percent change: 30.57; 95 percent confidence interval: From: 30.24; 95 percent confidence interval: To: 30.9. Average household income: Providence EC; 95 percent confidence interval: 1990 estimate: 28,593; 95 percent confidence interval: From: 27,525; 95 percent confidence interval: To: 29,661; 95 percent confidence interval: 2000 estimate: 32,616; 95 percent confidence interval: From: 31,229; 95 percent confidence interval: To: 34,004; 95 percent confidence interval: Percent change: 14.07; 95 percent confidence interval: From: 13.75; 95 percent confidence interval: To: 14.39. Average household income: Fayette-Haywood EC; 95 percent confidence interval: 1990 estimate: 32,560; 95 percent confidence interval: From: 31,008; 95 percent confidence interval: To: 34,111; 95 percent confidence interval: 2000 estimate: 45,353; 95 percent confidence interval: From: 43,249; 95 percent confidence interval: To: 47,457; 95 percent confidence interval: Percent change: 39.29; 95 percent confidence interval: From: 39.01; 95 percent confidence interval: To: 39.57. Average owner-occupied housing value: Atlanta EZ; 95 percent confidence interval: 1990 estimate: $55,883; 95 percent confidence interval: From: $52,688; 95 percent confidence interval: To: $59,077; 95 percent confidence interval: 2000 estimate: $117,869; 95 percent confidence interval: From: $106,218; 95 percent confidence interval: To: $129,519; 95 percent confidence interval: Percent change: 110.92; 95 percent confidence interval: From: 110.68; 95 percent confidence interval: To: 111.17. Average owner-occupied housing value: Atlanta EZ: Comparison; 95 percent confidence interval: 1990 estimate: 74,063; 95 percent confidence interval: From: 72,446; 95 percent confidence interval: To: 75,680; 95 percent confidence interval: 2000 estimate: 101,774; 95 percent confidence interval: From: 96,312; 95 percent confidence interval: To: 107,236; 95 percent confidence interval: Percent change: 37.42; 95 percent confidence interval: From: 37.15; 95 percent confidence interval: To: 37.68. Average owner-occupied housing value: Baltimore EZ; 95 percent confidence interval: 1990 estimate: 53,714; 95 percent confidence interval: From: 51,381; 95 percent confidence interval: To: 56,048; 95 percent confidence interval: 2000 estimate: 62,219; 95 percent confidence interval: From: 58,659; 95 percent confidence interval: To: 65,779; 95 percent confidence interval: Percent change: 15.83; 95 percent confidence interval: From: 15.48; 95 percent confidence interval: To: 16.19. Average owner-occupied housing value: Baltimore EZ: Comparison; 95 percent confidence interval: 1990 estimate: 55,966; 95 percent confidence interval: From: 54,113; 95 percent confidence interval: To: 57,819; 95 percent confidence interval: 2000 estimate: 62,514; 95 percent confidence interval: From: 59,920; 95 percent confidence interval: To: 65,108; 95 percent confidence interval: Percent change: 11.7; 95 percent confidence interval: From: 11.39; 95 percent confidence interval: To: 12.01. Average owner-occupied housing value: Chicago EZ; 95 percent confidence interval: 1990 estimate: 71,429; 95 percent confidence interval: From: 67,487; 95 percent confidence interval: To: 75,372; 95 percent confidence interval: 2000 estimate: 160,411; 95 percent confidence interval: From: 150,476; 95 percent confidence interval: To: 170,347; 95 percent confidence interval: Percent change: 124.57; 95 percent confidence interval: From: 124.38; 95 percent confidence interval: To: 124.77. Average owner-occupied housing value: Chicago EZ: Comparison; 95 percent confidence interval: 1990 estimate: 88,445; 95 percent confidence interval: From: 85,343; 95 percent confidence interval: To: 91,548; 95 percent confidence interval: 2000 estimate: 167,015; 95 percent confidence interval: From: 159,548; 95 percent confidence interval: To: 174,482; 95 percent confidence interval: Percent change: 88.83; 95 percent confidence interval: From: 88.64; 95 percent confidence interval: To: 89.03. Average owner-occupied housing value: Detroit EZ; 95 percent confidence interval: 1990 estimate: 23,114; 95 percent confidence interval: From: 22,153; 95 percent confidence interval: To: 24,075; 95 percent confidence interval: 2000 estimate: 52,234; 95 percent confidence interval: From: 49,362; 95 percent confidence interval: To: 55,106; 95 percent confidence interval: Percent change: 125.99; 95 percent confidence interval: From: 125.81; 95 percent confidence interval: To: 126.16. Average owner-occupied housing value: Detroit EZ: Comparison; 95 percent confidence interval: 1990 estimate: 28,598; 95 percent confidence interval: From: 27,620; 95 percent confidence interval: To: 29,575; 95 percent confidence interval: 2000 estimate: 61,160; 95 percent confidence interval: From: 58,688; 95 percent confidence interval: To: 63,632; 95 percent confidence interval: Percent change: 113.86; 95 percent confidence interval: From: 113.7; 95 percent confidence interval: To: 114.03. Average owner-occupied housing value: New York EZ; 95 percent confidence interval: 1990 estimate: 207,544; 95 percent confidence interval: From: 166,353; 95 percent confidence interval: To: 248,735; 95 percent confidence interval: 2000 estimate: 301,835; 95 percent confidence interval: From: 244,974; 95 percent confidence interval: To: 358,697; 95 percent confidence interval: Percent change: 45.43; 95 percent confidence interval: From: 44.89; 95 percent confidence interval: To: 45.98. Average owner-occupied housing value: New York EZ: Comparison; 95 percent confidence interval: 1990 estimate: 177,446; 95 percent confidence interval: From: 167,025; 95 percent confidence interval: To: 187,867; 95 percent confidence interval: 2000 estimate: 209,423; 95 percent confidence interval: From: 198,465; 95 percent confidence interval: To: 220,380; 95 percent confidence interval: Percent change: 18.02; 95 percent confidence interval: From: 17.66; 95 percent confidence interval: To: 18.38. Average owner-occupied housing value: Upper Manhattan; 95 percent confidence interval: 1990 estimate: 238,864; 95 percent confidence interval: From: 188,845; 95 percent confidence interval: To: 288,882; 95 percent confidence interval: 2000 estimate: 384,155; 95 percent confidence interval: From: 308,848; 95 percent confidence interval: To: 459,462; 95 percent confidence interval: Percent change: 60.83; 95 percent confidence interval: From: 60.32; 95 percent confidence interval: To: 61.33. Average owner-occupied housing value: Bronx; 95 percent confidence interval: 1990 estimate: 99,728; 95 percent confidence interval: From: 71,856; 95 percent confidence interval: To: 127,600; 95 percent confidence interval: 2000 estimate: 124,588; 95 percent confidence interval: From: 100,021; 95 percent confidence interval: To: 149,155; 95 percent confidence interval: Percent change: 24.93; 95 percent confidence interval: From: 24.23; 95 percent confidence interval: To: 25.63. Average owner-occupied housing value: Philadelphia-Camden EZ; 95 percent confidence interval: 1990 estimate: 29,899; 95 percent confidence interval: From: 28,060; 95 percent confidence interval: To: 31,739; 95 percent confidence interval: 2000 estimate: 37,780; 95 percent confidence interval: From: 35,895; 95 percent confidence interval: To: 39,664; 95 percent confidence interval: Percent change: 26.36; 95 percent confidence interval: From: 26.02; 95 percent confidence interval: To: 26.69. Average owner-occupied housing value: Philadelphia-Camden EZ: Comparison; 95 percent confidence interval: 1990 estimate: 42,045; 95 percent confidence interval: From: 39,630; 95 percent confidence interval: To: 44,461; 95 percent confidence interval: 2000 estimate: 51,159; 95 percent confidence interval: From: 44,926; 95 percent confidence interval: To: 57,392; 95 percent confidence interval: Percent change: 21.67; 95 percent confidence interval: From: 21.22; 95 percent confidence interval: To: 22.13. Average owner-occupied housing value: Philadelphia; 95 percent confidence interval: 1990 estimate: 28,288; 95 percent confidence interval: From: 26,263; 95 percent confidence interval: To: 30,313; 95 percent confidence interval: 2000 estimate: 37,353; 95 percent confidence interval: From: 35,178; 95 percent confidence interval: To: 39,528; 95 percent confidence interval: Percent change: 32.04; 95 percent confidence interval: From: 31.7; 95 percent confidence interval: To: 32.39. Average owner-occupied housing value: Camden; 95 percent confidence interval: 1990 estimate: 35,076; 95 percent confidence interval: From: 30,928; 95 percent confidence interval: To: 39,224; 95 percent confidence interval: 2000 estimate: 39,398; 95 percent confidence interval: From: 35,730; 95 percent confidence interval: To: 43,067; 95 percent confidence interval: Percent change: 12.32; 95 percent confidence interval: From: 11.8; 95 percent confidence interval: To: 12.84. Average owner-occupied housing value: Cleveland EZ; 95 percent confidence interval: 1990 estimate: 38,071; 95 percent confidence interval: From: 36,277; 95 percent confidence interval: To: 39,866; 95 percent confidence interval: 2000 estimate: 75,186; 95 percent confidence interval: From: 71,537; 95 percent confidence interval: To: 78,835; 95 percent confidence interval: Percent change: 97.49; 95 percent confidence interval: From: 97.29; 95 percent confidence interval: To: 97.69. Average owner-occupied housing value: Cleveland EZ: Comparison; 95 percent confidence interval: 1990 estimate: 46,972; 95 percent confidence interval: From: 45,966; 95 percent confidence interval: To: 47,979; 95 percent confidence interval: 2000 estimate: 70,161; 95 percent confidence interval: From: 68,649; 95 percent confidence interval: To: 71,674; 95 percent confidence interval: Percent change: 49.37; 95 percent confidence interval: From: 49.19; 95 percent confidence interval: To: 49.54. Average owner-occupied housing value: Los Angeles EZ; 95 percent confidence interval: 1990 estimate: 141,665; 95 percent confidence interval: From: 138,933; 95 percent confidence interval: To: 144,397; 95 percent confidence interval: 2000 estimate: 156,492; 95 percent confidence interval: From: 151,907; 95 percent confidence interval: To: 161,078; 95 percent confidence interval: Percent change: 10.47; 95 percent confidence interval: From: 10.21; 95 percent confidence interval: To: 10.72. Average owner-occupied housing value: Los Angeles EZ: Comparison; 95 percent confidence interval: 1990 estimate: 160,090; 95 percent confidence interval: From: 157,393; 95 percent confidence interval: To: 162,787; 95 percent confidence interval: 2000 estimate: 165,180; 95 percent confidence interval: From: 161,599; 95 percent confidence interval: To: 168,761; 95 percent confidence interval: Percent change: 3.18; 95 percent confidence interval: From: 2.94; 95 percent confidence interval: To: 3.42. Average owner-occupied housing value: Kentucky Highlands EZ; 95 percent confidence interval: 1990 estimate: 43,392; 95 percent confidence interval: From: 38,713; 95 percent confidence interval: To: 48,071; 95 percent confidence interval: 2000 estimate: 65,815; 95 percent confidence interval: From: 62,527; 95 percent confidence interval: To: 69,104; 95 percent confidence interval: Percent change: 51.68; 95 percent confidence interval: From: 51.35; 95 percent confidence interval: To: 52. Average owner-occupied housing value: Mid-Delta EZ; 95 percent confidence interval: 1990 estimate: 50,061; 95 percent confidence interval: From: 47,323; 95 percent confidence interval: To: 52,800; 95 percent confidence interval: 2000 estimate: 66,872; 95 percent confidence interval: From: 59,968; 95 percent confidence interval: To: 73,777; 95 percent confidence interval: Percent change: 33.58; 95 percent confidence interval: From: 33.19; 95 percent confidence interval: To: 33.97. Average owner-occupied housing value: Rio Grande Valley EZ; 95 percent confidence interval: 1990 estimate: 46,100; 95 percent confidence interval: From: 42,654; 95 percent confidence interval: To: 49,546; 95 percent confidence interval: 2000 estimate: 61,450; 95 percent confidence interval: From: 55,970; 95 percent confidence interval: To: 66,929; 95 percent confidence interval: Percent change: 33.3; 95 percent confidence interval: From: 32.91; 95 percent confidence interval: To: 33.69. Average owner-occupied housing value: Providence EC; 95 percent confidence interval: 1990 estimate: 124,339; 95 percent confidence interval: From: 118,190; 95 percent confidence interval: To: 130,489; 95 percent confidence interval: 2000 estimate: 116,698; 95 percent confidence interval: From: 99,200; 95 percent confidence interval: To: 134,196; 95 percent confidence interval: Percent change: -6.15; 95 percent confidence interval: From: -6.77; 95 percent confidence interval: To: - 5.52. Average owner-occupied housing value: Fayette-Haywood EC; 95 percent confidence interval: 1990 estimate: $68,945; 95 percent confidence interval: From: $65,765; 95 percent confidence interval: To: $72,125; 95 percent confidence interval: 2000 estimate: $103,619; 95 percent confidence interval: From: $97,144; 95 percent confidence interval: To: $110,094; 95 percent confidence interval: Percent change: 50.29; 95 percent confidence interval: From: 50.01; 95 percent confidence interval: To: 50.57. Source: GAO analysis of Census data. [A] Other variables used in the model and shown in the site visit descriptions that were based on percentages, such as the poverty rate, had confidence intervals of less than +/-5 percentage points. [End of table] In addition to sampling errors, Census data (both sampled and 100 percent data) are subject to nonsampling errors that may occur during the operations used to collect and process census data. Examples of nonsampling errors are not enumerating every housing unit or person in the sample, failing to obtain all required information from a respondent, obtaining incorrect information, and recording information incorrectly. Operations such as field review of enumerator's work, clerical handling of questionnaires, and electronic processing of questionnaires also may introduce nonsampling errors in the data. The Census Bureau discusses sources of nonsampling errors and makes attempts to limit them. Choosing Comparison Areas Using the Propensity Score: To provide context for the changes we observed in the EZs and ECs, we calculated the percent change of the designated areas as well as areas, called comparison areas, that most closely resembled the EZ/EC program areas. To select comparison areas for our analysis, we used a statistical matching method called the propensity score. The propensity score predicts the probability that a tract could have been designated based on having characteristics similar to those found in the tracts selected for the program. We used five factors to calculate the propensity scores, as shown in table 7. Table 7: Factors Selected for Choosing Comparison Tracts: Factor: 1990 poverty rate[ A]; Reason selected: * EZ/EC program eligibility criteria; * Factor considered in a similar study[B, E]. Factor: 1990 unemployment rate[C]; Reason selected: * EZ/EC program eligibility criteria; * Factor considered in a similar study[ B]. Factor: 1990 population density[ D]; Reason selected: * Calculation based on two EZ/EC program eligibility criteria, population and area; * Factor considered in a similar study[B]. Factor: 1990 average household income; Reason selected: * Factor considered in similar studies[ B, E]. Factor: Percentage of minority population in 1990[F]; Reason selected: * Factor considered in similar studies[ B, E]. Source: GAO. [A] Percent based on individuals for whom poverty status has been determined. [B] See Bondonio, Daniele and John Engberg (2000), "Enterprise Zones and Local Employment: Evidence from the States' Programs," Regional Science and Urban Economics, Vol. 30, No.5, pp. 519-549. [C] Percent based on individuals 16 years of age or older. [D] Individuals per square mile. [E] Hebert and others, Interim Assessment. [F] For the purposes of this report, we calculated minority population by subtracting the percent of white population from the total population. [End of table] To ensure that our comparison areas were similar to the designated areas in terms of geography, we explored two selection methods, one that included tracts in the same county as the EZ/EC and in adjacent counties, and another that selected tracts within a 5-mile radius of the EZ/EC.[Footnote 69] We excluded tracts that received a subsequent designation in the EZ/EC or Renewal Community programs in 1998 and 2002 in order to remove the possibility of tracts that may have received similar benefits affecting our analysis. After mapping the resulting comparison tracts using these two methods, we decided to use tracts selected within a 5-mile radius of the EZs and ECs because this method provided more contiguous areas, while the results of the county and the adjacent counties method yielded comparison tracts in other states where political structures and types of funds could differ. Using the computed propensity scores, we selected comparison tracts whose scores were greater than 0.1. This threshold was chosen because most EZ tracts had propensity scores of 0.1 or higher; therefore, comparison tracts with propensity scores of at least 0.1 were the most similar to the EZ tracts. This threshold also yielded approximately the same number of comparison tracts as EZ tracts in most of the eight urban EZs. In addition, we tested this threshold by running our models with comparison tracts whose propensity scores were greater than 0.05 or 0.15 and found that the results did not change significantly.[Footnote 70] Some limitations exist with this method. For example, since many of census tracts chosen for the program may have had the highest level of poverty, it was difficult to find tracts with the same level of poverty. Our Descriptive and Econometric Analyses: We calculated the percent changes at the program wide level for our four indicators of poverty, unemployment, and economic growth for both designated and comparison areas.[Footnote 71] We also calculated the changes for urban and rural designees and EZs and ECs separately, so that we could make comparisons between those groups. In addition, for the eight urban Round I EZs, we calculated the percentages separately for each EZ and EZ comparison area to show differences between zones. Although the comparison areas were sufficient to use in our program wide analyses, for rural EZs and urban and rural ECs, we did not use comparison areas for site-level analyses because there were too few comparison tracts. For example, the Providence, Rhode Island EC consisted of 13 tracts, but the area had only four eligible comparison tracts. We also completed an econometric analysis of the eight urban EZs. We used a standard econometric approach, the weighted least squares model, which allowed us to analyze the change from 1990 to 2000 and compare it with the 1990 value of several explanatory variables. The benefit of this approach is that the program, officially implemented in 1994, would not affect the 1990 values of the explanatory variables. In addition, we spoke with several experts in the urban studies field on our methodology. For more information on the methods used in our econometric analysis and a full discussion of our results, please see appendix II. [End of section] Appendix II: Methodology for and Results of Our Econometric Models: This appendix describes our efforts to isolate the effect of the EZ/EC program on the changes in poverty, unemployment, and economic growth, by conducting an econometric analysis of all urban EZ census tracts.[Footnote 72] In our analysis of percent changes, we found that poverty and unemployment had decreased and that some economic growth had occurred. However, when we used the econometric models to control for other area characteristics, our results did not definitively suggest that the observed changes in poverty and unemployment were associated with the EZ program in urban areas. In addition, our models did not adequately explain the observed changes in the proxy measures we used for economic growth; thus, the results did not allow us to conclude whether there is an association between the EZ program and economic growth. As mentioned in the report, there were several challenges that limited our ability to determine the effect of the program. First, data at the census tract-level for the program years were limited. We used data from the 1990 and 2000 decennial censuses to show the changes in poverty and unemployment. In addition, we primarily used two measures for economic growth--the number of businesses and the number of jobs from Claritas Business-Facts dataset for years 1995, 1999, and 2004--in our models of economic growth.[Footnote 73] Second, we were not able to account for the spillover effects of EZ designation into their neighboring areas. For example, if the EZ/EC program affected comparison tracts as well as the designated communities, our analyses would not find any significant differences between the designated and comparison tracts. The result may be an obscuring of the extent of the statistical association between the urban EZ program and the study variables. Third, the analyses did not account for the confounding effects of other public or private programs, such as those intended to reduce poverty or unemployment or increase the number of area jobs. As a result, estimates for the EZ program in our analyses may under or overstate the extent of EZ program's correlation with poverty, unemployment, and economic growth. Fourth, our estimations did not fully account for the economic trends that were affecting the choice of areas selected for the program. For example, if program officials tended to pick census tracts that were already experiencing gentrification prior to 1994, our estimations could overstate the effect of the EZ designation. Conversely, if officials tended to choose census tracts that were experiencing economic declines prior to 1994, such as those in which major employers had closed, we might understate the program's impact. We did include a variable from Census data--new housing construction between 1990 and 1994--that measured one dimension of economic trends prior to EZ designation, but we did not include other dimensions, such as employment trends at the tract level, in the models. Description of Our Models: We used a weighted least square regression for our analyses.[Footnote 74] Our dependent variables were (1) the difference in the poverty rate between 1990 and 2000, (2) the difference in the unemployment rate between 1990 and 2000 (3) the difference in the number of businesses between 1995 and 1999, and (4) the number of jobs between 1995 and 1999. For the basic model, we measured the difference in each dependent variable against the 1990 value of some explanatory variables. The benefit of this approach is that 1990 values of the explanatory variables would not have been affected by the program, which was implemented in 1994. We also ran an expanded version of the model that included variables for each of the EZs to determine whether there were differences among the EZs, and we included variables for the EZs and their surrounding areas to account for economic trends at the metropolitan level, such as the growing or declining output of local industries. Some of the explanatory variables for which we controlled included socioeconomic factors, such as percent of population with a high school diploma. In addition to these socioeconomic factors, we also considered the five factors we used to select the comparison tracts: * percent of minority population in 1990, * average household income in 1990, * population density in 1990, * poverty rate in 1990, and: * unemployment rate in 1990. We included these variables because the comparison tracts may not be perfectly matched to the EZ tracts; including these factors allowed us to further account for differences between EZ and comparison tracts. Moreover, we weighted the estimations by the geometric mean of 1990 and 2000 household counts of each tract to account for differences in the number of households in each tract. The purpose of this decision was to put more weight on the tracts with large numbers of households, because these tracts would tend to have smaller sampling errors. The coefficients for the EZ program variables represent the EZs with respect to the comparison areas, and the positive or negative values suggest whether the EZs fared better or worse than the comparison areas. For instance, a positive coefficient in the models for poverty and unemployment would mean that the EZs did not fare as well as the comparison areas--that is, they had either a greater increase or a smaller decrease in poverty or unemployment. See our discussion of the results of each model for more information. Results of Our Models for Poverty: Although our comparison of the percentage change between 1990 and 2000 showed that poverty decreased in most urban EZs, the results of our models did not conclusively suggest that the change in poverty was associated with the EZ program. Our analysis of the percentage changes showed that the poverty rate fell more in the EZs than in the comparison areas. But, when we controlled for other factors in our models, we found in the basic model that poverty decreased less in the EZs than in the comparison areas, although the difference was very small (table 8). In addition, many of the variables used in selection of comparison tracts were significant, suggesting that the choice of areas selected for the program might have affected the differences between the urban EZs and the comparison areas in the change in poverty. When accounting for the different urban EZs and their comparison tracts, the poverty rate decreased more in some urban EZs but less in others with respect to the comparison tracts, although the only significant result was in the Los Angeles EZ, which experienced a greater increase in poverty than the comparison areas. The differences among EZs may be a result of the local factors. In addition, one researcher found that there was a nationwide decrease in the number of people living in high poverty neighborhoods, defined as census tracts with poverty rates of 40 percent or higher, between 1990 and 2000--a trend that might be a factor affecting our results.[Footnote 75] Table 8: Estimates of the Association between the EZ Program and the Change in Poverty Rate, 1990-2000: Variables: EZ program; Basic model: Coefficient: 1.54; Basic model: Standard error: 0.67; Expanded model: Coefficient: [Empty]; Expanded model: Standard error: [Empty]. Variables: Atlanta EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 3.41; Expanded model: [Empty]; Expanded model: Standard error: 2.69. Variables: Baltimore EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -2.27; Expanded model: Standard error: 1.94. Variables: Chicago EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 1.63; Expanded model: Standard error: 1.50. Variables: Cleveland EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -2.43; Expanded model: Standard error: 2.10. Variables: Detroit EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 2.25; Expanded model: Standard error: 1.54. Variables: Los Angeles EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 3.07; Expanded model: Standard error: 1.20. Variables: New York EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -1.87; Expanded model: Standard error: 1.23. Variables: Philadelphia-Camden EZ; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -1.10; Expanded model: Standard error: 2.70. Variables: Atlanta EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -5.50; Expanded model: Standard error: 2.61. Variables: Baltimore [A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 0.60; Expanded model: Standard error: 2.49. Variables: Chicago EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -5.08; Expanded model: Standard error: 2.23. Variables: Cleveland EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -4.65; Expanded model: Standard error: 2.36. Variables: Detroit EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -7.73; Expanded model: Standard error: 2.35. Variables: Los Angeles EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 2.00; Expanded model: Standard error: 2.30. Variables: New York EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 0.56; Expanded model: Standard error: 2.40. Variables: Philadelphia-Camden EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: [ B]; Expanded model: Standard error: [B]. Variables: Percent of high school dropouts[C]; Basic model: Coefficient: 0.080; Basic model: Standard error: 0.026; Expanded model: Coefficient: 0.017; Expanded model: Standard error: 0.024. Variables: Percent of vacant housing units; Basic model: Coefficient: -0.064; Basic model: Standard error: 0.042; Expanded model: Coefficient: 0.046; Expanded model: Standard error: 0.044. Variables: Percent of female-headed households with children[D]; Basic model: Coefficient: 0.24; Basic model: Standard error: 0.045; Expanded model: Coefficient: 0.22; Expanded model: Standard error: 0.042. Variables: Percent employed in retail industry[E]; Basic model: Coefficient: -0.0073; Basic model: Standard error: 0.054; Expanded model: Coefficient: 0.040; Expanded model: Standard error: 0.052. Variables: Percent housing units built between 1990 and 1994[F]; Basic model: Coefficient: -0.17; Basic model: Standard error: 0.081; Expanded model: Coefficient: - 0.29; Expanded model: Standard error: 0.082. Variables: Percent minority population[G]; Basic model: Coefficient: 0.0045; Basic model: Standard error: 0.022; Expanded model: Coefficient: -0.021; Expanded model: Standard error: 0.021. Variables: Average household income (in 2004 dollars); Basic model: Coefficient: -0.00046; Basic model: Standard error: 0.00085; Expanded model: Coefficient: - 0.00060; Expanded model: Standard error: 0.000081. Variables: Population density[H]; Basic model: Coefficient: 0.000047; Basic model: Standard error: 0.0000087; Expanded model: Coefficient: 0.000024; Expanded model: Standard error: 0.000013. Variables: Poverty rate[I]; Basic model: Coefficient: -0.71; Basic model: Standard error: 0.052; Expanded model: Coefficient: -0.79; Expanded model: Standard error: 0.047. Variables: Unemployment rate[J]; Basic model: Coefficient: -0.047; Basic model: Standard error: 0.046; Expanded model: Coefficient: 0.085; Expanded model: Standard error: 0.054. Variables: Constant; Basic model: Coefficient: 30.39; Basic model: Standard error: 4.29; Expanded model: Coefficient: 40.79; Expanded model: Standard error: 4.83. Variables: Number of tracts; Basic model: Coefficient: [Empty]; Basic model: 851; Basic model: Standard error: [Empty]; Expanded model: Coefficient: [Empty]; Expanded model: 851; Expanded model: Standard error: [Empty]. Variables: R-sq; Basic model: Coefficient: [Empty]; Basic model: 0.40; Basic model: Standard error: [Empty]; Expanded model: Coefficient: [Empty]; Expanded model: 0.47; Expanded model: Standard error: [Empty]. Source: GAO analysis of Census data. Notes: Coefficients significant at the 5 percent level are in bold. All variables are from the 1990 Census unless otherwise noted. We weighted the regressions by the geometric mean of 1990 and 2000 household counts of each tract. [A] We defined the EZ area to include both the EZ tracts and comparison tracts that were selected from within a 5-mile boundary of the EZ. [B] Results for the Philadelphia-Camden EZ area are not listed, because we used them as a reference group for the other seven EZs and their surrounding areas. [C] Percent based on the civilian population between ages 16 and 19 who are not enrolled in school and are not high school graduates. [D] Percent based on households headed by females without husbands present in which there are children under 18 years of age. [E] Percent based on individuals 16 and over. [F] From the 2000 Census. [G] We calculated minority population by subtracting the percent of white population from the total population. [H] Individuals per square mile. [I] Percent based on individuals for whom poverty status has been determined. [J] Percent based on individuals 16 years of age or older in the labor force. [End of table] Results of Our Models for Unemployment: Like our models for poverty, our models for the unemployment did not conclusively suggest that the changes in unemployment were associated with the EZ program. The results of our basic model suggested that unemployment decreased more in the EZs than in the comparison areas, but the difference was very small and was not statistically significant (table 9). All five of the variables we used to select comparison tracts were statistically significant, suggesting that the choice of areas selected for the program might have affected the difference in the change in unemployment rate between EZ and comparison tracts. Like the model for poverty, our model showed that the unemployment rate decreased more in some urban EZs but less in others, although the only EZ that experienced a significant change was the Cleveland EZ, which showed a significantly greater decrease in unemployment than the comparison areas. As with poverty rate, local factors may have accounted for the difference between the various urban EZs with respect to the comparison tracts. Table 9: Estimates of the Association between the EZ Program and the Change in Unemployment Rate, 1990-2000: Variables: EZ program; Basic model: Coefficient: -0.065; Basic model: [Empty]; Basic model: Standard error: 0.50; Expanded model: Standard Error: [Empty]. Variables: Atlanta EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 2.56; Expanded model: Standard error: 1.95. Variables: Baltimore EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -0.71; Expanded model: Standard error: 1.77. Variables: Chicago EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -0.68; Expanded model: Standard error: 1.07. Variables: Cleveland EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -3.65; Expanded model: Standard error: 1.40. Variables: Detroit EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 1.76; Expanded model: Standard error: 1.09. Variables: Los Angeles EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 0.092; Expanded model: Standard error: 1.20. Variables: New York EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -1.57; Expanded model: Standard error: 0.88. Variables: Philadelphia-Camden EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -0.87; Expanded model: Standard error: 1.83. Variables: Atlanta EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -2.48; Expanded model: Standard error: 1.71. Variables: Baltimore EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 2.35; Expanded model: Standard error: 1.71. Variables: Chicago EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 2.66; Expanded model: Standard error: 1.57. Variables: Cleveland EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -1.89; Expanded model: Standard error: 1.58. Variables: Detroit EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -2.59; Expanded model: Standard error: 1.55. Variables: Los Angeles EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 3.61; Expanded model: Standard error: 1.62. Variables: New York EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 3.13; Expanded model: Standard error: 1.62. Variables: Philadelphia-Camden EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: [ B]; Expanded model: Standard error: [ B]. Variables: Percent of population of working age[C]; Basic model: Coefficient: 0.068; Basic model: [Empty]; Basic model: Standard error: 0.072; Expanded model: Coefficient: 0.093; Expanded model: Standard error: 0.068. Variables: Percent of population with a high school diploma[D]; Basic model: Coefficient: 0.11; Basic model: [Empty]; Basic model: Standard error: 0.039; Expanded model: Coefficient: 0.20; Expanded model: Standard error: 0.044. Variables: Percent of housing units built between 1990 and 1994[E]; Basic model: Coefficient: -0.18; Basic model: [Empty]; Basic model: Standard error: 0.062; Expanded model: Coefficient: -0.23; Expanded model: Standard error: 0.062. Variables: Percent minority population[F]; Basic model: Coefficient: 0.072; Basic model: [Empty]; Basic model: Standard error: 0.012; Expanded model: Coefficient: 0.054; Expanded model: Standard error: 0.014. Variables: Average household income (in 2004 dollars); Basic model: Coefficient: -0.00017; Basic model: [Empty]; Basic model: Standard error: 0.000068; Expanded model: Coefficient: -0.00032; Expanded model: Standard error: 0.000078. Variables: Population density[G]; Basic model: Coefficient: 0.000032; Basic model: [Empty]; Basic model: Standard error: 0.0000066; Expanded model: Coefficient: 0.0000064; Expanded model: Standard error: 0.000011. Variables: Poverty rate[H]; Basic model: Coefficient: 0.20; Basic model: [Empty]; Basic model: Standard error: 0.033; Expanded model: Coefficient: 0.15; Expanded model: Standard error: 0.034. Variables: Unemployment rate[I]; Basic model: Coefficient: -0.90; Basic model: [Empty]; Basic model: Standard error: 0.039; Expanded model: Coefficient: -0.86; Expanded model: Standard error: 0.044. Variables: Constant; Basic model: Coefficient: -0.88; Basic model: [Empty]; Basic model: Standard error: 4.54; Expanded model: Coefficient: 1.92; Expanded model: Standard error: 4.61. Variables: Number of tracts; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: 866; Expanded model: Standard error: 866. Variables: R-sq; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: 0.53; Expanded model: Standard error: 0.57. Source: GAO analysis of Census data: Notes: Coefficients significant at the 5 percent level are in bold. All variables are from the 1990 Census unless otherwise noted. We weighted the regressions by the geometric mean of 1990 and 2000 household counts of each tract. [A] We defined the EZ area to include both the EZ tracts and comparison tracts that were selected from within a 5-mile boundary of the EZ. [B] Results for the Philadelphia-Camden EZ area are not listed, because we used them as a reference group for the other seven EZs and their surrounding areas. [C] We defined "working age" as between 16 and 64 years of age. [D] Percent based on population 25 years of age and over. [E] From the 2000 Census. [F] For the purposes of this report, we calculated minority population by subtracting the percent of white population from the total population. [G] Individuals per square mile. [H] Percent based on individuals for whom poverty status has been determined. [I] Percent based on individuals 16 years of age or older in the labor force. [End of table] Results of Our Models for Economic Growth: To estimate the statistical relationship between the EZ program and economic growth, we used two proxy measures: (1) the number of businesses excluding establishments that were not eligible for program tax benefits such as nonprofit and governmental organizations and (2) the number of jobs in the EZ. In order to be consistent with our analyses of poverty rate and unemployment, which covered the time period between 1990 and 2000, we used 1995 and 1999 data for our models of economic growth.[Footnote 76] We also tested the model using Home Mortgage Disclosure Act data on the number of loan originations for new home purchases and the mean loan amount for new home purchases as other possible measures of economic growth, but found consistent results, which are not presented here. On the basis of the results of our models, we were not able to determine whether there is a statistical association between the EZ program and economic growth because the explanatory variables we used explained little of the variation in the changes in the number of businesses or jobs between 1995 and 1999 (tables 10 and 11).[Footnote 77] Not surprisingly, most explanatory variables were also not significant. The low explanatory power of our models could be the result of not having considered the right variables; however, we explored many combinations of variables, all of which yielded consistent results. This lack of explanatory power might also be the result of the fact that our proxy measures--the number of businesses and jobs--were not strongly representative of economic growth. Nevertheless, similar to the models of the change in poverty and unemployment, the models of the change in economic growth reflect variation between the EZs with respect to the comparison areas, but none of the results were statistically significant. Table 10: Estimates of the Association between the EZ Program and Economic Growth, Measured by the Change in the Number of Businesses, from 1995-1999A: Variables: EZ program; Basic model: Coefficient: -22.58; Basic model: [Empty]; Basic model: Standard error: 18.84; Expanded model: Standard error: [Empty]. Variables: Atlanta EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 24.59; Expanded model: Standard error: 26.62. Variables: Baltimore EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -0.84; Expanded model: Standard error: 12.12. Variables: Chicago EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 12.80; Expanded model: Standard error: 14.80. Variables: Cleveland EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 4.95; Expanded model: Standard error: 7.05. Variables: Detroit EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 13.68; Expanded model: Standard error: 13.05. Variables: Los Angeles EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -113.70; Expanded model: Standard error: 91.86. Variables: New York EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 10.80; Expanded model: Standard error: 10.92. Variables: Philadelphia-Camden EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 1.03; Expanded model: Standard error: 21.50. Variables: Atlanta EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -12.32; Expanded model: Standard error: 22.13. Variables: Baltimore EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -23.11; Expanded model: Standard error: 17.86. Variables: Chicago EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -17.65; Expanded model: Standard error: 16.25. Variables: Cleveland EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 11.51; Expanded model: Standard error: 29.68. Variables: Detroit EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 8.64; Expanded model: Standard error: 25.45. Variables: Los Angeles EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -38.46; Expanded model: Standard error: 27.71. Variables: New York EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -15.84; Expanded model: Standard error: 20.05. Variables: Philadelphia-Camden EZ area[B]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: [C]; Expanded model: Standard error: [ C]. Variables: Population; Basic model: Coefficient: -0.0027; Basic model: [Empty]; Basic model: Standard error: 0.0026; Expanded model: Coefficient: 0.0011; Expanded model: Standard error: 0.0014. Variables: Percent vacant housing units; Basic model: Coefficient: 0.10; Basic model: [Empty]; Basic model: Standard error: 0.27; Expanded model: Coefficient: -1.09; Expanded model: Standard error: 1.01. Variables: Percent of housing units built between 1990 and 1994[D]; Basic model: Coefficient: 0.95; Basic model: [Empty]; Basic model: Standard error: 0.58; Expanded model: Coefficient: 1.77; Expanded model: Standard error: 1.37. Variables: Percent minority population[E]; Basic model: Coefficient: 0.74; Basic model: [Empty]; Basic model: Standard error: 0.69; Expanded model: Coefficient: 0.80; Expanded model: Standard error: 0.69. Variables: Average household income (in 2004 dollars); Basic model: Coefficient: 0.0044; Basic model: [Empty]; Basic model: Standard error: 0.0051; Expanded model: Coefficient: 0.0056; Expanded model: Standard error: 0.006. Variables: Population density[F]; Basic model: Coefficient: 0.00027; Basic model: [Empty]; Basic model: Standard error: 0.00014; Expanded model: Coefficient: 0.000056; Expanded model: Standard error: 0.00011. Variables: Poverty rate[G]; Basic model: Coefficient: 1.67; Basic model: [Empty]; Basic model: Standard error: 2.21; Expanded model: Coefficient: 2.31; Expanded model: Standard error: 2.61. Variables: Unemployment rate[H]; Basic model: Coefficient: 0.13; Basic model: [Empty]; Basic model: Standard error: 0.35; Expanded model: Coefficient: -0.22; Expanded model: Standard error: 0.59. Variables: Constant; Basic model: Coefficient: -253.01; Basic model: [Empty]; Basic model: Standard error: 278.86; Expanded model: Coefficient: -300.97; Expanded model: Standard error: 314.17. Variables: Number of tracts; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: 860; Expanded model: Standard error: 860. Variables: R-sq; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: 0.042; Expanded model: Standard error: 0.11. Source: GAO analysis of Census and Claritas data. Notes: Coefficients significant at the 5 percent level are in bold. All variables are from the 1990 Census unless otherwise noted. We weighted the regressions by the geometric mean of 1990 and 2000 household counts of each tract. [A] Excluding establishments that were not eligible for the program tax benefits, such as nonprofit and governmental organizations. [B] We defined the EZ area to include both the EZ tracts and comparison tracts that were selected from within a 5-mile boundary of the EZ. [C] Results for the Philadelphia-Camden EZ area are not listed, because we used them as a reference group for the other seven EZs and their surrounding areas. [D] From the 2000 Census. [E] For the purposes of this report, we calculated minority population by subtracting the percent of white population from the total population. [F] Individuals per square mile. [G] Percent based on individuals for whom poverty status has been determined. [H] Percent based on individuals 16 years of age or older in the labor force. [End of table] Table 11: Estimates of the Association between the EZ Program and Economic Growth, Measured by the Change in the Number of Jobs, 1995- 1999: Variables: EZ program; Basic model: Coefficient: -68.86; Basic model: [Empty]; Basic model: Standard error: 196.83; Expanded model: Standard error: [Empty]. Variables: Atlanta EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 286.01; Expanded model: Standard error: 983.37. Variables: Baltimore EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 1199.74; Expanded model: Standard error: 715.74. Variables: Chicago EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -88.43; Expanded model: Standard error: 180.84. Variables: Cleveland EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 102.85; Expanded model: Standard error: 293.30. Variables: Detroit EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 199.16; Expanded model: Standard error: 181.98. Variables: Los Angeles EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -438.63; Expanded model: Standard error: 632.33. Variables: New York EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -288.08; Expanded model: Standard error: 231.26. Variables: Philadelphia-Camden EZ; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 197.87; Expanded model: Standard error: 631.54. Variables: Atlanta EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -318.84; Expanded model: Standard error: 883.10. Variables: Baltimore EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -819.50; Expanded model: Standard error: 810.54. Variables: Chicago EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 36.80; Expanded model: Standard error: 547.93. Variables: Cleveland EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 134.99; Expanded model: Standard error: 557.58. Variables: Detroit EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -61.44; Expanded model: Standard error: 576.67. Variables: Los Angeles EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: -239.53; Expanded model: Standard error: 571.81. Variables: New York EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: 270.91; Expanded model: Standard error: 553.59. Variables: Philadelphia-Camden EZ area[A]; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: [Empty]; Expanded model: Coefficient: [B]; Expanded model: Standard error: [ B]. Variables: Percent of population of working age[D]; Basic model: Coefficient: -21.04; Basic model: [Empty]; Basic model: Standard error: 28.00; Expanded model: Coefficient: -19.59; Expanded model: Standard error: 29.65. Variables: Percent of population with a high school diploma[D]; Basic model: Coefficient: 10.36; Basic model: [Empty]; Basic model: Standard error: 8.39; Expanded model: Coefficient: 4.06; Expanded model: Standard error: 11.77. Variables: Percent of housing units built between 1990 and 1994[E]; Basic model: Coefficient: 10.57; Basic model: [Empty]; Basic model: Standard error: 29.50; Expanded model: Coefficient: 5.09; Expanded model: Standard error: 29.33. Variables: Percent minority population[F]; Basic model: Coefficient: 3.41; Basic model: [Empty]; Basic model: Standard error: 4.63; Expanded model: Coefficient: 3.23; Expanded model: Standard error: 5.36. Variables: Average household income (in 2004 dollars); Basic model: Coefficient: 0.022; Basic model: [Empty]; Basic model: Standard error: 0.038; Expanded model: Coefficient: 0.032; Expanded model: Standard error: 0.048. Variables: Population density[G]; Basic model: Coefficient: 0.0016; Basic model: [Empty]; Basic model: Standard error: 0.0023; Expanded model: Coefficient: -0.0019; Expanded model: Standard error: 0.0029. Variables: Poverty rate[H]; Basic model: Coefficient: -1.30; Basic model: [Empty]; Basic model: Standard error: 20.33; Expanded model: Coefficient: 2.99; Expanded model: Standard error: 20.34. Variables: Unemployment rate[I]; Basic model: Coefficient: 3.10; Basic model: [Empty]; Basic model: Standard error: 15.33; Expanded model: Coefficient: -0.95; Expanded model: Standard error: 14.30. Variables: Constant; Basic model: Coefficient: -56.23; Basic model: [Empty]; Basic model: Standard error: 1483.38; Expanded model: Coefficient: -200.97; Expanded model: Standard error: 1695.03. Variables: Number of tracts; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: 859; Expanded model: Standard error: 859. Variables: R-sq; Basic model: Coefficient: [Empty]; Basic model: [Empty]; Basic model: Standard error: 0.016; Expanded model: Standard error: 0.043. Source: GAO analysis of Census and Claritas data. Notes: Coefficients significant at the 5 percent level are in bold. All variables are from the 1990 Census unless otherwise noted. We weighted the regressions by the geometric mean of 1990 and 2000 household counts of each tract. [A] We defined the EZ area to include both the EZ tracts and comparison tracts that were selected from within a 5-mile boundary of the EZ. [B] Results for the Philadelphia-Camden EZ area are not listed, because we used them as a reference group for the other seven EZs and their surrounding areas. [C] We defined "working age" as between 16 and 64 years of age. [D] Percent based on population 25 years of age and over. [E] From the 2000 Census. [F] For the purposes of this report, we calculated minority population by subtracting the percent of white population from the total population. [G] Individuals per square mile. [H] Percent based on individuals for whom poverty status has been determined. [I] Percent based on individuals 16 years of age or older in the labor force. [End of table] Other Variables Tested for Use in Our Econometric Models: In addition to the variables presented in the models above, we explored many alternative dependent variables and explanatory variables to test the robustness of the models we used (table 12). In particular, we experimented with several alternative measures for economic growth. To test how our results might change in response to the selection of comparison tracts, we also reestimated the models using comparison tracts selected with different propensity scores. We also ran the models excluding the Los Angeles and Cleveland EZs, because these EZs received a slightly different package of benefits when they were initially designated as Supplemental EZs. These tests all yielded results consistent with our models, so they are not presented here. Table 12: Alternative Variables Considered in Our Analyses: Definition of variables: Dependent variables: Change in per-capita income between 1990 and 2000; Rationale: An opposite measure of poverty; Data sources: 1990 and 2000 Census. Definition of variables: Dependent variables: Change in employment rate between 1990 and 2000; Rationale: An opposite measure of unemployment; Data sources: 1990 and 2000 Census. Definition of variables: Dependent variables: Percent change in number of businesses between 1995 and 1999; Rationale: Alternative measure of economic growth; Data sources: Claritas 1995, 1999. Definition of variables: Dependent variables: Percent change in the number of jobs between 1995 and 1999; Rationale: Alternative measure of economic growth; Data sources: Claritas 1995, 1999. Definition of variables: Dependent variables: Percent change in number of businesses between 1995 and 2004; Rationale: Alternative measure of economic growth; Data sources: Claritas 1995, 2004. Definition of variables: Dependent variables: Percent change in the number of jobs between 1995 and 2004; Rationale: Alternative measure of economic growth; Data sources: Claritas 1995, 2004. Definition of variables: Dependent variables: Change in jobs per business between 1995 and 1999; Rationale: Alternative measure of economic growth; Data sources: Claritas 1995, 1999. Definition of variables: Dependent variables: Change in aggregate sales volume of businesses at each tract between 1995 and 1999; Rationale: Alternative measure of economic growth; Data sources: Claritas 1995, 1999. Definition of variables: Dependent variables: Percent change in number of loan originations for new home purchases between 1995 and 1999; Rationale: Alternative measure of economic growth; Data sources: Home Mortgage Disclosure Act data 1995, 1999. Definition of variables: Dependent variables: Percent change in mean loan amount for new home purchases between 1995 and 1999; Rationale: Alternative measure of economic growth; Data sources: Home Mortgage Disclosure Act data 1995, 1999. Definition of variables: Explanatory variables: Percent foreign-born population; Rationale: Alternative indirect measure for minority population; Data sources: 1990 Census. Definition of variables: Explanatory variables: Adjusted per capita income in 2004 dollars; Rationale: Alternative indirect measure household income; Data sources: 1990 Census. Definition of variables: Explanatory variables: Percent of males aged 16 or greater; Rationale: Alternative measure for working population; Data sources: 1990 Census. Definition of variables: Explanatory variables: Percent of housing units built last 5 years before census; Rationale: Alternative measure to account for economic trend; Data sources: 1990 Census. Definition of variables: Explanatory variables: Percent of persons aged 25 or greater with some college; Rationale: Alternative measure for educational level; Data sources: 1990 Census. Definition of variables: Explanatory variables: Percent of employment in manufacturing industry; Rationale: Alternative measure of industry characteristics; Data sources: 1990 Census. Definition of variables: Explanatory variables: Percent of female- headed single households; Rationale: Alternative measure for household characteristics; Data sources: 1990 Census. Source: GAO. [End of table] [End of section] Appendix III: List of Communities Designated in Round I of the EZ/EC Program: Round I:Urban EZs (8): Atlanta, Georgia: Baltimore, Maryland: Chicago, Illinois: Cleveland, Ohio[A]: Detroit, Michigan: Los Angeles, California[A]: New York, New York: Philadelphia, Pennsylvania/Camden, New Jersey: Round I Urban ECs (65): Akron, Ohio: Albany, Georgia: Albany/Schenectady/Troy, New York: Albuquerque, New Mexico: Birmingham, Alabama: Boston, Massachusetts[B]: Bridgeport, Connecticut: Buffalo, New York: Burlington, Vermont: Charleston, South Carolina: Charlotte, North Carolina: Cleveland, Ohio[C]: Columbus, Ohio: Dallas, Texas: Denver, Colorado: Des Moines, Iowa: East St. Louis, Illinois: El Paso, Texas: Flint, Michigan: Harrisburg, Pennsylvania: Houston, Texas[B]: Huntington, West Virginia: Indianapolis, Indiana: Jackson, Mississippi: Kansas City, Missouri/Kansas City, Kansas[B]: Las Vegas, Nevada: Little Rock/Pulaski, Arkansas: Los Angeles, California: Louisville, Kentucky: Lowell, Massachusetts: Manchester, New Hampshire: Memphis, Tennessee: Miami/Dade County, Florida: Milwaukee, Wisconsin: Minneapolis, Minnesota: Muskegon, Michigan: Nashville/Davidson, Tennessee: New Haven, Connecticut: New Orleans, Louisiana: Newark, New Jersey: Newburgh/Kingston, New York: Norfolk, Virginia: Oakland, California[B]: Ogden, Utah: Oklahoma City, Oklahoma: Omaha, Nebraska: Ouachita Parish, Louisiana: Phoenix, Arizona: Pittsburgh, Pennsylvania: Portland, Oregon: Providence, Rhode Island: Rochester, New York: San Antonio, Texas: San Diego, California: San Francisco, California: Seattle, Washington: Springfield, Illinois: Springfield, Massachusetts: St. Louis, Missouri: St. Paul, Minnesota: Tacoma, Washington: Tampa, Florida: Waco, Texas: Washington, District of Columbia: Wilmington, Delaware: Round I Rural EZs (3): Kentucky Highlands, Kentucky: Mid-Delta, Mississippi: Rio Grande Valley, Texas: Round I Rural ECs (30): Accomack and Northampton County, Virginia: Arizona Border Region, Arizona: Beadle/Spink Counties, South Dakota: Central Appalachia, West Virginia: Central Savannah River Area, Georgia: Chambers County, Alabama: City of East Prairie, Missouri: City of Lock Haven, Pennsylvania: City of Watsonville, California: Crisp/ Dooly County, Georgia: East Arkansas, Arkansas: Fayette/Haywood County, Tennessee: Greater Portsmouth, Ohio: Greene-Sumter, Alabama: The Halifax/ Edgecombe/Wilson Empowerment Alliance, North Carolina: Imperial County, California: Jackson County, Florida: Josephine County, Oregon: La Jicarita, New Mexico: Lake County, Michigan: Lower Yakima County, Washington: Macon Ridge, Louisiana: McDowell County, West Virginia: Mississippi County, Arkansas: North Delta Mississippi, Mississippi: Northeast Louisiana Delta, Louisiana: Robeson County, North Carolina: Scott, Tennessee/McCreary, Kentucky: Southeast Oklahoma, Oklahoma: Williamsburg-Lake City, South Carolina: [A] Initially designated as a Supplemental EZ: [B] Also designated as an Enhanced EC: [C] Also designated as a Supplemental EZ: Source: HUD and USDA data. [End of section] Appendix IV: Description of the Empowerment Zones and Enterprise Communities We Visited: This appendix contains detailed information we gathered from our site visits to the 11 Round I EZs and 2 ECs. The appendix describes how the EZs and ECs were governed; the activities they implemented; changes in poverty, unemployment, and economic growth; and stakeholders' perceptions of factors influencing those changes. It also includes the percent changes in variables used in the econometric model. Atlanta Empowerment Zone: Figure 13: Map of the Atlanta EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] How the EZ Was Governed: The city of Atlanta established the nonprofit Atlanta Empowerment Zone Corporation to operate the EZ. The corporation had two boards: the Executive Board and the Community Empowerment Advisory Board, which included representatives of each of the EZ neighborhoods. According to EZ stakeholders we interviewed, the EZ Executive Board gave final approval on activities the EZ implemented. However, EZ stakeholders also mentioned that the intended process was not always followed and that the board was not always able to approve activities due to difficulties reaching consensus. Activities the EZ Implemented: According to HUD data, most of the Atlanta EZ's activities related to community development, but the EZ also implemented some activities related to economic opportunity, such as making loans to EZ businesses. Initiatives involving housing, public safety, and assistance to businesses were the most frequently implemented types of activities (fig. 14). In our interviews, EZ stakeholders mentioned initiatives they saw as particularly useful, including housing programs for seniors and low-income EZ residents--for example, a program that helped to repair code violations in homes of senior citizens. Stakeholders also said that the EZ provided funds to after-school and health-related programs, such as one that provided children and adults with asthma with needed resources and education. Some EZ stakeholders suggested that the loan program lacked positive results, because many of the businesses that received the loans failed.[Footnote 78] Figure 14: Activities Implemented by the Atlanta EZ: [See PDF for image] Sources: GAO (photos); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Poverty declined in the Atlanta EZ, but unemployment did not, and measures of economic growth did not show improvement. Atlanta had the highest poverty rate of any EZ in 1990 (55 percent). By 2000, this rate had fallen by around 10 percentage points, while the rate of its comparison area remained the same. Conversely, the unemployment rate went from one of the lowest of the urban EZs in 1990 to one of the highest in 2000, and the increase was greater than in its comparison area. Similarly, the Atlanta EZ and its comparison area experienced a large decline--more than 20 percent--in total number of businesses from 1995 to 2004. The Atlanta EZ had the second largest decline in the number of jobs of any EZ, which was also more than in its comparison area. Tables 13 and 14 show the changes in poverty, unemployment, and economic growth in the EZ and its comparison area. Table 13 also includes data on the changes in other variables included in our models. Table 13: Changes in Selected Census Variables Observed in the Atlanta EZ and Its Comparison Area: Poverty rate (%); 1990: EZ: 54.67; 1990: Comparison: 30.15; 2000: EZ: 44.82; 2000: Comparison: 28.02; Percent change[A]: EZ: -9.84[B]; Percent change[A]: Comparison: -2.12. Unemployment rate (%); 1990: EZ: 17.48; 1990: Comparison: 11.36; 2000: EZ: 23.44; 2000: Comparison: 11.88; Percent change[A]: EZ: 5.96[B]; Percent change[A]: Comparison: 0.52. Average household income; 1990: EZ: $18,343; 1990: Comparison: $30,567; 2000: EZ: $28,552; 2000: Comparison: $39,500; Percent change[A]: EZ: 55.66[B]; Percent change[A]: Comparison: 29.23[B]. Percentage of single female headed households with children; 1990: EZ: 24.62; 1990: Comparison: 20.02; 2000: EZ: 21.26; 2000: Comparison: 19.95; Percent change[A]: EZ: -3.36[B]; Percent change[A]: Comparison: -0.07. Total population; 1990: EZ: 49,966; 1990: Comparison: 65,809; 2000: EZ: 45,931; 2000: Comparison: 64,022; Percent change[A]: EZ: -8.07; Percent change[A]: Comparison: - 2.71. Total individuals per square mile; 1990: EZ: 5,408; 1990: Comparison: 2,671; 2000: EZ: 4,972; 2000: Comparison: 2,756; Percent change[A]: EZ: -8.07; Percent change[A]: Comparison: 3.16. Percentage of households that moved in the last 5 years; 1990: EZ: 50.87; 1990: Comparison: 46.01; 2000: EZ: 53.32; 2000: Comparison: 52.52; Percent change[A]: EZ: 2.45[B]; Percent change[A]: Comparison: 6.51[B]. Percentage of population of working age (16-64); 1990: EZ: 60.16; 1990: Comparison: 61.68; 2000: EZ: 63.42; 2000: Comparison: 64.59; Percent change[A]: EZ: 3.26[B]; Percent change[A]: Comparison: 2.92[B]. Percentage of population with a high school diploma (or equivalent); 1990: EZ: 43.10; 1990: Comparison: 60.53; 2000: EZ: 58.96; 2000: Comparison: 69.3; Percent change[A]: EZ: 15.86[B]; Percent change[A]: Comparison: 8.78[B]. Percentage of high school dropouts; 1990: EZ: 19.12; 1990: Comparison: 19.1; 2000: EZ: 21.48; 2000: Comparison: 21.19; Percent change[A]: EZ: 2.36[B]; Percent change[A]: Comparison: 2.08[B]. Percentage of vacant housing units; 1990: EZ: 20.79; 1990: Comparison: 14.65; 2000: EZ: 13.30; 2000: Comparison: 7.43; Percent change[A]: EZ: -7.48[B]; Percent change[A]: Comparison: -7.22[B]. Average owner occupied housing value; 1990: EZ: $55,883; 1990: Comparison: $74,063; 2000: EZ: $117,869; 2000: Comparison: $101,774; Percent change[A]: EZ: 110.92[B]; Percent change[A]: Comparison: 37.42[B]. Source: GAO analysis of Census data. Note: There are 23 census tracts in the designated area and 16 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 14: Changes in Selected Economic Growth Variables Observed in the Atlanta EZ and Its Comparison Area: Number of businesses; 1995: EZ: 1,930; 1995: Comparison: 3,980; 1999: EZ: 1,549; 1999: Comparison: 3,380; 2004: EZ: 1,529; 2004: Comparison: 3,248; Percent change 1995-2004[A]: EZ: -20.78; Percent change 1995-2004[A]: Comparison: -18.39. Number of jobs; 1995: EZ: 36,888; 1995: Comparison: 71,346; 1999: EZ: 31,470; 1999: Comparison: 79,580; 2004: EZ: 28,672; 2004: Comparison: 69,140; Percent change 1995-2004[A]: EZ: - 22.27; Percent change 1995-2004[A]: Comparison: -3.09. Source: GAO analysis of Claritas data. Note: There are 23 census tracts in the designated area and 16 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, stakeholders said that changes in the poverty rate may have been due to changes in the EZ population and the demolition of public housing. They explained that residents with lower incomes had left the EZ and that households with higher incomes were moving in because of changes in the EZ as a result of development from the Olympics and the demolition of public housing through the HOPE VI program.[Footnote 79] Commenting on unemployment, stakeholders suggested that EZ residents had benefited from EZ job training and placement programs but that a mismatch still existed between residents' skills and some of the new jobs available in the EZ. Although our economic growth data suggested a decrease in the number of businesses and number of jobs, stakeholders suggested that the EZ had helped to foster economic growth in some of the commercial corridors by helping to fund neighborhood plans. Two stakeholders also mentioned the 1996 Olympics as a factor in bringing jobs and development to the EZ and the city of Atlanta, although one stakeholder noted that several businesses had closed down after the Olympics. This loss of businesses potentially helps explain the significant decrease in the number of businesses between 1995 and 2004. Baltimore Empowerment Zone: Figure 15: Map of the Baltimore EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] How the EZ Was Governed: The nonprofit Empower Baltimore Management Corporation was created specifically to manage Baltimore's EZ program. The EZ was governed by a board composed of community leaders, three committees (one for each core strategic goal), an executive committee of the three committee chairs, and an advisory council of individuals from all areas of the EZ. Governance of the Baltimore EZ also included six "Village Centers"- -community groups that applied to be the implementing agencies of EZ programs in their local communities. EZ activities were vetted through the advisory council and sent to the executive committee and full board for final approval. Activities the EZ Implemented: Unlike most EZs, the Baltimore EZ implemented a higher number of economic opportunity activities than community development activities. The three types of activities most often implemented were workforce development, access to capital, and assistance to businesses (fig. 16). Most stakeholders described the EZ's workforce training activities, such as the customized training program that provided EZ residents with individualized instruction and a stipend during the training period. In addition, the EZ operated several loan funds and partially funded the Bank One check processing center and the Montgomery Park business incubator, two business developments. The EZ also ran a lead paint abatement program and a homeownership program. The Baltimore EZ received a grant extension through June 2006. Figure 16: Activities Implemented by the Baltimore EZ: [See PDF for image] Source: GAO (photo); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Poverty decreased in the Baltimore EZ and economic growth improved somewhat, but unemployment stayed the same. The poverty rate in the EZ fell between 1990 and 2000, while its comparison area stayed about the same. However, the unemployment rate, which was one of the lowest of the urban EZs in 1990, stayed the same between 1990 and 2000, while the rate in its comparison area increased. In terms of economic growth, the results were mixed, with the EZ doing somewhat better than its comparison area. The number of businesses in the EZ fell from 1995 to 2004, but the number of jobs increased. In its comparison area, the number of businesses also fell, but the number of jobs fell substantially. Tables 15 and 16 show the changes in poverty, unemployment, and economic growth in the EZ and its comparison area. Table 15 also includes data on the changes in other variables included in our models. Table 15: Changes in Selected Census Variables Observed in the Baltimore EZ and Its Comparison Area: Poverty rate (%); 1990: EZ: 41.81; 1990: Comparison: 41.17; 2000: EZ: 35.66; 2000: Comparison: 39.74; Percent change[A]: EZ: - 6.16[ B]; Percent change[A]: Comparison: -1.43. Unemployment rate (%); 1990: EZ: 15.00; 1990: Comparison: 14.55; 2000: EZ: 16.48; 2000: Comparison: 17.58; Percent change[A]: EZ: 1.49; Percent change[A]: Comparison: 3.03[ B]. Average household income; 1990: EZ: $28,185; 1990: Comparison: $27,931; 2000: EZ: $35,059; 2000: Comparison: $31,367; Percent change[A]: EZ: 24.39[B]; Percent change[A]: Comparison: 12.30[B]. Percentage of single female headed households with children; 1990: EZ: 22.50; 1990: Comparison: 23.15; 2000: EZ: 19.49; 2000: Comparison: 19.64; Percent change[A]: EZ: -3.01[ B]; Percent change[A]: Comparison: -3.51[B]. Total population; 1990: EZ: 72,725; 1990: Comparison: 150,507; 2000: EZ: 54,657; 2000: Comparison: 113,052; Percent change[A]: EZ: -24.84; Percent change[A]: Comparison: -24.89. Total individuals per square mile; 1990: EZ: 10,460; 1990: Comparison: 16,934; 2000: EZ: 7,890; 2000: Comparison: 12,923; Percent change[A]: EZ: -24.57; Percent change[A]: Comparison: -23.69. Percentage of households that moved in the last 5 years; 1990: EZ: 41.07; 1990: Comparison: 42.98; 2000: EZ: 41.00; 2000: Comparison: 44.90; Percent change[A]: EZ: -0.07; Percent change[A]: Comparison: 1.92[B]. Percentage of population of working age (16-64); 1990: EZ: 58.55; 1990: Comparison: 60.31; 2000: EZ: 60.04; 2000: Comparison: 61.63; Percent change[A]: EZ: 1.48; Percent change[A]: Comparison: 1.32. Percentage of population with a high school diploma (or equivalent); 1990: EZ: 45.69; 1990: Comparison: 49.86; 2000: EZ: 56.44; 2000: Comparison: 58.50; Percent change[A]: EZ: 10.74[B]; Percent change[A]: Comparison: 8.64[B]. Percentage of high school dropouts; 1990: EZ: 32.36; 1990: Comparison: 26.43; 2000: EZ: 19.55; 2000: Comparison: 20.61; Percent change[A]: EZ: -12.81[ B]; Percent change[A]: Comparison: -5.81[B]. Percentage of vacant housing units; 1990: EZ: 17.59; 1990: Comparison: 12.67; 2000: EZ: 26.22; 2000: Comparison: 23.63; Percent change[A]: EZ: 8.62[B]; Percent change[A]: Comparison: 10.96[B]. Average owner occupied housing value; 1990: EZ: $53,714; 1990: Comparison: $55,966; 2000: EZ: $62,219; 2000: Comparison: $62,514; Percent change[A]: EZ: 15.83[B]; Percent change[A]: Comparison: 11.7[B]. Source: GAO analysis of Census data. Note: There are 25 census tracts in the designated area and 41 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 16: Changes in Selected Economic Growth Variables Observed in the Baltimore EZ and Its Comparison Area: Number of businesses; 1995: EZ: 2,797; 1995: Comparison: 3,481; 1999: EZ: 2,399; 1999: Comparison: 2,930; 2004: EZ: 2,487; 2004: Comparison: 3,005; Percent change 1995-2004[A]: EZ: -11.08; Percent change 1995-2004[A]: Comparison: -13.67. Number of jobs; 1995: EZ: 41,837; 1995: Comparison: 61,519; 1999: EZ: 53,732; 1999: Comparison: 35,268; 2004: EZ: 47,504; 2004: Comparison: 36,860; Percent change 1995-2004[A]: EZ: 13.55; Percent change 1995-2004[A]: Comparison: -40.08. Source: GAO analysis of Claritas data. Note: There are 25 census tracts in the designated area and 41 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, several stakeholders from the Baltimore EZ said that changes in the population of the zone had influenced the change in poverty rate. They said that a local HOPE VI project had relocated many of the original EZ residents and that rising property values may have caused some original residents to move out of the zone. Four stakeholders also mentioned lower crime rates in the EZ, which three of them linked to the decrease in poverty. Two stakeholders mentioned trends in the national economy that influenced the change in unemployment, and some said that population changes in the zone had affected unemployment as well as poverty. Stakeholders cited both EZ-related and external factors as affecting economic growth. For example, some said that the EZ created economic growth with its entrepreneurial programs, loan funds, and businesses developments, such as the Montgomery Park business incubator. Stakeholders offered mixed perceptions on the impact of the EZ tax benefits on economic growth. Some believed that tax benefits were helpful to economic growth, while others did not. In addition, one stakeholder said that the waterfront area of the EZ was a natural place for development and that the designated area might have experienced economic growth in the absence of the program. Chicago Empowerment Zone: Figure 17: Map of the Chicago EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] How the EZ Was Governed: The city of Chicago operated its EZ program, incorporating an EZ coordinating council and advisory subgroups called "community clusters." Both the coordinating council and community clusters were made up of EZ residents and local officials. All proposals for EZ activities were submitted through a request-for-proposal process, made available for comment by the coordinating council, and were reviewed and approved by the Chicago City Council. Activities the EZ Implemented: The Chicago EZ implemented more community development than economic opportunity activities. The activities it implemented most often were related to workforce development, education, and human services, and stakeholders said that the EZ was also active in the area of housing development (fig. 18). EZ stakeholders also noted that the EZ had helped to improve health care for individuals without insurance by contributing to the renovation or expansion of local medical facilities. In addition, businesses in the Chicago EZ used six program tax-exempt bonds. The Chicago EZ received a grant extension through 2009. Figure 18: Activities Implemented by the Chicago EZ: [See PDF for image] Sources: GAO (photo); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Our analyses showed improvements in the Chicago EZ in the poverty and unemployment rates, but not in economic growth. Both the EZ and its comparison area saw a decrease in poverty from 1990 to 2000. The EZ also experienced a decrease in unemployment that was considerably greater than that of its comparison area in that time period. In terms of economic growth, the Chicago EZ and its comparison area saw decreases in the numbers of businesses and jobs between 1995 and 2004, with the EZ seeing a larger decline in the number of jobs but less of a decline in the number of businesses than its comparison area. Tables 17 and 18 show the changes in poverty, unemployment, and economic growth in the EZ and its comparison area. Table 17 also includes data on the changes in other variables included in our models. Table 17: Changes in Selected Census Variables Observed in the Chicago EZ and Its Comparison Area: Poverty rate (%); 1990: EZ: 49.10; 1990: Comparison: 40.38; 2000: EZ: 39.32; 2000: Comparison: 33.49; Percent change[A]: EZ: - 9.77[ B]; Percent change[A]: Comparison: -6.89[B]. Unemployment rate (%); 1990: EZ: 24.57; 1990: Comparison: 20.52; 2000: EZ: 19.34; 2000: Comparison: 18.97; Percent change[A]: EZ: -5.23[ B]; Percent change[A]: Comparison: -1.54[B]. Average household income; 1990: EZ: $23,097; 1990: Comparison: $28,431; 2000: EZ: $34,718; 2000: Comparison: $39,985; Percent change[A]: EZ: 50.31[B]; Percent change[A]: Comparison: 40.64[B]. Percentage of single female headed households with children; 1990: EZ: 25.64; 1990: Comparison: 23.07; 2000: EZ: 21.59; 2000: Comparison: 19.69; Percent change[A]: EZ: -4.05[B]; Percent change[A]: Comparison: -3.38[B]. Total population; 1990: EZ: 200,182; 1990: Comparison: 377,580; 2000: EZ: 177,309; 2000: Comparison: 369,343; Percent change[A]: EZ: -11.43; Percent change[A]: Comparison: -2.18. Total individuals per square mile; 1990: EZ: 13,967; 1990: Comparison: 15,523; 2000: EZ: 12,380; 2000: Comparison: 15,752; Percent change[A]: EZ: -11.36; Percent change[A]: Comparison: 1.47. Percentage of households that moved in the last 5 years; 1990: EZ: 37.52; 1990: Comparison: 40.03; 2000: EZ: 39.21; 2000: Comparison: 39.68; Percent change[A]: EZ: 1.69[B]; Percent change[A]: Comparison: -0.35. Percentage of population of working age (16-64); 1990: EZ: 53.51; 1990: Comparison: 57.87; 2000: EZ: 55.63; 2000: Comparison: 59.53; Percent change[A]: EZ: 2.12[B]; Percent change[A]: Comparison: 1.66[B]. Percentage of population with a high school diploma (or equivalent); 1990: EZ: 44.04; 1990: Comparison: 54.24; 2000: EZ: 54.30; 2000: Comparison: 63.58; Percent change[A]: EZ: 10.26[B]; Percent change[A]: Comparison: 9.35[B]. Percentage of high school dropouts; 1990: EZ: 22.46; 1990: Comparison: 19.50; 2000: EZ: 22.05; 2000: Comparison: 15.23; Percent change[A]: EZ: -0.41; Percent change[A]: Comparison: -4.27[ B]. Percentage of vacant housing units; 1990: EZ: 19.69; 1990: Comparison: 13.54; 2000: EZ: 18.23; 2000: Comparison: 12.44; Percent change[A]: EZ: -1.45[B]; Percent change[A]: Comparison: -1.1[B]. Average owner occupied housing value; 1990: EZ: $71,429; 1990: Comparison: $88,445; 2000: EZ: $160,412; 2000: Comparison: $167,015; Percent change[A]: EZ: 124.57[B]; Percent change[A]: Comparison: 88.83[B]. Source: GAO analysis of Census data. Note: There are 96 census tracts in the designated area and 146 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 18: Changes in Selected Economic Growth Variables Observed in the Chicago EZ and Its Comparison Area: Number of businesses; 1995: EZ: 5,089; 1995: Comparison: 10,567; 1999: EZ: 4,614; 1999: Comparison: 9,582; 2004: EZ: 4,496; 2004: Comparison: 9,211; Percent change 1995-2004[A]: EZ: -11.65; Percent change 1995-2004[A]: Comparison: -12.83. Number of jobs; 1995: EZ: 83,935; 1995: Comparison: 183,369; 1999: EZ: 80,294; 1999: Comparison: 169,741; 2004: EZ: 69,767; 2004: Comparison: 162,541; Percent change 1995-2004[A]: EZ: - 16.88; Percent change 1995-2004[A]: Comparison: -11.36. Source: GAO analysis of Claritas data. Note: There are 96 census tracts in the designated area and 146 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: Asked about factors influencing the change in poverty, stakeholders pointed to both EZ activities and external factors. Among the EZ activities they cited were projects promoting homeownership or providing educational training. However, some stakeholders mentioned changes in the EZ population as an external factor that may have affected the changes in poverty, noting the demolition of several public housing buildings in the EZ and the addition of individuals with higher incomes moving into new housing built on those sites. Some stakeholders attributed a decrease in unemployment to the zone's focus on creating jobs and the requirement that subgrantees demonstrate that they had created jobs for EZ residents. Some stakeholders also noted that the EZ's provision of services, such as childcare, after- school programs, and job training, provided opportunities for more residents to obtain jobs. But some stakeholders believed that the decreases in unemployment were due to external economic forces, such as changes in the population of the EZ and more jobs available due to changes in the national economy. In terms of economic growth, some EZ stakeholders observed that the EZ had provided some initial investment in the zone and that private investment had followed. However, some stakeholders noted that the EZ had not done enough in the area of economic development. In addition, stakeholders said that not all the jobs from new businesses in the zone had gone to zone residents and that the number of new businesses did not meet the zone's employment needs. Detroit Empowerment Zone: Figure 19: Map of the Detroit EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] How the EZ Was Governed: The nonprofit Detroit Empowerment Zone Development Corporation ran the Detroit EZ and included an executive committee, a board made up of residents and other local officials, and three neighborhood review panels representing neighborhoods in the EZ. Each review panel had an advisory role in determining how a portion of the EZ funds would be spent. The EZ was required to obtain the approval of the Detroit City Council and Mayor for many EZ-funded activities. Activities the EZ Implemented: The Detroit EZ implemented mostly community development activities. The two most common types of activities were in the areas of human services and education (fig. 20). In addition, EZ stakeholders explained that the EZ had helped to spur housing development in the east and southwest areas of the zone by providing funds to community development corporations. Detroit EZ stakeholders also highlighted a business façade improvement program during our tour of the EZ. Although they focused mainly on community development, the Detroit EZ did implement some economic opportunity activities. Some EZ stakeholders said that the Financial Institutions Consortium, which set lending goals within the EZ, had helped EZ businesses. The Detroit EZ did not request a grant extension because it had used nearly all of the EZ grant funds. Figure 20: Activities Implemented by the Detroit EZ: [See PDF for image] Source: GAO (photo); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: The Detroit EZ experienced positive changes in poverty, unemployment, and one measure of economic growth. Of the urban EZs, the Detroit EZ had the largest decrease in poverty and the second largest decrease in unemployment from 1990 to 2000. Although the decrease in the EZ was slightly greater than in its comparison area in poverty, the decrease in unemployment was less than in its comparison area. Between 1995 and 2004, the EZ generally fared better than its comparison area in our measures of economic growth; however, the changes were not always positive. The number of businesses declined slightly, but the decrease was notably smaller than the decline in its comparison area. In addition, the EZ saw a greater increase in the number of jobs than in either its comparison area or most urban EZs. Tables 19 and 20 show the changes in poverty, unemployment, and economic growth in the EZ and its comparison area. Table 19 also includes data on the changes in other variables included in our models. Table 19: Changes in Selected Census Variables Observed in the Detroit EZ and Its Comparison Area: Poverty rate (%); 1990: EZ: 47.63; 1990: Comparison: 42.72; 2000: EZ: 36.73; 2000: Comparison: 32.38; Percent change[A]: EZ: -10.90[B]; Percent change[A]: Comparison: -10.34[B]. Unemployment rate (%); 1990: EZ: 28.41; 1990: Comparison: 26.01; 2000: EZ: 18.83; 2000: Comparison: 15.54; Percent change[A]: EZ: -9.58[ B]; Percent change[A]: Comparison: -10.47[ B]. Average household income; 1990: EZ: $22,644; 1990: Comparison: $25,609; 2000: EZ: $33,751; 2000: Comparison: $36,200; Percent change[A]: EZ: 49.05[B]; Percent change[A]: Comparison: 41.36[B]. Percentage of single female headed households with children; 1990: EZ: 17.30; 1990: Comparison: 20.94; 2000: EZ: 15.88; 2000: Comparison: 18.77; Percent change[A]: EZ: -1.43; Percent change[A]: Comparison: -2.18[B]. Total population; 1990: EZ: 103,346; 1990: Comparison: 256,371; 2000: EZ: 88,707; 2000: Comparison: 229,536; Percent change[A]: EZ: -14.16; Percent change[A]: Comparison: -10.47. Total individuals per square mile; 1990: EZ: 5,547; 1990: Comparison: 6,923; 2000: EZ: 4,762; 2000: Comparison: 6,200; Percent change[A]: EZ: -14.15; Percent change[A]: Comparison: -10.45. Percentage of households that moved in the last 5 years; 1990: EZ: 40.80; 1990: Comparison: 37.39; 2000: EZ: 42.20; 2000: Comparison: 38.47; Percent change[A]: EZ: 1.40; Percent change[A]: Comparison: 1.08. Percentage of population of working age (16-64); 1990: EZ: 57.65; 1990: Comparison: 56.96; 2000: EZ: 60.34; 2000: Comparison: 57.01; Percent change[A]: EZ: 2.69[B]; Percent change[A]: Comparison: 0.05. Percentage of population with a high school diploma (or equivalent); 1990: EZ: 49.34; 1990: Comparison: 53.63; 2000: EZ: 58.06; 2000: Comparison: 60.87; Percent change[A]: EZ: 8.72[B]; Percent change[A]: Comparison: 7.24[B]. Percentage of high school dropouts; 1990: EZ: 23.29; 1990: Comparison: 20.47; 2000: EZ: 20.83; 2000: Comparison: 18.89; Percent change[A]: EZ: -2.46[B]; Percent change[A]: Comparison: -1.57[B]. Percentage of vacant housing units; 1990: EZ: 18.26; 1990: Comparison: 11.22; 2000: EZ: 17.46; 2000: Comparison: 13.98; Percent change[A]: EZ: -0.79; Percent change[A]: Comparison: 2.76[B]. Average owner occupied housing value; 1990: EZ: $23,114; 1990: Comparison: $28,598; 2000: EZ: $52,234; 2000: Comparison: $61,160; Percent change[A]: EZ: 125.99[B]; Percent change[A]: Comparison: 113.86[B]. Source: GAO analysis of Census data. Note: There are 49 census tracts in the designated area and 86 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 20: Changes in Selected Economic Growth Variables Observed in the Detroit EZ and Its Comparison Area: Number of businesses; 1995: EZ: 3,723; 1995: Comparison: 5,343; 1999: EZ: 3,650; 1999: Comparison: 5,282; 2004: EZ: 3,621; 2004: Comparison: 4,770; Percent change 1995-2004[A]: EZ: -2.74; Percent change 1995-2004[A]: Comparison: -10.72. Number of jobs; 1995: EZ: 95,172; 1995: Comparison: 86,500; 1999: EZ: 99,480; 1999: Comparison: 73,770; 2004: EZ: 124,172; 2004: Comparison: 66,179; Percent change 1995-2004[A]: EZ: 30.47; Percent change 1995-2004[A]: Comparison: -23.49. Source: GAO analysis of Claritas data. Note: There are 49 census tracts in the designated area and 86 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: An EZ stakeholder said that the population of the zone had changed, possibly affecting the changes in the poverty rate. For example, the stakeholder noted that many of the initial EZ residents had moved out of the zone since designation, and that other individuals with higher incomes had moved into the zone. Stakeholders noted that EZ programs in job training, youth education, supportive services, and health care had helped some EZ residents to gain employment. However, some stakeholders also mentioned the lack of a skilled workforce in the EZ and the need for more job training. In addition, some stakeholders thought that the changes in the zone's population might also have influenced the change in unemployment. In terms of economic growth, EZ stakeholders noted that their façade improvement program had contributed to business growth in the EZ. One stakeholder also suggested that the EZ tax benefits and financing from the Financial Institutions Consortium had provided incentives to attract businesses to locate in the EZ. Another stakeholder mentioned external challenges to economic growth that included the loss of the automobile industry and the poor national economy over the time period of the EZ. New York Empowerment Zone: Figure 21: Map of the New York EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] How the EZ Was Governed: The New York EZ was governed by three boards: an overarching board and two subzone boards representing the Upper Manhattan and Bronx neighborhoods. The overarching board, which included officials from the city, state, and each subzone board, as well as local congressional representatives, provided final funding approval for all EZ projects.[Footnote 80] However, the program was managed locally by the two subzones, which had separate management organizations, boards, and budgets and made decisions about the activities that would be funded and the organizations that would implement them. The Upper Manhattan subzone received the bulk of the EZ grant ($83 million), and the Bronx portion received the remaining $17 million. The New York EZ also received matching funds from the city and state, bringing the total funding for the EZ to $300 million. The EZ created the nonprofit Upper Manhattan Empowerment Zone to manage the Upper Manhattan portion of the zone. This EZ is governed by a board that includes community members, at-large members selected for their expertise, and representatives from city community planning boards. The board also has seven committees. Activities proposed in this portion of the EZ were reviewed by the committees, approved by the Upper Manhattan board, and finally approval by the overarching EZ board. The Bronx Overall Economic Development Corporation, a part of the Bronx Borough President's office, managed the Bronx portion of the New York EZ. The board of the Bronx Overall Economic Development Corporation covered both EZ and non-EZ activities but included an EZ committee. Although the board did not include any EZ residents, an EZ stakeholder explained that it included some residents of other areas of the Bronx. In general, the board decided on activities, encouraged local nonprofits to submit proposals, and chose the organizations to implement the activities. Then the activities went before the New York EZ board for final approval. Activities the EZ Implemented: Unlike most EZs, both portions of the New York EZ implemented more economic opportunity activities than community development activities. However, the Upper Manhattan and Bronx portions of the EZ differed somewhat in the types of activities they implemented. The New York EZ as a whole received a grant extension until 2009. The types of activities most commonly implemented by the Upper Manhattan portion were assistance to businesses, workforce development, access to capital, and infrastructure (fig. 22). Several EZ stakeholders mentioned the business developments in Harlem USA or along 125TH Street as major accomplishments of their program. Stakeholders also noted that the EZ had assisted small businesses, successfully sponsored a restaurant initiative that provided local restaurants with loan capital and technical assistance, and facilitated the use of an EZ tax-exempt bond to finance a new car dealership. In addition, an EZ stakeholder said that the EZ fostered job growth by requiring recipients of EZ grants and loans to employ a certain number of EZ residents. Figure 22: Activities Implemented by the Upper Manhattan portion of the New York EZ: [See PDF for image] Source: GAO (photo); GAO analysis of HUD data(charts). [End of figure] The types of activities most commonly implemented by the Bronx portion of the New York EZ included workforce development, education, access to capital, and human services (fig. 23). EZ stakeholders explained that the EZ had funded several workforce training activities, such as a program to train women to become childcare providers. However, several stakeholders also said that as the program progressed more funds were used to provide loans to EZ businesses, an activity that was felt to provide the best return on investment. Figure 23: Activities Implemented by the Bronx portion of the New York EZ: [See PDF for image] Sources: GAO (photos); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Overall, the New York EZ saw poverty fall and economic growth improve, but unemployment increase. The changes in the Upper Manhattan portion of the EZ followed this pattern, but the Bronx portion of the EZ also showed a decrease in one of the measures of economic growth, the number of total jobs. Tables 21 and 22 show the changes in poverty, unemployment, and economic growth in the EZ, the Upper Manhattan and Bronx portions of the EZ, and the EZ comparison area. Table 21 also includes data on the changes in other variables included in our models. Indicators for the Upper Manhattan portion of the New York EZ were mixed compared with the New York comparison area. The poverty rate in the Upper Manhattan portion of the EZ fell between 1990 and 2000, while the unemployment rate stayed the same. The New York comparison area stayed about the same in poverty, and its unemployment rate rose. In economic growth, between 1995 and 2004 the Upper Manhattan portion of the EZ had the largest increase in total number of businesses and the second-largest increase in jobs of any urban EZ. The comparison area saw a slightly smaller increase in businesses and a larger increase in jobs. Like the Upper Manhattan portion of the New York EZ, the Bronx portion showed mixed results relative to the New York comparison area. Its poverty rate stayed the same between 1990 and 2000 as did the New York comparison area. Between 1990 and 2000, it experienced a greater increase in unemployment than the New York comparison area. In terms of economic growth, the area did show an increase in the number of businesses from 1995 to 2004, but its comparison area showed a larger increase. However, in the same time period, the Bronx experienced a slight decrease in the number of jobs, while the comparison area experienced a large increase. Table 21: Changes in Selected Census Variables Observed in the New York EZ, the Bronx and Upper Manhattan (UM) Portions, and the EZ Comparison Area (Comp.) Poverty rate (%); 1990: Entire EZ: 42.68; 1990: Bronx: 44.2; 1990: UM: 42.38; 1990: Comp.: 42.5; 2000: Entire EZ: 38.62; 2000: Bronx: 41.59; 2000: UM: 38.02; 2000: Comp.: 41.75; Percent change[A]: Entire EZ: -4.07[B]; Percent change[A]: Bronx: -2.61; Percent change[A]: UM: -4.35[B]; Percent change[A]: Comp.: -0.75. Unemployment rate (%); 1990: Entire EZ: 17.45; 1990: Bronx: 15.36; 1990: UM: 17.86; 1990: Comp.: 17.17; 2000: Entire EZ: 19.46; 2000: Bronx: 20.96; 2000: UM: 19.18; 2000: Comp.: 20.04; Percent change[A]: Entire EZ: 2.00[B]; Percent change[A]: Bronx: 5.6[B]; Percent change[A]: UM: 1.32; Percent change[A]: Comp.: 2.87[B]. Average household income; 1990: Entire EZ: $26,518; 1990: Bronx: $26,294; 1990: UM: $26,559; 1990: Comp.: $26,993; 2000: Entire EZ: $33,557; 2000: Bronx: $30,842; 2000: UM: $34,041; 2000: Comp.: $31,247; Percent change[A]: Entire EZ: 26.54[B]; Percent change[A]: Bronx: 17.29[B]; Percent change[A]: UM: 28.17[B]; Percent change[A]: Comp.: 15.76[B]. Percentage of single female headed households with children; 1990: Entire EZ: 20.19; 1990: Bronx: 25.55; 1990: UM: 19.2; 1990: Comp.: 25.91; 2000: Entire EZ: 19.6; 2000: Bronx: 23.15; 2000: UM: 18.97; 2000: Comp.: 25.26; Percent change[A]: Entire EZ: -0.59; Percent change[A]: Bronx: -2.40[B]; Percent change[A]: UM: -0.23; Percent change[A]: Comp.: -0.64. Total population; 1990: Entire EZ: 199,983; 1990: Bronx: 34,266; 1990: UM: 165,717; 1990: Comp.: 638,776; 2000: Entire EZ: 219,324; 2000: Bronx: 36,886; 2000: UM: 182,438; 2000: Comp.: 672,826; Percent change[A]: Entire EZ: 9.67; Percent change[A]: Bronx: 7.65; Percent change[A]: UM: 10.09; Percent change[A]: Comp.: 5.33. Total individuals per square mile; 1990: Entire EZ: 31,890; 1990: Bronx: 11,651; 1990: UM: 49,763; 1990: Comp.: 58,404; 2000: Entire EZ: 35,286; 2000: Bronx: 12,553; 2000: UM: 55,672; 2000: Comp.: 67,150; Percent change[A]: Entire EZ: 10.65; Percent change[A]: Bronx: 7.73; Percent change[A]: UM: 11.88; Percent change[A]: Comp.: 14.97. Percentage of households that moved in the last 5 years; 1990: Entire EZ: 31.93; 1990: Bronx: 31.96; 1990: UM: 31.93; 1990: Comp.: 32.5; 2000: Entire EZ: 34.07; 2000: Bronx: 33.4; 2000: UM: 34.2; 2000: Comp.: 33.64; Percent change[A]: Entire EZ: 2.14[B]; Percent change[A]: Bronx: 1.45; Percent change[A]: UM: 2.28[B]; Percent change[A]: Comp.: 1.15[B]. Percentage of population of working age (16-64); 1990: Entire EZ: 59.67; 1990: Bronx: 60.09; 1990: UM: 59.59; 1990: Comp.: 58.97; 2000: Entire EZ: 61.3; 2000: Bronx: 60.36; 2000: UM: 61.49; 2000: Comp.: 58.49; Percent change[A]: Entire EZ: 1.63[B]; Percent change[A]: Bronx: 0.28; Percent change[A]: UM: 1.9[B]; Percent change[A]: Comp.: -0.48. Percentage of population with a high school diploma (or equivalent); 1990: Entire EZ: 47.74; 1990: Bronx: 44.43; 1990: UM: 48.37; 1990: Comp.: 48.4; 2000: Entire EZ: 55.16; 2000: Bronx: 51.25; 2000: UM: 55.9; 2000: Comp.: 51.66; Percent change[A]: Entire EZ: 7.42[B]; Percent change[A]: Bronx: 6.82[B]; Percent change[A]: UM: 7.53[B]; Percent change[A]: Comp.: 3.26[B]. Percentage of high school dropouts; 1990: Entire EZ: 19.85; 1990: Bronx: 18.12; 1990: UM: 20.19; 1990: Comp.: 20.18; 2000: Entire EZ: 15.59; 2000: Bronx: 17.75; 2000: UM: 15.2; 2000: Comp.: 17.7; Percent change[A]: Entire EZ: -4.26[B]; Percent change[A]: Bronx: -0.37; Percent change[A]: UM: - 4.99[B]; Percent change[A]: Comp.: -2.48[B]. Percentage of vacant housing units; 1990: Entire EZ: 8.81; 1990: Bronx: 3.24; 1990: UM: 9.77; 1990: Comp.: 3.99; 2000: Entire EZ: 11.09; 2000: Bronx: 7.35; 2000: UM: 11.73; 2000: Comp.: 5.99; Percent change[A]: Entire EZ: 2.28[B]; Percent change[A]: Bronx: 4.11[ B]; Percent change[A]: UM: 1.96[B]; Percent change[A]: Comp.: 2.00[B]. Average owner occupied housing value; 1990: Entire EZ: $207,544; 1990: Bronx: $99,728; 1990: UM: $238,864; 1990: Comp.: $177,446; 2000: Entire EZ: $301,835; 2000: Bronx: $124,588; 2000: UM: $384,155; 2000: Comp.: $209,423; Percent change[A]: Entire EZ: 45.43[B]; Percent change[A]: Bronx: 24.93[B]; Percent change[A]: UM: 60.83[B]; Percent change[A]: Comp.: 18.02[B]. Source: GAO analysis of Census data. Note: There are 65 census tracts in the designated area and 160 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 22: Changes in Selected Economic Growth Variables Observed in the New York EZ, the Bronx and Upper Manhattan (UM) Portions, and the EZ Comparison Area (Comp.) Number of businesses; 1995: Entire EZ: 5,415; 1995: Bronx: 1,738; 1995: UM: 3,677; 1995: Comp.: 8,294; 1999: Entire EZ: 6,203; 1999: Bronx: 1,750; 1999: UM: 4,453; 1999: Comp.: 9,400; 2004: Entire EZ: 6,691; 2004: Bronx: 1,840; 2004: UM: 4,851; 2004: Comp.: 10,719; Percent change 1995- 2004[A]: Entire EZ: 23.6; Percent change 1995-2004[A]: Bronx: 5.87; Percent change 1995-2004[A]: UM: 31.9; Percent change 1995-2004[A]: Comp.: 29.2. Number of jobs; 1995: Entire EZ: 96,228; 1995: Bronx: 32,243; 1995: UM: 63,985; 1995: Comp.: 108,785; 1999: Entire EZ: 101,462; 1999: Bronx: 28,696; 1999: UM: 72,766; 1999: Comp.: 122,447; 2004: Entire EZ: 121,550; 2004: Bronx: 30,137; 2004: UM: 91,413; 2004: Comp.: 162,360; Percent change 1995- 2004[A]: Entire EZ: 26.3; Percent change 1995-2004[A]: Bronx: -6.53; Percent change 1995-2004[A]: UM: 42.9; Percent change 1995-2004[A]: Comp.: 49.3. Source: GAO analysis of Claritas data. Note: There are 65 census tracts in the designated area and 160 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: Many Upper Manhattan stakeholders we interviewed attributed the change in poverty to the higher incomes from the jobs the EZ helped create. In addition, several stakeholders discussed changes in the zone's population, as low-income residents were displaced by increases in property values and rental costs and employed residents with higher incomes moved into the area. One stakeholder attributed some of the decrease in poverty to welfare reform. For unemployment, some stakeholders said that it was difficult to improve the unemployment rate in the Upper Manhattan portion of the EZ due to a lack of residents with needed job skills. Stakeholders also noted that the change in the zone's population had affected unemployment as well as poverty. Several stakeholders observed that the Upper Manhattan EZ had helped foster economic growth, citing its role in the creation of retail areas and real estate development as examples. They also said that it had helped small businesses by providing technical assistance, training, and loans. Bronx stakeholders noted that the program had helped to influence poverty and unemployment through the resident employment requirements it had for businesses that received loans and the center it had created to match residents to jobs. However, some EZ stakeholders said that the EZ had had trouble getting EZ residents jobs, since there were few residents living in the Bronx portion of the EZ and many of them lacked necessary job skills. Further affecting the changes in poverty and unemployment, one stakeholder perceived that some original EZs residents had relocated and that new residents had moved into the EZ. Bronx stakeholders also said that access to capital for businesses resulted in an increase in businesses moving into the EZ, but one stakeholder noted that few jobs had been created. Another stakeholder said that some zone businesses were downsizing as a result of changes in the national economy. Philadelphia-Camden Empowerment Zone: Figure 24: Map of the Philadelphia-Camden EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] How the EZ Was Governed: Although Philadelphia and Camden were designated as one EZ and their original strategic plan envisioned a central board to oversee the EZ operations, the Philadelphia and Camden portions operated completely independently. Of the $100 million EZ grant, the Philadelphia portion of the EZ received about $79 million, and the Camden portion of the EZ received about $21 million. The Philadelphia Empowerment Zone Office that oversees the EZ program is part of the city government. The EZ created three subzones, each of which had its own Community Trust Board to identify the needs of the community, select activities to implement, and allocate resources to the activities. However, the EZ did not create an overarching board to oversee the three Philadelphia subzones. To select entities to implement activities, the EZ issued requests for proposals. A panel of community members, experts, and officials selected the best applications, and the city approved the funding. The mayor required that more than half of EZ/EC grant be spent on economic development (including job training) and retained the right to veto decisions by the Community Trust Boards, although this rarely happened. The Camden portion of the EZ was managed by a nonprofit entity called the Camden Empowerment Zone Corporation and was governed by a board, which included residents as well as "block captains" who were residents that had been elected to represent their communities, and other individuals from the business, cultural, religious, and nonprofit community.[Footnote 81] Under the board, there was a subcommittee structure. The EZ issued requests for proposals to identify organizations to implement the programs, which were reviewed by a subcommittee and then forwarded to the full board for approval. Activities the EZ Implemented: Both portions of the Philadelphia-Camden EZ implemented more community development activities than economic opportunity activities. Officials from the Philadelphia portion of the EZ explained that, while they spent more than half of their program grant funding on economic development as required by their mayor, the number of community development programs implemented was greater than the number of economic opportunity programs. In addition, both portions of the EZ received grant extensions until 2009. Activities related to education, access to capital, and assisting businesses were the most common in the Philadelphia portion of the Philadelphia-Camden EZ (fig. 25). The two activities most often cited by stakeholders in our interviews were the program to clean up vacant lots and the community lending institutions. EZ stakeholders also noted that the EZ had helped to organize the business community in each neighborhood. Figure 25: Activities Implemented by the Philadelphia Portion of the Philadelphia-Camden EZ: [See PDF for image] Sources: GAO (photo); GAO analysis of HUD data (charts). [End of figure] The activities most frequently implemented in the Camden portion of the EZ were housing, capacity building, access to capital, and infrastructure (fig. 26). In addition, most EZ stakeholders described the U.S.S. New Jersey, a new tourist attraction for which the EZ funded a portion of the application and the visitors' center, as a success of the EZ program. During our interviews, stakeholders also pointed out activities such as a summer youth program, the refurbishing of a local park, physical improvements to the streets and sidewalks, and an EZ program designed to make loans and grants to EZ businesses. Figure 26: Activities Implemented by the Camden Portion of the Philadelphia-Camden EZ: [See PDF for image] Source: GAO (photo); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Overall, the Philadelphia-Camden EZ was the only urban EZ to see positive changes in poverty, unemployment, and both measures of economic growth. However, changes in the Philadelphia and Camden portions varied, and were not always positive. Tables 23 and 24 show the changes in poverty, unemployment, and economic growth in the EZ, the Philadelphia and Camden portions of the EZ, and the EZ comparison area. Table 23 also includes data on the changes in other variables included in our models. The Philadelphia portion of the EZ experienced decreases in poverty and unemployment and little change or a decrease in our measures of economic growth. Its declines in the poverty and unemployment rates from 1990 to 2000 outpaced those in the Philadelphia-Camden comparison area.[Footnote 82] In economic growth, the Philadelphia portion of the EZ experienced little change in the number of businesses between 1995 and 2004, while its comparison area experienced a large increase. Both the Philadelphia portion of the EZ and the EZ comparison area saw a similar decline in the number of jobs available. In contrast, the Camden portion of the Philadelphia-Camden EZ experienced little change in poverty or the unemployment rate, but it experienced positive changes in both measures of economic growth. Its poverty rate from 1990 to 2000 stayed about the same, while its comparison area decreased. Both the Camden portion of the EZ and the EZ comparison area saw little change in the unemployment rate. For economic growth, the Camden portion of the EZ had one of the highest increases in the number of businesses of any EZ from 1995 to 2004, slightly better than its comparison area. It also saw a large increase in the number of jobs. In contrast, the EZ comparison area experienced a large decline in number of jobs over the same time period. Table 23: Changes in Selected Census Variables Observed in the Philadelphia-Camden EZ, the Camden (Cam.) and Philadelphia (Phila.) Portions, and the EZ Comparison Area (Comp.) Poverty rate (%); 1990: Entire EZ: 50.14; 1990: Cam.: 43; 1990: Phila.: 52.1; 1990: Comp.: 43.07; 2000: Entire EZ: 42.98; 2000: Cam.: 40.55; 2000: Phila.: 43.68; 2000: Comp.: 37.97; Percent change[ A]: Entire EZ: -7.16[B]; Percent change[ A]: Cam.: -2.45; Percent change[ A]: Phila.: -8.42[B]; Percent change[ A]: Comp.: -5.1[B]. Unemployment rate (%); 1990: Entire EZ: 22.21; 1990: Cam.: 18.09; 1990: Phila.: 23.6; 1990: Comp.: 18.81; 2000: Entire EZ: 19.3; 2000: Cam.: 19.08; 2000: Phila.: 19.38; 2000: Comp.: 17.91; Percent change[ A]: Entire EZ: -2.91[B]; Percent change[ A]: Cam.: 0.99; Percent change[ A]: Phila.: -4.22[B]; Percent change[ A]: Comp.: -0.9. Average household income; 1990: Entire EZ: $23,188; 1990: Cam.: $26,742; 1990: Phila.: $22,269; 1990: Comp.: $27,292; 2000: Entire EZ: $28,562; 2000: Cam.: $31,158; 2000: Phila.: $27,851; 2000: Comp.: $31,318; Percent change[ A]: Entire EZ: 23.17[B]; Percent change[ A]: Cam.: 16.52[B]; Percent change[ A]: Phila.: 25.07[B]; Percent change[ A]: Comp.: 14.8[B]. Percentage of single female headed households with children; 1990: Entire EZ: 22.73; 1990: Cam.: 24.13; 1990: Phila.: 22.37; 1990: Comp.: 21.01; 2000: Entire EZ: 21.93; 2000: Cam.: 23.22; 2000: Phila.: 21.57; 2000: Comp.: 20.66; Percent change[ A]: Entire EZ: -0.81; Percent change[ A]: Cam.: -0.92; Percent change[ A]: Phila.: -0.8; Percent change[ A]: Comp.: -0.36. Total population; 1990: Entire EZ: 52,440; 1990: Cam.: 13,332; 1990: Phila.: 39,108; 1990: Comp.: 38,520; 2000: Entire EZ: 45,725; 2000: Cam.: 12,749; 2000: Phila.: 32,976; 2000: Comp.: 35,827; Percent change[ A]: Entire EZ: -12.81; Percent change[ A]: Cam.: -4.37; Percent change[ A]: Phila.: -15.68; Percent change[ A]: Comp.: -6.99. Total individuals per square mile; 1990: Entire EZ: 12,248; 1990: Cam.: 7,342; 1990: Phila.: 15,861; 1990: Comp.: 7,601; 2000: Entire EZ: 10,698; 2000: Cam.: 7,034; 2000: Phila.: 13,396; 2000: Comp.: 7,064; Percent change[ A]: Entire EZ: -12.66; Percent change[ A]: Cam.: -4.2; Percent change[ A]: Phila.: -15.54; Percent change[ A]: Comp.: -7.06. Percentage of households that moved in the last 5 years; 1990: Entire EZ: 34.71; 1990: Cam.: 41.93; 1990: Phila.: 32.25; 1990: Comp.: 40.69; 2000: Entire EZ: 34.05; 2000: Cam.: 44.42; 2000: Phila.: 30.04; 2000: Comp.: 44.05; Percent change[ A]: Entire EZ: -0.66; Percent change[ A]: Cam.: 2.49; Percent change[ A]: Phila.: -2.21; Percent change[ A]: Comp.: 3.36[B]. Percentage of population of working age (16-64); 1990: Entire EZ: 57.58; 1990: Cam.: 63.02; 1990: Phila.: 55.72; 1990: Comp.: 59.76; 2000: Entire EZ: 58.85; 2000: Cam.: 66.86; 2000: Phila.: 55.75; 2000: Comp.: 61.93; Percent change[ A]: Entire EZ: 1.27; Percent change[ A]: Cam.: 3.84[B]; Percent change[ A]: Phila.: 0.03; Percent change[ A]: Comp.: 2.17. Percentage of population with a high school diploma (or equivalent); 1990: Entire EZ: 42.79; 1990: Cam.: 44.51; 1990: Phila.: 42.21; 1990: Comp.: 54.04; 2000: Entire EZ: 51.82; 2000: Cam.: 50.43; 2000: Phila.: 52.38; 2000: Comp.: 63.19; Percent change[ A]: Entire EZ: 9.04[B]; Percent change[ A]: Cam.: 5.92[B]; Percent change[ A]: Phila.: 10.17[B]; Percent change[ A]: Comp.: 9.15[B]. Percentage of high school dropouts; 1990: Entire EZ: 25.58; 1990: Cam.: 22.67; 1990: Phila.: 26.56; 1990: Comp.: 23.07; 2000: Entire EZ: 16.02; 2000: Cam.: 12.54; 2000: Phila.: 17.25; 2000: Comp.: 14.05; Percent change[ A]: Entire EZ: - 9.56[B]; Percent change[ A]: Cam.: -10.13[B]; Percent change[ A]: Phila.: -9.3[B]; Percent change[ A]: Comp.: -9.02[B]. Percentage of vacant housing units; 1990: Entire EZ: 21.45; 1990: Cam.: 22.37; 1990: Phila.: 21.21; 1990: Comp.: 14.48; 2000: Entire EZ: 24.94; 2000: Cam.: 25.7; 2000: Phila.: 24.73; 2000: Comp.: 14.91; Percent change[ A]: Entire EZ: 3.49[B]; Percent change[ A]: Cam.: 3.33[B]; Percent change[ A]: Phila.: 3.52[B]; Percent change[ A]: Comp.: 0.43. Average owner occupied housing value; 1990: Entire EZ: $29,899; 1990: Cam.: $35,076; 1990: Phila.: $28,288; 1990: Comp.: $42,045; 2000: Entire EZ: $37,780; 2000: Cam.: $39,398; 2000: Phila.: $37,353; 2000: Comp.: $51,159; Percent change[ A]: Entire EZ: 26.36[B]; Percent change[ A]: Cam.: 12.32[B]; Percent change[ A]: Phila.: 32.04[B]; Percent change[ A]: Comp.: 21.7[B]. Source: GAO analysis of Census data. Note: There are 18 census tracts in the designated area and 16 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 24: Changes in Selected Economic Growth Variables Observed in the Philadelphia-Camden EZ, the Camden (Cam.) and Philadelphia (Phila.) Portions, and the EZ Comparison Area (Comp.) Number of businesses; 1995: Entire EZ: 2,064; 1995: Cam.: 730; 1995: Phila.: 1,334; 1995: Comp.: 2,631; 1999: Entire EZ: 2,078; 1999: Cam.: 720; 1999: Phila.: 1,358; 1999: Comp.: 2,768; 2004: Entire EZ: 2,150; 2004: Cam.: 806; 2004: Phila.: 1,344; 2004: Comp.: 2,821; Percent change 1995- 2004[A]: Entire EZ: 4.17; Percent change 1995-2004[A]: Cam.: 10.41; Percent change 1995-2004[A]: Phila.: 0.75; Percent change 1995-2004[A]: Comp.: 7.22. Number of jobs; 1995: Entire EZ: 35,867; 1995: Cam.: 14,430; 1995: Phila.: 21,437; 1995: Comp.: 55,071; 1999: Entire EZ: 35,904; 1999: Cam.: 16,702; 1999: Phila.: 19,202; 1999: Comp.: 53,702; 2004: Entire EZ: 36,789; 2004: Cam.: 21,032; 2004: Phila.: 15,757; 2004: Comp.: 39,720; Percent change 1995-2004[A]: Entire EZ: 2.57; Percent change 1995-2004[A]: Cam.: 45.75; Percent change 1995-2004[A]: Phila.: -26.5; Percent change 1995-2004[A]: Comp.: -27.87. Source: GAO analysis of Claritas data. Note: There are 18 census tracts in the designated area and 16 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: Stakeholders in the Philadelphia portion of the EZ mentioned the change in the zone's population as having an effect on the changes in the poverty and unemployment rates. They noted that the number of poor households had decreased, in part due to the HOPE VI housing program, which had demolished some area public housing, and in part because some individuals who had obtained jobs had also moved out of the zone neighborhoods. In addition, some stakeholders noted that the change in welfare policy over the course of the EZ/EC program had an effect on poverty and unemployment by moving former welfare recipients into jobs. In describing the changes in economic growth, Philadelphia stakeholders said that community lending institutions had provided loans to businesses and that the vacant lot improvement program had helped retain and attract businesses to the EZ. However, one stakeholder noted that EZ lending had not resulted in many new jobs. EZ stakeholders in the Camden portion of the EZ said that EZ programs such as the Battleship New Jersey and housing and after school initiatives may have contributed to the slight decrease in poverty rate. They also noted that there had probably been a change in the EZ population, since Camden's population was transient and individuals often left the area when they found a job. In addition, some stakeholders also mentioned changes in the national economy and a high homeless population as challenges to improving the area's poverty and unemployment rates. Camden stakeholders noted that the EZ had influenced economic growth through the improvements it had made to the physical appearance of certain commercial corridors, as well as through its loans and grants to small businesses. Stakeholders also said that the development of market-rate housing had helped to increase the customer base for local businesses. Finally, one stakeholder said that the state's expansion of the light rail to Camden had influenced economic growth by improving transportation to the area. Cleveland Empowerment Zone: Figure 27: Map of the Cleveland EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] HUD initially designated Cleveland as a Supplemental EZ, which provided it with Economic Development Initiative grants and Section 108 Loan Guarantees rather than EZ/EC grant funds. The area received full Round I EZ status in 1998, and businesses in the EZ could claim the program tax benefits starting in 2000. How the EZ Was Governed: The EZ was operated by the city of Cleveland Department of Economic Development and included the Community Advisory Committee, an advisory board made up of EZ residents, business owners, bank representatives, and representatives from four local community development corporations. Although the Community Advisory Committee was involved in the decision- making process, the mayor and Cleveland City Council made all final decisions about EZ funding. Activities the EZ Implemented: Cleveland EZ stakeholders said that they focused mainly on economic development activities, largely because of the type of benefits they received with the Supplemental EZ designation.[Footnote 83] Stakeholders explained that most of their funds had been used to fund loans to EZ businesses. They also said that the EZ also worked to build the capacity of four community development corporations helping each of them complete a major project in their neighborhood--for example, the Quincy Place building in the Fairfax neighborhood (fig. 28). EZ stakeholders noted that the EZ had implemented some successful job training programs. The EZ received an extension of its grants and loan guarantees through 2009. Figure 28: Activity Implemented by the Cleveland EZ: [See PDF for image] Source: GAO. Note: We were not able to determine specific types of activities the Cleveland EZ implemented, because reliable data were not available. [End of figure] Changes in Poverty, Unemployment, and Economic Growth: The Cleveland EZ experienced positive changes in poverty, unemployment, and one measure of economic growth. From 1990 to 2000, Cleveland had one of the sharpest reductions in both poverty and unemployment of the urban EZs, and these changes outpaced those of its comparison area. Between 1995 and 2004, the EZ experienced an increase in economic growth as measured by the number of businesses, while its comparison area experienced a decrease. However, the EZ experienced a decrease in the number of jobs that was greater than the decrease experienced in its comparison area. Tables 25 and 26 show the changes in poverty, unemployment, and economic growth in the EZ and its comparison area. Table 25 also includes data on the changes in other variables included in our models. Table 25: Changes in Selected Census Variables Observed in the Cleveland EZ and Its Comparison Area: Poverty rate (%); 1990: EZ: 46.85; 1990: Comparison: 39.41; 2000: EZ: 36.03; 2000: Comparison: 35.70; Percent change[A]: EZ: -10.82[B]; Percent change[A]: Comparison: -3.71[B]. Unemployment rate (%); 1990: EZ: 25.44; 1990: Comparison: 20.63; 2000: EZ: 15.48; 2000: Comparison: 17.29; Percent change[A]: EZ: -9.96[B]; Percent change[A]: Comparison: -3.34[B]. Average household income; 1990: EZ: $20,535; 1990: Comparison: $24,688; 2000: EZ: $28,781; 2000: Comparison: $30,311; Percent change[A]: EZ: 40.16[B]; Percent change[A]: Comparison: 22.78[B]. Percentage of single female headed households with children; 1990: EZ: 19.07; 1990: Comparison: 23.24; 2000: EZ: 18.24; 2000: Comparison: 23.19; Percent change[A]: EZ: -0.83; Percent change[A]: Comparison: -0.05. Total population; 1990: EZ: 50,724; 1990: Comparison: 153,578; 2000: EZ: 43,694; 2000: Comparison: 141,465; Percent change[A]: EZ: -13.86; Percent change[A]: Comparison: -7.89. Total individuals per square mile; 1990: EZ: 8,319; 1990: Comparison: 7,231; 2000: EZ: 7,168; 2000: Comparison: 6,532; Percent change[A]: EZ: -13.84; Percent change[A]: Comparison: -9.66. Percentage of households that moved in the last 5 years; 1990: EZ: 36.13; 1990: Comparison: 36.04; 2000: EZ: 39.55; 2000: Comparison: 39.70; Percent change[A]: EZ: 3.41[B]; Percent change[A]: Comparison: 3.66[B]. Percentage of population of working age (16-64); 1990: EZ: 55.01; 1990: Comparison: 56.80; 2000: EZ: 53.08; 2000: Comparison: 55.54; Percent change[A]: EZ: -1.93; Percent change[A]: Comparison: -1.26. Percentage of population with a high school diploma (or equivalent); 1990: EZ: 47.14; 1990: Comparison: 54.84; 2000: EZ: 61.82; 2000: Comparison: 65.13; Percent change[A]: EZ: 14.68[B]; Percent change[A]: Comparison: 10.29[B]. Percentage of high school dropouts; 1990: EZ: 18.35; 1990: Comparison: 14.83; 2000: EZ: 13.30; 2000: Comparison: 16.38; Percent change[A]: EZ: -5.05[B]; Percent change[A]: Comparison: 1.55[B]. Percentage of vacant housing units; 1990: EZ: 14.68; 1990: Comparison: 14.70; 2000: EZ: 18.82; 2000: Comparison: 15.13; Percent change[A]: EZ: 4.14[B]; Percent change[A]: Comparison: 0.43. Average owner occupied housing value; 1990: EZ: $38,071; 1990: Comparison: $46,972; 2000: EZ: $75,186; 2000: Comparison: $70,164; Percent change[A]: EZ: 97.49[B]; Percent change[A]: Comparison: 49.37[B]. Source: GAO analysis of Census data. Note: There are 32 census tracts in the designated area and 68 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 26: Changes in Selected Economic Growth Variables Observed in the Cleveland EZ and Its Comparison Area: Number of businesses; 1995: EZ: 1,766; 1995: Comparison: 4,883; 1999: EZ: 2,067; 1999: Comparison: 4,889; 2004: EZ: 1,899; 2004: Comparison: 4,602; Percent change 1995-2004[A]: EZ: 7.53; Percent change 1995-2004[A]: Comparison: -5.75. Number of jobs; 1995: EZ: 42,087; 1995: Comparison: 87,334; 1999: EZ: 58,679; 1999: Comparison: 102,996; 2004: EZ: 38,023; 2004: Comparison: 84,064; Percent change 1995-2004[A]: EZ: - 9.66; Percent change 1995-2004[A]: Comparison: -3.74. Source: GAO analysis of Claritas data. Note: There are 32 census tracts in the designated area and 68 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: When asked about factors that had affected the changes observed in the Cleveland EZ, stakeholders said that factors related to poverty and unemployment were intertwined. For example, EZ stakeholders felt that EZ training programs had helped prepare residents for jobs, potentially affecting both poverty and unemployment. Stakeholders also cited changes in the zone population that had affected both factors, noting that as residents obtained jobs, they left the zone, and that some individuals with higher incomes had moved in, particularly in areas where new housing had been built. EZ stakeholders also mentioned the effect of general economic trends on poverty and unemployment. In terms of economic growth, EZ stakeholders noted that the majority of the businesses that had received EZ loans were still operating, that the number of businesses had increased in some areas of the EZ, and that these businesses had brought new jobs to the community. However, some EZ stakeholders commented that the EZ's strict underwriting standards made it less successful in helping new or less sophisticated businesses. In addition, although the EZ had helped to create some jobs, some stakeholders felt that the jobs created were going to new residents rather than to original EZ residents. EZ staff also observed that regional trends such as the overall loss of jobs in the city of Cleveland had an effect on economic growth in the Cleveland EZ. Los Angeles Empowerment Zone: Figure 29: Map of the Los Angeles EZ and Its Comparison Area: [See PDF for image] Source: GAO analysis of Census and HUD data. [End of figure] HUD initially designated Los Angeles as a Supplemental EZ, which provided it with Economic Development Initiative grants and Section 108 Loan Guarantees rather than EZ/EC grant funds. The area received full Round I EZ status in 1998, and businesses in the EZ could claim the program tax benefits starting in 2000. How the EZ Was Governed: The Los Angeles EZ created the Los Angeles Community Development Bank to administer its EZ program. The Community Development Bank was a "wholesale" rather than a conventional bank that entered into partnerships with other economic development entities that were already delivering services and operating loan programs. The EZ was autonomous from the city and had its own board of directors predominately made up of private sector members with one seat for a community representative. However, EZ stakeholders told us that this seat usually remained vacant. The board had a committee structure that included an audit committee, credit committee, and venture capital committee. The EZ board made all funding decisions. Any transaction over $1 million required full board approval, but smaller amounts could be approved by a committee of the board. In an effort to involve community members, the city created an advisory council called EZ Oversight Committee, which was filled through appointments made by the mayor and county board of supervisors. However, EZ stakeholders said that the EZ oversight committee never had a formal role in decision making or oversight. Activities the EZ Implemented: Los Angeles EZ stakeholders said that they focused mainly on economic development activities, largely due to the type of benefits they received with the Supplemental EZ designation.[Footnote 84] Stakeholders noted that the job requirements attached to loans from the EZ and the six tax-exempt bonds had helped create jobs in the zone (fig. 30). In addition to providing loans to several businesses, the EZ helped fund a shopping complex and other development. One stakeholder felt the EZ did not lend enough funds to small businesses and pointed out that some of the loans to large businesses, such as a large dairy, had defaulted. The EZ bank filed for bankruptcy in 2002 due to a high level of loan defaults and the remaining funds were transferred to the city of Los Angeles. The city of Los Angeles received an extension for the grant and loan guarantees through 2009. Figure 30: Activity Implemented by the Los Angeles EZ: [See PDF for image] Source: GAO. [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Unlike the other EZs, both poverty and unemployment in the Los Angeles EZ largely remained the same between 1990 and 2000, and measures of economic growth declined from 1995 to 2004. The comparison area also saw little change in poverty and unemployment, but economic growth in the comparison area increased in that time period. Tables 27 and 28 show the changes in poverty, unemployment, and economic growth in the EZ and its comparison area. Table 27 also includes data on the changes in other variables included in our models. Table 27: Changes in Selected Census Variables Observed in the Los Angeles EZ and Its Comparison Area: Poverty rate (%); 1990: EZ: 40.24; 1990: Comparison: 31.52; 2000: EZ: 41.49; 2000: Comparison: 33.14; Percent change[A]: EZ: 1.25; Percent change[A]: Comparison: 1.61. Unemployment rate (%); 1990: EZ: 18.39; 1990: Comparison: 15.07; 2000: EZ: 18.61; 2000: Comparison: 15.47; Percent change[A]: EZ: 0.22; Percent change[A]: Comparison: 0.40. Average household income; 1990: EZ: $28,801; 1990: Comparison: $34,087; 2000: EZ: $32,631; 2000: Comparison: $37,843; Percent change[A]: EZ: 13.30[B]; Percent change[A]: Comparison: 11.02[B]. Percentage of single female headed households with children; 1990: EZ: 18.32; 1990: Comparison: 18.64; 2000: EZ: 16.90; 2000: Comparison: 17.40; Percent change[A]: EZ: -1.43[B]; Percent change[A]: Comparison: -1.24. Total population; 1990: EZ: 211,365; 1990: Comparison: 221,657; 2000: EZ: 225,591; 2000: Comparison: 219,001; Percent change[A]: EZ: 6.73; Percent change[A]: Comparison: -1.20. Total individuals per square mile; 1990: EZ: 11,082; 1990: Comparison: 12,918; 2000: EZ: 11,836; 2000: Comparison: 13,170; Percent change[A]: EZ: 6.81; Percent change[A]: Comparison: 1.95. Percentage of households that moved in the last 5 years; 1990: EZ: 46.12; 1990: Comparison: 44.06; 2000: EZ: 43.53; 2000: Comparison: 41.20; Percent change[A]: EZ: -2.59[B]; Percent change[A]: Comparison: -2.86[B]. Percentage of population of working age (16-64); 1990: EZ: 57.60; 1990: Comparison: 58.11; 2000: EZ: 57.53; 2000: Comparison: 57.28; Percent change[A]: EZ: -0.06; Percent change[A]: Comparison: -0.83. Percentage of population with a high school diploma (or equivalent); 1990: EZ: 38.40; 1990: Comparison: 50.30; 2000: EZ: 37.46; 2000: Comparison: 49.53; Percent change[A]: EZ: -0.94; Percent change[A]: Comparison: -0.77. Percentage of high school dropouts; 1990: EZ: 32.14; 1990: Comparison: 23.50; 2000: EZ: 23.09; 2000: Comparison: 17.04; Percent change[A]: EZ: -9.04[B]; Percent change[A]: Comparison: -6.46[B]. Percentage of vacant housing units; 1990: EZ: 6.33; 1990: Comparison: 6.31; 2000: EZ: 9.65; 2000: Comparison: 8.11; Percent change[A]: EZ: 3.31[B]; Percent change[A]: Comparison: 1.80[B]. Average owner occupied housing value; 1990: EZ: $141,665; 1990: Comparison: $160,090; 2000: EZ: $156,493; 2000: Comparison: $165,180; Percent change[A]: EZ: 10.47[B]; Percent change[A]: Comparison: 3.18[B]. Source: GAO analysis of Census data. Note: There are 41 census tracts in the designated area and 43 in the comparison area. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 28: Changes in Selected Economic Growth Variables Observed in the Los Angeles EZ and Its Comparison Area: Number of businesses; 1995: EZ: 15,746; 1995: Comparison: 4,248; 1999: EZ: 12,315; 1999: Comparison: 3,986; 2004: EZ: 13,853; 2004: Comparison: 4,662; Percent change 1995-2004[A]: EZ: -12.02; Percent change 1995-2004[A]: Comparison: 9.75. Number of jobs; 1995: EZ: 165,457; 1995: Comparison: 52,973; 1999: EZ: 153,340; 1999: Comparison: 55,627; 2004: EZ: 156,793; 2004: Comparison: 66,783; Percent change 1995- 2004[A]: EZ: -5.24; Percent change 1995-2004[A]: Comparison: 26.07. Source: GAO analysis of Claritas data. Note: There are 41 census tracts in the designated area and 43 in the comparison area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: Los Angeles EZ stakeholders we interviewed suggested that the EZ was not as likely as other factors to have effected changes in poverty and unemployment because they could not address those issues directly with the benefits they received. One stakeholder did not believe that the EZ had met its goals of increasing job training and employment opportunities, but other stakeholders believed that it had helped to assist and retain businesses and redevelop the area. Stakeholders mentioned external factors that influenced changes in poverty and unemployment, such as shifts in demographics with the influx of new immigrants and the outmigration of EZ residents as they obtained jobs or their incomes increased. In addition, some said that the EZ's high concentration of homeless individuals and the lack of available public transportation in the EZ could be additional reasons that poverty and unemployment rates did not improve. One stakeholder noted that, because the original strategic plan was designed to focus on social services, the census tracts chosen were not well-suited for economic development. However, stakeholders mentioned that the EZ had helped to stabilize the area, since a large number of businesses had been leaving the Los Angeles area for advantages offered in other locations. Kentucky Highlands Empowerment Zone: Figure 31: Map of the Kentucky Highlands EZ: [See PDF for image] Source: GAO analysis of USDA data. [End of figure] How the EZ Was Governed: The EZ was managed by the Kentucky Highlands Investment Corporation, a nonprofit that had been operating in the area for over 25 years. There were subzone boards in each of the three counties that became separate nonprofit entities and had funds to hire staff, manage the board, and conduct fiscal oversight. An overarching steering committee, which included representatives of the subzone boards, directed the EZ's activities in the entire zone by providing oversight, making financial decisions, and implementing certain activities, such as the revolving loan fund. EZ stakeholders suggested that most of the decision making occurred at the subzone level, although the steering committee gave final approval to all projects. The EZ used about half of the available funds, and the rest was distributed among the three subzones. Activities the EZ Implemented: Almost two-thirds of the EZ's activities involved community development. Initiatives involving business development and job training; resources for communities, youth and families; and education were the most common activities (fig. 32). In addition, each county implemented different types of activities from the strategic plan. For example, stakeholders from the Clinton County subzone funded a library, a learning center, and health care initiatives--such as helping to fund the expansion of an emergency room and surgical wing at the local hospital--and attracted businesses from the houseboat industry. Stakeholders from the Jackson County subzone said that they had provided funds for a community center, which housed vocational training classes and a community theatre. Stakeholders from the Wayne County subzone said that they completed a water infrastructure project that they said was critical to attracting businesses and brought in jobs in the houseboat industry. The Kentucky Highlands EZ received a grant extension until 2009. Figure 32: Activities Implemented by the Kentucky Highlands EZ: [See PDF for image] Sources: GAO (photo); GAO analysis of USDA data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: Not only did the Kentucky Highlands EZ experience positive changes in all indicators, it experienced the largest decrease in unemployment between 1990 and 2000 and the largest increases in the number of businesses and jobs between 1995 and 2004 of any rural EZ. Tables 29 and 30 show the changes in poverty, unemployment, and economic growth in the EZ. Table 29 also includes data on the changes in other variables included in our models of the urban EZs. Table 29: Changes in Selected Census Variables Observed in the Kentucky Highlands EZ: Poverty rate (%); 1990: 37.88; 2000: 27.76; Percent change[A]: - 10.12[B]. Unemployment rate (%); 1990: 9.76; 2000: 7.75; Percent change[A]: - 2.01[B]. Average household income; 1990: $23,304; 2000: $31,064; Percent change[A]: 33.3[B]. Percentage of single female headed households with children; 1990: 4.64; 2000: 5.73; Percent change[A]: 1.09. Total population; 1990: 27,212; 2000: 30,464; Percent change[A]: 11.95. Total individuals per square mile; 1990: 36; 2000: 40; Percent change[A]: 11.96. Percentage of households that moved in the last 5 years; 1990: 32.46; 2000: 31.45; Percent change[A]: -1.01. Percentage of population of working age (16-64); 1990: 59.04; 2000: 61.98; Percent change[A]: 2.93[B]. Percentage of population with a high school diploma (or equivalent); 1990: 42.82; 2000: 55.1; Percent change[A]: 12.28[B]. Percentage of high school dropouts; 1990: 15.80; 2000: 16.47; Percent change[A]: 0.67. Percentage of vacant housing units; 1990: 16.74; 2000: 18.97; Percent change[A]: 2.23[B]. Average owner occupied housing value; 1990: $43,392; 2000: $65,815; Percent change[A]: 51.68[B]. Source: GAO analysis of Census data. Note: There are seven census tracts in the designated area; we did not use comparison areas for rural EZs. For more information on our methodology, see appendix I. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 30: Changes in Selected Economic Growth Variables Observed in the Kentucky Highlands EZ: Number of businesses; 1995: 609; 1999: 691; 2004: 810; Percent change: 1995-2004[A]: 33. Number of jobs; 1995: 5,327; 1999: 7,691; 2004: 8,941; Percent change: 1995-2004[A]: 67.84. Source: GAO analysis of Claritas data. Note: There are seven census tracts in the designated area; we did not use comparison areas for rural EZs. For more information on our methodology, see appendix I. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, stakeholders said that changes in the poverty rate may have been the result of new jobs created by EZ projects, many of which offered benefits such as health insurance that helped to stabilize families. However, EZ staff and other stakeholders acknowledged that external factors, such as welfare reform and general economic trends, also could have contributed to poverty reduction. Stakeholders also attributed the reduction in unemployment to the job creation efforts, saying that the EZ had helped stabilize the area when a key employer, a sewing plant, closed prior to designation. In terms of economic growth, stakeholders felt that the EZ had played a role in the change in economic growth, citing infrastructure improvements and zone workshops on how to start new businesses. In addition, some EZ stakeholders noted that the economic growth that had occurred was due in part to the EZ program tax benefits, although not all stakeholders agreed. Mid-Delta Mississippi Empowerment Zone: Figure 33: Map of the Mid-Delta EZ: [See PDF for image] Source: GAO analysis of USDA data. [End of figure] How the EZ Was Governed: The nonprofit Mid-Delta Empowerment Zone Alliance was created to manage the EZ. It included a board that consisted of city and county elected officials and representatives from community organizations, plus subzones boards in each of the six counties. Most decisions were made by the committees and brought to the full board for approval. However, several stakeholders noted that this formal process was not always followed and that some board decisions appeared to favor large businesses over community groups. Activities the EZ Implemented: Most of the activities the Mid-Delta EZ implemented were related to community development. Initiatives involving business development and job training; resources for communities, youth, and families; education; and housing accounted for the bulk of the activities (fig. 34). In our interviews, stakeholders noted that EZ funds were used for a variety of community-and family-oriented projects. These included helping a small municipality purchase needed police and fire equipment, partially funding a mortgage assistance program that moved 20 people into houses, and implementing some health care programs, such as a substance abuse treatment center for women. Also, one business in the Mid-Delta EZ used a program tax-exempt bond. In addition, stakeholders mentioned that EZ funds were used to attract major corporations, such as an automobile parts manufacturer and retail distribution center. However, several stakeholders also noted that some programs were unsuccessful, and an EZ official said that approximately 16 projects were under review for possible misuse of funds. The Mid-Delta EZ received a grant extension until 2009. Figure 34: Activities Implemented by the Mid-Delta EZ: [See PDF for image] Sources: GAO (photo); GAO analysis of USDA data (charts). [End of figure] - graphic text: Changes in Poverty, Unemployment, and Economic Growth: The Mid-Delta EZ saw positive changes in two indicators: poverty and economic growth. Between 1990 and 2000, the poverty rate in the Mid- Delta EZ decreased more than any rural EZ. However, the Mid-Delta EZ experienced a small increase in unemployment over that time period. For economic growth, the EZ saw an increase in both measures from 1995 to 2004, but the changes were significantly less than in the other two rural EZs. Tables 31 and 32 show the changes in poverty, unemployment, and economic growth in the EZ. Table 31 also includes data on the changes in other variables included in our models of the urban EZs. Table 31: Changes in Selected Census Variables Observed in the Mid- Delta EZ: Poverty rate (%); 1990: 46.35; 2000: 35.67; Percent change[A]: - 10.68[B]. Unemployment rate (%); 1990: 14.31; 2000: 17.38; Percent change[A]: 3.07[B]. Average household income; 1990: $25,872; 2000: $3,559; Percent change[A]: 37.44[B]. Percentage of single female headed households with children; 1990: 16.51; 2000: 17.31; Percent change[A]: 0.80. Total population; 1990: 29,494; 2000: 29,770; Percent change[A]: 0.94. Total individuals per square mile; 1990: 30.06; 2000: 30.34; Percent change[A]: 0.95. Percentage of households that moved in the last 5 years; 1990: 34.75; 2000: 31.00; Percent change[A]: -3.75[B]. Percentage of population of working age (16-64); 1990: 51.71; 2000: 57.36; Percent change[A]: 5.65[B]. Percentage of population with a high school diploma (or equivalent); 1990: 49.09; 2000: 60.52; Percent change[A]: 11.43[B]. Percentage of high school dropouts; 1990: 14.66; 2000: 12.62; Percent change[A]: -2.04[B]. Percentage of vacant housing units; 1990: 8.08; 2000: 9.41; Percent change[A]: 1.33. Average owner occupied housing value; 1990: $50,061; 2000: $66,872; Percent change[A]: 33.58[B]. Source: GAO analysis of Census data. Note: There are eight census tracts in the designated area; we did not use comparison areas for rural EZs. For more information on our methodology, see appendix I. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 32: Changes in Selected Economic Growth Variables Observed in the Mid-Delta EZ: Number of businesses; 1995: 634; 1999: 838; 2004: 733; Percent change 1995-2004[A]: 15.62. Number of jobs; 1995: 9,415; 1999: 12,694; 2004: 9,884; Percent change 1995-2004[A]: 4.98. Source: GAO analysis of Claritas data. Note: There are eight census tracts in the designated area; we did not use comparison areas for rural EZs. For more information on our methodology, see appendix I. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, EZ stakeholders credited the EZ with improving poverty and unemployment by helping bring in higher paying jobs with benefits. However, some suggested that increases in unemployment were not the same for each county, and added that the Mississippi Delta region overall had a low educational level that limited some residents' ability to participate in the workforce. In terms of economic growth, EZ stakeholders noted the EZ's efforts to attract large businesses through grants and loans had brought in new companies that provided jobs with relatively high wages and benefits. One stakeholder said that the EZ's efforts helped to stabilize the area during a period when several large manufacturing plants relocated to other countries. Rio Grande Valley, Texas Empowerment Zone: Figure 35: Map of the Rio Grande Valley EZ: [See PDF for image] Source: GAO analysis of USDA data. [End of figure] How the EZ Was Governed: The EZ was managed by the nonprofit Rio Grande Valley Empowerment Zone, which was created specifically for the EZ. EZ stakeholders explained that the EZ board included an executive committee of members representing each of the four counties in the EZ and four subzone boards, one for each county. Both the EZ and the subzone boards were involved in selecting activities for implementation. Subzone members reviewed proposals and then forwarded their recommendations to a project review committee, which reviewed the activities for feasibility and sustainability. Once this process was complete, the activity was sent to the full board for approval. Activities the EZ Implemented: The Rio Grande Valley EZ implemented mostly community development activities, most commonly education, public infrastructure, and business development and job training initiatives (fig. 36). In our interviews, stakeholders mentioned that the EZ had provided funds to several projects sponsored by the school districts, focusing its funding on improving the well-being of children. For example, the EZ provided computers and technical assistance to local Boys and Girls Clubs. EZ stakeholders also cited several infrastructure improvements, such as a water plant, a water tower, the expansion of a fire department facility, and the creation of community centers. In terms of economic opportunity initiatives, three counties provided loans through a revolving loan program, and one county created a small business incubator. In addition, the EZ provided funding to a community-based organization to provide low-skilled workers with training for jobs in the health care field. Figure 36: Activities Implemented by the Rio Grande Valley EZ: [See PDF for image] Sources: GAO (photo); GAO analysis of USDA data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: The Rio Grande Valley EZ experienced positive changes in poverty and economic growth. The EZ had the highest poverty and unemployment rates in 1990 of any of the rural EZs. Between 1990 and 2000, the EZ experienced a decrease in poverty; however, the unemployment rate did not show a significant change. For economic growth, the EZ experienced an increase in the number of businesses and jobs between 1995 and 2004. Tables 33 and 34 show the changes in poverty, unemployment, and economic growth in the EZ. Table 33 also includes data on the changes in other variables included in our models of the urban EZs. Table 33: Changes in Selected Census Variables Observed in the Rio Grande Valley EZ: Poverty rate (%); 1990: 49.65; 2000: 42.34; Percent change[A]: - 7.31[B]. Unemployment rate (%); 1990: 14.94; 2000: 13.82; Percent change[A]: - 1.12. Average household income; 1990: $25,093; 2000: $32,763; Percent change[A]: 30.57[B]. Percentage of single female headed households with children; 1990: 9.44; 2000: 10.38; Percent change[A]: 0.95. Total population; 1990: 29,817; 2000: 37,044; Percent change[A]: 24.24. Total individuals per square mile; 1990: 131; 2000: 159; Percent change[A]: 21.82. Percentage of households that moved in the last 5 years; 1990: 30.11; 2000: 34.41; Percent change[A]: 4.29[B]. Percentage of population of working age (16-64); 1990: 55.47; 2000: 55.48; Percent change[A]: 0.01. Percentage of population with a high school diploma (or equivalent); 1990: 41.51; 2000: 46.80; Percent change[A]: 5.29[B]. Percentage of high school dropouts; 1990: 20.1; 2000: 16.38; Percent change[A]: -3.72[B]. Percentage of vacant housing units; 1990: 14.37; 2000: 16.80; Percent change[A]: 2.43[B]. Average owner occupied housing value; 1990: $46,100; 2000: $61,450; Percent change[A]: 33.3[B]. Source: GAO analysis of Census data. Note: There are six census tracts in the designated area; we did not use comparison areas for rural EZs. For more information on our methodology, see appendix I. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner-occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 34: Changes in Selected Economic Growth Variables Observed in the Rio Grande Valley EZ: Number of businesses; 1995: 551; 1999: 688; 2004: 710; Percent change 1995-2004[A]: 28.86. Number of jobs; 1995: 6,025; 1999: 6,548; 2004: 7,427; Percent change 1995-2004[A]: 23.27. Source: GAO analysis of Claritas data. Note: There are six census tracts in the designated area; we did not use comparison areas for rural EZs. For more information on our methodology, see appendix I. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, EZ stakeholders suggested that EZ programs may have helped to improve residents' quality of life through programs that provided employment opportunities or taught residents skills to improve their income. One stakeholder mentioned a health clinic that was partially funded by the EZ that had helped to provide additional jobs in the area. However, another stakeholder mentioned the large number of migrant farm workers in the area make tracking these changes difficult. In terms of changes in economic growth, EZ stakeholders noted the initial lack of public infrastructure in the zone and mentioned that the EZ infrastructure development helped to prepare the area for future economic development and growth. Stakeholders credited EZ activities with helping to attract tourism to areas of the EZ and said that efforts to help businesses through revolving loan funds in some of the EZ counties had fostered economic growth. Some EZ stakeholders added that some of the growth of cities surrounding the EZ also might be due to an increase in trade across the border with Mexico. Providence, Rhode Island Enterprise Community: Figure 37: Map of the Providence EC: [See PDF for image] Source: GAO analysis of HUD data. [End of figure] How the EC Was Governed: The Providence EC was managed by the nonprofit Providence Plan and included a board called the Oversight Committee that included EC residents from each neighborhood, small business owners, and two city council members. Unlike many of the EZs, the EC allocated most of its grant funds during the strategic planning process, so there were few funds for the board to approve during the course of the program. However, in those cases, the board reviewed background information on the organizations that requested funds, discussed the applicants at their meetings, and then chose the applicants to fund. The board also reviewed the routine reporting by subgrantees and participated in site visits. Activities the EC Implemented: The Providence EC implemented four types of activities--workforce development, assistance to businesses, access to capital, and human services--most of which were related to economic opportunity (fig. 38). According to stakeholders, the EC's largest subgrantee was a community development corporation, which implemented workforce training, a summer youth program, and business development programs. It also funded the renovation and development of some small business incubators that offered space and technical assistance to new small businesses. In addition, stakeholders noted that the EC implemented some Community Opportunity Zones, which were designed to provide integrated access to education, health, and social services for families with children. An EC official explained that most of the EC funds were spent in the first 5 years of the program and that all EC funds had been spent by June 2004. Figure 38: Activities Implemented by the Providence EC: [See PDF for image] Sources: GAO (photo); GAO analysis of HUD data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: In the Providence EC, poverty and unemployment stayed about the same from 1990 to 2000 and the number of businesses and jobs decreased between 1995 and 2004.[Footnote 85] Tables 35 and 36 show the changes in poverty, unemployment, and economic growth in the EC. Table 35 also includes data on the changes in other variables included in our models of the urban EZs. Table 35: Changes in Selected Census Variables Observed in the Providence EC: Poverty rate (%); 1990: 35.36; 2000: 37.58; Percent change[A]: 2.22. Unemployment rate (%); 1990: 13.63; 2000: 11.90; Percent change[A]: - 1.73. Average household income; 1990: $28,593; 2000: $32,616; Percent change[A]: 14.07[B]. Percentage of single female headed households with children; 1990: 21.64; 2000: 21.96; Percent change[A]: 0.31. Total population; 1990: 48,789; 2000: 53,845; Percent change[A]: 10.36. Total individuals per square mile; 1990: 9,179; 2000: 10,141; Percent change[A]: 10.48. Percentage of households that moved in the last 5 years; 1990: 52.23; 2000: 50.85; Percent change[A]: -1.38. Percentage of population of working age (16-64); 1990: 55.45; 2000: 58.00; Percent change[A]: 2.55[B]. Percentage of population with a high school diploma (or equivalent); 1990: 48.69; 2000: 53.51; Percent change[A]: 4.82[B]. Percentage of high school dropouts; 1990: 26.17; 2000: 19.16; Percent change[A]: -7.01[B]. Percentage of vacant housing units; 1990: 14.55; 2000: 10.43; Percent change[A]: -4.13[B]. Average owner occupied housing value; 1990: $124,339; 2000: $116,698; Percent change[A]: -6.15[B]. Source: GAO analysis of Census data. Note: There are 13 census tracts in the designated area; we did not use comparison areas for individual ECs. For more information on our methodology, see appendix I. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner- occupied housing estimates, see appendix I. [A] Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. [B] The change in estimates from 1990 to 2000 is statistically significant. [End of table] Table 36: Changes in Selected Economic Growth Variables Observed in the Providence EC: Number of businesses; 1995: 2,714; 1999: 2,426; 2004: 2,200; Percent change 1995-2004[A]: -18.94. Number of jobs; 1995: 37,724; 1999: 34,763; 2004: 33,545; Percent change 1995-2004[A]: -11.08. Source: GAO analysis of Claritas data. Note: There are 13 census tracts in the designated area; we did not use comparison areas for individual ECs. For more information on our methodology, see appendix I. We excluded establishments that were not eligible for the program tax benefit, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, stakeholders cited several factors that they thought had influenced changes in poverty in the EC, including the increased costs of housing and utilities, growth in the foreign-born population, the loss of manufacturing jobs, and changes to welfare reform. In addition, one EC stakeholder noted that many residents were working but not earning high enough incomes to move them out of poverty. Although the EC experienced a decline in unemployment, stakeholders noted that barriers to employment remained, including limited job and language skills and records of incarceration. With respect to economic growth, EC stakeholders said that businesses began working together as a result of the EC. However, one stakeholder suggested that the EC was influenced by the slow Rhode Island economy and that the EC should have done more to foster economic growth. Fayette-Haywood, Tennessee Enterprise Community: Figure 39: Map of the Fayette-Haywood EC: [See PDF for image] Source: GAO analysis of USDA data. [End of figure] How the EC Was Governed: Three entities shared responsibility for operating the Fayette-Haywood EC. Haywood County administered the EC grant funds, a local development district was in charge of the EC's reporting to USDA, and a board that represented both counties in the EC made funding decisions.[Footnote 86] To make decisions about what activities to fund, each county held separate meetings to discuss projects that pertained to their community and sought final approval at a meeting of the full board. EC stakeholders mentioned that USDA officials played an active role in the EC and attended most board meetings. Activities the EC Implemented: The majority of the activities implemented by the Fayette-Haywood EC were in community development, mainly in the areas of health care and housing (fig. 40). In our interviews, stakeholders mentioned benefits of the EC that included health care-related activities, such as recruiting doctors and nurses to the area and the reopening a medical clinic that had been closed for 10 years. Stakeholders also noted that new housing projects had been also built with the help of EC funds. The EC also conducted activities related to public infrastructure, such as helping to build a YMCA in Haywood County and other community centers in both counties. The EC did not request a grant extension, because it had used all of its grant funds. Figure 40: Activities Implemented by the Fayette-Haywood EC: [See PDF for image] Source: GAO (photo); GAO analysis of USDA data (charts). [End of figure] Changes in Poverty, Unemployment, and Economic Growth: The Fayette-Haywood EC experienced positive changes in poverty and unemployment between 1990 and 2000 and both measures of economic growth between 1995 and 2004. Tables 37 and 38 show the changes in poverty, unemployment, and economic growth in the EC. Table 37 also includes data on the changes in other variables included in our models of the urban EZs. Table 37: Changes in Selected Census Variables Observed in the Fayette- Haywood EC: Poverty rate (%); 1990: 28.37; 2000: 19.30; Percent change[A]: - 9.07[B]. Unemployment rate (%); 1990: 9.75; 2000: 7.02; Percent change[A]: - 2.73[B]. Average household income; 1990: $32,560; 2000: $45,353; Percent change[A]: 39.29[B]. Percentage of single female headed households with children; 1990: 10.97; 2000: 11.33; Percent change[A]: 0.36. Total population; 1990: 29,080; 2000: 30,551; Percent change[A]: 5.06. Total individuals per square mile; 1990: 44; 2000: 46; Percent change[A]: 5.07. Percentage of households that moved in the last 5 years; 1990: 34.49; 2000: 36.48; Percent change[A]: 1.99. Percentage of population of working age (16-64); 1990: 55.39; 2000: 58.82; Percent change[A]: 3.43[B]. Percentage of population with a high school diploma (or equivalent); 1990: 53.57; 2000: 65.81; Percent change[A]: 12.24[B]. Percentage of high school dropouts; 1990: 18.26; 2000: 12.77; Percent change[A]: -5.49[B]. Percentage of vacant housing units; 1990: 6.46; 2000: 6.85; Percent change[A]: 0.39. Average owner occupied housing value; 1990: $68,945; 2000: $103,619; Percent change[A]: 50.29[B]. Source: GAO analysis of Census data. Note: There are eight census tracts in the designated area; we did not use comparison areas for individual ECs. For more information on our methodology, see appendix I. Differences in poverty rate, unemployment rate, and other variables shown as percentages are based upon percentage point differences. Differences for average household income, population, individuals per square mile, and average housing value are calculated as percent changes. Estimates for all census variables based on percentages had 95 percent confidence intervals of plus or minus 5 percentage points or less. For the confidence intervals for average household income and average owner- occupied housing estimates, see appendix I. [End of table] Table 38: Changes in Selected Economic Growth Variables Observed in the Fayette-Haywood EC: Number of businesses; 1995: 892; 1999: 921; 2004: 1,128; Percent change 1995-2004[A]: 26.46. Number of jobs; 1995: 9,556; 1999: 10,128; 2004: 11,240; Percent change 1995-2004[A]: 17.62. Source: GAO analysis of Claritas data. Note: There are eight census tracts in the designated area; we did not use comparison areas for individual ECs. For more information on our methodology, see appendix I. We excluded establishments that were not eligible for the program tax benefit, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [A] Differences for the number of businesses and the number of jobs are calculated as percent changes. [End of table] Stakeholder Perceptions of the Factors Influencing Changes in Poverty, Unemployment, and Economic Growth: In our interviews, stakeholders said that changes in the poverty rate may have been due to changes in demographics as higher-income residents from neighboring counties moved into the EC, which had lower property taxes. In addition, stakeholders suggested that EC residents benefited from new affordable housing partially funded by the EC. When discussing changes in unemployment and economic growth, stakeholders mentioned that one factor was the designated area's proximity to a growing city 25 miles away that provided additional job opportunities. In addition, stakeholders mentioned that the EC designation had helped the Haywood county government win grants to build infrastructure, such as a rail spur that attracted large industries to the EC. These industries offered jobs with higher wages and provided water lines with potable water for EC residents. [End of section] Appendix V: Comments from the Department of Health and Human Services: Department Of Health & Human Services: Office of Inspector General: Washington, D.C. 20201: AUG 18 2006: Mr. William B. Shear: Director, Financial Markets and Community Investment: U.S. Government Accountability Office: Washington, DC 20548: Dear Mr. Shear: Enclosed are the Department's comments on the U.S. Government Accountability Office's (GAO) draft report entitled, "Empowerment Zone And Enterprise Community Program-Improvements Occurred in Communities, but the Effect of the Program Is Unclear" (GAO-06-727), before its publication. These comments represent the tentative position of the Department and are subject to reevaluation when the final version of this report is received. The Department provided several technical comments directly to your staff. The Department appreciates the opportunity to comment on this draft report before its publication. Sincerely, Signed by: Daniel R. Levinson: Inspector General: Enclosure: The Office of Inspector General (OIG) is transmitting the Department's response to this draft report in our capacity as the Department's designed focal point and coordinator for U.S. Government Accountability Office reports. OIG has not conducted an independent assessment of these comments and therefore expresses no opinion on them. Comments Of Department Of Health And Human Services On The U.S. Government Accountability Office's (GAO) Draft Report Entitled, "Empowerment Zone And Enterprise Community Program-Improvements Occurred In Communities, But The Effect Of The Program Is Unclear" (GAO- 06-727): The Department of Health and Human Services (HHS) appreciates the opportunity to comment on the U.S. Government Accountability Office's (GAO) draft report. GAO Observations: The EZ/EC program, one of the most recent large-scale federal programs aimed at revitalizing distressed urban and rural communities, resulted in a variety of activities intended to improve social and economic conditions in the nation's high poverty communities. As of March 31, 2006, all but 1 S percent of the $1 billion in program grant funds provided to Round I communities had been expended, and the program is reaching its end. All three rounds of the EZIEC program are scheduled to end no later than December 31, 2009. However, given our findings from this evaluation of Round I EZs and ECs, the following observations should be considered if these or similar programs are authorized in the future. Based on our review, we found that oversight for Round I of the program was limited because the three agencies HHS, HUD, and USDA-did not collect data on how program funds were used, and HHS did not provide state and local entities with guidance sufficient to ensure monitoring of the program. These limitations maybe related in part to the design of the program, which offered increased flexibility in the use of funds and relied on multiple agencies for oversight. However, limited data and variation in monitoring hindered federal oversight efforts. In addition, the lack of data on the use of program grant funds, the extent of leveraging, and extent to which program tax benefits were used also limited our ability and the ability of others to evaluate the effect of the program. The lack of data on the use of tax benefits is of particular concern, since the estimated amount of the tax benefits was far greater than the amount of grant funds dedicated to the program. In response to the recommendation in our 2004 report, HUD, IRS, and USDA discussed options for collecting additional data on program tax benefits and determined two methods for collecting the information-through a national survey or the modification of tax forms. The three agencies, however, did not reach agreement on a cost- effective method for collecting the additional data. In our and others' prior attempts to obtain this information using surveys, survey response rates were low and thus did not produce reliable information on the use of program tax benefits. We acknowledge that the collection of additional tax data by IRS would introduce additional costs to both IRS and taxpayers. Nonetheless, a lack of data on tax benefits is significant given that subsequent rounds of the EZ/EC program and the Renewal Community program rely almost exclusively on tax benefits, and other federal economic development programs, such as the recent Gulf Opportunity Zone initiative, involve substantial amounts of tax benefits. Furthermore, the nation's current and projected fiscal imbalance serves to reinforce the importance of understanding the benefits of such tax expenditures. If Congress authorizes similar programs in the future that rely heavily on tax benefits, it would be prudent for federal agencies responsible for administering the program to collect information necessary to determine whether the tax benefits are effective in achieving program goals. HHS Comments: The GAO report of the EZ/EC program asserts that HHS did not provide States and designated communities with clear guidance regarding how to monitor the program. It also noted that HHS did not receive or review required financial data, nor specify or require that financial and program information should be segregated by specific activities. We believe the statement concerning monitoring unfairly represents the relationship between HHS and the other Federal agencies in the administration of the EZ/EC program. HHS is cited (on pages 20 and 21) along with HUD and USDA for not collecting data on how program funds were spent: "Although HHS regulations require States, EZs, and ECs to maintain fiscal control of program grant funds, the agency also did not provide guidance detailing the steps state and local authorities should take to monitor the program." Based on our review of a number of documents, including a Policy Memorandum from HUD dated July 16, 1996; Memoranda of Agreement between HUD and EZ/EC Communities; EZ/EC Terms and Conditions; an August 5, 1998, ACF Memorandum; and the Guidance for Auditors with respect to the EZ/EC program issued February 12, 1996, HUD and USDA are indicated as the lead Federal agencies with programmatic oversight and management responsibilities for the overall EZ/EC program, while HHS has fiscal responsibility for the EZ/EC Social Services Block Grant (SSBG) program. [End of section] Appendix VI: Comments from the Department of Housing and Urban Development: U.S. Department Of Housing And Urban Development: Washington, D.C. 20410-7000: Office Of The Assistant Secretary: For Community Planning And Development: AUG 17 2006: Mr. Charles Wilson: Assistant Director: Financial Markets & Community Investment Team: U.S. Government Accounting Office: 441 G. Street, NW, Room 2B17: Washington, D.C. 20548: Dear Mr. Wilson: Thank you for the opportunity to review and comment on the Government Accounting Office's (GAO) proposed report entitled, Empowerment Zones and Enterprise Community Program: Improvements Occurred in Communities, but the Effect of the Program is Unclear (GAO-06-727). We disagree with GAO's observations that there was an "absence of data on the use of program grant funds, the amount of funds leveraged, and the use of tax benefits." While we acknowledge that the division of program authority and responsibilities among HUD, USDA, HHS and Treasury has resulted in several unforeseen and unintended consequences, HUD collected as much data as possible on the use of program grant funds and leveraging data in the Performance Measurement System. Thank you for your work on reviewing the Round I EZs/ECs. We believe that Appendix IV provides valuable information on changes to poverty, unemployment, and economic growth rates occurring in a sample of EZ/EC designated areas as well as stakeholders' perception of factors influencing those changes. This appendix through its onsite case studies will be useful in providing Congress an understanding of how Round I EZs/ECs were governed. Since many of our enclosed comments touch upon and, in certain instances, overlap in several report areas, we decided that for the sake of clarity, we would present them by the report's four objectives: 1. Describes how the designated communities implemented Round I of EZ/ EC program. 2. Evaluates the extent of federal, state, and local oversight of the program. 3. Examines the extent to which data are available to assess the use of program tax benefits. 4. Analyzes the effects that the Round I EZs and ECs had on poverty, unemployment, and economic growth in their communities. In you have questions or would like to discuss our comments, please contact Pamela Glekas, Director, Office of Community Renewal, at (202) 708-6339. Sincerely yours, Signed by: Nelson R. Bregon: General Deputy Assistant Secretary For CPD: Enclosure: Comments to Draft Report # GAO-06-727, Entitled "Empowerment Zone And Enterprise Community Program, Improvements Occurred In Communities, But The Effect Of The Program Is Unclear" The following HUD comments on the above GAO draft report are categorized by the report's four objectives. Objective #1: Describes how the designated communities implemented Round I of EZ/EC program. * The report needs to make clear that each Round 11 urban EZs received a total amount of $25.6 million in EZ HUD grant funds plus tax incentives (much less than that of Round 1) and Round III urban EZs received no funding, but did receive tax incentives for their EZ businesses. Reference page l, end of the first paragraph of the report. * Four Key Principles: Highlighting the Round I achievements in meeting the key principles would provide Congress with a balanced view of the effectiveness of the Round I EZ/EC program that is more consistent with the statute. Reference CFR 24 Section 597.200 (c) (1)(2)(3) and (4). It would be useful for the report to address the Round I successes in meeting the four key principles of 1) economic development, 2) sustainable community development, 3) community-based partnerships and 4) strategic vision for change. Highlighting the key principles also helps to characterize in a more positive light, one of the report's conclusions that in general, EZs and ECs used program grants to implement a larger number of community development activities in areas such as education, housing and infrastructure than economic opportunity activities, such as job training and assistance to businesses. Reference page 13 of the report. The report's subtle message is that Round I EZ/ECs inappropriately focused on community development rather than economic opportunity activities. The report does not point out that the key principles allow for a wide range of SSBG supported activities to be undertaken and that the implementation of community development activities meets the principle of sustainable community development. Recommendation: Given the fact that the four key principles are central to the development of the strategic plan, HUD believes an evaluation of how well the EZ/ECs did in meeting the key principles of the strategic plan would be a significant performance standard for GAO evaluation of the Round II and Round III EZs. Reference the Omnibus Budget Reconciliation Act of 1995, Sec. 1391 (f)(2)(A) through (F), and CFR 24 Section 597.200 (c) of the Round I EZ/EC governing regulation. * Strategic Plan: Statutorily, the quality of an application's strategic plan was the essential document in the rating and ranking of EZ/EC applications and in the selection of the Round I designees. Because of the key principles significance and strong influence in the development of the strategic plan, a discussion in the report of the designees' success or lack thereof in meeting these principles would be a reasonable addition to the report. Examples of the urban Round I EZs and ECs successes in carrying out their programs are captured in HUD's 2005 publication "Spotlight on Results." Section 6 of the publication consists of interviews with key administrators of Round I urban designated areas that demonstrate the achievements of these cities in empowering their residents to act on the four key principles. Reporting on how well Round I EZ/ECs did in implementing their strategic plan is also a critical factor in measuring Round I performance although the Community Renewal Tax Relief Act of 2000, Sec. 1400J, requires a Congressional report only on the EZs/ECs impact on poverty, unemployment and economic growth. It is clear that GAO's Round I EZ/EC report evaluating the program performance in conjunction with these indices was not achievable. However, we disagree that this failure was due to a lack of program data but rather from a need for a more inclusive methodology, such as more emphasis on the strategic plan and a better method to track performance. * PERMS & Strategic Plans: In view of the GAO conclusion that it was unable to determine the impact because of the lack of financial data to effectively tie the projects/activities to the SSBG funding source, HUD recommends an approach to compensate for the lack of data on the use of program funds. This approach would be based on information emanating from the strategic plan and HUD's Performance Measurement System (PERMS), consisting of budgets, implementation plans (IPs) and annual reports. A separate discussion on PERMS is presented under Objective #2. Recommendation: HUD asserts that an assessment of designees' performance in carrying out their strategic plans be included in the reports on evaluating Round II and Round III EZs as a standard of performance. Objective #2: Evaluates the extent of federal, state, and local oversight of the program. * PERMS & Leveraging: A function of HUD's PERMS allows for ad hoc reports on a variety of funding and program analyses, including the extent and degree of leveraging occurring in Round I EZ/EC designated areas. For example in the 2001 reporting year alone, Round I urban EZs leveraged approximately $4.155 billion in non-Social Service Block Grants (SSBG) for $365 million in SSBG assisted projects in the EZ designated areas. Thus, for every $1 dollar SSBG investment there was approximately $11+ million in leveraged funds. Table #5 "Coding of Data Reliability of HUD/USDA Performance Systems." Reference page 55 of the report indicates that the systems received a code of #2.0 indicating, "items had strong documentation." This suggests a contradiction in information and that the leveraging information is more reliable than the last two sentences on page 55 suggests. Recommendation: The amount of leveraging for each of the urban Round I EZs and ECs can be tracked through PERMS. HUD would like the opportunity to demonstrate how the information is recorded and tracked in PERMS and believe that a demonstration of this system may help to alleviate GAO's concern that "reliable data on the extent of leveraging were not available." Reference page 3, "Results in Brief' and page 55 of the report. * HUD PERMS: HUD's belief is that the report did not adequately address PERMS and its role in HUD's oversight of the urban Round I EZs/ECs. PERMS primary function allows the user to enter transmit and share their program information easily and in a consistent fashion. HUD recommends that PERMS be fully addressed in the next GAO reports. Among PERMS informational and data elements are: executive summaries for field review and coordination with designees for needed review changes and additional information. For example, the summaries narrate the accomplishments and concerns in the areas of community-based partnerships, economic opportunity, sustainable community development and Round III and Renewal Community tax incentive utilization plans. Recommendation: HUD requests that the report acknowledge that HUD met its obligation to compile program data through PERMS (that includes a description of the activities implemented, program outputs and budgets for projects) and include a statement such as the following, in the final report on Round I EZs/ECs: PERMS allow Round I EZ/EC to submit annual reports on their progress in achieving goals, milestones, outputs, and implementing their activities and projects. Although PERMS is not considered a financial system per se, it is considered an informational management system consisting of elements useful for measuring program effectiveness. HUD is congressionally mandated to obtain performance reports from the EZs/ECs to evaluate their performance and to undertake program oversight/ monitoring. Furthermore, HUD notes that activities and the fiscal data discussed on page 20 of the report were HHS responsibility. Reference page 20 of the report. HUD also requests that the following statement be added to the final report: Reliability of PERMS Data: Designees are responsible for providing accurate and complete PERMS data that enables HUD to determine whether designees are carrying out its strategic plan. HUD relies on the designee's written assurances certifying that it will carry out its strategic plan in accordance with the provisions of CFR Section 597.200 (c) and (d), including a certification that the designee will provide periodic reports on the use of and how funds will be allocated/ budgeted. * Monitoring and Performance Reviews: The report criticizes the lack of monitoring guidance pointing out that to some degree the lack of reporting requirements may be an outcome of program design. This point is significant because it provides insight about the nature and extent of the federal, state, and local attitude that existed at the time of the first Round of EZ/ECs. That attitude was based on the premise of maximum local flexibility and control and limited government intrusion. Monitoring: HUD did not conduct monitoring of the SSBG funds because monitoring those funds came under the purview of HHS since this agency was responsible for the allocation, tracking and monitoring of those funds. You may recall we responded to GAO in a written document, dated December 8, 2004, on questions submitted by the GAO team. The following points were made: 1. HUD Office of Inspector General, (OIG) audited a sample of urban Round I EZs consisting of Atlanta, Philadelphia, Los Angeles, Cleveland, and Detroit. 2. HHS informed HUD that it acts as a pass through for the SSBG mandatory grants to the states and provides technical assistance on compliance. 3. HUD conducted a "standard" grants monitoring procedure consisting of a field review rating the grantees in their jurisdiction by risk analysis resulting in a selection of high risk grantees for monitoring The risk analysis for Round I would have been limited to EDI/108 funds awarded to Round II EZ which compete with all other CPD competitive grants for monitoring. 4. The HUD OIG found disallowed and questioned costs in some Round I EZ's use of SSBG funds and EDI/108 funds in audits performed in 1998- 1999 and 2002-2003. The OIG also raised issues regarding management controls, lack of progress and benefit of the program to residents. In 1997, HUD warned five EZs/ECs of possible revocation of their designation regarding deficiencies after field visits. The warnings related to adequacy of progress in implementing the strategic plans and did not involve the use of funds. 5. Round I urban challenges involved having one-agency award funds, another agency track progress and having different Federal benefits in different types of designations for which different rules apply. In the case of Round I urban communities large funding sources included SSBG funds, HUD's 108 grantees, and Economic Development Initiative funds. Periodic Performance Reviews: HUD's oversight role stems from Section 1391 (d) (2) of the Omnibus Budget Reconciliation Act of 1993 and CFR Part 597.400, 597.401, 597.402, and 597.403, all of which provide for reporting, periodic performance reviews, validation of designation and revocation. According to limited legislative provisions, performance reviews required only an evaluation of progress in carrying out an EZ/ EC strategic plan. In particular, the funding aspects of the strategic plan were confined to identifying the "funding requested" under any Federal program in support of the proposed economic, human, community and physical development and related activities. Reference Omnibus Budget Reconciliation Act of 1993, Sec. 1391, Designation Procedures. Furthermore, for the purpose of the strategic plan "funding requested" was identified as amounts budgeted for the proposed projects and activities and did not include provisions for obligations, and/ expenditures. The limited statutory provisions were the basis for determining the degree and extent for Round I EZ/EC monitoring and the related collection of data as well. HUD agrees with the GAO's report statement on page 6 urging, "more should have been done." We appreciate GAO surfacing this issue. HUD, too believes, that it is important that Congress be aware of the program's imperfections associated with the oversight of Round I EZ/EC federal expenditures, particularly when developing legislation for future programs. Recommendation: 1. The report's observation that "more should have been done" should also make clear that "more" was not allowed in Round I. 2. In conjunction with this issue, HUD requests that the report include an accompanying statement that HUD met its agency requirement to undertake periodic performance reviews on Round I urban EZ/ECs and did so based on legislative provisions. * Division of Authority and Responsibilities: As mentioned in the cover letter, a substantial flaw in the administration of the Round I was the division of authority and responsibilities among HUD, USDA, HHS, and Treasury. This division has resulted in unforeseen and unintended consequences that affected financial oversight of the Social Service Block Grant (SSBG) funds. HUD recognized in its administration of Round I urban EZ/ECs that we had no control over the State's distribution of SSBG funds leading to a vacuum of information on how the funds were being expended. After numerous attempts to collect SSBG data, it was not until in 2004/2005 that we were provided a report of the expenditures and remaining Round I urban EZ unallocated/unspent funds. * Line of Credit Control System (LOCCS) Administration: The issues originating from the division of authority among the four agencies led HUD to closely examine its in-place financial and management systems when Congress instituted a second round of urban EZs in 1998. In its direct responsibility of administering the 15 Round II urban EZs and their HUD EZ grant funds, HUD coupled PERMS and LOCCS. HUD further made changes to these systems so that Round II HUD EZ grant funds could be tracked and tied to Zone activities and projects. LOCCS Administrator: To ensure better oversight of the HUD EZ grant funds, HUD assigned a LOCCS administrator at the level of senior professional, and set forth the duties assigned to the administrator as well as Desk Officers in drawing and reconciling EZ grant funds on a monthly basis. Additionally, HUD has assigned a PERMS administrator to manage the timely execution of annual performance reviews. Another significant HUD action was developing a Round II Zone guidebook containing post designation policies and procedures with sections covering drawdowns and tracking the use of HUD EZ grants. Objective #3: Examines the extent to which data are available to assess the use of program tax benefits. * HUD and Treasury Discussions on Tax Incentives: It should be noted that HUD took immediate action in response to the GAO recommendations in the March 2004 report "Federal Revitalization Programs Are Being Implemented, but Data on the Use of Tax Benefits are Limited." The specific GAO recommendations were: 1. Identify the data needed to assess the use of the tax benefits: 2. Determine the cost effectiveness of collecting these data: 3. Document the finds of their analysis: 4. If necessary seek the authority to collect data, if a cost-effective means is available: * HUD Actions: In addressing the above, HUD initiated actions that included a May 10, 2004, meeting with Treasury and IRS managers to discuss the GAO recommendations. At that meeting, HUD suggested changes to the IRS form 8844, compiled by Zip Codes by adding a line to the form that would allow an understanding of the utilization of the EZ/ Renewal Community (RC) employment credits, the largest used incentive out of the total incentives provided in the $11 billion tax incentive package. In a follow-up email in November 8, 2004, from the IRS and in a Memorandum GAO-04-306, HUD was informed that changing form 8844 was not an option largely because of the costs associated with such a change and the number of other tax forms that would have to be changed in meeting HUD's suggestion requiring additional appropriated funds. IRS also stated in that email that changes to these forms would result in at least $1 million hours of taxpayer burden annually. The memorandum also stated that surveying business located in RCs/EZs would require that HUD request that grant recipients compile lists identifying businesses that benefited from the credits. IRS further stated, "No data are collected on additional 179 expensing and capital gains relief." * HUD Survey: The IRS response lead HUD to undertake its own survey through contractor resources. This resulted in the development of a methodology, business tax incentive questionnaire and a random sample of businesses selected from the latest Dun & Bradstreet CD ROM identifying and providing relevant information on approximately 300,000 EZ/RC businesses. Because the survey had not been included in the latest HUD budgets, the completion of the survey is on hold. HUD's Office of Policy, Development & Research have included the completion of the survey methodology as part of its 2007 research agenda. * Treasury/HUD Partnership: Through this partnership, Treasury has provided assistance and data on the use of the New Markets Tax Incentives (NMTC) and on the Commercial Revitalization Deductions available to RCs. In December 2006, Treasury will provide HUD with national level data on RC/EZ wage credits from the IRS Form 8844 relating to individual and partnership data for 2004 with corporation data expected in September 2007. Businesses benefit from the NMTC program, as private sector resources are made available to them to better meet their short and long term financial needs through increased loans and financial technical assistance. * HUD Action on Tax Incentives includes a recent HUD publication "Spotlight on Results, Capturing Successes in Renewal Communities and Empowerment Zones." This publication contains anecdotal evidence of the utilization of tax incentives by EZ/RC businesses in the areas of wage credits, expensing/deductions, bond financing, & capital gains. Recommendation: To meet the GAO recommendations made in its March 2004 report, report needs to include information on the above HUD efforts. * Round I versus Rounds II & III use of Tax Incentives: Although there was little programmatic emphasis on the use of tax incentives until the second round of EZs, there are at least two studies that addressed the likelihood and significance of their use in the first round, one GAO and the other a HUD-Policy Development and Research study. Reference pie 45 of the report. Both studies suggested significant limitations on the use of tax incentives in the first round, which led HUD to provide much more explicit emphasis on and in support of the incentives by aggressively marketing them. HUD held workshops conducted by IRS representatives and a nationally known tax attorney, national hearings by the Secretary's Advisory Council where businesses, community leaders, and designees gave testimonials on how tax incentives were being used and their economic impact in job creation and business expansion in their communities. These groups published three publications on tax incentives with the "Tax Incentive Guide for Businesses" distributing 60,000 copies nationwide, and a grass root campaign representing the best efforts by 16 EZ/RC in promoting tax incentives. * HUD asserts that the policy issue to consider in the remaining GAO reports is what impact more explicit emphasis on tax incentives has had on the Rounds II and III and on RCs. The Round I urban expectation was an assumption that the use of tax incentives would proceed more or less automatically without federal encouragement or technical support. Objective #4: Analyzes the effects that the Round I EZs and ECs had on poverty, unemployment and economic growth in their communities: * Appendix IV: We believe that the report's Appendix IV should be useful in assisting Congress to make judgments on how individual Round I EZs/ ECs did in changing poverty rates, unemployment rates and increasing and economic growth. Moreover, Appendix IV presents a valuable assessment of individual urban Round I EZs and ECs governance, the activities implemented and other related operational issues of the program. * The indices of poverty, unemployment, and economic growth were used in the application process as eligibility thresholds in identifying the most distressed communities. These indices were never intended to be used as a performance measure, except in the broadest sense of comparing indice changes over a particular timeframe. * You may want to consider HUD's parallel research and related methodology used in its econometric model. As we understand, the GAO model used 1994 to 2000 as a single time period to measure the changes in poverty, unemployment and economic growth rather than a time line of measuring change over two five-year periods, starting five years before designation (1990-1995) and five years into designation (1995-2000) as HUD's research had done. It is not surprising that it is tough to tease out changes in neighborhoods over time. The successes of EZs varied by things like staff expertise, institutionalization of administrative structure, and whether the program was government administered. For example, the groups named in Figure #5 pg. 19 (Managed by another type of organizations; Managed by local government; Managed by nonprofit organizations) might suggest a way to cut the outcome data. Reference pages 18 and 19 of the report. Theory of Change Approach: Certainly having data on the use of program grants and tax incentives would have allowed for a richer assessment of the program, seemingly as such data could be tied to the induces of poverty, unemployment and economic growth. On the other hand, it might be worth citing some other methodological issues. This was intended to be a ten-year intervention, which has been statutorily extended for another five years for the purposes of utilizing tax incentives and the remaining unspent funds. How activities are sequenced over ten years is undoubtedly important as well as the intention of policy makers. Even with so called "good data on expenditures", a longitudinal case study approach might be the best way to assess the effectiveness of this type of intervention. There are scholars who think a "theory of change" approach to this subject matter is more compelling than using an artificially constructed comparison area even in circumstances where "good data" is available. Reference pages 28-29. The following are GAO's comments on the Department of Housing and Urban Development's letter dated August 17, 2006. GAO Comments: 1. HUD commented that GAO should include details on the amount of funding and tax incentives provided for Rounds II and III of the EZ/EC program. We noted in our report that communities designated in Rounds II and III received a smaller amount of funding and more tax benefits than those designated in Round I. Our statement does not provide further details on Rounds II and III because the focus of the report is Round I. 2. We recognize that Round I designees were required to address four key principles as part of their strategic plans. However, our mandate was to assess the effectiveness of the EZ/EC program on poverty, unemployment, and economic growth. Assessing the extent to which communities addressed the key principles would not have been useful in meeting our mandate because, among other things, there is not a clear relationship between the key principles and poverty, unemployment, and economic growth. Further, while the report did not evaluate the extent to which communities met the key principles, it included many examples of activities carried out under them. The report also indicated that communities had implemented a larger percentage of community development activities than economic opportunity activities but did not comment on the appropriateness of the distribution of activities. 3. Our mandate was to assess the effects of the EZ/EC program on poverty, unemployment, and economic growth. Our report stated that communities were required to submit strategic plans that addressed the four key principles. However, because communities were able to modify their strategic plans over time, it would have been difficult to establish set criteria for assessing performance. Nonetheless, our report does contain numerous examples of activities undertaken by the communities, including examples mentioned in a separate appendix focusing on the 13 designated communities we visited. 4. HUD commented that because GAO found that a lack of data on how program funds were used was a limiting factor in determining the effectiveness of the EZ/EC program, we should make use of information in the agency's performance reporting system and in communities' strategic plans. However, we reported that our file review to determine the accuracy of data in HUD's performance reporting system found that the data were not sufficiently reliable for our purposes. For example, we found evidence that communities had undertaken certain activities with program funding, but we were often unable to find documentation of the actual amounts allocated or expended. As a result, we were unable to rely on information contained in the agency's performance reporting system on the amounts of program funds allocated or expended on specific activities. 5. We found that data in HUD's performance reporting system on the amounts of funds used and the amounts leveraged were not reliable. For example, we found that HUD's system included information on the amount of funds leveraged. But for the sample of activities we reviewed, the supporting documentation either showed an amount conflicting with the reported amount or was not available. Moreover, we found that the definition of "leveraging" varied across EZ and EC sites. HUD further commented that Table 5 in the report showed that the agency's performance reporting system received a code of 2.0, showing that leveraging data had strong documentation. However, HUD appears to have misinterpreted the information we presented on this matter. We found that HUD's data on leveraging received an average code of 1.0, indicating that such information had weak documentation. Lastly, HUD recommended that it be allowed to alleviate GAO's concerns about the reliability of its leveraging data by demonstrating how the data were tracked and recorded in its performance reporting system. However, the data reliability problems we found during the course of this work were due not to concerns about the system used to track and record the data, but rather to the frequent lack of supporting documentation for the data entered into the system. 6. HUD commented that our report did not adequately address HUD's performance reporting system and its role in HUD's oversight of the urban Round I EZ and ECs. We acknowledge that HUD established the system in response to an earlier GAO recommendation and has since used it to oversee Round I EZs and ECs. Moreover, we agree that the system contains a variety of information and data elements, including activities implemented and program outputs. We also acknowledge that the performance reporting system is not intended to be a financial system for Round I. However, as discussed in our report, we found that because the system did not always contain information on what was spent on activities and did not always contain reliable information, HUD and the other federal agencies were limited in their ability to oversee the program. 7. HUD commented that the program's design was significant because it provided insight about the nature and extent of the federal, state, and local attitudes that existed at the time of the first Round of EZs/ECs. HUD also stated that it did not conduct monitoring of the SSBG funds because monitoring those funds was the responsibility of HHS. HUD's statement further supports our discussion on the limitation in the oversight of the EZ/EC program that may have resulted from the program's design. Although we found program oversight was hindered, we also reported that no single federal agency had sole responsibility for oversight. We do not agree with HUD's recommendation that we make clear that more oversight was not allowed in Round I. For example, early in the program HUD and HHS made some efforts to share information. Specifically, HUD officials said that they had received fiscal data from HHS and reconciled that information with their program data on the activities implemented, but these efforts to share information were not maintained. Regarding the second recommendation, although HUD described some of its efforts to monitor the program according to applicable regulations, the oversight concerns we identified in the report remain. 8. We reported that limitations in the oversight of the EZ/EC program may have resulted from the design of the program. 9. We stated in the report that the concerns raised about program oversight for the Round I EZ/EC program may not apply to future rounds of the EZ/EC program. We also acknowledge that HUD may have made changes in its oversight of later rounds of the program. However, an evaluation of later rounds of the EZ/EC and Renewal Community programs is beyond the scope of this report. 10. In our report, we acknowledged HUD's as well as the other agencies' response to the recommendation in our 2004 report to identify a cost- effective means of collecting the data needed to assess the use of the tax benefits. 11. Our report acknowledged the collaboration among HUD, IRS, and USDA in addressing our previous recommendation and summarizes the outcome of their discussions, including the identification of two data collection methods--through a national survey or by modifying the tax forms. In addition, our report also acknowledged that IRS did not have any data for some program tax benefits. The lack of data on the use of tax benefits continues to be a source of concern that limits an assessment of the effect of the EZ/EC program. 12. We agree that HUD's efforts to develop a methodology to administer a survey to businesses to assess the use of the program tax benefits is a useful step in gathering such information. 13. We recognize the efforts between HUD and Treasury on sharing national-level data on EZ businesses' use of tax credits for employing EZ residents. However, as we mention in our report, data on the EZ employment tax benefit were limited because they could not be linked to the specific EZ claiming the benefit. 14. In the absence of other data, we acknowledge HUD's efforts to capture anecdotal information on the use of program tax benefits by EZ businesses. 15. We recognize HUD's efforts to market the EZ/EC program tax benefits. 16. We appreciate HUD's suggestion on how to approach evaluations of later rounds of the EZ/EC and Renewal Community programs and welcome the opportunity to discuss these ideas. 17. We appreciate HUD's comments on the descriptive information on EZs and ECs we visited that are discussed in appendix IV. 18. HUD commented that the measures used in our report---poverty, unemployment and economic growth--were used in the application process and were not intended to be used as performance measures. However, as mentioned earlier our mandate was to assess the effects of the EZ/EC program on poverty, unemployment, and economic growth. 19. HUD suggested that we consider additional methodologies for measuring the effects of the EZ/EC program, such as trend analysis using data from 1990 through 1995 and 1995 through 2000. To conduct our work, we used 1990 and 2000 data to measure changes in poverty and unemployment and 1995, 1999, and 2004 data to measure changes in economic growth. We chose these dates because data were available at the census tract level for these years. Moreover, in designing our methodology for our econometric analysis, we conducted a literature review and discussed our methodology with several experts in the urban studies field and determined that the approach presented in this report was effective in answering the objectives of our mandate. As mentioned in Appendix II, we also conducted different tests to ensure the robustness of our models, which all yielded results consistent with our models. The approach that HUD suggested controlled for trends that began before the EZs were designated in 1994. Because we did not have data on poverty or unemployment for 1995 we were unable to use this approach. However, our use of housing trends between 1990 and 1994 in our econometric model controlled for some trends that were in place prior to EZ designation. HUD also suggested a longitudinal case study approach might be the best way to assess the effectiveness of this type of program. Although a longitudinal case study approach would be informative, it is unlikely that a successful retrospective longitudinal study could be designed at the end of the program. As HUD noted, this intervention was intended to be implemented over a ten-year period. However, a longitudinal case study approach would necessitate data collection beginning at the inception of the program and continuing for the duration of the program as well as some period of time after it ends. [End of section] Appendix VII: Comments from the U.S. Department of Agriculture: United States Department of Agriculture: Office of the Secretary: Washington, D.C. 20250: AUG 22 2006: William B. Shear: Director, Financial Markets and Community Investment: United States Government Accountability Office: 441 G Street, NW, Room 2A10: Washington, DC 20548: Dear Mr. Shear: Thank you for providing the United States Department of Agriculture (USDA) and Rural Development with your Government Accountability Office (GAO) draft report entitled, "Empowerment Zone and Enterprise Community Program: Improvements Occurred in Communities, but the Effect of the Program is Unclear," Report Number GAO-06-727. For your consideration, Rural Development offers the following comments to the draft report and requests that a copy of these comments be included in your final report. Rural Development's response is limited to a discussion of the rural Empowerment Zones and Enterprise Communities (EZ/ECs). USDA concurs that data and analyses on the effectiveness of programs, such as the EZ/EC, are useful. Our long experience working with rural communities, operating economic development programs, and attempting to evaluate those programs may be instructive to future evaluations. Rural Development suggests three general areas to consider: First, rural communities are substantially different from each other and from urban and metropolitan areas, and many performance metrics have been designed with the urban model and data in mind. Rural measurements need to acknowledge different realities. For instance, small populations mean that some statistically meaningful economic, demographic, and education data are not always available in targeted rural communities. Currently, USDA is developing a methodology that focuses on economic impacts using county-level economic data and captures the short-term Gross Domestic Product changes in the impacted rural counties. However, this system can not be applied retroactively to communities receiving grants or other economic stimulation in the past. In addition, this system can not be used to measure impacts at the sub-county level. Second, it is especially important in rural areas to build in up front a clear and adequately funded data collection process for program evaluation. When that data collection process is not in place at the start, the appropriate resources and attention required will often be difficult to obtain. This problem is likely to be particularly acute in rural areas, where local governments and key economic development institutions frequently have limited financial capacity and professional staff. The collection of baseline and on-going data should enable construction of a program evaluation that will examine the program's success in achieving intended results. Beyond data on the participating communities, if a comparison is desired, the appropriate data should be obtained for comparison communities from the outset. Third, performance metric design is dynamic, and Rural Development encourages an expansion of the discussion, as well as the use of appropriate methodologies that recognize data conditions at the time grants are made. The new collaborations at the local level among key economic and community development institutions, which were required by the EZ/EC program, are worthy of careful examination within a full program evaluation. We commend GAO for looking at program impacts in critical areas such as poverty, unemployment, and economic growth. Program evaluation might, however, go beyond these measures to look at the construction of stronger economic development capacity and more effective collaborative networks. Such a broader perspective on program results might be particularly pertinent for rural EZ/ECs since GAO's econometric analysis was applied only to urban EZ/ECs. A major portion of the report is devoted to comparing the change in the number of jobs and businesses between EZ/ECs and comparison areas. We do not have a good understanding of how these comparison areas were chosen; therefore, we can not comment on this portion of the report. GAO's report states on page 4, "Data were not collected on program benefits for specific activities." and also ".USDA - did not collect data on how program funds were used." USDA did, however, design and implement a Benchmark Management System (BMS), intended as a management tool for both USDA and the individual EZ/ECs. The BMS requires each of the EZ/EC's major activities to be tracked over time by the community and verified by USDA. The BMS was not designed to be an accounting tool, but it has proven useful for providing a good picture of each community's achievements. GAO also states in its report that, "USDA encouraged rural EZ and ECs to report all investment in the EZ as leveraged funds, not only those projects that received EZ/EC funds." In fact, USDA encouraged each community to report all investments that contributed to accomplishing the community's strategic plan, the guiding document for the community's revitalization strategy. Most EZ/ECs developed ambitious plans that could be realized only by leveraging resources beyond the core EZ/EC funding. In addition to the resources marshaled through leveraging, the communities were able to build a strong network of partners, with a collaborative track record that will likely be in place long after the last core EZ/EC dollar has been spent. There are also some technical details that need correction and/or further explanation: * Page 10. The statement, "In addition, all designated communities reported leveraging additional resources, though a lack of reliable data prevented us from determining how much." USDA's BMS does include data on leveraged resources, as self-reported by the communities. Anecdotal evidence indicates that communities generally underreport such leveraged resources. * Page 64. The first paragraph concludes, ".the results did not allow us to conclude whether there is an association between the EZ program and economic growth." The word "urban" should be inserted before the EZ, as the econometric analysis only dealt with urban EZ and ECs. As GAO notes in its report, there are differences in how the program was implemented and the results achieved. Rural EZ and ECs generally had a greater per capita number of jobs created and new businesses formed than urban EZ and ECs. * Page 133. Rio Grande Valley EZ, Texas, received an extension, according to information USDA received from the Department of Health and Human Services. Thank you for this opportunity to comment on the report. If you have any questions, please contact John Dunsmuir, Acting Director, Financial Management Division, at (202) 692-0080. Sincerely, Signed by: Thomas C. Dorr: Under Secretary Rural Development: [End of section] Appendix VIII GAO Contact and Staff Acknowledgments: GAO Contact: William B. Shear (202) 512-8678: Acknowledgments: In addition to the individual named above, Charles Wilson, Jr., Assistant Director, Carl Barden, Mark Braza, Marta Chaffee, Emily Chalmers, Nadine Garrick, Kenrick Isaac, DuEwa Kamara, Austin Kelly, Terence Lam, John Larsen, Alison Martin, Denise McCabe, John McGrail, John Mingus, Jr., Marc Molino, Gretchen Maier Pattison, James Vitarello, and Daniel Zeno made key contributions to this report. (250221): FOOTNOTES [1] Since its enactment in 2000, the Renewal Community program has focused on providing tax benefits to businesses in designated communities to attract or retain jobs and businesses. [2] For the purposes of this report, EZ/EC stakeholders include EZ/EC officials, board members, subgrantees, local chamber of commerce representatives, and other local officials recommended by EZ or EC officials as having a role in the program. See appendix I for information on the types of stakeholders we interviewed at each site. [3] Survey results can be viewed at [Hyperlink, http://www.gao.gov/cgi- bin/getrpt?GAO-06-734SP]. [4] Program data include information on how activities were implemented and outputs. [5] We calculated confidence intervals around estimates derived from Census data. In this report, all estimates shown for a percentage have a 95 percent confidence interval of less than plus or minus 5 percentage points, unless otherwise noted. [6] GAO, Standards for Internal Control in the Federal Government, GAO/ AIMD-00-21.3.1 (Washington, D.C.: November 1, 1999). [7] HUD's performance reporting system is known as Performance Measurement System while USDA's is called the Benchmark Management System. For the purposes of this report, we refer to them as performance reporting systems. [8] GAO, Government Performance and Accountability: Tax Expenditures Represent a Substantial Federal Commitment and Need to Be Reexamined, GAO-05-690 (Washington, D.C.: September 23, 2005). [9] GAO, Community Development: Federal Revitalization Programs Are Being Implemented, but Data on the Use of Tax Benefits Are Limited, GAO- 04-306, (Washington, D.C.: March 5, 2004). [10] Our earlier recommendation was not directed to HHS because its role was limited to distribution and oversight of the EZ/EC grant funds. [11] We were able to use statistical modeling techniques for the eight Round I urban EZs only, because the rural EZs were made up of too few census tracts to perform these analyses, and because the ECs received such small amounts of money over the 10-year period that we could not separate the program's effects from those of other programs. [12] GAO-04-306. [13] Multiagency teams also included officials from the Department of Justice, the Environmental Protection Agency, and the Small Business Administration, among other agencies. [14] Two urban EZs--Cleveland and Los Angeles--were originally designated as Supplemental EZs and received a combination of Economic Development Initiative grants and Section 108 Loan Guarantees, both of which could only be used for certain economic development or revitalization projects. The Supplemental EZs were given full Round I EZ status in 1998, and local businesses were allowed to use the program tax benefits starting in 2000. However, they did not receive the grants the other Round I EZs received. Four urban ECs also received Enhanced EC designations, which provided them with some Economic Development Initiative grants and Section 108 Loan Guarantees. [15] This number represents businesses' use of tax benefits in all Round I EZs and ECs, as well as 20 Round II EZs, 10 Round III EZs, and 40 Renewal Communities. [16] The Cleveland EZ received $87 million in grants and an equal amount of loan guarantees, while the Los Angeles EZ received $125 million in grants and $325 million in loan guarantees. [17] EZs and ECs self-determined the categories for their activities, so it is possible that a similar activity implemented in two sites could be categorized differently. [18] The federal Community Development Block Grant program supports a wide array of local community development activities that are primarily to benefit low-and moderate-income individuals. [19] However, past reported amounts remain in their performance data in HUD's system. [20] This finding is consistent with findings of the HUD and USDA Offices of Inspector General. [21] GAO/AIMD-00-21.3.1. [22] GAO/AIMD-00-21.3.1. [23] Because data on the amount of funds used for specific activities was not reliable, this report only includes information on the number of activities implemented. We were able to find complete documentation for our sample of activities for the amount of EZ grant funding reported by the Baltimore and Detroit EZs and the Camden portion of the Philadelphia-Camden EZ. [24] Neither HHS nor USDA officials told us that they had made any efforts to reconcile fiscal and program data on the EZ/EC program. [25] GAO-05-690. [26] GAO-04-306. [27] GAO-05-690. [28] GAO-04-306. [29] GAO-05-690; GAO, 21ST Century Challenges: Reexamining the Case of the Federal Government, GAO-05-325SP (Washington, D.C.: February 1, 2005); and GAO, Tax Policy: Tax Expenditures Deserve More Scrutiny, GAO/GGD/AIMD-94-122 (Washington, D.C.: June 3, 1994). The New York Liberty Zone was created in response to the September 11, 2001 terrorist attacks and created seven tax benefits designed to assist in the recovery efforts, including a tax credit for hiring employees who are Liberty Zone residents. [30] GAO-04-306. Our earlier recommendation was not directed to HHS because its role was limited to distribution and oversight of the EZ/EC grant funds. [31] GAO, Community Development: Businesses' Use of Empowerment Zone Tax Incentives, GAO/RCED-99-253 (Washington, D.C.: September 30, 1999) and Scott Hebert and others, Interim Assessment of the Empowerment Zones and Enterprise Communities (EZ/EC) Program: A Progress Report, prepared for the U.S. Department of Housing and Urban Development (Washington, D.C.: November 2001). [32] GAO/RCED-99-253 and Hebert and others, Interim Assessment. [33] For more information on our survey methods, see appendix I. [34] GAO/RCED-99-253 and Hebert and others, Interim Assessment. [35] The EZ/EC tax benefits were nonrefundable, which means that a business can only claim them if it has a tax liability. [36] An Industrial Revenue Bond is a bond used to finance the construction of manufacturing or commercial facilities for a private user. This bond is available to businesses in general and is not part of the EZ program benefits package. [37] A bond cap is the maximum dollar amount available for issuing the tax exempt bond. Prior to January 1, 2002, the cap for bonds was $3 million per borrower for activities in any EZ or EC, with a nationwide limit of $20 million per borrower. After 2002, the bond cap was removed for all EZs. [38] GAO-05-690 and GAO-04-306. [39] The Joint Committee on Taxation estimated that businesses in Round I communities would claim about $2.5 billion in tax benefits in the first four years of the program alone. Information for later years of the Round I program is not available because subsequent estimates by the committee did not separate benefits claimed by Round I communities from benefits claimed by later round communities. [40] The Gulf Opportunity Zone Act of 2005 provides tax benefits to assist in the recovery and rebuilding of areas affected by Hurricanes Katrina, Rita, and Wilma. [41] GAO-04-306. [42] The indicators of poverty, unemployment, and economic growth were specifically identified in our mandate. [43] We selected comparison areas through a statistical technique called the propensity score, which allowed us to identify the census tracts that were most similar to the census tracts selected in the original EZ and EC designations based on a set of factors, such as 1990 poverty and unemployment rates. For more information on how we chose the comparison areas, see appendix I. [44] Gentrification or economic displacement refers to the transformation of a relatively low-income neighborhood into a more affluent neighborhood through redevelopment, usually in conjunction with changing demographics and an influx of wealthier residents. [45] Our analysis of changes in economic growth only included 94 ECs, since we did not have data for the Miami/Dade County, Florida EC. [46] Changes in poverty rate are based upon percentage point differences. [47] We were able to use statistical modeling techniques only for the eight Round I urban EZs, because the rural EZs were made up of too few census tracts to perform these analyses and because the ECs received such small amounts of money that we could not separate the program's effects from those of other programs. In addition, we could not isolate the effects of the program on individual EZs because the number of census tracts in some urban EZs was not large enough to provide reliable results. Appendix II describes our methodology and the results of the econometric analysis. [48] These stakeholders did not comment on the changes that occurred in our comparison areas. [49] The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 instituted a change in welfare policies, establishing work requirements and time limits for receiving benefits. [50] Changes in unemployment rate are based upon percentage point differences. [51] These stakeholders did not comment on the changes that occurred in our comparison areas. [52] We obtained these data for 1995, 1999, and 2004 from a private data vendor, Claritas. Changes in the number of businesses and number of jobs are based upon percent changes. [53] Data were not available for the Miami/Dade County, Florida EC. [54] These stakeholders did not comment on the changes that occurred in our comparison areas. [55] GAO-04-306. [56] GAO-04-306. [57] GAO-05-690 and GAO/GGD/AIMD-94-122. [58] Of the 61 ECs that had not received additional designations, we selected a judgmental sample of 8 ECs, 4 urban and 4 rural, that represented geographic diversity and also factored in each EC's combined poverty and unemployment rates in 1990. From this sample, we chose two ECs to visit--Providence, Rhode Island (urban) and Fayette- Haywood, Tennessee (rural)--that were similar in terms of their poverty and unemployment rates. [59] In the New York and Philadelphia-Camden EZs, we implemented two separate site visit protocols, due to the split governance structures in those locations. [60] The four urban pretests were conducted in the Akron, Ohio EC; Bridgeport, Connecticut EC; Providence, Rhode Island EC; and the Albuquerque, New Mexico EC. The two rural pretests occurred in the Central Appalachia EC located in West Virginia and the Williamsburg/ Lake City EC in South Carolina. [61] There were 61 ECs that did not receive subsequent designations, but we excluded 1 urban EC from our sample because it was no longer in operation as of June 2005. [62] We selected the lesser of 10 activities or half of an EZ's or EC's activities using a systematic random sample. We use the term "activity" to describe the units of information reported in the HUD and USDA systems, called "implementation plans" in HUD's system and "benchmarks" in USDA's system. [63] We excluded the Cleveland and Los Angeles EZs from our discussion, because they did not receive EZ/EC grant funds. [64] GAO/RCED-99-253 and Hebert and others, Interim Assessment. [65] Businesses in the Atlanta EZ were excluded from the analysis, since the same area received a Renewal Community designation in 2002 and we did not want businesses to confuse the different tax benefits available for each designation. [66] We also excluded private businesses that were not eligible for the tax benefits, such as gambling establishments and liquor stores. [67] The points in time for Census data were 1990 and 2000 and for the business data were 1995 and 2004. [68] We determined this definition based on research done on the EZ program and similar programs, such as the state Enterprise Zone initiative. Similar to the EZ program, state Enterprise Zones are state run programs that offer certain tax benefits within an established area. We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those businesses in our analysis of the change in the number of jobs. [69] Other researchers have used similar approaches, such as Greenbaum and Engberg in "The Impact of State Urban Enterprise Zones on Business Outcomes," December 1998, Carnegie Melon University. p. 24. [70] Generally, increasing the propensity score for selecting comparison tracts has the effect of reducing the sample of comparison tracts and decreasing the propensity score has the effect of increasing the sample of comparison tracts. [71] Due to changes in the census tract boundaries for some EZs and ECs from 1990 to 2000, we used the 2000 census block group to recreate the initial 1990 boundaries and ensure that our analysis remained consistent. For the census variables based on percentage characteristics, like poverty and unemployment rates, we calculated the change in percentage points by finding the difference between the 1990 sample estimate and the 2000 sample estimate. For census variables based on average characteristics, such as average household income, we calculated the percent change by finding the difference between the 1990 sample estimate and the 2000 sample estimate and then dividing the difference by the 1990 sample estimate. Census variables using dollar amounts like average household income were adjusted for inflation to 2004 dollars. We calculated the percent change for our economic growth measures. [72] We completed econometric analyses of the eight urban EZs only, because the amount of program grant funding for ECs was too small to separate the program's effects from other programs. In addition, we excluded the rural EZs because they are made up of too few census tracts to perform these analyses. [73] We excluded establishments that were not eligible for program tax benefits, such as nonprofit and governmental organizations, from our analysis of the change in the number of businesses. However, we included jobs at those establishments in our analysis of the change in the number of jobs. [74] We also tested use of a fixed effect model, which allowed us to account for some tract-specific factors that may not vary over time but might be correlated with the designation of a tract, called "fixed effects." Other researchers used fixed effects regression techniques to control for area-specific unchanging factors, such as industry composition. (See, for example, Leslie E. Papke, "What Do We Know About Enterprise Zones?" NBER Working Paper No. 4251. Cambridge, Mass., National Bureau of Economic Research, 1993.) The analysis in this report predicts first differences in the dependent variables, that is, the difference between the value in the 2000 Census and the value in the 1990 Census. When only 2 years of data are analyzed, regressions based on first differences are equivalent to fixed effect models. (See Zvi Griliches and Jerry Hausman, "Errors in Variables in Panel Data," Journal of Econometrics, 31 (1986), pp. 93-118.) [75] Paul A. Jargowsky, Stunning Progress, Hidden Problems: The Dramatic Decline of Concentrated Poverty in the 1990s, (Washington, D.C.: The Brookings Institution, May 2003). [76] We also tested the models using the longer time period of 1995 to 2004, but the results were consistent with those using the time period from 1995 to 1999. [77] This low explanatory power is indicated in the low R-square statistics in tables 10 and 11. [78] In 2002, the Atlanta EZ was designated as a Renewal Community, and by 2003 the Atlanta EZ was no longer in operation. HHS transferred the remaining EZ/EC funds (over $53 million) to the Renewal Community, but as of March 2006, these funds had not been used. [79] Congress established the HOPE VI program in 1992 to revitalize severely distressed public housing by demolition, rehabilitation, or replacement of sites. [80] The overarching board also included a nonvoting HUD official. [81] The Camden portion of the EZ was initially managed by the city of Camden, but HUD officials fostered the change to a nonprofit to deal with tensions between the city of Camden and state of New Jersey. [82] All tracts that qualified as comparison tracts for Philadelphia- Camden were located in Philadelphia. [83] The grants and loan guarantees the EZ received could only be used for certain economic development or revitalization projects. When Cleveland received the Supplemental EZ instead of the regular EZ designation, the officials modified their strategic plan. [84] When Los Angeles received the Supplemental EZ instead of the regular EZ designation, the officials modified their strategic plan from having some social service initiatives to focusing only on activities directly related to economic development. [85] We did not use comparison areas for individual ECs. For more information on our methodology, see appendix I. [86] Local development districts are unique to the state of Tennessee, and the majority of their funding comes from the Tennessee General Assembly. GAO's Mission: The Government Accountability Office, the investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through the Internet. GAO's Web site ( www.gao.gov ) contains abstracts and full-text files of current reports and testimony and an expanding archive of older products. The Web site features a search engine to help you locate documents using key words and phrases. You can print these documents in their entirety, including charts and other graphics. Each day, GAO issues a list of newly released reports, testimony, and correspondence. GAO posts this list, known as "Today's Reports," on its Web site daily. The list contains links to the full-text document files. To have GAO e-mail this list to you every afternoon, go to www.gao.gov and select "Subscribe to e-mail alerts" under the "Order GAO Products" heading. Order by Mail or Phone: The first copy of each printed report is free. Additional copies are $2 each. A check or money order should be made out to the Superintendent of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a single address are discounted 25 percent. Orders should be sent to: U.S. Government Accountability Office 441 G Street NW, Room LM Washington, D.C. 20548: To order by Phone: Voice: (202) 512-6000: TDD: (202) 512-2537: Fax: (202) 512-6061: To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov Automated answering system: (800) 424-5454 or (202) 512-7470: Public Affairs: Jeff Nelligan, managing director, NelliganJ@gao.gov (202) 512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149 Washington, D.C. 20548:

The Justia Government Accountability Office site republishes public reports retrieved from the U.S. GAO These reports should not be considered official, and do not necessarily reflect the views of Justia.