Methods the Bureau of Labor Statistics Uses to Arrive at Unemployment Percentages

Gao ID: HRD-77-39 January 26, 1977

Actual and probable inaccuracies are contained in the data and mathematical relationships the Bureau of Labor Statistics uses to arrive at unemployment percentages.

The Bureau of the Census makes a monthly survey of about 47,000 households throughout the country which provides statistically valid monthly and annual estimates of national unemployment. The survey data are statistically valid for estimating average annual unemployment for 27 large states and the 30 largest metropolitan areas. Inaccuracies in the data and mathematical relationships used to develop unemployment estimates involve: (1) handbook estimates for state and local areas on a pro-rata basis; (2) state counts of unemployment insurance claimants; and (3) estimates of the noncovered unemployed. Additional criticisms concern the definition of the unemployed. Current or proposed actions of the Department of Labor include expansion of the population survey and improvement of state unemployment insurance data.



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.