ABSTRACTEach year Congress allocates billions of dollars to states on the basis of unemployment. The stated objectives include assisting unskilled, economically disadvantaged, and dislocated workers; yet unemployment is a poor proxy for identifying such individuals. This paper examines the consequences of altering allocation formulas and concludes that switching from unemployment statistics to alternative measures of need, e.g., low wages, low income, or poverty, would have a dramatic impact on the geographic distribution of funds. Even switching from one unemployment statistic to another (e.g., from aggregate unemployment to excess unemployment or long‐term unemployment) would substantially alter the distribution of fund