QUARTERLY JOURNAL OF ECONOMICS
work of 1.07 for any employment during the year and 0.83 for work in an average week. We also estimated specications with separate coefcients on state and federal income taxes, although for brevity these full estimates are not reported here. The results for federal taxes were similar to all taxes, while the derivative (standard error) for state income taxes was a large and signicant
0.0336 (0.0083) in the ORG sample and a smaller and insigni-
0.0165 (0.0139) in the March sample. Thus, while the state
tax estimates are much less precise and differ in the two samples, they give the same message as the other tax coefcients; i.e., that
the labor supply of single mothers responds to taxes.
The full sample specications of columns (1) and (5) also indicate substantial effects of welfare on employment. A one thousand dollar reduction in the annual Welfare Maximum Bene-
t (the AFDC plus Food Stamp benet a women receives if she
does not work) increases employment last week by 3.4 percentage points, and increases employment last year by 3.0 percentage points. This calculation holds constant the other welfare vari- ables, Welfare Benets if Work and Probability of AFDC Receipt if Work, that generally change with the maximum benet. The Welfare Benets if Work effect is sizable, implying that a one thousand dollar increase in benets when one works increases employment last week by 7.2 percentage points and last year by 5.7 percentage points. These estimates suggest substantial posi- tive employment effects of reductions in implicit tax rates and increases in earnings disregards.
The transaction costs or stigma of welfare receipt as mea- sured by the Probability of AFDC Receipt if Work variable is negative and signicantly different from zero as expected (see equation (5)). The magnitude of this coefcient can be gauged by comparing it with the coefcients on the variables denominated in thousands of dollars. Such comparisons suggest a transaction cost of several thousand dollars, with the exact number depend- ing on the employment measure and the income variable used. For example, using the Welfare Benets if Work coefcient in the ORG sample yields a transaction cost estimate of $2571, while the March sample implies an estimate of $3051. This result agrees with past studies as well as ethnographies that have tended to nd substantial transaction costs or stigma of welfare receipt.