Effect of Nonemployment Income and Wage Rates on the Work Incentives of the Poor: Reply

Abstract

N and those without are identical (Green and Tella, p. 404). At the very minimum, the results of the Green and Tella study should have been tested for their sensitivity to this assumption. Going beyond the conceptual issues discussed above, we now turn to a criticism of the method Green and Tella used in estimating the substitution and income effects they posited. We recall that we are dealing with a income tax package, i.e., an income grant which has an income effect (as N does), and a negative tax rate on income, which reduces the net wage rate and thereby has both an income effect and a substitution effect. Consider first the substitution effect. Green and Tella want to rationalize AH He2H1 as the substitution effect which results from Lw= (w2 -w1). They do not have an observed final equilibrium associated with Y1 = W1, but instead one associated with Y W1 + N, that is H1. However, this equilibrium (at point F in their figure 2) is at a higher level or well-being than that of the initial (pre-wage decline) equilibrium (at point E). This must be so because Y is the same at E as at F, but leisure is greater at F than at E. Thus, the substitution effect they have measured is overcompensated.6 Consider next Green and Tella's measure of the income effect. They observe a new set of hours, assume wage rates are equal, ascribe the observed difference in hours to a difference in N, and identify this difference with the income effect of a tax plan. Their method, however, involves an upward bias because, while the income grant portion of the NIT package produces an income effect of increased leisure, the attendant net wage rate decline involves an income effect toward decreased leisure (which their measure does not include). In concluding, we note that Green and Tella explicitly rejected a priori the possibility of a positive labor supply response to a lump-sum income supplement. But if the poor are allowed to share in the concepts of a minimum income and minimum leisure, then their labor supply response may well be positive, as Leuthold earlier suggested and preliminary results from the New Jersey NIT experiment have documented.7