Let Y=(Y 1 ,Y 2 , .Yn) -0 be an income distribution pattern ton-"income receiving units" which may be n-persons, n-farnilies, n-states of the same country or n-countries.Abstractly, a fiscal program is a course of action, undertaken by social conseasus, under which portions of the incomes of certain receiving units are transferred to other receiving units to render the income distribution more equitable.The most familiar example of such a fiscal program is the collection of taxes from individuals (or individual families) with the revenue being paid out as.welfare payments by the government.As another example, the Federal government may collect taxes from the states only to give some of the revenue back to the states under a "revenue sharing" program.An international consortium or the World Bank may work out a formula under which contributions will be solicited from the wealthy countries or "donors" to provide foreign aid or make concessionary loans to the poor countries.This paper is concerned with the principles governing the design of such equity oriented fiscal programs.The first general principle concerns the "rationality" of the fiscal program.Suppose the income level of "i" is higher that that of "j".On the one hand, a principle of "minimally progressive" suggests that, in case "i" and "j" are taxpayers, "i" should pay no less taxes than "j" and, in case "i" and "j" are recipients of welfare payments "i" should receive no more than "j".On the other hand, a principle of "incentive perservation" suggests that the disposable income of "i" should be no less than that of "j"--i.e. the fiscal program clearly should not reverse their relative income ranks in order to preserve the incentive for the individuals .,. ..

This paper presents a precise characterization of the bias of least squares in two limited dependent variable models, the Tobit model and the truncated regression model. For the cases considered, the method of moments can be used to correct the bias of OLS. For more general cases, the results provide approximations which appear to be relatively robust. 13, and o, . In this paper we present a precise characterization of that bias for the particular case in which xt, as well as -,, is normally distributed. We also show that the bias of the OLS slope estimator can be corrected by dividing each estimate by the sample proportion of nonlimit observations. Other structural parameters can be consistently estimated in a similar fashion. We present some evidence on the effect of nonnormality with respect to the predictions obtained in the normal model. The case in which the sample contains only nonlimit observation (the truncated regression model) is considered elsewhere (Olsen (7)). We analyze the relationship between his results and ours, and derive some predictions of the normal model with respect to the seriousness of truncation bias.