Measuring the Well‐Being of Farm Households: Farm, Off‐Farm, and In‐Kind Sources of Income: Discussion

Abstract

The papers presented at this session reflect a slow but steady evolution in our thinking and our analytical approaches to the economic well-being of households. Twelve years after Becker posited the full-income constraint, Gronau and Wales and Woodland fleshed out its definition to include all money income, the value of household production, and the opportunity cost of leisure time. Most empirical models still do not explicitly estimate the latter, but estimating and adding the value of household production to money income is becoming a common practice where the wellbeing of households is at issue. All three papers in this session identified and measured some type of nonmoney income which increased the well-being of individual households and changed their relative position in the income distribution. The studies by Ahearn and Johnson and Bryant and Zick both estimated the value of household services, though in different ways. They both estimated Gini coefficients for the distribution of money income and for a (quasi-) full income. The latter was found to be distributed more evenly than money income. Household production contributed greatly to the well-being of households, especially in rural areas. Bryant and Zick's work focused on the distributional impacts of household production activities by white rural, and urban, husbands and wives in 1976 and 1980. Ahearn and Johnson's households were headed by farm operators regardless of residence, race, or marital status. The different sampling criteria help explain the lower Gini coefficients among rural households in the Bryant and Zick study (.31 to .24). Ahearn and Johnson's sample included several households with negative incomes from large losses in farm business income. They contributed mightily to a higher Gini coefficient (.60) depicting more inequality among farm operators than rural households in general. Differences in the estimated Gini coefficients in these two papers serve to caution us about the ambiguities of this measure and its underlying assumptions. This measure of inequality assumes that the underlying Lorenz curves do not cross. Noncrossing Lorenz curves are assured by three axioms, assumed to hold when the Gini coefficient is used to compare relative inequality among two or more distributions. The axioms are important for the interpretation of results and bear repeating. The first is mean independence; the distribution does not depend on the level of income. The second is the principle of transfers; income transferred from the rich to the poor always decreases inequality. The third is anonymity; the model is indifferent to transfers between individuals that leave the overall distribution unchanged. These axioms allow us to interpret a Gini coefficient of .5 as representing income less equally distributed than .4, but it tells us nothing about how each distribution is skewed or who won and who lost. For political and welfare purposes, it is often the winners and losers who are important. Lorenz curves generally will cross if the rich and the poor lose relative to the middleincome groups. When Lorenz curves cross, the Gini ranking is ambiguous because you can always find other functions that will rank the distributions differently. It then becomes necessary to specify the form of the social welfare function to achieve a complete ranking of inequality (Atkinson). With negative incomes, the Lorenz curve crosses the horizontal axis unless incomes are truncated at zero, as was done in the Ahearn and Johnson study. This has the effect of underestimating the inequality. Figure 1 illusJean Kinsey is an associate professor, Department of Agricultural and Applied Economics, University of Minnesota.