IQ and Income Inequality in a Sample of Sibling Pairs from Advantaged Family Backgrounds

Abstract

The Bell Curve (Richard Herrnstein and Murray, 1994) presented data on the independent effect of IQ on a wide variety of social and economic outcomes for members of the National Longitudinal Survey of Youth (NLSY). To control for socioeconomic background, we constructed an index using the standard three indicators: parental education, occupation, and income. Among the many threads in the response to The Bell Curve, the following question arose: How much would the independent effect of IQ have been attenuated if a broader set of family background variables had been used as controls? To test this, Sanders Korenman and Christopher Winship conducted a fixedeffects analysis of the large number of siblings within the NLSY, in effect controlling not just for socioeconomic status, but for everything in the shared environment of the family. The results were that [w]ith a few exceptions, the fixed-effects for AFQT [the cognitive test used in the NLSY] are remarkably similar to the standard OLS and logit estimates (Korenman and Winship, 2000 p. 146). The independent effect of IQ is robust across methods. I subsequently conducted my own analysis of the NLSY siblings for a publication on income inequality and IQ (Murray, 1998). In the course of that work, another aspect of sibling analysis struck me. The procedure, which is used in this paper as well, begins by identifying every sibling pair in which one sibling had scored in the normal range, defined as a tested IQ of 90-109, and the other member of the pair had scored somewhere outside that range. Those in the normal range were used as the reference group against which sibling outcomes were compared. I then limited the sample to full biological siblings, to take genetics out of the picture, and further limited the sample to sibling pairs in which the siblings had lived with both biological parents for at least seven years after birth, to minimize differential family backgrounds arising from divorce and remarriage. Both limits on the sample were essential to preserve the virtues of the sibling comparison. But notice what had been done in the process. In applying those conditions, the sibling sample represented a population in which all parents were wed when the child was born (zero illegitimacy) and all young children were brought up by both biological parents during their most formative years (zero early divorce). Such a population is one that has achieved much of what we ideally want to achieve through social policy. The thought occurred to me: why not complete the process? Having already created a sample without illegitimacy and early divorce, why not slay as well the great beast of social policy, poverty? To achieve that, I lopped off the sibling pairs whose parents were anywhere in the bottom 25 percent of the income distribution as of 1978-1979, when the NLSY began. This produced a sample of 733 sibling pairs who grew up in households which, by 1978-1979, had a median parental family income of $64,586 and a minimum income of $30,486 (expressed in 2000 dollars). For practical purposes, I had by this process nearly achieved the utopia of income distribution as defined by many, one in which the lowest income is half the median income. I hereby dub these 733 pairs the utopian sample. It is utopian not just because it has virtually no illegitimacy, divorce, and poverty. The way it has been selected has also necessarily effected drastic improvements in the neighborhoods, peers, and educational systems which the youths attended. The members of the utopian sample had a big edge in their potential access to college, both economically and because the sample is highly selected for the kind of parents who actively encourage their children to continue their educations. The same selection factors mean that I have created a sample in * American Enterprise Institute, 1150 17th Street, N.W., Washington, DC 20036.