I welcome Christine Sypnowich’s powerful corrective to the ongoing drift in egalitarian circles toward opportunities and away from outcomes. I have argued in my own writings as an economist that this drift is not merited on conceptual, empirical, and policy grounds. Consider each domain in turn.
The opportunity argument is that while outcomes that flow from factors outside the control of individuals are legitimate targets for correction, outcomes over which individuals have had choice and control do not warrant such correction. But the distinction faces conceptual problems right from the start.
Sypnowich rightly highlights the difficulty of drawing a line between circumstance and choice. We should consider as well the many cases where one person’s choice becomes another person’s circumstance: think of parents and children, teachers and pupils, or corporation executives and their workers. Moreover, think of the myriad choices in markets that shape outcomes for individuals at large. Free choices by high-income individuals in the property market push up property values and rents and push out low-income tenants; the circumstance many renters face is caused by the free choices of high earners. And as Sypnowich argues, it is furthermore poor sociology to ignore the influences of community and society—and “poor sociology makes for poor ethics.”
Sypnowich touches briefly on the opportunity dimension of the capability approach developed by Martha Nussbaum and Amartya Sen. I agree with her comments; elsewhere I have presented a detailed critique of the capability approach, lauding its “broadening of the evaluation space from the instrumental means such as income to the intrinsic ends of beings and doings, or functionings” but strongly cautioning against “the further broadening of evaluation from achievement of ends to opportunity to achieve those ends—from functionings to capabilities.” At a normative level, it is essential that ends be valued in themselves.
Over the last two decades there has been an explosion in attempts to quantify “inequality of opportunity,” particularly for developing countries. The exercise took off after the World Bank embraced the effort in its 2006 World Development Report, “Equity and Development.”
Though this work continues to advance in econometric sophistication, the procedure is quite straightforward at its heart. First, identify a set of circumstance variables. In an influential early study by Paes de Barros and colleagues on Latin America and the Caribbean, these were selected as gender, race/ethnicity, birthplace, mother’s education, father’s education, and father’s occupation. (Note how few variables make the cut!) Second, identify an outcome variable: say, income. Third, quantify the fraction of variation in the outcome variable that is accounted for by the circumstance variables and their intersections—and call this “inequality of opportunity.” In the Barros study, the numbers turn out to be of the order of 20 to 50 percent; recent reviews put the range at 10 to 60 percent across countries.
This method faces many obstacles and limitations. For cross-country comparisons or assessments over time, the circumstance variables need to be comparable. If data on the level of father’s education is given in ten categories in one country and five categories in another, the ten categories have to be aggregated to match the five. Similarly, if a circumstance variable of interest is totally missing from a country’s data source, the calculation for that country is not comparable to other countries. And data on many circumstance variables of interest simply does not exist.
As a result, we are effectively driven to the lowest common denominator of empirical information, leading to the coarsest possible picture. The smaller the number of variables, the lower will be the overall variation accounted for by these variables—and thus the lower will be the measured inequality of opportunity we present to the world and to policymakers! One defense that is often mounted is that this measure can and should be presented as a lower bound: inequality of opportunity is “at least” 30 percent, we might say. But without an upper bound, technically we could not be contradicted in saying that the number may be as high as 70 percent, 80 percent, or even 100 percent. It is disconcerting that such an important calculation should be so dependent on mundane properties of the data.
In another major strand of empirical work, the Human Development Report’s Human Development Index (HDI) and its Multidimensional Poverty Index (MPI) are both explicitly modeled on the capability perspective. But in fact they are largely based on achievements, not opportunity for achievements. At the national level, for example, the HDI is a weighted sum of per capita income, years of schooling, and life expectancy. The broadening from a sole focus on income is certainly welcome, but years of schooling is a measure of a functioning, not capability or opportunity for schooling; the same goes for life expectancy. (Of course, schooling—and life itself—also provide opportunities for other kinds of functionings, but we are back to a conceptual difficulty or normative debate: Why should we think of education or life expectancy solely as opportunities for functioning, instead of as outcomes valuable in themselves?)
It is indeed difficult to see how we could go beyond functionings with the available data. The same is true of the ten components of the MPI—as its originators have admitted. In practice, then, we are back to measuring variation in outcomes.
In policy circles, opportunity arguments are often couched in terms of “pre-distribution” being superior to “redistribution.” A standard formulation says that it is better to provide equal public education (pre-distribution) and then let the chips fall where they may through effort and choice rather than have a progressive tax regime that takes income from some and gives it to others or invests in public goods (redistribution). But again, it is difficult to separate pre-distribution from redistribution, especially when parental inputs combine with publicly provided education to determine quality of education and its market outcomes.
The idea of pre-distribution is also present in the work of those such as Raj Chetty and colleagues who propose targeted fixes to get people out of areas where intergenerational mobility is low. But as with education, two issues arise. First, the resources needed for pre-distribution will have to come from taxation, which raises the question of how progressive the taxation should be. Second, as Dylan Matthews notes in Vox, the fixes do not “dismantle the structural causes of segregation, or prevent rich families from using their political power to keep out poor families.”
It is sometimes argued that equal opportunity talk is more palatable in policy circles than talk of equalizing outcomes—so egalitarians should embrace the former if they want a seat at the table. But if this is the price of admission, egalitarians may find themselves boxed into limits on policies like substantial redistribution and progressive taxation. Moreover, saying things like “inequality of opportunity accounts for at least 30 percent of observed inequality” is likely to prompt a dismissive eye roll from policymakers: “But that means equality of opportunity could be 70 percent, which is not at all bad, and I have many other things to worry about.” At a minimum, economists with egalitarian commitments should embrace freedom from poverty as an essential corrective to equality of opportunity.
Putting all these considerations together, I would rather that there was a strong health warning for egalitarians who are drawn to equality of opportunity: “Use with extreme caution!”
We’re interested in what you think. Submit a letter to the editors at email@example.com. Boston Review is nonprofit, paywall-free, and reader-funded. To support work like this, please donate here.