Late last year, The Economist published an article, "Meritocracy in America," which discussed social mobility in the United States. One metric used to measure social mobility was the percentage of sons who ended up in different income brackets than their fathers:
Some researchers claim that social mobility is actually declining. A classic social survey in 1978 found that 23% of adult men who had been born in the bottom fifth of the population (as ranked by social and economic status) had made it into the top fifth. Earl Wysong of Indiana University and two colleagues recently decided to update the study. They compared the incomes of 2,749 father-and-son pairs from 1979 to 1998 and found that few sons had moved up the class ladder. Nearly 70% of the sons in 1998 had remained either at the same level or were doing worse than their fathers in 1979. The biggest increase in mobility had been at the top of society, with affluent sons moving upwards more often than their fathers had. They found that only 10% of the adult men born in the bottom quarter had made it to the top quarter.
Now, there's a lot to criticize in this paragraph. The startling implication that children of the poor were overrepresented among the wealthy in the earlier study is presented without comment, and the fact that just a bit over 30% of sons in the second study ended up in a higher income quartile than their fathers isn't terribly disturbing when you consider that that figure would be only 37.5% even if the sons were randomly assigned to income quartiles.
But there's a more interesting fallacy at play here: the assumption that the metric used measures social mobility. It doesn't. More importantly, there's good reason to expect this metric to peak, decline, and finally level off, even in a perfectly meritocratic economy.
To understand why, we must draw a distinction between mobility and movement. To me, at least, the idea that social mobility might be declining is highly counterintuitive. Productivity continues its inexorable upward trend. College education is heavily subsidized, and student loans are readily available. We can debate the extent to which racial discrimination still impedes meritocracy, but that it's become dramatically less important in recent decades is well-nigh incontrovertible. In short, material barriers to escaping poverty are at an all-time low.
What this metric shows us is not mobility, but movement. That is, it does not measure the potential for the children of the poor to reach the upper income brackets, but rather the rate at which they actually do so. Counterintuitive though it may seem, we can construct a plausible model of a perfectly meritocratic economy in which such movement peaks and then declines.
Let's define a perfectly meritocratic economy as follows: Assume a natural ordering of all persons in terms of productivity. Since this is likely a product of both genetic and environmental factors, we'll assume that this ordering is established when they reach adulthood. A perfectly meritocratic economy is one in which incomes are directly proportional to productivity, so that the most productive citizen will have the highest income, and the least productive will have the lowest income. Obviously this isn't a perfect model of any real economy---the point is that it represents an economy with an ideal level of mobility.
The model has two further assumptions:
- A person's productivity is to some degree positively correlated with his parents'. This may be due to genetic factors (physical attributes, intelligence, etc.), to environmental and cultural factors, or to some combination of the two. I don't think that anyone will deny that this holds in the real world.
- Financial advantages or disadvantages related to a person's parents' income have no effect on his productivity. Obviously this doesn't hold in the real world, but my goal is to demonstrate that the phenomenon I'm describing occurs even in a perfectly meritocratic economy.
We begin the thought experiment with an economy governed by a rigid class system. Income is totally independent of productivity, so in the first generation there will be some very unproductive people at or near the top of the income distribution, and conversely some highly productive people at or near the bottom.
Just as the second generation reaches adulthood, we liberalize the economy (or progressivize it, or whatever it is you believe will create a perfectly meritocratic economy as defined above). Since we now have a perfectly meritocratic economy, there's a great deal of movement among income brackets in this generation, and each person's income is completely independent of his parents'.
In the third generation, we make no institutional changes. The economy is still perfectly meritocratic. But since the correlation between productivity and parents' income is now positive, the correlation between a person's income and his parents' is also positive. Social mobility has not decreased, but the economy has reached a sort of semi-equilibrium, and movement among income brackets has decreased. As long as a productive parents tend to have productive children, movement among income brackets will never again reach the level it reached in the second generation.
The takeaway message isn't that the United States is a perfectly meritocratic society; it's that increases in social mobility throw an economy into disequilibrium, which may create an unsustainable level of movement among income brackets as the most economically fit take advantage of newly available opportunities. This movement may well have declined in recent years, but that in itself isn't sufficient reason to believe that social mobility isn't as high in the US as it's ever been.