Questions About Income Mobility

by Will Wilkinson on May 17, 2005

Questions about income mobility and inequality are much in the news. I’m pretty dissatisfied with what I’m seeing so far, so I thought I’d try to start getting straight on how to think about it.

(1) What is income mobility?

Income mobility is generally measured in terms of jumping quintiles, i.e., changing relative position in the income distribution.

(2) Why should we care?

Suppose, starting next year, that everyone’s real income would double every January forever. Suppose, however, that it turns out that nobody ever changes quintiles ever again. There’s no force keeping anyone from jumping quintiles, but it just never happens. Income mobility, change in relative position, has ground to a halt. But everyone now making $10,000 a year will be making over a million in eight years. Do you care that there is no income mobility? I don’t.

We do in fact care about mobility because we want to know that our system of institutions facilitates opportunity and rewards initiative and enterprise. A decent number of people changing relative position shows us that our system is amenable to upward aspirations.

(3) Is income mobility A LOT less important than high levels of economic growth that increase everyone’s absolute wealth.

Yes.

(4) Should we expect less bottom to top, number one with a bullet, mobility as an economy grows wealthier overall?

Yes. People are constantly confused by the growing gap between the rich and poor. This is good thing, not a bad thing. If the bottom is fixed, at zero income, and the top keeps going higher, you’ve got a bigger gap. But lots of people are better off and nobody is worse off. Similarly, if the lowest quintile is anchored by a fixed bottom, and the top is untethered and rising, the distance from the bottom to the top will increase. The distance from the bottom to the middle will increase. So it will take longer to get there. If today’s middle is equivalent in real terms to yesteryear’s top, people who are going from the bottom to the middle are doing no worse than people of yore who went from the bottom to the top (even if we assume, counterfactually, that there has been no change in quality of life for people at the bottom.)

We should be AIMING at a system where the middle of the middle is, say $500,000 per annum, and so the trip from the bottom of the bottom to the top of the bottom, much less to the middle of middle, is a VERY BIG trip indeed.

This is a priori theorizing. So perhaps a real economist will pipe up.

(4) What don’t studies about income mobility tell you?

Among other things, they don’t tell you how many people really TRIED to improve their relative economic position but couldn’t. The relevant normative question about a system of institutions here is how easy it is to do well if one really TRIES to acquire human capital, takes initiative, exerts effort, etc.

We might assume that everyone tries equally hard across cultures, but I think this would be mistaken. I think it would also be mistaken to take widespread cultural beliefs about income mobility to predict effort. Salient cultural figures like pop stars and athletes often have rags-to-riches stories, and probably reinforce the sense that rags-to-riches is in principle possible, even if one has no athletic or musical ability, and does not believe in one’s particular case that effort is going to be very worthwhile.

And if you’re looking at things in terms of quintiles, you won’t be seeing significant movement within the quintile. As the economy grows, the range of each fifth should widen, and movement from the bottom to the top of a fifth becomes more significant. At a certain point, dividing things into fifths is going to mask a lot of dramatic income mobilitty, and so it becomes more useful to look at things at a finer grain of division. But this is trivially easy to do.

(4) Does the New York Times have an irrational fetish for ominous tales about relative position?

Yes.

[Update: Ron Bailey has some good thoughts on the NYT series.]

  • One thing many libertarians don't consider wrt income inequality is that absolute income levels are not really what most people care much about, at least above some minimal level.

    Rather, people are evolved to care about relative income levels. We naturally envy the rich and despise the poor. Women are attracted to rich men. Men want to be rich to attract women. Both sexes prefer having rich friends.

    What does not change in the doubling income scenario is the relative position that people in society have; and given that relative position is what people use for important social decisions, it would remain a sore spot for the "losers", no matter how absolutely rich they became.
  • monkyboy
    Javier, the picture doesn't change much between pretax and after-tax income.

    Between 1980 and now, the average income of the lowest quintile is unchanged, pretax. After-tax it grew only 1 percent.

    The average income of top quintile grew, pretax, 43% in the same period. It grew 45% after-tax.
  • Javier
    As a side note, does anyone know of any good philosophical work on the value of absolute incomes/resources versus the importance of fair equality of opportunity? For example, Rawls is very hard on inequality because, among other reasons, inequality may harm fair equality of opportunity. People might have less equal opportunities for achieving "desired social position." However, if everyone's prospects are absolutely increasing along the manner that Will suggests, should we care about equality of opportunity?

    As far as I can tell, there is a huge gap in the philosophical literature on this topic.
  • Javier
    I don't know the issues in detail, but Olaf Gersemann in Cowboy Capitalism provides a useful discussion. Gersemann shows that the pattern of redistribution is probably more equitable in the US than in Germany, France, and Italy. In the US, 41.4 percent of cash transfers go to the poorest 30 percent of the population. In Italy for instance, the poorest 30 percent recieve 20.5 percent of total cash transfers.

    The poorest 30 percent in the US also probably pay fewer taxes relative to the rich than in many Western European countries. This is especially true when we look at consumption taxes, which are higher in Europe (think of the taxes on gas there) and have a greater impact on the poor.

    Gersemann uses this study as his main source of evidence.
  • Will Wilkinson
    Javier, Thanks for the link the Jencks paper. And good points. I would like to see total inequality taking into account taxation, redistribution, and benefit from public goods, etc. But this is a bit hard to do. I'm not sure if things would turn out better for the poor. Social Security, for example, appears to increase inequality by creating a barrier to investment for the poor, and reducing inheritances among the poor by in effect annuitizing what otherwise could have been saved and passed on.
  • Javier
    Will, I found this piece to be an extremely honest and fair-minded assessment of inequality in the United States. It turns out that many of the bad things people associate with inequality (such as lower health outcomes, lower political participation, crime) apparently are only weakly correlated with inequality in the transnational comparisons.

    The author also notes a possibility that John raises--a leading cause of inequality is immigration. One way to decrease inequality would simply be to let fewer people into the US.

    There are also other relevant measures of inequality that the NYT article omits. Increasing income inequality is not parralleled by increasing consumption inequality. That is, incomes may have become more unequal, but not people's patterns of spending. Why this is so remains a mystery. It might be that the official statistics underestimate the income of the poor.

    Also, another point that the NYT article only briefly touches upon: the driving cause of inequality is the high education premium in this country. The average income of a person with a college degree is more than double that of someone with a high school diploma, and this gap will probably keep rising.

    I could be wrong, but my understanding is that many or most studies on income inequality look at pre-tax income inequality. What's more interesting is the income distribution after taxes and redistributions. On first glance, an after-tax distribution of income would probably look considerably more favorable to the poor.
  • Anton
    Seems to me, Anton, that you've vitiated your own argument. Aren't you arguing here that if there are plenty of laws, then they'll inevitably be twisted to benefit the rich and not the poor? That's very similar to what libertarians and economists believe about regulatory capture. In the long run, lots of regulations tend to favor incumbents and squeeze out competition and change.

    I agree with the libertarians and economists on the point about regulatory capture, but there are degrees. To the extent to which income inequality can be mitigated, the relative influence of the lower classes will be increased. Mitigating income inequality requires fighting the interests of the powerful, and that's hard, but if we can do it in this particular case (thereby re-arranging the structural incentives of politicians) we make it easier next time. If we just stand by and say "inequality - just fine with me" as Will wants to, we make the problem even worse.

    I don't have a plan for turning the government into a level playing field where the interests of the poor can be fairly represented, but we can make things less unfair than they are. Will says they should be even more unfair.
  • Will Wilkinson
    John, Thanks! That's great stuff!
  • John Thacker
    Aggregate statistics for inequality are very difficult to accurately read. Take the increasing number of people getting higher education. More education causes one to accept a lower student income for more years in exchange for a higher income once one gets out. Compare it to the flatter income over one's life of a blue-collar worker. The result of more people getting higher education is greater income inequality over one's own lifespan. The result of that, in aggregate, is greater social income inequality, because at any one time different people are in a different stage of their lives. Viewed over their whole lifetimes inequality has not increased, even as social inequality has.

    Second complicating factor: Marriage. All the data gathered by the Census and cited is about households. (Check whenever you see something reported.) That means that when a family gets divorced it creates two households with lower income from one household. When a marriage happens one household is created from two. The lowest decile of household income is inhabited by the single and divorced as well. Increasing divorce rates in a society where marriage is still the norm create a widening gap between richer (the married) and poorer (the single) even without changing per capita income of inequality at all. Certainly boosting marriage among the poor would help in many ways, but the household inequality statistics perhaps overestimate the true inquality involved.

    Complication factor three: Women. Women and education and marriage. In the "old days," middle and upper middle class women were less likely to
    work than poorer women; the higher the husband's salary, the less likely the wife worked. This flattened inequality. The eighties and on have seen a great change in that. There are a lot more double income middle and upper middle class (and lower upper class) households now. (The massively upper end has seen less change, though.) Once you add in the tendency to marry people with similar incomes (I know a lot of two-doctor marriages, for example), it does affect household inequality. I certainly don't want to turn back the clock or change things, but the complicating factor I want to note.

    Complicating factor four: Immigration. At each stage we measure the people who are in the country now. Immigrants overwhelmingly inhabit the lower income brackets; they're also overwhelmingly wealthier than they were in their home country. This is an exaggerated version of the income mobility problem; the lowest quintile in 1985 is not the same people as in 2005. A lot of the lowest quintile in 2005 wasn't even in the country in 1985, and were much poorer elsewhere. We've decreased global inequality by accepting immigrants and raising their income levels, but increased our individual country's inequality by doing so. To be sure, importing an "underclass" has important social ramifications, but on a global scale it's a definite good thing. Indeed, if you use Census data and just look at people who were in the country in both 1985 and 2005, the poorest quintile gained income and inequality did not increase.

    Don't tell me that this can be solved by making the government weaker in libertopia. If libertopia is characterized by inequality and lack of mobility, poor areas will be systematically deprived of vital and legitimate state services such as police, while the wealthy and powerful will have plenty of police protection... Under inequality and non-mobility, those laws will shield the assets of the rich and expose the assets of the poor.

    Seems to me, Anton, that you've vitiated your own argument. Aren't you arguing here that if there are plenty of laws, then they'll inevitably be twisted to benefit the rich and not the poor? That's very similar to what libertarians and economists believe about regulatory capture. In the long run, lots of regulations tend to favor incumbents and squeeze out competition and change.
  • Anton
    You're missing the important question about the shape of the distribution. If "inequality" means the difference between the person at the top and the person at the bottom, then inequality will probably keep rising forever. But there are lots of other measures of something one might reasonably call "inequality": for example, the difference (or ratio) between the median person in the top decile and the median person in the bottom decile. That should remove the homeless and others who have absolutely nothing from the calculations.

    In any case, inequality is bad for other than standard-of-living reasons. As long as wealth gives you more access to, and influence on, powerful institutions such as the government, (i.e. forever and in all circumstances) inequality and lack of mobility mean that the levers of power will be controlled by one culture, and the people affected by those institutions will be a different culture. The people in those cultures will have systematically different desires, preferences, and goals, and inequality means that those at the bottom will be *oppressed*, not merely underrepresented.

    Don't tell me that this can be solved by making the government weaker in libertopia. If libertopia is characterized by inequality and lack of mobility, poor areas will be systematically deprived of vital and legitimate state services such as police, while the wealthy and powerful will have plenty of police protection. Other laws, while remanining within the scope of what's permissible under libertarianism, will be systematically biased in favor of the interests of the powerful. For example, libertarian doctrine doesn't tell us exactly what the bankruptcy laws should be like. Under inequality and non-mobility, those laws will shield the assets of the rich and expose the assets of the poor.

    Come to think of it, the U.S. is a lot like that today. . .
blog comments powered by Disqus

Previous post:

Next post: