Economic Mobility in America

The US Department of the Treasury has released a "new" study on intra-generational income mobility in the United States from 1996 to 2005, which a recent Wall Street Journal editorial describes as "a careful, detailed piece of research by professional economists that avoids political judgments." The Journal says that the results "show beyond doubt that the U.S. remains a dynamic society marked by rapid and mostly upward income mobility." Is this true? A closer look at the actual study suggests that there is neither nothing new about the findings, nor anything particular careful or helpful about them. Sadly, as economist Tom Hertz, among numerous others, have found, both intra-generational and intergenerational mobility are distressingly low in the US when compared both to what a society with true equal opportunity might look like, as well as to mobility in other prosperous economies. For example, economists Katherine Bradbury and Jane Katz have found that "In the 1990s...53 percent of families that began the decade in the poorest quintile were still there ten years later... [and that] 40 percent of families ended the 1990s where they began..." Bradbury and Katz note that mobility in the 1990s was lower than in previous decades, giving "cause to worry." Chicago Federal Reserve economist Bhashkar Mazumder finds that some studies of intergenerational income mobility overstate mobility by up to 30 percent, and that, "Given the rising evidence from studies in other countries it appears that the U.S. may be among the most immobile countries." Nevertheless, the good news is that the US is not a rigid caste-based society. Table 2 in the Treasury study does somewhat adequately highlight the fact that there is substantial movement in and out of various income quintiles, though certainly not to the extent one might expect.

It's important to understand what Americans generally understand as "income mobility," as well as what a helpful study of income mobility might look like, thus revealing the flaws in analyses like the one just released by the Treasury Department.

How is income mobility defined?
First, income mobility is generally though about as either mobility within a lifetime (i.e. intra-generational) or mobility relative to one's parents and even grandparents (i.e. intergenerational). Second, income mobility is generally thought about as either absolute (i.e. increased or decreased real income over time) or relative (i.e. increased or decreased position in the income distribution). A society can have high levels of one without the other, with absolute mobility stemming largely from shared economic growth and relative mobility stemming largely from efforts to ensure equal opportunity. Thus, a prosperous nation that values equal opportunity would have high levels of both.

When should we care about absolute mobility versus relative mobility?
Because those on top have high initial incomes, we might expect small increases or even decreases in absolute mobility over 10 years, and because those at the bottom have low initial incomes, we might expect large increases in absolute mobility over 10 years. This is in part the statistical phenomenon of regression towards the mean. The Treasury study does confirm this phenomenon in Tables 3 and 8. Still, broadly shared absolute mobility is important for any society interested in continuously improving living standards, but it tells us little about whether or not our "society [is] stratified by more or less permanent income differences...[which] can breed class resentments and unrest." Relative mobility, on the other hand, tells us about the extent to which opportunity to move up and down within a society is equal or unequal for all.

How is income mobility related to income inequality?
There are at least three ways income mobility and income inequality relate to each other directly. First, income mobility might assuage some concern about income inequality. If high or low income status are temporary and people have an equal chance of finding themselves with high or low incomes (excluding the importance of merit in determining income), then the difference between high and low incomes, i.e. income inequality, might not matter so much. As the Treasury study argues,

"Economic historian Joseph Schumpeter compared the income distribution to a hotel where some rooms are luxurious, but others are small and shabby. Important aspects of fairness are that those in the small rooms have an opportunity to move to a better one, and that the luxurious rooms are not always occupied by the same people. The frequency with which people move between rooms is a crucial aspect of the trends in income inequality in the United States."

Second, the meaning of mobility depends in part on the extent on inequality; this is the other side of the same coin. The value of moving up or down depends in part on the income differentials in a society.

Related to that point is the third reason why one might care about income inequality even in a highly mobile society: income inequality can distort occupational choices in ways that are inefficient and harmful society as a whole. For example, Robert Frank points out that there are currently many "winner-take-all markets," in which the very best and brightest in particular fields are paid exorbitant sums relative to others. Robert Frank goes on to note that,

"The lure of the top prizes in winner-take-all markets has also steered many of our most able graduates toward career choices that make little sense for them as individuals, and still less sense for the nation as a whole. In increasing numbers, our best and brightest graduates pursue top positions in law, finance, consulting, and other overcrowded arenas, in the process forsaking careers in engineering, manufacturing, civil service, teaching, and other occupations in which an infusion of additional talent would yield greater benefit to society.

One study estimated, for example, that whereas a doubling of enrollments in engineering would cause the growth rate of GDP to rise by half a percentage point, a doubling of enrollments in law would actually cause a decline of three-tenths of a point. Yet the number of new lawyers admitted to the bar each year more than doubled between 1970 and 1990, a period during which the average standardized test scores of new public school teachers fell dramatically."

Are all changes in relative income also examples of relative income mobility?
No, and to answer otherwise would misunderstand what we mean when we think of income mobility. To understand clearly this apparent paradox--i.e. of some income changes not equaling income mobility--consider what we might expect to see and experience with regard to intra-generational mobility among those age 25 and over-- the focus of the Treasury study. First, most Americans would probably expect absolute and relative incomes to start off low around age 25; increase over a few decades, possibly plateauing; and then decrease nearing and during retirement. The fact that nearly all people's incomes follow this sort of trajectory is hardly a sign of meaningful relative income mobility.

What does income mobility look like then?
Rather, Americans interested in income mobility are interested in comparing people over time to others of similar age. The question we really care about is, "How likely am I to move up or down relative to others my age over my lifetime?" Thus, a thoughtful study of income mobility might track 30-year olds against each other over a decade. The Treasury study fails to do so. As University of Chicago labor economist Kevin Murphy pointed out 15 years ago in response to a horribly similar Treasury study, "This isn't your classic income mobility...This is the guy who works in the college bookstore and has a real job by his early 30's." Unsurprisingly, the Treasury study finds "There was considerable income mobility of individuals in the U.S. economy during the 1996 through 2005 period with roughly half of taxpayers who began in the bottom quintile moving up to a higher income group within 10 years."

Unlike the Treasury study, Americans do not consider the movement of a recent college or graduate student into a substantive job 10 years later to be a significant sign of upward mobility in society.

Unlike the Treasury study, Americans do not consider the movement of a high-earning 65-year old into retirement to be a sign of downward mobility in society.

Unlike the Treasury study, Americans do not consider the entry of new workers into the labor force to be a sign of upward mobility for everyone else in society.

Unlike the Treasury study, Americans are interested in mobility among all, not just those who file taxes.