Adjusted for inflation, the US economy has more than doubled in real terms since 1975.
How much of that growth has gone to the average person? According to many economists, the answer is none or close to none.
In a recent gloomy study of the American economy, Thomas Piketty, Emmanuel Saez, and Gabriel Zucman find that between 1980 and today, almost none of the gains from economic growth accrued to the bottom half of the population. They write, “Looking first at income before taxes and transfers, income stagnated for bottom 50% earners: for this group, average pre-tax income was $16,000 in 1980 — expressed in 2014 dollars, using the national income deflator — and still is $16,200 in 2014.”¹ Piketty, Saez, and Zucman also found that incomes of the top 1% tripled over the same time period.
New York Times columnist David Leonhardt, reacting to this work, concluded “the very affluent, and only the very affluent, have received significant raises in recent decades.”
Last month, Nobel Laureate Joseph Stiglitz wrote in the New York Times:
Some 90 percent have seen their incomes stagnate or decline in the past 30 years.
Claims like these are not new. Nobel Laureate Paul Krugman wrote in 2014 that “Wages for ordinary workers have in fact been stagnant since the 1970s.”² Jared Bernstein wrote in 2014 in the New York Times, “for middle-income households earnings have declined in real terms 7 percent from 1979 to 2010.”³
Writing in Vanity Fair in 2011, Stiglitz wrote: “All the growth in recent decades — and more — has gone to those at the top.”
All. And more — implying that everyone except those at the top is losing ground. No wonder people are losing faith in capitalism and considering alternatives like socialism, whatever flavor they have in mind.
But these depressing conclusions rely on studies and data that are incomplete or flawed. They understate economic growth for the poor and the middle class because they use measures of prices that mis-measure inflation. Some studies leave out important components of compensation such as fringe benefits which have become increasingly important in recent years. And some studies include the elderly which lowers measured progress because the elderly are an increasing share of the population and they are less likely to be working full-time if at all. I look at some of the problems with the standard stories about middle-class stagnation in the first video of my series, The Numbers Game:
Many of the most pessimistic studies about the fate of the American middle class ignore the changes in the American family since the 1970s and the effects have had on the way we measure changes in household income. I look at the issue of demographic change in the second episode of the Numbers Game:
But the biggest problem with the pessimistic studies is that they rarely follow the same people to see how they do over time. Instead, they rely on a snapshot at two points in time. So for example, researchers look at the median income of the middle quintile in 1975 and compare that to the median income of the median quintile in 2014, say. When they find little or no change, they conclude that the average American is making no progress.
But the people in the snapshots are not the same people. You can’t use two snapshots to conclude that only the rich have made progress. It’s possible that everyone from the earlier snapshot has actually gotten richer and then been replaced by different people whose incomes will also rise. This is especially true when there is immigration. If new immigrants are disproportionately less skilled than Americans already here, measured incomes can fall even when those who are already here have steadily improving economic prospects.
And when marriage rates are falling and people are increasingly living on their own, household income can fall while every individual is doing better. Estimate of economic progress based on household income are distorted by these effects.
What the snapshots show is that the rich today are richer than the rich of yesterday. If the rich people are the same people as yesterday, then one’s class determines one’s fate. But if they are not the same people, the snapshots tell you that the dispersion of income has increased. That may or may not bother you, but it doesn’t necessarily mean that there is a distinct group called “the rich” who are capturing all the gains while the rest of us tread water.
How important are these issues? One way to find out is to follow the same people over time.
When you follow the same people over time, you get a very different story from the standard one.
When you follow the same people over time, the largest gains over time often go to the poorest workers; the richest workers often make no progress.
The most dramatic claims by the pessimists that no one is making progress other than the rich are wrong.
A study by Leonard Lopoo and Thomas DeLeire for the Pew Charitable Trusts uses the Panel Study of Income Dynamics (PSID) and compares the family incomes of children to the income of their parents.⁴ Parents income is taken from a series of years in the 1960s. Children’s income is taken from a series of years in the early 2000s. As shown in Figure 1, 84% earned more than their parents, corrected for inflation. But 93% of the children in the poorest households — those in the bottom 20% — surpassed their parents. Only 70% of those raised in the top quintile exceeded their parent’s income.
The poor may find it easier to do better than their parents. But how much better off do they end up? Julia Isaacs’s study for the Pew Charitable Trusts looking at the late 1960’s up to 2002 finds that children raised in the poorest families made the largest gains as adults relative to children born into richer families.⁵
The children from the poorest families ended up twice as well-off as their parents when they became adults. The children from the poorest families had the largest absolute gains as well. Children raised in the top quintile did no better or worse than their parents once those children became adults.
One explanation of these findings is there is regression to the mean — if your parents are particularly unlucky, they may find themselves at the bottom of the economy. You, on the other hand, can expect to have average luck and will find it easier to do better than your parents. At the other end of the income distribution, one reason you might have very rich parents is that they have especially good luck. You are unlikely to repeat their good fortune, so you will struggle to do better than they did.
But that doesn’t change what actually happened in the last three decades of the 20th century in the Isaacs study: the children from the poorest families added more to their income than children from the richest families. The pessimistic claims I mentioned at the beginning of this essay deny there is any regression to the mean. They argue that only the richest Americans have benefited from economic growth over the last 30–40 years. Or that only the richest Americans have gotten raises. The pessimistic story based on comparing snapshots of the economy at two different points in time misses the underlying dynamism of the American economy and does not accurately measure how workers at different places in the income distribution are doing over time.
Moving forward in the time of the analysis, Gerald Auten, Geoffrey Gee, and Nicholas Turner of the Office of Tax Analysis in the Treasury Department used tax returns to see how rich and poor did between 1987 and 2007. They find the same encouraging pattern: poorer people had the largest percentage gains in income over time:⁶
The study looks at people who were 35–40 in 1987 and then looks at how they were doing 20 years later, when they are 55–60. (Unfortunately, Auten, Gee, and Turner only report percentage changes.) The median income of the people in the top 20% in 1987 ended up 5% lower twenty years later. The people in the middle 20% ended up with median income that was 27% higher. And if you started in the bottom 20%, your income doubled. If you were in the top 1% in 1987, 20 years later, median income was 29% lower.
Yes, these are only percentages and a large percentage gain for a poor person can be less in absolute terms than a small percentage gain for a rich person. Unfortunately, Auten, Gee, and Turner did not report the absolute gains; I also wrote them but they can’t find the original numbers. But the magnitude of the percentage gains and the results from the Isaac paper above show that it’s at least possible that the gains to the poor and middle need not be small or insignificant in absolute terms when the percentage gains are this large.
A recent study by David Splinter of the Joint Committee on Taxation looks at a narrower definition of income using tax data.⁷ Splinter has an estimate of the impact on growth on the different parts of the income distribution where, like many other estimates, he takes a snapshot of the income distribution in an early period, in this case, 1980, and compares it to the distribution at a later time, 2014. That is, he follows the quintiles over time, rather than the people, the standard approach of the pessimistic studies. And he finds that the bottom quintile in 2014 has less income than the bottom in 1980. Like Piketty, Saez, and Zucman, he finds essentially no gain for the bottom half of the income distribution.
Only the people at the top gain much of anything between 1980 and 2014. The average income of the top 1% (not shown in the figure) went from $189,000 to $843,000, which seems to confirm the view that most of the gains from economic growth go to the richest of the rich while people in the middle or the bottom make no progress at all. But the people in the top 1% in 2014 are not the same people in 1980. What happens when you follow the same people? Splinter makes that calculation as well:
As in the other panel studies, when you follow the same people, the biggest gains go to the poorest people. The richest people in 1980 actually ended up poorer, on average, in 2014. Like the top 20%, the top 1% in 1980 were also poorer on average 34 years later in 2014.
Splinter reminds readers that the pattern here is more important than the size of the changes — he points out that tax returns grossly understate actual income, capturing only about 60% of the total in recent decades. And some of what he is capturing may be life-cycle effects that differ by quintile. But his findings are a dramatic example of the potential of cross-sections — two snapshots in time — to mislead compared to a panel approach where the same people are followed over time.
Splinter’s numbers don’t control for age. Presumably the lowest quintile workers in 1980 are going to be younger so perhaps the growth we see is just normal raises as workers gain experience. So I asked Yonatan Berman⁸ who has been working with the same data to do a similar calculation but restricting the sample to workers who start out 25–30 in 1980. Here is what he finds:
The starting level of income — the solid bars — is the median for workers centered in 1980 — it averages data from 1978, 1980, and 1982 and then puts people into quintiles based on those averages. Using data across four years reduces measurement error and transitory effects. The shaded bars are centered on 2012 averaging from 2010, 2012, and 2014. So these data cover most of a person’s working life from young ages to near-retirement. The time period is roughly the same as Piketty, Saez, and Zucman which covered 1980 to 2014. It uses the same deflator to correct for inflation as they do — the national income deflator.
Using four years of data biases growth downward. To show up in the bottom quintile you have to be there over a four year period. But the results are still pretty cheerful. All groups, and not just the richest, gain over their lifetime. The poorest workers make the largest percentage gains. The two lowest quintiles see their incomes more than double over their working life. The absolute gains for the three lowest quintiles are $18,800, $22,300, and $23,800. While the rich have larger absolute gains, the variance is much larger.
Finally, this is labor income. It does not include any measures of fringe benefits, benefits that became larger part of compensation over this period. No government transfers. No capital gains. And I would argue that the national income deflator imperfectly controls for improved quality of the goods we consume. This biases it upward meaning real growth is understated for all groups. Going in the other direction — these numbers are for workers — people who had earnings at both the beginning and the end of the period, so this probably overstates the increase in labor income of people at the bottom.
The American economy isn’t just helping the richest Americans. Prosperity is being enjoyed widely. Here is the third episode of the Numbers Game that looks at some of the cheerier evidence I’ve mentioned above and gives more of the intuition for what is going on:
There is one study of progress over time that follows parents and children that is gloomy and that is “The Fading American Dream: Trends in Absolute Mobility Since 1940” by Raj Chetty, David Grusky, Maximilian Hell, Nathaniel Hendren, Robert Manduca, Jimmy Narang (Chetty et al)⁹ They find that if you were born in 1940, you had a 92% chance of surpassing your parents income. But if you were born in 1984, the number is a depressing 50%. Chetty et al control for age — this is for parents and children when they are both 30. This does suggest that the American dream is dead or at least dying — half of the children do better than their parents but half do worse, suggesting no progress over time. And because the data are from more recent years, they cast doubt on the cheerier results I reported at the top of this essay which looked at older data and may not apply to more recent times.
But Chetty et al’s estimate of 50% is quite sensitive to various assumptions they make as they report. (And they are to be saluted for reporting these results which is rarely done.) For example, the 50% measure does not correct for changes in household size over time, the issue that I mentioned above that distorts progress over time. When Chetty et al correct for household size, the proportion of children who as adults are better off than their parents is between 60% and 67% depending on how you choose to correct for family size. Many economists believe that the CPI-U-RS, the inflation measure used by Chetty et al in their gloomy 50% number, overstates inflation. If the overstatement is .8% percentage points annually, Chetty et al report that 59% of children born in 1984 are better off than their parents. If CPI-U-RS is off by 2% points annually, they report that 69% of children do better than their parents.
Chetty et al do not report the sensitivity of their results to changes in more than a single assumption — what if you correct for household size and the overstatement of inflation by the CPI-U-RS? But it’s clear that the true measure of how children born in 1984 are doing relative to their parents may actually be in the range of 70–90% rather than 50%.
[Edit: Chetty et al do report results with changing multiple assumptions in an appendix to the published version of the paper in Science— I thank Maximillian Hell, one of the authors, for pointing this out to me. For example, if you include taxes and transfers, assume the CPI-U-RS is overstated by .8% per year and use one of the corrections for family size, then 72% of children born in 1984 do better than their parents. I’d like to see the number when inflation is overstated by 2% and without taxes and transfers to isolate market effects only…]
And even under the gloomiest assumptions that leads to the 50% figure, Chetty et al find that the children growing up in the poorest part of the income distribution have a better chance of surpassing their parents than children growing up with rich parents: 70% of children born in 1984 into the bottom decile exceed their parents’ income in 2014. For those born in the top 10%, only 33% exceed their parents’ income.
All of the studies I have mentioned in this essay are about absolute mobility — how much income people gain over time. Relative mobility is much less dramatic — the richest people have an easier time staying in the top quintile than others have of getting there. While rich people in 1980 actually lost in some of the studies I’ve discussed, they still have a much higher income on average than the people who were poor in 1980.
There’s a lot more to study and understand. But what the studies above show is that the economic growth of the last 30–40 years has been shared much more widely than is generally found in the cross-section studies that compare snapshots at two different times, following quintiles rather than people. No one of these studies is decisive. They each make different assumptions about income (see the footnotes below), which people to include, how to handle inflation. Together they suggest the glass isn’t as empty as we’ve been led to believe. It’s at least half-full.
Of course, the fact that people across the income distribution make progress over time doesn’t change the fact that people at the top today make more than the people at the bottom. Income inequality is larger than it once was. The snapshot approach captures that. But it does not capture the impact of economic growth on people’s material well-being or provide evidence that the rich or the poor are static categories no one ever escapes. That the rich today are richer than the rich of yesterday is a very different finding than that the rich are getting all the gains. Too many economists have treated these as identical.
This does not mean that everything is fine in the American economy. There are special privileges reserved for the rich that help them reduce their risk of downward mobility — financial bailouts are the most egregious example. There are too many barriers like occupational licensing and the minimum wage that handicap the disadvantaged desperately trying to succeed in the workplace. And the American public school system is an utter failure for too many children who need to acquire the skills needed for the 21st century. But the glass is at least half-full. If we want to give all Americans a chance to thrive, we should understand that the standard story is more complicated than we’ve been hearing.
The videos embedded in this series are a creative partnership I have been privileged to share with Shana Farley, Chris Dauer, and Tom Church. Working with them has been such a pleasure. Their help and insights have been invaluable. I am grateful to Alicia Reece of Motion 504, the chief animator on the three episodes to date for her creativity and superb work. I want to thank David Splinter and Yonatan Berman for conversations and their work presented here.
One of the reactions to this essay is that of course household income has gone up — that’s because there are so many more households where both the husband and wife is working. Another version of this is that in the old days, one person working could support a family, now you need two, so we’re falling behind. Actually, the cheerier studies I’ve mentioned above look at individuals, not households. But it is worth noting that the proportion of households with two earners has actually decreased since 1980. (See Table H-12 here from the Census Bureau.) That’s because while more married couples are households where both spouses are working, the marriage rate has fallen. So in 1980 (when the Census data starts), 33% of households had two earners but in 2017 it is only 31%. Since 1980, the number of households with zero or only one earner has risen from 56% to 60%.
Some readers of this essay have asked why in Splinter’s work, the poorest and middle quintile is worse off in 2014 than in 1980. Doesn’t that mean the economy is hurting the poor? That depends. As Splinter points out, income based on tax returns leaves a lot of income unreported, maybe 40%. A surprising amount of that unreported income did not go to the rich. Piketty, Saez, and Zucman assume that unreported income is a proportion of reported income. Auten and Splinter use tax audit data to get what they hope is a better measure of where the unreported income shows up. That has a huge effect on measured inequality:
I suspect there is more unreported income in 2014 than in 1980. Another possibility is that 2014 was just an unrepresentative year as 1980 was, as a measure of lifetime prospects.
¹“Distributional National Accounts: Methods and Estimates for the United States,” by Thomas Piketty, Emmanuel Saez, and Gabriel Zucman, Quarterly Journal of Economics, December 2016: https://eml.berkeley.edu/~saez/PSZ2018QJE. See also “Income Inequality in the United States: Using Tax Data to Measure Long-term Trends by Gerald Auten and David Splinter, August 23, 2018 http://davidsplinter.com/AutenSplinter-Tax_Data_and_Inequality.pdf. Auten and Splinter make different assumptions about under-reported income and find a dramatically smaller increase in inequality.
²“On Income Stagnation,” by Paul Krugman, New York Times, November 12, 2014.https://krugman.blogs.nytimes.com/2014/11/12/on-income-stagnation/
³“Piketty’s Arguments Still Holds Up, After Taxes,” by Jared Bernstein, May 9, 2014: https://www.nytimes.com/2014/05/10/upshot/pikettys-arguments-still-hold-up-after-taxes.html
⁴“Pursuing the American Dream: Economic Mobility Across Generations,” by Leonard Lopoo and Thomas DeLeire, Pew Charitable Trusts, July 2012. The study defines income “as the total income derived from the taxable income (such as earnings, interest, and dividends) and cash transfers (such as Social Security and welfare) of the head, spouse, and other family members. The PSID definition of family used in this analysis includes single-person families and unmarried cohabiting couples who share resources, in addition to families related by blood, marriage, or adoption. Family income does not include the value of non-cash compensation such as employer contributions to health insurance and retirement benefits, nor does it include the effect of taxes or non-cash benefits such as food stamps.More information here: https://www.pewtrusts.org/~/media/legacy/uploadedfiles/pcs_assets/2012/pursuingamericandreampdf.pdf
⁵“Economic Mobility of Families Across Generations,” by Julia Isaacs, Economic Mobility Project, Pew Charitable Trusts.Parents income is from 1967–1971. Children’s income as adults is taken from 1995–2002. Isaacs writes: “Family cash income is the focus of the analysis, including taxable income (such as earnings, interest and dividends) and cash transfers (such as Social Security and welfare) of the head, spouse and other family members… As discussed in Appendix B, family cash income does not include the value of non-cash compensation such as employer contributions to health insurance and retirement benefits, nor does it include the effect of taxes or non-cash benefits such as food stamps. All incomes are reported in 2006 dollars, using the CPI-U-RS to adjust for inflation.” https://www.brookings.edu/wp-content/uploads/2016/06/11_generations_isaacs.pdf
⁶“New Perspectives on Income Mobility and Inequality,” by Gerald Auten, Geoffrey Gee, and Nicholas Turner, National Tax Journal, December 2013. “The measure of cash income used in this paper starts with total income as reported on individual income tax returns and then adds known sources of non-taxable income and adjusts for several items where the tax treatment differs from what might be considered a better measure of the current income realized by a taxpayer. In particular, tax exempt interest, non-taxable Social Security benefits, non-taxed unemployment compensation (2009 only), excluded foreign wages and housing benefits, excluded capital gains on small business stock, and net operating loss carryovers reflecting prior year losses are added. State tax refunds, alimony paid, the itemized deduction for gambling losses (up to the amount of gambling income reported), and disallowed current year passive losses are deducted.” https://www.ntanet.org/NTJ/66/4/ntj-v66n04p893-912-new-perpectives-income-mobility.pdf
⁷“Income Mobility and Inequality in the United States: Evidence from Tax Data since 1979,” by David Splinter, September 13, 2018. “The first income definition is fiscal income including capital gains. This is defined the same as tax return-based market income in Piketty and Saez (2003) — adjusted gross income (AGI), plus adjustments and excluded Schedule D capital gains before 1987, less government transfers in AGI (unemployment and taxable Social Security benefits) — but capital losses reported on Form 1040 are replaced with losses before limitations. Unfortunately, fiscal income is limited to income reported on tax returns, and therefore only captures 60 percent of national income in recent decades (Auten and Splinter, 2018; Piketty, Saez, and Zucman, 2018)… For absolute mobility estimates [the ones shown in the charts here], fiscal income excluding capital gains is used to limit sensitivity to business cycles.” http://www.davidsplinter.com/Splinter-Mobility_and_Inequality.pdf. The results shown in the charts are taken from his online spreadsheet: http://davidsplinter.com/Splinter-Mobility_and_Inequalitya.xlsx
⁸See Berman’s paper here that finds that the chances of the poor having gains in income over short periods of 2–4 years are greater than that of the rich.
⁹“The Fading American Dream: Trends in Absolute Income Mobility Since 1940” by Raj Chetty, David Grusky, Maximilian Hell, Nathaniel Hendren, Robert Manduca, Jimmy Narang. http://www.equality-of-opportunity.org/papers/abs_mobility_paper.pdf. Another discussion of the sensitivity of their results to assumptions is here.