Scott Winship on Income Inequality
Is America growing more economically unequal? And if so, does it really matter? On this episode of Ricochet’s Money & Politics podcast, my guest is Scott Winship, a senior fellow at the Manhattan Institute. Previously a fellow at the Brookings Institution, Winship’s areas of expertise include living standards and economic mobility, inequality, and insecurity. Before that, he was research manager of the Economic Mobility Project of The Pew Charitable Trusts and a senior
policy advisor at Third Way.
Pethokoukis: The Census Bureau recently came up with some updated income statistics looking at median household income and how things have been going since the Great Recession, and where we’re at now. According to the statistics, the median household income is still down – despite the fact we’re four years into a recovery- about 8 percent from the pre-recession high back in 2007. In a very real way, the recession hasn’t ended because we’re not back to where we were.But that’s not the whole story, is it, as you recently wrote?
Winship: I started with official income figures but then I dove in more and made some improvements using the state of the art in cost of living adjustments, also taking into account things that the official measure does not. The official measure includes cash benefits from the federal government, things like unemployment insurance, social security and temporary assistance for needy families, but doesn’t include non-cash benefits, like food stamps or Medicaid and Medicare and it doesn’t take into account taxes. So when part of the federal response to the recession was to reduce taxes, particularly the payroll tax, that isn’t incorporated into the official measure.
When you do take into account those things, you find that disposable income for the middle class was back to its 2007 level in 2011. We’ve actually fully recovered. If you look at the bottom, that too looks like it’s recovered to its 2007 level. It depends on how you value Medicaid and Medicare, but it’s a more positive story than the official figures convey.
You also made an inflation adjustment. What did you do there?
The official measures use the CPI-U-RS, which doesn’t take into account the fact that if some prices go up, people switch to buying things that are less expensive so they’re actually able to maintain their living standards. They’re not switching and ending up worse off because of it, they’re switching to take advantage of price changes.
Are you creating your own index or you’re using a different one that the government also tracks?
I’m not creating my own. I’m using one that a different part of the federal government has put together. The Bureau of Economic Analysis puts together the national accounts that tell us how much economic growth has increased over time. They have their own measure called the Personal Consumption Expenditures Deflator, updated at the end of July. It’s the measure that the Federal Reserve Board prefers when they look at it. The Census Bureau isn’t using it, but they are developing the chained CPI that you hear about, which is controversial because it would lower cost of living increases in entitlements. But that measure tracks this other one that Bernanke prefers pretty well for the years that the two overlap.
The number that people read about in the paper says incomes still have not recovered. You’re saying that’s pre-tax, pre-transfer, what they’ll call market income, but when you take into account this other stuff, things have not been great. If the economy had kept on growing, we’d be much better off, but we’re back to where we were. Let me read what you wrote at the end of the piece.
The left worries that if things are getting better, public support for helping the poor will wane and its “middle-out agenda” will look misguided. The right does not want to concede that the safety net may have achieved income security. In response, the left might reply that things would be even better today if we had more stimulus while the right might say that safety net programs, the work disincentives prevented even greater improvement. It is undoubtedly the case that policy could have produced at least a somewhat better result, though, if we are honest, we have no idea how or by how much.
We have the left and the right looking at this number and drawing two different conclusions. What’s your conclusion?
I tend to be pretty agnostic about what might have happened if we had done something differently. It’s hard to develop credible numbers around thought experiments. The best numbers we’ve got, in my view, are the comprehensive ones that I put together. If you think that government ought to be redistributing money or providing a safety net during downturns, that’s exactly what it did in this case.
This is a result that the left ought to trumpet, that the safety net did what it’s supposed to do. If you’re on the right, you might take issue with that. Casey Mulligan made this argument that, in fact, the safety net sucks people out of the labor force. They took disability benefits or they stayed on unemployment longer than they should have.
So there are quibbles that you can make there. But it does reinforce a conservative case that the economy –even in the worst downturn that we’ve seen in 50 years – has not wrecked the middle class or living standards in general.
This is the anniversary of the financial crisis and of the Occupy Wall Street movement. Looking at the best numbers that we have, to what extent have incomes grown more unequal in the U.S. over the past three or four decades? Is it the 1 percent versus the 99? Is it the 10 versus the 90? Is this trend accelerating, decelerating?
Within the 99 percent or for sure within the bottom 90 percent, inequality rose from the late 1960s through the late 1980s. Since the late 1980s, it’s moved up and down a bit, but inequality between the middle class and the poor has not grown in a couple of decades. When it was growing, particularly in the 1970s and early 1980s, there’s evidence to suggest that it was reflecting changes in family structure. The increase in single parenthood increased inequality between the poor and the middle class.
When you look at the top, I would have said a couple of months ago that since the late 1970s, the top 1 percent has pulled far away from everyone else. The research that’s come out recently that throws the monkey wrench into what was a pretty safe summary of what we knew is by Richard Burkhauser and Jeff Larrimore and Philip Armour, who took the estimates of Emmanuel Saez and Thomas Piketty, the famous economists …
If you’re a New York Times reader and there’s a story about inequality, the chart is going to show the share of income grab by the top 1 percent going through the roof. That chart will be by Thomas Piketty and Emmanuel Saez. They’re the go-to guys on inequality.
Right. Saez updates it every year and he makes these great spreadsheets available so you can really dig into the numbers. Those numbers have shown dramatic increases in inequality.
Burkhauser and his colleagues start with those numbers, which include the capital gains that people realize when they sell assets like stocks. That’s a big part of the story because the stock market has done well over these past few decades. When the stock market is at a peak, people sell and realize these big gains that have been accumulating over time. Instead of showing gradual increases in income over time, they show up as these spikes when the market peaks and people sell. It will also show up as spikes when tax policy changes affect the returns to come of these investments.
Rich and his colleagues said, “Let’s take into account year by year capital gains that are accruing to people, many of which are behind the scenes because people are not actually selling these assets.” By the way, in some years, there are capital losses that we don’t see because of the 2008 financial crisis where there are these huge losses at the top. When they attempted to account for the accrued gains and losses behind the scenes, they found that inequality has not risen from 1989 to 2007.
This is between the 1 percent and the 99?
Right. Between either the top 1 percent and say the middle fifth, or the top 1 percent and the bottom fifth, inequality has not risen because the returns to capital have not been as strong as they were in the 1980s.
What would be the response to looking at the data that way?
It’s not straightforward to account for these behind-the-scenes gains that are accruing. It’s certainly not the case that what Rich and his colleagues did is unimpeachable. I think they’ve done the best that you can do, but it’s more difficult than what Piketty and Saez have done to look at the gains that people actually realized.
When people realize the gains, they show up on tax returns and we know what people actually received. We don’t know year to year how much wealthier or less wealthy people are getting because we don’t know the portfolios of the folks at the top. We only see the money when they cash in.
The response has been, well, these estimates are unreliable. Some people have criticized the idea that you should count these year-to-year gains or losses as income. The bottom line is that, if what Rich and his colleagues did turns out to be an accurate way of getting at these behind the scenes gains, it will overturn the conventional wisdom that everyone’s been talking about since Piketty and Saez first developed their estimate.
The president talks about inequality as one of the big problems of America. Would you say to all those New York Times readers and to President Obama the story might be wrong; high-end inequality might not actually be increasing?
That’s a possible conclusion if the Burkhauser research stands the test of time.
Do you think there’s value in looking at inequality the way that Piketty and Saez look at it, looking at that market income?
Piketty and Saez have this other set of estimates that doesn’t include realized capital gains. They show increases over time in inequality too. Apart from investment income even, Piketty and Saez have yet another set of estimates that only looked at earnings – there has been this big increase in inequality over time.
But it’s a more complicated story than looking at the headline numbers. When people choose to save and invest their money rather than consume, they’re taking a risk. There’s a chance that they’re not going to see that money again. They’re doing it to make the economy more productive and to create jobs for other people. In cases where people are foregoing consumption to save and getting high returns on that later on, the higher returns are the cost of convincing people to not spend the money when they get it.
We can’t simply look at these big returns that are coming in at the top and say, this is terrible. We ought to tax them away and there won’t be any consequences from taxing it away.
These Burkhauser numbers, which are looking at capital gains, are those also pre-tax, pre-transfer?
Rich and his colleagues are able to account for taxes and transfers as well. It’s pretty much a comprehensive income measure that you could put together.
Even with his numbers, we’re still seeing a lot happening in market income, pre-tax – where the numbers I’ve seen show the bottom 20 percent suffering huge losses over the past three decades and other income groups doing much better, which may be the result of technology and globalization. Do I have that story right?
What he shows is that, for market income, inequality has increased dramatically. He doesn’t show that the bottom half is doing worse over time. He actually shows that, so long as you take into account taxes and the value of health insurance, either from employers or for Medicare/Medicaid, that both the middle and the bottom have seen gains over time of at least 25 percent.
But pre-tax, pre-transfer not so much? What I’m trying to get at is that it’s the same story that we saw in your analysis of the Census Bureau data in which, if you’re looking in the market income, you’re seeing one thing and then those numbers are changed by taking into account taxes and government transfers where they end up better– and also in the case of Burkhauser, looking at benefits. You have two sets of numbers. Are they providing answers to different questions?
They do answer different questions so it’s not a matter of people picking the ones that make the case that they want to make. If your question is, how far is the labor market improving living standards for the poor or the middle class at the top, then it might be more appropriate to look at market incomes rather than after transfers and taxes.
Even there the story over the short run is not great. Since 2007, market incomes are down at the middle by about 10 percent. For the bottom 20 percent, they’re down by over 20 percent.
Going back through to 1979, the Congressional Budget Office has another set of inequality estimates and income estimates. Market income at the middle rose by about 30 percent between 1979 and 2007. That’s a much bigger gain than the loss since the recession hit. At the bottom, they show big gains as well that are more imprecise. Over the long run, even market income has done well, unless your benchmark is “how has the top 1 percent done,” or, “what happened in the 1950s and ’60s ,” that period with this amazing income growth.
To the extent that the high-end have done well, there’s a debate as to why. Is it because the union power has weakened so entrepreneurs and CEOs have been able to get bigger salaries and take advantage of workers? We see bigger discrepancies in this country between what CEOs make and what workers make than in Japan, for instance. Is it that kind of thing or is it more technology, globalization, companies are bigger? Which is the more persuasive case to you?
You again have to distinguish, are you talking about top 1 percent versus 99 percent inequality or just within the 99 percent inequality?
Globalization is part of the story in either case but, for the top 1 percent, globalization has increased the importance of having the best possible CEO or having the artist that’s going to generate the most revenue possible. The top performers in all of these industries have pulled away, not only from middle or the upper middle class but the bottom half of the top 1 percent. The levels of inequality within the top 1 percent are mindboggling.
Below the top, it’s a story about the increasing importance of skill, the demand for which has outpaced the supply. In particular, college graduation rates for men have stalled. There’s some evidence that that’s changing in the last few years, they kind of peaked during the Vietnam War when people were trying to avoid the draft.
This is all reflected in the college wage premium, right?
Right. We live in a global economy now and companies can’t pay these high family wages, the union pay where people were getting paid more than what their marginal productivity dictated. That’s an area in the past that we’re not going back to and that has been reflected in the inequality statistics.
Let’s just concede that inequality has increased. Why do we care? The White House has said and some folks on the left say, here’s why you should care. Let’s put aside ideas about fairness. They make the link that this inequality is hurting income mobility, making it harder for people to go up that ladder, and they come up with the Gatsby Curve, which shows some relationship between rising inequality and lower social mobility. What do you make of that relationship between the two? Is one worse than the other?
That is one of the most important policy questions of our time. The case on the left is that they’re strongly related to each other, which is a weak case. You’re right, what people tend to go to for evidence is this “Great Gatsby” curve which shows that, when you look at Western industrialized countries, there is a pretty strong correlation between inequality and mobility in a country.
But, for anyone who’s taken an entry level stats class in college, one of the first things you learn is that the correlation doesn’t equal causation. I’m doing some research that summarizes some alternative “Great Gatsby” curves.
There’s new mobility data, released by Harvard economist Raj Chetty, for local areas in the United States. When you look at the share of income within a local area received by the top 1 percent and you look at their mobility levels, there’s little relationship between those two.
Another example- if you look at the relationship between educational mobility and educational inequality across countries, that’s a strong relationship, except the U.S. in that context, is a high-mobility low-inequality country. It’s flimsy evidence that everyone on the left has latched onto to make this case that inequality has hurt opportunity.
Inequality of opportunity is something that we ought to be concerned about, but we shouldn’t be focusing on income inequality. There are other inequalities in how nurturing your neighborhood is or how good your parents are, you name it – that’s where we ought to focus.
How does U.S. income mobility stack up against other nations?
Most research that compares mobility across countries uses this measure that says, “How much does an additional 10 percent parental income buy you in terms of your own income when you become an adult?” By that measure, the U.S. looks pretty bad versus other countries. English-speaking countries look worse than continental European countries, but the U.S. is an outlier even among English-speaking countries.
That measure is influenced by inequality. The U.S. has more inequality and inequality has grown more than in other countries. That means that a given difference between parent incomes is going to translate into a bigger difference in the incomes of adult children, even if everyone is moving around at the same rate across countries.
If you look at whether kids who start in the bottom are able to make it out of the bottom or whether kids who start in the middle can make it to the top or fall to the bottom, there’s one study that compared the U.S. to some Scandinavian countries and to the U.K. What that revealed is that we have less upward mobility from the bottom, specifically among men. That’s where the U.S. does lag other countries.
Is there more or less mobility than there was 30 years ago, even for those groups you just mentioned?
People tend to jump from the finding that the U.S. has lower mobility than other countries to the conclusion that mobility has been declining in the U.S. Actually, the evidence of what’s happened in the U.S. over time is mixed. It tends not to show big changes one way or another. My assessment of the totality of the evidence is that we may not have enough mobility now, but it’s always been the case that we haven’t had enough mobility.
For middle-class people, has the standard of living stalled or has it continued to go up? Is it going up at a slower pace? What can we expect as far as the increase in middle-class standard of living going forward? Obviously, there’s concern about stagnation and there’s been this debate, is the era of economic growth over?
The income growth for the middle class has been stronger than people realize. Since about 1979 it’s gone up about 40 percent. I think we can continue to expect that growth over the next 35 years. Between the end of World War II and the end of the 1960s, middle-class income doubled. We’re not going to see that kind of growth again. But these are real gains over time that we’re seeing, and can continue to see. So a lot of the alarmism about stagnation is misplaced.
A lot of fears about the robots, Scott.
We’re entering a transitional period where certain jobs are going to be disrupted by automation and technology. But what’s going to happen is that robots are going to allow us to work less. A lot of two-worker families with kids are feeling frazzled. Increasingly, we’re actually going to have more time for leisure and family time. Then there are people who prefer to pursue professional aims, and to make as much money as they can, and they’re going to choose that, and they’re going to enrich everybody else in the economy as a whole.
So we’re going to see inequality continue to rise. People on the left will continue to think that’s a terrible thing. But the result of it will enrich society and allow people more time.
If you’re talking about a 25 percent shorter average work week, you’re not talking about a 25 percent smaller paycheck?
Right. When people think about the robots, they think that our pay is going to go down by a quarter, but it’s just as likely that everyone will choose to work 25 percent less. At the bottom there are folks – particularly if they continue to only get a high school diploma or not even that- whose incomes are not going to go up. But we’ll be in a better position as a society to support them too. If conservatives can develop a more robust opportunity agenda in the coming years, we can develop these safety nets that don’t have the perverse incentives and work disincentives that the current safety net does.
I’m an optimist. I think we’re going to get richer over time, and, with conservatives getting increasingly interested in mobility at the bottom, we can craft a new policy agenda that’s going to help folks there as well.