Ashok Rao on Has Rising Inequality Really Been a Problem Over the Past Generation?: The Honest Broker for the Week of December 8, 2013
Attention Conservation Notice: 2700 words mostly by Ashok Rao reviewing--and dismissing--the arguments against worrying that rising inequality has been a major problem over the past generation.
The near-consensus view over here at Equitable Growth and at the Equitablog is that U.S. economic growth over the past generation has been very disappointing. Too-much of our economic growth has been wasted producing the wrong stuff and delivering it to the wrong people, and we have failed to properly and productively invest at the rate we could in people, machines and buildings, ideas and organizations, and institutions. The hunch around here is that these two are tightly coupled: that the rapid rise in inequality as a result of the derangement of incentives has both decoupled the links between higher measured real GDP and human economic welfare and material well-being, and has also slowed the growth of our potential to produce real GDP.
But we could be wrong. And many argue that we are. What do we think of their arguments at their best?
Overwhelmed with work, I asked the smart, thoughtful, and enthusiastic Ashok Rao--along with Evan Soltas one of the leading lights in the next generation of webloggers--to take a look at the arguments of Scott Winship (formerly of Brookings and now of the Manhattan Institute) and others. He was not persuaded: cherry-picking and tendentious shifting of the burden of proof was his assessment.
So let me turn the microphone over to Ashok Rao:
Ashok Rao: Many of us on the left are in the business of arguing that rising inequality reduces economic welfare, and that something must be done. To promote this business we want--if we can--to refute the best arguments to the contrary, including but not limited to those from Scott Winship. He can serve as a representative agent of those making the best arguments that progressives overstate the costs of inequality.
(1) Burden of Proof: Scott Winship makes a multilayered argument, asserting all of the following:
- Econometric data is inconclusive on the causal link between inequality and immobility.
- There has been no great stagnation in wages for the poor.
- It is unclear that income growth directed to the richest one percent could have been realized by anyone else.
- Consumption inequality matters more, and trends in it are not as worrisome as trends in income inequality.
I conclude that Winship's contribution to the debate is a contribution of a new framing, not of a different readings of the facts. His arguments present themselves as empirical studies--additions to the body of facts and analyses about inequality and its effects. Unpacked his argument looks much more like an assertion that our ex ante priors before we look at the evidence should be different, and that the burden of proof should lie 100% on those who worry about inequality. This looks to me much more on the "value" than the "fact" side of David Hume's divide--a values-based redistribution of the burden of proof, placing the entire burden of proof on those of us who think that inequality has been a problemm. And in my judgment, the assignment of 100% of the burden of proof to the left that Winship makes is unreasonable.
Taken together, their errors severely undermine the liberal case about the consequences of inequality. True, a careful examination of the evidence does not establish that inequality is harmless, or that it has nothing to do with our other economic problems. Economic data cannot prove a negative. But they can fail to prove a positive, and they do fail to prove the claims that underlie the left's basic economic narrative. They reveal little basis for thinking that inequality is at the root of our economic challenges, and therefore for believing that reducing inequality would meaningfully address our lagging growth, enable greater mobility, avert future financial crises, or secure America's democratic institutions...
We do live in a world of scarcity. But what is scarce is not just resources, but quality economic data and quality interpretation thereof. That data and interpretive resources are scarce matters. Why? Because it seems to me that the major part of Winship's argument is it its core a dismissal of the mammoth’s share of the progressive argument with little more than:
- Assertions that progressives have failed to reject the null hypothesis that inequality does not increase immobility.
- Assertions that, because correlation does not imply causation, many significant studies are in fact unconvincing.
It seems like Winship would accept nothing short of randomized control trial of deep macroeconomic phenomena before he would accept the claim that inequality does hurt immobility.
Now, if I were arguing, say, that promoting atheism was a uniquely effective means towards mathematical achievement, such would be a fair requirement to impose. That claim is testable given current data capacity. That claim is also extraordinary: so extraordinary ex ante that it makes sense to require that it meet a high burden of proof.
Thus Winship's argument is, at its core, a philosophical one. It is an argument about base cases, about what we should presume before we look at the evidence. These assumptions about the base care are unfair – it makes more sense to require conservatives to reject our null--that inequality-reducing measures should be presumed to benefit the poor, and should not be presumed to not harm growth simply. Our theoretical underpinnings are much stronger. Moreover, our ex ante prior that inequality-reducing measures do harm growth should be updated for the slew of evidence that this is not the case. At an abstract level a society with more inequality from causes like low estate-tax levies and segregated educational opportunities is likely to be a society that allows the rich to pass on a large share of their relative success to their children. The children of the relatively rich of today then disproportionately crowd into the space of the relatively-rich tomorrow. Almost by definition, this makes it more difficult for the poor and middle to reach up. This is nearly arithmetic.
There are reasons why this arithmetic may not hold. Perhaps credentialism from Ivy League universities (or education at all), consistent support from one’s parents, and quality primary education do not actually help future earning potential at all. But no fair arbiter would let Winship set the burden of proof in the debate the way he does on this one.
Thus Winship’s argument boils down not just to the idea that correlation does not prove causation, but that correlation does not even suggest causation. Knowing that correlation -> causation is a logical fallacy is fine and dandy. I wish schools would also remind students more often of the equal fallacy that statistical failure to reject the null hypothesis -> the null hypothesis is true.
That is not to say Alan Krueger's Great Gatsby Curve is dispositive. Indeed, I do not think any real economists would suggest that it was. It is an empirical correlation, nothing more and nothing less. It leaves the work of proof yet to be done.
That brings us back to square one: Winship's argument against inequality being a problem is a value judgement--a value judgement that those of us with a strong theoretical ex ante case must amass an enormous amount of high-quality, expensive, detailed evidence that rising inequality has in fact had the obvious consequences before we are allowed to even have a conversation about the subject, let alone make the case for policy changes. This does strike me as similar to some of the most sophisticated forms of climate-change denialism, which also focus not on reading the evidence we have differently but rather setting up a very different prior from the rest of us, and so redistributing the burden of proof.
(2) Not Complete "Stagnation": A secondary argument that Winship makes--a common talking-point of the more sophisticated conservatives today--is that wages have not really stagnated. To some degree this is the fault of liberals who have overstated the case. The right claim is that the U.S. economy has done a lousy job at turning increased productive potential into human material welfare outside the top 1%, and that over the past generation income gains have been extraordinarily maldistributed in favor of the upper class. The wrong claim is that material welfare outside the top has not increased at all over the past generation. But mostly Winship's claim seems misguided. Does it really mean that the economy has performed well and we should be happy with it just because median income has improved marginally rather than stagnated in an era in which top incomes have soared? Especially when that widening divide comes with increasing social segregation and, yes, declining mobility?
Some--including Winship--stress that the liberal argument of stagnant income is predicated on using the Consumer Price Index (CPI) instead of the superior Personal Consumption Expenditures (PCE) deflator. But the PCE is bad in this case for precisely the reason it is good when it comes to measuring macroeconomic or monetary phenomenon--it is broad. It includes everything from spending by nonprofit institutions to government expenditures on healthcare. It is not especially relevant when we want to compare the disposable cash income that a median household has received over time (an OECD statistic some like cite because it puts the United States right behind Luxembourg). The CPI places more weight on germane expenditures, like rent.
And remember, much of meager increase came from a substantial increase in total working hours, despite smaller household sizes.
Nor does the case against inequality falter noting that many of the gains realized by the richest one percent could not have gone to the poor. This point is usually made in the process of acknowledging some form of capital-biased technological change:
But is it likely that this dramatically increased prosperity for the wealthiest Americans has suppressed income growth for everyone else? Some of the gains made by those at the top could not have accrued to people lower on the income ladder in any event: For instance, had Chinese investors not enriched bankers in New York, their money would surely have gone to bankers in London or Frankfurt--not to workers in middle America. Moreover, by promoting economic efficiency and broad growth, at least some gains at the top actually make non-rich families better off...
This is a straw-man argument. Find me a single liberal who believes that making the rich poorer per se would make anyone else better off. The absence of the gains to the rich received as a consequence of the Great Financialization would not have accrued to the poor in its absence, any more than a tighter market for fry-cooks would have directly benefited Lloyd Blankfein. But there is not just redistribution, there is redistribution: taxing Blankfein and Co. more and handing that money out to the poor. And the argument is that that would make the poor better off. Does anyone want to argue that the welfare cost of redistributing a dollar from rich to poor is higher than a dollar and so would leave the recipient worse off?
(3) Trickle Down: In fact, Winship goes to extreme lengths teasing out a tenuous story of how inequality might help the poor.
Frank's hypothesis is, however, supported by one additional study, conducted by economists Marianne Bertrand and Adair Morse. Bertrand and Morse find that, within states and over time, higher incomes at the top of the distribution are associated with more bankruptcies and more people saying their financial situation has worsened over the past year throughout the income distribution. Bertrand and Morse also find that, within states, a 10% rise in expenditures by households in the top fifth of the income distribution is associated with increases of 1.3% to 3.6% in expenditures by everyone else, even after accounting for increases in income below the top fifth. (The effect is smaller and not statistically meaningful for poor households.)
There is some evidence, however, that part of this phenomenon is caused not by increased financial pressure on the non-rich to "keep up" with the rich, but rather by greater housing wealth and the borrowing against home equity that that increased housing wealth induces. In other words, rising inequality may increase the demand for housing among those at the top of the income scale, in turn boosting the home values of the non-rich, and then leading those newly wealthy home owners to tap into their increased home equity...
What is a more likely story? In states where the economy is doing relatively well, everyone happens to do at least somewhat well? Or: in states where the rich are doing well demand for booming real estate percolates into a wealth effect for the poor so that the real growth in consumption among all classes is not from simple business-cycle dynamics but a complicated chain of events?
(4) Income and Consumption Inequality: We usually cite income inequality: data that’s relatively clean and easy to gather. Some conservatives, like Scott Sumner, insist that income inequality is a useless metric. For life cycle reasons, consumption inequality is what we really ought to watch. Such a claim is often coupled with studies showing an insignificant increase--or even a decrease--in consumption inequality over the past several decades.
Yet income inequality has increased within pretty much any age group. Importantly, to the extent life cycle factors matter--that is people are richer later in life than they were older--income matters, and the effects of today's income inequality on consumption inequality will be seen in the future. Much of my future wealth, and hence my future consumption, is predicated on my ability to save today requiring income above consumption. With defined benefit pensions now visible only in the distance via the rear-view mirror, income at any stage becomes increasingly relevant to welfare. And to the extent that this is not true for me, it is because my savings will be transferred down boosting the opportunities of my kids, further depressing mobility.
But more importantly, the claim that consumption inequality has not increased has melted away. The most recent and careful estimates of the increase in consumption inequality place it as of the same order of magnitude as that in income inequality. Consumption inequality is not the holy grail for inequality skeptics.
(5) The Last Argument: This completes my tour of the best arguments against the belief that rising inequality has been a serious problem over the past generation, and is a serious problem now. I conclude that the arguments are statistically self-serving--have to work hard to cherry-pick the data--and possess a ridiculous double standard in their allocation of the burden of proof. Good evidence linking equality and mobility, backed by unimpeachable theory, is thrown out with little more than Latin quotes. And we are asked to accept contrived links between household wealth effects and consumption of the rich as solid.
And there is a last argument:
It may be true that the government could redistribute at least some income between rich and poor Americans such that the benefits to the poor exceeded the harm done to the rich. But justifying such redistribution requires making a persuasive and empirically grounded case that the status quo is unfair.
Here is a two-sentence introduction to the mountain of evidence that there is both a persuasive and empirically grounded case for redistribution: The bottom decile of Americans live in conditions not too much better than poor Italians or Greeks. Poor, middle class, and rich households spend approximately $200,000, $300,000, and $500,000 on raising their children respectively, and much of that difference is channeled towards better human and network capital investment. This is not a story of equal opportunity or consistently unequal opportunity. It is one of increasingly unequal opportunity.
Let us hope the future debate about inequality and its consequences does not mirror that on climate change.
More on Scott Winship:
Noah Smith: Scott Winship fails to show that Alan Krueger is a liar: "Winship presents this as a refutation, but it is actually a value judgment. Krueger is talking about inequality, and he presents numbers on inequality. Winship is saying 'No, you shouldn't care about inequality, you should only care about absolute income levels and the upward mobility of the poor.'...
"Just because Krueger chose not to focus on what Winship believes is 'the relevant figure' and the 'only important' issue does not mean that Krueger is 'dishonest'. A value judgment is not a refutation!... But there's another problem with Winship's claim. Yes, it is true that from 1970 to 2011, the shrinking of the middle class, on net, reflected (unequal) upward mobility out of the middle class. That is because absolute incomes themselves increased a lot between 1970 and 2011. But if you look at the period from 2001-2011, it is a different story entirely..."
Justin Wolfers: Is Higher Income Inequality Associated with Lower Intergenerational Mobility?: "There’s a broader lesson here for what we learn from these sorts of data debates. Basically, Winship has a bunch of complaints about how the data are constructed--and many are valid. He says that it’s difficult to construct internationally comparable measures of income inequality--that the chart should use inequality from an earlier era, and that only some types of inequality would generate immobility. He also points out that mobility is difficult to measure....
"It’s a standard play from the wonk-fight playbook: throw lots of mud at the data, and hope that this leads people to mistrust the conclusions that follow. Here’s the thing: his criticisms actually strengthen the original finding. Think about it. Imagine how strong the 'true' relationship must be if it shows up even when using only rough proxies for the 'true' levels of inequality and immobility..."
Miles Corak: The Economics of the Great Gatsby Curve: "Winship’s article does a disservice to a well-established literature on generational mobility by suggesting that the basic information Krueger used is in some sense invalid. Krueger’s Great Gatsby Curve is in fact well-rooted in the labour economics literature, and debate would be better placed addressing the policy implications he draws than to suggest that President Obama’s top economist feels compelled to create his own facts..."
The Economics of the Great Gatsby Curve: a picture is worth a thousand words: "A quick post to thank Scott Winship for his response.... But I am afraid they do not advance the discussion.... The internationally comparable estimates I offered account for these concerns.... The best part of Scott’s response is the suggestion that we should start thinking about how inequality, and what kind of inequality, influences opportunity. And that is why I suggested those interested in public policy give the Haskins=Sawhill book a careful read. It is written in a bi-partisan way, faithful to the evidence, and linked to policy levers..."