Tuesday, November 4, 2014

Is Student Debt Only a Problem for Dropouts?

In a recent article in the otherwise-excellent Washington Monthly issue on inequality, Rachel Fishman argues that college debt isn't a problem. Instead, low graduation rates are. Her article follows a Brookings Institution report that took a similar line from June, greeted with applause for its willingness to question the conventional wisdom about a flood of student debt drowning young workers.

The facts, unfortunately, aren't so optimistic. While it's true that college dropouts carrying higher education debt loads face a dire labor market situation, they're not the only ones. The happy talk about student debt for those who do manage to graduate is misplaced, as is the idea that the labor market's problems could be solved if more people manage to complete their BAs.

The key evidence is from a recent paper by Beaudry, Green, and Sand. Those authors show that demand for skilled labor has declined substantially since 2000, even during the expansion that ended in 2007 and even as the supply of college graduates increased. As a result, college graduates take jobs that didn't previously require a BA, "filtering down" the labor market hierarchy in what my colleague Elisabeth Jacobs calls "a cruel game of musical chairs." As a result, those with fewer educational credentials get forced out of employment and the labor market altogether, and those lucky enough to have jobs aren't in a position to demand higher wages.

Fishman, on the other hand, writes "at current graduation rates, the United States will fall five million credentials short of meeting labor market demands by 2020." I have no idea where that forecast of labor demand growth come from, but it would be fair to say it's completely at odds with reality. She writes that "according to research by Georgetown University, over half of the added jobs since the recession have gone to those who have a bachelor's degree or better." It's unclear what research Fishman is referring to or whether that factual claim is about job creation or total hires (that is, new jobs plus continuously-existing jobs whose occupants have turned over), but regardless: the claim is false. I've plotted cumulative job creation and hires since the first quarter of 2009 for all workers and for those with a BA or higher credential below, as reported by the Quarterly Workforce Indicators. In fact, insofar as there has been hiring and job creation, it's been among low-wage, low-education-qualification jobs.

Let's consider Fishman's main point. She writes "college degrees generally provide students enough of a bump in income to cover the cost of paying back loans." College grads do indeed earn significantly more than those who don't manage to graduate. But is the difference enough to make up for the increase in the cost of higher education? That depends, of course, on how much education costs have increased, and I haven't yet assembled illustrative data on that question. As for simple college wage premiums, however, it certainly doesn't look like the wages graduates earn make up for the increased cost of college.

I would argue that the implication of this data is that while there is a persistent gap in average earnings among those who do gain employment at different educational qualification levels, that gap is essentially static since 1990.

What would be especially informative here is wage premia weighted by access to jobs. We do know that the more educational credentials workers have, the lower their unemployment rate and the easier it is to find a job. If that's true, the real benefit to a college degree in a persistently slack labor market is not wages, but job security. The implication of that fact is stark and entirely contrary to Fishman's argument that increasing the share of workers with college degrees would reduce inequality. On the contrary, if there were more college grads, inequality would be worse, because lower-wage workers would be pushed out of the labor market entirely.

Additionally, there's evidence that recent college grads entering a labor market with depressed demand for their skills don't enjoy as large a wage gain as previous cohorts.

In the near future, I will present evidence on wages by education group and age cohort, as well as education costs, to investigate what's happened to the relative rate of return to college. Given the dire story I've been telling, it's something of a puzzle that educational attainment hasn't declined. I suspect that puzzle is explained by the job-security/musical chairs mechanism, but the data will tell. One thing we can say for sure, however, is that there's very little good news for workers to be found in the wage data, no matter how good their educational credentials. And pushing college students to saddle up with debt in a push for their BA, even if it does make sense as an individual decision, would probably cause yet worse outcomes for the uncredentialed and result in higher inequality overall.

Tuesday, October 14, 2014

Rising Inequality in the United States

Gabriel Zucman and Emmanuel Saez released a paper today tracking rising inequality in the United States (slides here). Unfortunately, the inequality-related development that sucked up most of the day's coverage was an Institute on Global Markets poll of prominent economists on the topic of "Piketty and Inequality." More on that in a moment.

Zucman and Saez is a far more detailed empirical investigation of wealth inequality in the US than appears in Capital in the 21st Century. It documents several facts:

1. The wealth distribution has become significantly more skewed, with the top 0.1% owning in excess of 20% of total wealth. That number was between 5 and 10% in the 1970s.

2. Wealth concentration is due to several phenomena:
  • High labor incomes. Saez and Zucman document that a much larger share of labor income is accruing to households in the top 0.1% of the wealth distribution than was true in the 1960s. 
  • High savings rates by the wealthy. The authors provide further support for the declining asset position of the bottom 90% of the wealth distribution, including savings rates that dipped into negative territory for the entire boom of the 2000s. At the same time, the savings rate for the top 1% was around 40%.
  • Returns on wealth that increase in the stock of wealth. Saez and Zucman link individual capital income data with total wealth in each of nine asset classes from what's known as the Flow of Funds data. That enables them to determine a rate of return for each asset class. Because fractiles of the wealth distribution differ by their exposure to different asset classes, the authors are able to say that on average, the wealthier you are, the higher the return you enjoy on your wealth.
Crucially, all of this additional data Saez and Zucman bring to bear on the case of the United States confirms the analysis and predictions in Capital in the 21st Century. My paper responding to the book's critics divided the so-called capital channel for rising inequality into two arguments: that the overall wealth-to-income ratio in the economy would rise, and along with it capital's share of income, and that within the wealth distribution, savings rates and rates of returns would become increasingly stratified by household wealth.

Piketty spends a great deal of time analyzing the supposed anomaly of the US, where inequality has skyrocketed due to measured labor, not capital, income. His explanation for the "Rise of the Supermanager" is that executives have been increasingly able to ensure that the surplus earned by the firms they run accrues to themselves, and perhaps their shareholders, but certainly not to their employees. What Saez and Zucman tell us is that, to a surprising extent, these supermanagers are the same people as the very wealthy, and to the extent that today's wealth inequality is due to yesterday's labor market inequality, now and in the future, inequality will be driven by the very different wealth dynamics of different parts of the wealth distribution.

A few further points about Saez and Zucman: it absolutely puts the lie to supposed flaws in Piketty's reasoning. The rise of the wealth-to-income ratio is thanks to a housing bubble? Here's their Figure 2:

 The return on capital adjusts down as capital is accumulated, ensuring a constant factor share?

 The saving rate of the representative household will decline to zero as the economy-wide growth rate declines?

It's abundantly clear that in a world where the top 0.1% controls 20% of the wealth, a model of capital formation premised on the assumed behavior of a consumption-smoothing representative agent is a hopeless anachronism.

 So if the facts are so friendly to Piketty, what's up with that IGM poll? It asked thirty or so economists whether the rise of wealth inequality in the US to date is due to the empirical inequality r > g, what Piketty calls the Third Law of Capitalism. Emmanuel Saez said it wasn't. So did nearly everyone else. So is it Piketty against the world, or at least the economics world? No, or rather, not exactly. Piketty himself doesn't argue that r > g explains the dynamics of wealth in the US to date. What he argues is a less detailed version of the story Saez and Zucman tell in their paper. And in that story, what dominates the future is accumulated, inherited wealth and r > g, with inequality of savings rates and rates of return on wealth layered on top.

Some of the respondents to the poll understand both Piketty's argument and the facts, and could only disagree with the cartoonish version of Piketty's argument contained in the poll question. But many didn't, and predictably, the headline result that "no one agrees with Piketty" was picked up as a sort of death blow to the faddish theory du jour from some radical Parisian dilettante and the mindless t-shirt-wearing throngs who worship him. Many poll respondents offered up their own empirically-vacant, faddish theories: that biographies of the super-wealthy show they didn't get it all from Daddy, that fancifully-named-residual "technology" explains rising inequality, even that wealth inequality hasn't increased (!!). Caroline Hoxby decried Piketty's "negligible empirics," a subject she knows intimately.

So what to make of the dichotomy between a brilliant new working paper that brings tons of additional empirical insight to bear on the question of why inequality has risen so much in the US and a poll of economists that reveals, at best, ignorance of the facts, and at worst, an ideologically-motivated, dangerously groupthinking desire to close this debate as soon as possible? Simply that rising inequality and its causes poses a challenge to the economics profession at the deepest level, a challenge that isn't simply going to go away when all the t-shirts are sold out.