Why Education Does Not Fix Poverty

Brookings and the American Enterprise Institute claim to have hatched a bipartisan consensus plan for reducing poverty. As exciting as that sounds, the details of the plan, unfortunately, won't be available until David Brooks unveils them at an event on December 3rd. Nonetheless, it's clear from the materials they have released that the consensus plan will focus on three things: education, marriage, and work.

In the next few posts, I will attack all three focuses as misguided. Today's focus will be on education, easily the most misguided of the three.

1. Rehearsing the Education Poverty Argument

To see where the education poverty argument goes wrong, it's helpful to explain what that argument is first. In this post, I am going to do that by pretending initially that we are in the year 1991. That year has no special significance other than that it's the year the modern Census education questions begin.

So imagine you are an education-focused poverty person living in 1991. You peer out into the world of basic social statistics and you see this graph of adult poverty rates broken down by education:

You notice something very striking about the graph: the higher the education, the lower the poverty rate.

You go back out into the social statistics universe and you see this graph breaking down the distribution of adults across the various educational groups:

You combine this graph with the poverty rate graph in your mind and you have an epiphany. Because the lower educational bins have higher poverty rates and the higher educational bins have lower poverty rates, if we change the composition of adults such that a greater percentage of them wind up in the higher educational bins, that will mean lower overall poverty.

So, for instance, if we could move 9 points off the "less than high school" bar and on to the "associate" (or better) bars, we would definitely see lower poverty. After all, you are moving people out of a high poverty bin and into low poverty bins. Similarly, if you could move 6.5 points of the "high school" bar and on to the "associate" (or better) bars, you'd see lower poverty for the same reason.

2. It Didn't Work

Since 1991, we have done precisely what the education-focused poverty people said to do. Between 1991 and 2014, we steadily reduced the share of adults in the "less than high school" and "high school" bins and increased the share of adults in every other bin:

By 2014, the share of adults in the "less than high school" bin declined 9 points from 20.6% to 11.6%. The share of adults in the "high school" bin declined 6.5 points from 36% to 29.5%. Meanwhile, the share of adults with an Associate degree went up 3.9 points, the share with a Bachelor's degree went up 8.3 points, and the share with a post-Bachelor's degree went up 4.8 points.

If the poverty rates for each educational bin remained the same, then the upward redistribution of adults from the lower bins to the higher bins would have led to lower overall poverty. But that's not what happened.

Instead, the poverty rate for each educational bin went up over this time and overall poverty didn't decline at all. In fact it went up.

By 2014, the "less than high school" poverty rate had increased 3.7 points. The "high school" poverty rate increased 4.6 points. "Some college" went up 4.1 points, "associate" went up 3.8 points, "bachelor's" went up 2.1 points, and "post-bachelor's" went up 1.7. Despite the educational gains, overall adult poverty in 2014 was actually 1.1 points higher than in 1991.

As the adults migrated up the educational bins, they took the poverty into the higher educational bins with them:

Over this period, the share of poor adults with "less than high school" education plummeted 20.1 points from 48.3 points to 28.2 points. Every other educational bin saw share gains of 2.6 to 5 points.

Adults these days are as educated as they have ever been, but poverty is no lower than it was in 1991. This is not because the few lingering people with "less than high school" have soaked up all the poverty. Quite the contrary: poverty has simply moved up the educational scale. The poor in 2014 were the most educated poor in history.

3. Why It Doesn't Work

There are a number reasons why aggregate education gains do not necessarily translate into aggregate poverty declines. I will discuss three here.

First, handing out more high school and college diplomas doesn't magically create more good-paying jobs. When more credentials are chasing the same number of decent jobs, what you get is credential inflation: jobs that used to require a high school degree now require a college degree; jobs that used to require an Associate degee now require a Bachelor's degreee; and so on. Obviously the supply of good-paying jobs is not a fixed constant of nature, but there is no reason to think that the supply will automatically go up to match the number of people with the necessary credentials. The types of jobs available in a society, and their level of compensation, is determined by many factors (demand, worker power, technology, global competition, natural resources, etc.) that have little to do with the number of degrees that society is minting.

Second, having more education does not necessarily increase people's productive capacity. Those in the know will identify this as the old "signaling v. human capital" point. The short of it is that, even if jobs did automatically pop into existence to match people's level of productive ability, it's not at all clear that college education necessarily does a lot to increase people's productive ability. Instead, what college education does (at least in part) is signal to employers that you have a certain level of relative "quality" over others in society. As more people get degrees, the value of this signal declines, but more importantly, the point is that the degree was always a signal, not a productivity enhancer.

Third, poverty is really about non-working people: children, elderly, disabled, students, carers, and the unemployed. The big things that cause poverty for adults over the age of 25 in a low-welfare capitalist society—old-age, disability, unemployment, having children—do not go away just because you have a better degree. These poverty-inducing circumstances are social constants that could strike anyone of us and do strike many of us at some point in our lives. To the extent that education does nothing to provide better income support for those who do find themselves in these vulnerable situations, its effect on overall poverty levels will always be weak, or, as with the US in the last 23 years, totally nonexistent.

Technical note. For this post, I used the Official Poverty Metric (OPM) to measure poverty. This is mainly because it's the only metric for which publicly accessible microdata exists back to 1991. The OPM is deeply flawed because it excludes from its calculation taxes, tax credits, and non-cash benefits like WIC, Section 8, and Food Stamps. Because the alleged poverty-reducing mechanism of higher educational attainment is that it increases market income (not welfare income or income from refundable tax credits), the OPM's flaws are not really relevant here. In short, the OPM, despite its problems, works perfectly fine here.