Economists Do It With Models

Warning: “graphic” content…

Bookmark and Share
New Income Data Available, With Bonus Educational Arms Race Analysis…

September 17th, 2013 · 6 Comments
Fun With Data · Game Theory

In an economist’s world, the census people releasing updated data on income, poverty, and health insurance coverage is likely to be, by far, the most interesting thing to happen today, if not all week. (Joking aside, the slides that are linked to there are in fact pretty interesting.)

Overall, the numbers look pretty grim. The report begins by basically saying that there’s been no statistically significant change in median income, the poverty rate (i.e. the percentage of people in households below the poverty line), or the proportion of people with health insurance in 2012 as compared to 2011. Taken by itself, this information isn’t so bad- things haven’t gotten better, but they haven’t gotten worse either, and it’s only one year. But let’s take a look at a few graphs that tell a broader story…

I can’t help but get a little ragey when I look at this graph, since my knee-jerk takeway is that we rolled over to a new millennium and prosperity for normal and poor households started going in the crapper. Neil Irwin makes a closely-related point about a lost generation of economic growth, since the numbers show that the average household, in inflation-adjusted terms, had a lower income in 2012 than in 1989. As if this wasn’t frustrating enough, I can’t help but notice that we are at the only time in recent (or not even that recent) memory that median income has continued to decline significantly after a recession has ended.

If you are the sort of person who gets riled up about income distribution, I recommend that you not look directly what I will call the “non-money shot” in the census deck:

Now, I can guess the immediate reaction that this graph gets from people, but I want you to just go with me for a second. If you look at the period from the turn of the millennium to the present, it appears that households at all sorts of different income levels have stagnated, not just those in the middle or bottom of the income distribution. In other words, the vast majority of us appear to be in a holding pattern together, and this really underscores the point that the vast majority of economic gains in recent years have gone not only to the upper end of the distribution but to the upper end of the upper end. (I am guessing that the Occupy Wall Street people were somewhat arbitrary when they chose their 1% figure, but they appear to be surprisingly accurate. In any case, it’s quite interesting to me that a household in the 95th percentile would fit in fairly nicely at Zuccotti park, except perhaps for the designer shoes.) I suppose this all makes the finding that 95% of income gains in recent years have gone to the wealthiest 1% not particularly surprising, though it’s worth noting that the top earners were also those hit hardest financially during the recession. (Getting a res at Per Se got significantly easier at some point in 2008.)

There’s a lot more interesting stuff to look at in the census slides, and I may come back and do another post on them. For now, however, I want to take a detour and think about education a bit. No, wait- I want to think about prisoners and weapons arsenals first. So, the key feature of the prisoner’s dilemma (see here or here for a review) is that each player doing what is individually better for him results in an outcome that is worse for both players than if they had cooperated. The logic of the prisoner’s dilemma is often carried over to arms races since, regardless of whether your enemy has more weapons, it is better for you to have more weapons, but this implies that both you and your enemies try to stockpile weapons and basically cancel out the effects of the increased arsenal. This leaves both you and your enemy worse off, since neither strategic position is better than it was before but you both spent a shit ton of cash on guns.

Now, education…I’ve hypothesized for a while now that we are in the midst of an educational arms race. Think about it:

These numbers are hypothetical, but go with me here. If most people don’t go to college, we’ve seen historically that there are significant returns to going to college, so the numbers above seem reasonable if you count “College” as college cost ammortized over years of work and the numbers represent yearly income. Similarly, if a lot of people do go to college, there is an incentive to go to college in order to not get left behind. That said, it’s entirely possible that most people going to college just means that companies start requiring that even their file clerks have four-year degrees. (The phrase “the college degree is becoming the new high school diploma” at the beginning of that article is spot on in describing an arms race, in fact.)

So is this what is actually happening? In general, it’s entirely possible that companies are taking advantage of the more educated workforce and creating jobs that put the skills that the college-educated workers have learned to good use (and paying the workers more as a result). But, if this were the case, wouldn’t we see median income increasing as the median household gets more educated? (Empirically, we do see increases in educational attainment over time, so it stands to reason that the median earning household is more educated now than in, say, 1989.) Instead, we see incomes stagnating and tuition skyrocketing. Even if we hypothesize that the people at the bottom of the income distribution are pretty consistent over time in not going to college and vice versa for the top of the income distribution, the evidence suggests that there is a decent amount of education flux in the middle of the distribution that is not being matched by increases in economic prosperity.

I don’t think this is what my mom had in mind when she warned me about the dangers of keeping up with the Joneses, and the frustrating part is that this outcome results from individually rational decisions, so the solution is not as simple as just telling people to not go to college. In fact, it still appears to an individual that college is still worth the investment, but, rather than the investment earning a college premium, the investment appears to help people avoid a non-college penalty.

Tags: Fun With Data · Game Theory

6 responses so far ↓

  • 1 Jeff // Sep 17, 2013 at 8:09 pm

    ” In general, it’s entirely possible that companies are taking advantage of the more educated workforce … (and paying the workers more as a result)”

    Interesting example, but the above continuation is a non-sequitur. Worker pay is based on value creation only in a comparative sense, not an aggregate sense for many of the same reasons you are outlining in this article. If you double your output there is no a-priori reason that would double your pay. If you increased your output significantly compared with the unemployed people trying to take your job from you, however, you deserve a raise regardless of what happened to your physical output. This difference can usually be ignored in micro, but can cause massive errors in macro.

    Nevertheless, it is a good article, thanks.

  • 2 Jeff // Sep 17, 2013 at 9:07 pm

    I forgot to mention. The graph gets a lot more interesting if you start at the end of WWII.

    The median income close to doubled in 20 years from 1950-1970. Any changes in the 1970-present time range are trivial by comparison.

  • 3 dan // Sep 17, 2013 at 11:42 pm

    nice post. It is sad, and of course, devastating to economists as policies that would produce superior outcomes have been well known to economists for more than two generations.
    I don’t think this means that economics has failed, I think it means that economics is part of the real world. Time to put the political back in political economy, explicitly.

  • 4 J.D. // Sep 18, 2013 at 11:26 am

    Didn’t the “globalization” craze in the U.S. really pick up in the late 90s? Seems to time well with the stagnation. Just sayin’.

  • 5 Trevor // Sep 18, 2013 at 1:04 pm

    From looking at the real household income graph, I’m curious if there is a correlation between the length of a recession and the amount of time before incomes start to see an increase. You mentioned “I can’t help but notice that we are at the only time in recent (or not even that recent) memory that median income has continued to decline significantly after a recession has ended.” but it looks like something similar happened between 1991 and 1993.

    Great article; thanks!

  • 6 econgirl // Sep 22, 2013 at 1:54 am

    @ Jeff: That is actually decently consistent with what economists have to say about the concept of convergence.

Leave a Comment