The Other Income Inequality: The College Wage Gap

While Piketty focuses on how the really rich are pulling away from everyone else in terms of wealth, others have made the case that the college wage gap is the most pressing form of income inequality.  Here is a good and detailed article on the subject.  His claim is that while the 1% disproportionate increase in wealth is alarming, the growth of the college wage gap, if redistributed, would increase the welfare of the poorest people more than any gains from redistributing the increase in wealth of the economic elite.  This follows mostly because there are just so few economic elite.  The fact that it is only 4 times as great when the percentage of the population with college degrees is 25 times higher should be slightly worrying.

While I agree that this is worrying, I am not sure it is AS worrying.  After all, the physics professor making $100k is not in any danger of twisting political power around his finger to get what he wants.  A person with a billion dollars that is going to keep growing is far more menacing; look at all the activism by the Koch brothers.  Secondly, the gains of the 1% seem far less justified than those accrued by people investing in education and reaping the returns.

Furthermore, much of the gap is explained by an actual decline in the real wages of low education workers.  The loss of manufacturing jobs  with unions bolstering wages due to offshoring, automation and just plain efficiency gains has pushed low skill workers into a service sector with no organized labor movements and very low wages.  Also automation is starting to erode their value in that sector as well and will likely continue to do so.  For instance if driverless cars are perfected say goodbye to taxi and truck drivers.  A minimum wage that hasn’t kept pace with inflation hasn’t helped their cause either.

The article linked also shows a stagnation in the growth of wages for college graduates in the last decade that I also find worrying and that he cannot explain with his skill supply-demand model.  I say this because the slow growth of the last decade probably shifted people into college since the opportunity cost of wages is lower and there is evidence that entering the workforce in a bad economy depress yours lifetime wages.  Also, every education cohort has stagnating wages over this period.  It seems to me that the economic elite may have finally managed to siphon off the earning power of the one segment of workers that was able to avoid their vampiric gaze and garner some wage increases over the last few decades.  We will need more data to confirm that, however.

Now inequality stemming from educational attainment is a tricky subject.  At some level I do think that college imparts skills to those who attend.  However, unlike the Science article I am not convinced that learned skills explain the entirety of the wage differential.  After all, the most popular majors for a Master’s degree are Education and Business.  Yet these are exactly the majors that have been shown to learn the least.  Recall that much of the differential is driven by the wage gains of post-graduate degree holders and that Master’s degrees are far more prevalent now than ever before.  Now the article posts a chart showing wages as a function of skill in various countries as if that were enough evidence that it is skills acquired in college driving the wage gap.  Unfortunately, you need two more pieces of evidence: that colleges improve skills and then, together with something like the aforementioned chart, show that the skill increase explains the wage increase.

However, that is not the trickiest part.  I am fairly certain that people that graduate college are in fact more skilled.  But that doesn’t mean that college gave them that skill.  The problem is that the group of college graduates is drawn from a very selective part of the population, at least in the U.S.  Increasingly, students from the top quintile of the income distribution are dominating attendance at good four year universities and almost nobody from the poorest quintile is getting in.  Yes, if you are poor and get into Harvard you get a free ride, but the chances of that are miniscule compared to your wealthier peers.

With that knowledge we arrive at a simple explanation for why the college wage gap has increased, but may not be attributable to college actually teaching anything.  It’s well established that children from families with higher socioeconomic status do better at nearly every conceivable measure.  They are smarter, more ambitious and determined, they have more emotional support (not to mention financial) and they are usually just better adjusted and have all the bourgeois values we find exemplary as a society.  Basically all the perks of being born in a family not just eeking out an existence.  We expect these people to do better in life and earn more income and these are the same people that overwhelmingly fill college commencements.  Therefore, college may have done little to improve these people and yet we would expect graduates to make more money than their peers without degrees.  Thus, the truly worrying fact here is not the income inequality derived from collegiate attainment but perhaps how we have failed families long before their children are of age to attend college.

But lets say that college does impart skills that explain the wage gap.  This would still suggest to me that the far more worrying fact is how we have gated college admissions by wealth of families.  If only wealthy kids can make it past admissions (on account of their already enumerated advantages) and pony up the money to attend college then we will live in an increasingly stratified society with little income mobility.  If you are poor you can’t go to college to enable you to find a higher paying job and if you are rich, well you get to spend 4 years getting drunk and partying and then earn more money than your less well-endowed peers.  In fact, among the OECD the U.S. has one of the lower income mobilities and this inequality of access to higher education is surely contributing to that sad statistic.

In order to truly settle this question we really need to look at the same cohort of people, with similar grades, family wealth and all the other factors that go into your success as a human being.  Then we need to divide them into a group that goes to college and those that do not and compare their wages.  At which point we can finally settle if college is in fact the source of the college wage gap.  It would still not sort out whether that is because of skills or credentials, that would require the research path I already outlined above.


Piketty, the New Marx

The parallels between Marx and Piketty are quite obvious.  Both men are trying to grapple with the endgame of unfettered capitalism.  The big difference here is that Piketty has data and a century of violent and turbulent experiments in economic systems behind him and Marx mostly just had ambiguous models rarely compared with facts from which he jumped to conclusions.  That in some cases Marx was approximately correct does not change the fact that Piketty’s empirical approach is much preferred.

Now I turn to what Piketty has to say about the future of inequality.  As I said inequality of wealth is reaching its historic Belle Epoque levels and labor income inequality is at historic highs.  Given this backdrop Piketty introduces the simple inequality r>g, with r the rate of return on capital and g the total growth rate.  If this inequality holds then inequality of wealth will continue to grow.  This follows because if I reinvest all my wealth then my wealth increases by r, but the total economy only grew by g, and so a larger amount of the economic pie is now mine.  The larger the discrepancy, the faster a few people come to own everything.

You can’t argue with the logic of this inequality, so instead Piketty’s critics question whether this inequality will hold for much longer.  Now my first order response is yes.  There are two things keeping this inequality true: first growth is slowing and second the rate of return seems fairly constant.

Growth has been slowing in the developed countries for awhile now due to lower population and productivity growth and developing countries will likely end their catch-up growth soon and reduce their birth rate.  Thus, we expect g to be lower in the future.

As Piketty shows, the rate of return on capital, r, has been a fairly constant 4-5% for much of human civilization.  But even on the shorter, more relevant time scale of post-WW2 this is approximately correct despite huge changes in the C/I ratio.  Or more specifically, the C/I ratio was 2-3 after WW2 in Europe and has now doubled to 4-6 and yet if anything, the rate of return on capital has increased from those dark days of war and triumphant social democracy.  Yet, simple supply and demand should tell you that if the amount of capital increases then, all other things being equal, its price, r, should go down.

We can put this more concretely by following Brad DeLong’s model.  He posits a warranted rate of return, r*, of g+w where w is the wedge, a fudge factor to encapsulate war, taxes, nationalization, stupidity, etc. that would raise the price of capital.  The actual rate of return is then r=ρ(C/I)^(-λ) where ρ is the efficacy of capital in influencing politics and society to its benefit and λ describes how r varies with C/I.  If it is unity then they follow each other in lockstep and nothing of the wealth accumulation Piketty prognosticates can happen.  Instead we need a λ<1 so that increases in C/I do not cause a commensurate decrease in r.

Now as I already stated r has been remarkably constant over the last half century despite a doubling in C/I.  This requires a mixture of a λ less than one and a ρ that has increased.  Now it is conceivable that increases in ρ explain the entire effect and I would not discount such an idea.  After all it is in actuality a function of the C/I ratio.  As capital increases we would expect its influence to increase as well.  If this dependence goes as (C/I)^(λ)  then it would also stabilize r and Piketty’s increasing inequality would come into play.

I think more worrying is the empirical result from Piketty that the wealthiest investors also get the highest rates of return.  If the global average is 4-5%, then Harvard and people on the Forbes list of wealthiest people are getting 10% or more and it seems to decrease as you get less rich.  This is at odds with standard economic theory that says every investor should get the marginal product of a unit of capital, that everyone’s rate of return should be the same.  That the rate of return is an increasing function of your existing capital means that existing inequalities of wealth will be amplified as time goes on.  Even if r does fall to the growth rate on average, if the already rich can secure better than average rates of return then inequality will continue to grow.  In terms of our model, ρ appears to be a function of an individual’s capital and a not insignificant one.

This is frightening not only for its implications in terms of future inequality, but also because it suggests our democracies are increasingly being arrogated for the objectives of the wealthy.  One of the best articles on the recently launched Vox news site was on this very fact but from a political science perspective.  The study showed that the correlation between policies favored by the average citizen and their chance of being enacted was essentially zero.  Meanwhile it was a meaty .76 for economic elite and only .56 for net interest group alignment.  You would have to posit that the economic elite know what’s best for the average citizen for this to make any amount of sense.  I, however, do believe that they know what is best for themselves.  Anyways it is interesting that two different disciplines can come to similar conclusions about the power of wealth in modern society.

Finally, one more note about  λ.  If the robot future does come to pass as is increasingly likely we would expect capital to become more of a substitute for labor than it has been in the past.  This would increase the demand for capital and further staunch any decrease in the rate of return.  In fact the ultimate endgame of a robot society has been explored before.  If robots can create more robots then production is essentially solved for all tasks that can be automated and the marginal product of labor is zero.  At this point the only wealth left is ownership of natural resources.  It seems to me that without government interference this is the most unequal society imaginable, not dissimilar to the Russian oligarchs that profited massively off the natural resources of Russia while the average citizen is still quite poor.

Now Piketty doesn’t outright say that the economic elite has taken control but he does blame a conservative rise to power since the 80s that resulted in less progressive taxes and reduced regulation and nationalization.  All of these things reduced the wedge which should increase the C/I ratio.  You can read about the Doom Loop of Oligarchy, but as should be obvious as the rich get richer they get ever more influence that allows them to get even richer and so on.

So what should fill the gap of the Reagan Revolution?  Well, bringing back much higher progressive income and estate taxes is one step.  But for Piketty, the best solution is a global progressive tax on wealth.  He suggests this for data keeping purposes and to ameliorate the increasing concentration of wealth.  As an economist, I am almost convinced the former benefit excites him far more than the latter.  As he points out, Earth runs a trade deficit with Mars.  That is, there is an amount of capital seemingly unaccounted for after adding up all the net capital flows from every country on Earth.  However, estimates suggest a large part of that is rich people sheltering their money from taxes, wives, etc.  This is why it has to be a global tax and why the extra data would be useful apart from further analyzing trends in wealth.

Now the level of this tax would be at most a few percent per a year.  Ideally, in my mind it would be set such that it redresses the greater rate of return afforded by increasing capital.  So if your wealth class pulls 8% then it is a 3% tax and if your class pulls 10% it is a 5% tax.  This encourages people to find the best return on capital and should concentrate wealth in the hands of people that are actually better investors while the flighty heir squanders all of his money.

Now much has been made about the feasibility of such a tax since it requires global cooperation which is not a particularly successful field of endeavor.  However, I think that if the U.S. and the Eurozone implemented their own capital tax and then shared financial data we would be well on our way to implementing such a tax.  Their combined influence could certainly crack a lot of tax havens as the U.S. already did with Switzerland.  Yes an all in one step is impossible, but if the big players do it then it will already have great efficacy and others will surely follow.

Before I end my discussion I want to talk about inheritance.  If you recall in my previous post, Piketty talks about inheritance in Belle Epoque literature.  Everyone knew that it was better and easier to marry into wealth than earn it.  Immediately following WW2 this was not true at all because most wealth had been eradicated.  However, with C/I once again reaching Belle Epoque levels inheritance is poised to become very important again.  Brad DeLong estimates 30% of national income could eventually be inheritance under current conditions.  Piketty has estimates in the double digit percentages for the number of people that will inherit an amount of money greater than the amount of money that an average person will earn in their entire lives.  Realize though that the average person is unlikely to be the beneficiary of such an inheritance.  It will probably be bequeathed to someone with good earnings potential from a college degree or any of the other perks associated with being raised in a family with wealth saved larger than the average person earns their entire life.

On top of the absurdly wealthy pulling away from everyone else, we see a just as pertinent inequality in the patrimonial middle class pulling away from the bottom half of society.  Remember the very sad fact that half the population owns basically nothing and be grateful if your bank account (which many poor people do not have) contains an amount distinguishable from zero.  Not only will you recognize the plight of so many, but being grateful about things is a great way to improve your happiness.  You can thank me later.

Piketty, Economic Historian

I just finished Piketty’s book, Capital in the Twenty First Century.   The one thing that everyone, including his critics, can agree on is that he has elevated the data entering the discourse on inequality.  That his critics then ignore his data does not deter them from assuring us that it is very good.

First let me summarize the historical part of the book so that you can pretend to have read it just as well as everyone else at your dinner party.  Then I will talk about the implications and his policy prescriptions in a later post.

Growth and the Capital/Income Ratio

Now all of the trends revealed by Piketty can be divided into three eras: before WW1, the intrawar period and post-WW2.  The world was economically extremely stable before the industrial revolution, demographic growth was minuscule, less than .1% with productivity growth (due primarily to technological invention) at a similarly low level.  Obviously data from this period is quite scarce, but such estimates follow from a basic analysis of subsistence level economics and the cumulative effects of population growth.  That is, if population growth were any higher then there would be a lot more people around by modern times considering what we know about population in antiquity (ancient Rome for instance).

The point is that this is hugely different from what we have seen from the industrial revolution onward.  With industry came productivity growth an order of magnitude higher than that seen at any previous point in human civilization.  Until the World Wars, this productivity growth came with a concomitant increase in demographic growth.  It was only after WW2 that productivity growth really outpaced demographic growth in the Western nations, a consequence of historically unprecedented productivity growth and a declining birth rate, especially among European countries.

However, all estimates of the future suggest that demographic growth will continue to trend downward across the world.  Countries on the world technological frontier, i.e. the developed nations, will probably only experience about 1-2% productivity growth a year, which is a bit lower than experience in the recent past, but still amazing over the scope of human history.  As more countries reach the technological frontier global productivity growth will fall to this level as well.  This means that overall growth, the sum of population and productivity growth, will likely be lower than we are accustomed to, though again still quite high compared to before the Industrial Revolution.

So the basic trajectory is that the world was quite economically stagnant until the last two hundred years.  This burst of growth though will likely turn out to be anomalous as going forward we can expect growth to fall.

Piketty first informs us of the history of the capital/income ratio which he claims is better for historical analysis than the more prominent labor/capital income split.  As he notes, the World Wars caused the latter ratio to vary quite a lot, but it has now settled back at similar levels to the Belle Epoque. Because the Industrial Revolution changed the nature of capital fundamentally from land to industrial and financial assets, there must be something else at play stabilizing the labor/capital income ratio.

Now, the capital/income ratio is essentially the ratio of the amount of stuff a nation has versus how much it produces in a given year.  At an individual level if the only thing you owned was a house worth 300k and you made 100k a year then your capital/income ratio would be 3.

Piketty fastidious data taking allows him to track the capital income ratio in a handful of  countries, though throughout the book he focuses heavily on France since they have the best and longest record of economic data.  Anyways, the capital/income ratio is 6-7 in Europe in the Belle Epoque and about 4-5 in the U.S.  It then tumbles, reaching a low point after WW2 before again reaching similar levels in contemporary society.

Now you can get the labor/capital income split from the capital/income (C/I from now on) ratio by multiplying the latter by the rate of return on capital.  If you have a rate of return of 5% and a C/I ratio of 6 then 30% of income is from capital with the rest accruing to labor.  Thus, we can see that if the C/I ratio is similar to its level in the past then if the rate of return on capital is also stable that this would explain the stability of the labor/income ratio.

And indeed, the return on capital seems to always be 4-5% on average.  And by always I mean throughout human history it is always about the same.  There is no economic reason it should be this rate.  This is born out because even with the vast shifts in technology and society in the last two hundred years the rate of return on capital is still approximately the same as it was in Julius Caesar’s time.  Instead it must be some kind of universal human psychological effect, a combination of our discount rate (i.e. how much we value present versus future money) and what we deem fair.  An interesting discussion for another time, perhaps.

Since the rate of return on capital is relatively constant, we see that the determinant of the labor/capital income ratio is drive by the C/I ratio.  So what determines this ratio?  Asymptotically it approaches the ratio of savings to growth.  This can be visualized easily by realizing that savings is the rate of increase of capital and growth is the rate of increase of income so their ratio is the level at which C/I remains constant.

Thus, we see why the future trajectory of growth is so important as it will at some level determine the C/I ratio.  It  does not take a large change in growth to alter the C/I ratio.  For instance if we take a representative savings rate of 12% and growth rate of 2% we recover a C/I of 6, very near what we now observe.  However, if growth decreases to 1.5%, as is expected, then the ratio becomes 8.  We might also expect the numerator, savings, to continue to increase as people prepare for longer retirements.

A good case study of this relationship is Japan.  They have an elderly population with little in the way of demographic growth.  Thus they have a very high savings rate and a very low growth rate so we should expect a high C/I.  Indeed, Japan’s C/I is 7-8 among the highest in the world.  This simple formula has a high degree of predictive power.  It also explains why in the low growth past the C/I ratios were much higher than we saw in much of the 20th until very recently.

The Distribution of Capital and Income Within the Population

Let me now transition to a topic that is far more interesting to most of us, how wealth and income are distributed.  The C/I ratio tells us absolutely nothing about this topic so Piketty compiles more data to get at this highly pertinent question.

What he uncovers is that in the Belle Epoque inequality was extremely high in all of Europe.  Something like the top decile of wealth controlled 90% of the wealth.  Furthermore, the wealth distribution was much like a step function, you were either impoverished or you were among the lucky few with land that put you in the upper echelons of society.  There was no middle class as we conceive of it today.  The U.S. due to its high levels of demographic growth from mostly poor immigrants and lack of existing capital had a much more egalitarian distribution of wealth.  The one exception was the South where the establishment of large plantations and the ownership of slaves mimicked the conglomeration of capital in the hands of the few seen in Europe.

The other important characteristic to note was that all the very wealthiest people derived their income almost exclusively from capital.  Income from labor was paltry in comparison.  Apart from a very lucky few merchants, you inherited or married your way into real wealth.  In an interesting aside, Piketty uses a story from Balzac’s Le Pere Goriot where the character Vautrin lectures us on how marrying a rich heiress is the easiest and only way to secure a good income; that putting in a lifetime of service as a relatively lucrative lawyer would still pale in comparison for much more toil.  As a result, income from labor is naturally far more egalitarian than that of income from capital in this period.

In the immediate post-WW2 period inequality of wealth reached a nadir.  Capital was decimated by the wars and Europe went through a period of rapid catch-up growth.  The  new socialized state emerging from the ashes of war greatly ensured that the spoils of this growth were distributed throughout the populace.  A patrimonial middle class, as Piketty calls it, emerged for the first time ever.  That is, a substantial fraction, that Piketty identifies with the 4 deciles below the top decile of wealth, of the population actually managed to accrue some amount of wealth.  For once in human history it seemed that your economic status really was mostly due to the fruits of your labor.  Of course, this ignores the fact that the bottom 50% of the population still had less than 5% of national wealth to their name.

Sadly, this period ended around the 80s due to a complex mix of political and economic factors.  The share of national capital in the hands of the top decile is again above 50% if not close to 70% in the developed nations and the rate of increase seems to be, if anything, accelerating.  This has mostly come at the expense of the patrimonial middle class as the poorest half of the population essentially had nothing to lose.

Now while capital income (derived from wealth) is still far more unequally distributed than labor income, the latter has seen a dramatic decrease in its egalitarianism.  While all the very richest people derive the bulk of their income from capital, you can now go very high up the income scale and still find people that derive most of their income from labor.  This is, according to Piketty, due to the rise of the supermanager, as I discussed in a previous post.

Now in standard economic theory, productivity growth should translate into wage growth.  What Paul Krugman pointed out two decades ago was that 70% of the gains of productivity growth have gone to the top 1% in the income distribution through the 80s with the trend continuing today.  And you can show that the top .1% garnered an even more outsized proportion of the gains.  Therefore the really really rich are pulling away from the really rich who are pulling away from the just plain old rich and so on.

The War Years and Their Consequences

Before I continue let me convey the economic shock the world went through during both World Wars.  As I said earlier, inequality hit a trough after the wars and this coincided with a complete eradication of most capital in mainland Europe with lesser shocks in the U.K. and U.S.  Recall that the top incomes are all based on capital and particularly so in this period and therefore the trajectory of capital is the trajectory of the economic elite.  Some of this was just the plain destructive nature of war.  However, there was also a Great Depression and the government responses to it.  Things like rent control, nationalization and progressive income taxes with very high marginal rates came into being during this period.  All of these reduced the power of the rentier class.

One other fact that might be unfamiliar to most people is that nature of inflation in the past.  We live in a world of low stable inflation, but this is actually a modern fact of life.  Before the wars inflation was essentially zero.  Five thousand British Pounds were worth about the same amount for centuries.  Compared to this even the inflation rate of 2% we have now is enormous.

Now after the World Wars pretty much every country had a fiat currency and used that power to engage in high levels of inflation to get rid of their debt.  Combined with historically unprecedented tax rates they managed to overcome their war debt relatively quickly.

Contrast with lets say Great Britain after the Napoleonic Wars that racked up debt equal to 200% of GDP.  They did not have inflation as a tool and taxes were extremely low compared to modern standards.  As such they borrowed money and the only people to borrow from were the small subset of people that actually had capital.  And unlike today, government debt had very good yields, comparable to other asset classes and remember there is no inflation in this period.  So the owners of capital were also beneficiaries of government spending at this time.  In the end it took a century for Great Britain to bring its debt levels down.

I bring all this up to show how the inflation engaged in during and after the World Wars was a significant factor in reducing capital.  This isn’t always true, any assets that appreciate with inflation are fine, but government debt is almost entirely denoted in nominal terms and so inflation is particularly ruinous to anyone holding that particular asset.  While the owners of capital were becoming more sophisticated at this time, they still invested heavily in government debt and were thus wiped out.  Inflation would not necessarily curb inequality today as a middle class now exists with non-negligible wealth and the truly rich are much more sophisticated investors in a wide variety of financial assets.  In fact I would suspect government debt is mostly held by said middle class since the very low yields are very unattractive for the knowledgeable investor.

Now at the same time labor income remained relatively stable.  So again, the collapse of inequality was really the collapse of capital.  Then Europe had its post-war boom which further mitigated the benefits of previous stocks of capital.  After all if the pie is getting bigger every year then the little bit of pie you saved from last year starts to look smaller.  So we had a novel period where the distribution of wealth was more egalitarian and mirrored the distribution of labor income quite closely.  One could conceivably describe it as a meritocracy.  That period, however, is over and the sooner we get over the euphoria of that age the better.

Now that is most of the background history that is necessary.  From here we look at the future of capital and income accumulation and how to deal with its increasing agglomeration among a small part of the population.