Review Of Auerbach And Kotlikoff On Income Inequality

In the current issue of The New Republic, economists Alan Auerbach and Laurence Kotlikoff have an article titled "We've Been Measuring Income Inequality Wrong."   The article is a short version of a longer scholarly paper that appeared in February as part of the NBER Working Paper series.  Several readers wrote in to ask me to take a look at A&K work and provide my thoughts, so here goes.

I should begin by summarizing my own views on income inequality.  (You can read all you want by following my Income Inequality tag.)  Basically, I accept that income inequality on the low end, otherwise known as poverty and/or extreme poverty, is at least potentially a problem.  I do not accept that income inequality on the high end, otherwise known as rich people and/or very rich people, is a problem. It's the hope and possibility that one might get rich that drive human ambition and creativity, and that have enabled countries like the United States, that allow private wealth accumulation, to become wealthy, to the benefit of all.  Allow the worst of human instincts like envy and jealousy to get the better of you, and you're going down the road of Venezuela.

Anyway, if you take as your measure of income inequality the numbers put out by, for example, the Census Bureau, you get what appears to be a dire picture for a not-small number of those at the bottom.  Using the yardstick of "income" as Census defines and measures it, there appear to be lots of people down there with little to no income at all.  How are they even eating?  The simple answer is that the Census Bureau doesn't count all kinds of things that represent real  resources to low income people but aren't included in the artificial Census definition of "income."  For lots of detail, read this article I wrote back in 2014.

A&K have undertaken what I would call a serious effort to correct the completely phony Census "income inequality" numbers to account for as many as possible of the redistributions already built in to existing major government programs.  At the top, they make big subtractions, mostly for taxes paid; and at the bottom, they make big additions, mainly for government redistribution programs that Census systematically fails to count.  The results:

First, spending inequality—what we should really care about—is far smaller than wealth inequality. This is true no matter the age cohort you consider. . . .  The fact that spending inequality is dramatically smaller than wealth inequality results from our highly progressive fiscal system, as well as the fact that labor income is distributed more equally than wealth. . . .  To be clear, spending power remains extremely unequal. . . .  Another key finding is that U.S. fiscal policy acts as a serious disincentive to work longer hours or harder for more pay.    

Fair enough.  I would note that people who follow the government numbers on these things would clearly not be surprised that this would be the result, particularly because the BLS puts out numbers independent of the Census Bureau, including something called the Consumer Expenditure Survey, broken down by income decile.  And those numbers have long shown that, for example, those in the bottom decile spend about 4 times as much as their "income."  Here is a BLS chart for 2014.  I would still applaud A&K for doing a lot of detailed and quantitative work, and particularly for quantifying the effects at the bottom of many of the government handout programs that Census systematically omits.

And yet, for all their good and hard work, I would still say that the A&K numbers way overstate income inequality, particularly in under-counting the resources available to those at the bottom.  To get the best indication of their methodology, I went to the NBER paper, which at page 38 contains a lengthy list of all the adjustments they made to government income data:

TFA [the A&K adjustment methodology] also handles all major transfer programs taking full account of their assorted incomes and asset tests. The list here includes Social Security retirement, spousal, child-in- care spousal, divorcee spousal, widow(er), divorcee widow(er), mother (father), child, and child survivor benefits (all subject to the family benefit maximum), Medicare, Medicaid, Food Stamps, Supplemental Security Income, Transitional Aid to Families with Dependent Children, and Social Security Disability benefits. The program also considers in fine detail Social Security’s earnings test, the adjustment of the reduction factor, early benefit reductions for retirees, spouses, widow(ers), the delayed retirement credit, deeming, the re-computation of benefits, the Windfall Elimination Provision, the Government Pension Offset provision, and all other major and minor factors associated with Social Security’s benefit provisions. . . .  

Whew!  That sounds rather comprehensive, doesn't it?  Actually, it's not even close.  The government has so many zillions of handout programs that literally nobody can keep track of them, no matter how hard they try.  If I might point out a few things that A&K do not mention:

  • Housing subsidies are of course my favorite, although that may be because I live in Manhattan.  As I pointed out just Tuesday, we have lots of people in Manhattan who get annual housing subsidies of $50,000 and up per family, and some in my own neighborhood who get in excess of $100,000 per family.  I'm measuring this by rental value of nearby apartments, not by the size of the HUD and NYCHA budgets.  
  • A&K point out that they have taken account of the Food Stamp (or SNAP) program.  But that's only the start of federal nutrition programs.  SNAP is about $80 billion per year.  How about WIC ("Women, Infants, Children"), which is another $8 billion; and the school lunch program, which is another $13 billion; and the school breakfast program, which is another $4 billion?  And there are plenty more food distribution programs, many funded by state and local governments and/or private charities.  
  • How about job training programs?  According to a big government study in 2011 that I cited here, there were 47 such programs in 2009, costing some $18 billion.  Supposedly Joe Biden was going to reform them.  Have you heard anything about that since?  Do these things count as resources available to low income people?  If they don't, why do we have the programs?
  • Education aid?  Perhaps it's fair not to count the $80 billion spent by the federal Department of Education on subsidies to state and local K-12 education.  But how about the interest subsidies on student loans for college?  With over $1 trillion of such loans now outstanding, the interest subsidies have to be at least $50 billion annually.  Of course, if you are a low income student, you get a Pell grant.  That's another about $30 billion a year, mostly for the low-income.  Should we mention the value of privately-funded need-based scholarships and fellowships?
  • Energy assistance, clothing assistance, etc., etc., etc.
  • Dare we mention private charity?  Why doesn't that count as resources available to low-income people.

And yes, I still haven't gotten to the big one, the one that is clearly bigger by a multiple than all the others put together.  I'm talking, of course, about that fact that A&K use as the basis for all their calculations the officially reported government numbers.  And these numbers have two gigantic, enormous, overwhelming holes:

  • When the government surveys the people for their income and/or expenditures, they just take the numbers that people put down on the forms, without any kind of double check.  Of course people at the low end massively understate their incomes.  Wouldn't you?  Would you really trust the government not to use your survey response to take away some benefit you are receiving?
  • And then there's the whole subject of the illegal and underground economy.  That is completely omitted from all government numbers.  In this post I cited a figure of $2 trillion per year for the illegal and underground economy.  Divide it up equally among all families in the bottom half of the income distribution, and it's around $20,000+ per family. 

My bottom line:  Unlike many people who call themselves "economists" but are really just shills advocating for the growth of the government (Paul Krugman comes to mind), I think that A&K are serious people trying to do a serious job.  But the fact remains that their work is entirely derived from government numbers, and for that reason is subject to the inherent flaws in the government numbers.  And the government and everybody in it is a shill advocating for the growth of the government.  And all government numbers on income and poverty are intentionally designed and constructed to massively overstate poverty and understate income and resources available at the low end in order to further the advocacy for the growth and expansion of government redistribution programs.

So A&K have done a huge amount of detailed work with a great deal of detailed calculations on effects of complex things like the tax structure and the Social Security program, but then they've left out a few trillion of adjustments that are clearly essential to getting the full picture but are not available in any government numbers.  Yes their work is a little useful, but how about getting to the big stuff?  Yes, you need to guesstimate it.  There's no other way.  You might as well admit that the apparent precision of your other calculations is an illusion.

As to the top end of the income distribution, I have no major reasons to believe that the A&K calculations are wildly off.  Basically, the bottom line is, after taxes the rich are only about half as rich as they appeared before, but they are still very rich.  Well, that's what makes capitalism go.  It doesn't bother me.