Category Archives: Taxes

I’m Very Skeptical About That Yale Law School Study

Screen Shot 2015-09-25 at 3.56.13 PM

If you want to understand people’s policy views, here’s three ways you could get this information:

  1. You could conduct an experiment in a controlled setting that is somewhat related to the policy issue at hand.
  2. You could ask people their policy views.
  3. You could look at their voting records.

Ray Fishman, Daniel Markovitz, Pamela Jakiela and Shachar Kariv chose the first route, conducting a study that ran a redistribution experiment on average Americans (from a Rand data set), somewhat elite Americans (Berkeley students), and very elite Americans (Yale Law Students).

They then extrapolated their findings, which showed Yale Law students preference efficiency over redistribution in the experiment, and wrote in article in Slate with the byline:

Rich elites—even rich liberal elites—don’t believe in redistributing wealth.

I’ll assume they didn’t write that over exaggerated headline, but they did write this:

Yale Law students’ overwhelming, indeed almost eccentric, commitment to efficiency over equality is all the more astonishing given that the students self-identified as Democrats rather than Republicans …. Our results thus shine a revealing light on American politics and policy. They suggest that the policy response to rising economic inequality lags so far behind the preferences of ordinary Americans for the simple reason that the elites who make policy—regardless of political party—just don’t care much about equality. [emphasis mine]


I attended Yale Law School from 2003 to 2007 (I took a semester off to work in Sierra Leone so I graduated in January of 2007).

At least back then, Yale Law School was a very, very, very liberal place. I’m a Democrat, and I remember feeling, at times, feeling very conservative in some of my views.

So I was surprised to see these results (note: I read the Slate article and the article abstract, but didn’t have access to the full study).

The students I knew at Yale Law School, myself included, were generally very in favor of redistributing income.

My experiences aside, what makes me skeptical of the author’s claims is that the Democratic Party is currently pursuing very redistribute policies: health insurance paid for by taxing the rich, a $15 minimum wage, universal pre-k paid for by taxing the rich, etc.

This seems like an odd thing to do if you don’t care about redistribution.

If elite liberals were able to enact their policy preferences, we would likely see a significant shift towards equity and away from efficiency.

Now, who is voting for these Democratic elites, it’s Yale Law students! And who is not voting for these Democratic elites, average Americans living in red states!

To say, because of the results of artificial experiment, that the average American favors redistribution more than Yale Law students seems somewhat odd given that Yale Law students (I’m fairly confident) are much more likely to vote people who support significant redistribution efforts.

When it comes to policy preference, voting patterns are much better evidence than artificial experiments.


One other point: there seems to be a big difference between valuing efficiency when the person you’re sharing with is anonymous and valuing efficiency when the person you’re sharing with is in major need of medical, educational, or other necessary services.

It was unclear to me that the study replicated the wealth differences we consider in real life tax transfers.

In a world of generally equal wealth, efficiency (increasing the size of the pie) is much more important than equality (redistributing wealth). Of course, one could argue that this is always the case, but it seems especially true when equality is already at reasonable levels.

My guess is that the study results would have been different if there had been a narrative revealing that the person the YLS student was sharing money with was in great need.


Even if the authors draw weak conclusions from their data, the data does raise an interesting question: why is there an inverse correlation between private (charity / sharing) and public (voting / policy preference) redistributive actions?

I don’t know the answer, but, in practice, it seem unimportant.

If you want to understand who wants to redistribute the most resources, find the people who are willing to support the greatest amount of resource redistribution.

Given that charity and sharing, right now at least, are dwarfed in total amount by tax redistribution, it’s tax policy that really matters.

What people do in experiments (or in their private lives) is interesting, but in many ways it’s irrelevant.

Who they vote for is not.

Poverty, the Brain, Tax Rates…. a Way to End Childhood Poverty?


A new study links poverty to brain development.

The findings:

Screen Shot 2015-04-01 at 3.26.57 PM

I’m not an expert in brain science, so I will not attempt to evaluate the science. I’ll just note that the findings are based on correlation, not causation.

This findings mirror other findings linking poverty to reduced IQ.


Let’s say that we prove that poverty negatively affects brain development.

Furthermore, let’s assume that the negative effects can be completely eradicated by increasing the wealth of those living in poverty.

The question then becomes one of wealth transfers: what tax rates would we need to give enough wealth to those living in poverty in order to eradicate the negative brain development effects?


The Laffer Curve is a simple way to understand the trade offs between tax rates and revenue. The key insight of the curve is that there is some point at which tax rates are so high that they reduce tax revenue. For example, at a 100% tax rate, presumably no one would work, which would mean no revenues for government.

So, if you want to to raise taxes to fund wealth transfers, you want to make sure you stay on the “good” side of the curve (once you cross to the other side increased taxes do not increase revenue).

There’s not universal agreement on where the curve peaks, but estimates generally fall in the 50-70% range.


There is much debate about how to calculate the number of adults living in poverty. Similar debates are had about the number of children living in poverty. Estimates are around 16 million children. Let’s assume this is roughly right and let’s also assume an average of two children per household.

That would be 8 million households with children living in poverty.

Let’s say we wanted to transfer $20,000 worth of wealth to each of these households, which would presumably lift most of these households out of poverty. For example, someone with children earning minimum wage probably makes around $20,000 a year post tax transfers, so this proposal would double her earnings.

To annually provide $20,000 in additional funds to 8 million households, we would need to raise an additional $160,000,000,000 (160 billion) in tax revenue per year.

Currently, the federal government raises ~3.3 trillion in taxes.

Raising another 160 billion would require a ~5% increase in federal tax revenue.

This is all back of the envelop math, but I think it’s roughly right. Let me know if I’ve made any major errors.


A 5% increase in federal tax revenue will likely keep us on the “good” side of the Laffer Curve, especially if the revenue is drawn from both middle and high income taxpayers (the broader the base, the less likely any specific group is going to see their tax rates skyrocket, though the top 10% of tax payers produce the majority of income tax revenue, so the impact would likely be concentrated at the top).

And if you don’t think $20,000 is enough, make it $30,000; I think this would still keep us on the “good” side of the curve (based on eyeballing this somewhat outdated chart).

The next question would be how to transfer this wealth. Direct transfers could reduce the incentive to work, while wage subsidies would only impact those who do work. Direct transfers that are predicated on behavior (child attendance in school, etc.) could also be an option.

My preference would be to emphasize a mix of wage subsidies and conditional transfers, but I’m not an expert on the mechanics of wealth transfers, so others might come up with better ideas. There is probably some trade off between covering all households in poverty and incentivizing work. We may also want to keep the tax “reward” for having children from being too high.


Given my philosophical beliefs, I’m generally inclined to raise taxes as high as they can optimally go. No one chooses her genes; no one chooses her environment; and I’m skeptical of any strong versions of free will. To paraphrase Obama: you didn’t build you.

So I find hard to justify why some people should have way more money than others.

My philosophical beliefs aside, there may be strong scientific reasons to conduct wealth transfers.

Ensuring everyone’s brains develop to their full potential seems like a good enough reason to me.

Lastly, given my lack of expertise in taxes and transfers, all feedback especially welcome.