Gini coefficient for U.S. universities

Science Magazine

For the pointer I thank J.O.


The article suggests that the increase in inequality in research funding is an artefact of more institutions trying to do research: "If we restrict our analysis to a limited set of universities that have always been active in research, there is no clear trend." See (gated)

Sure, just like if you limit your analysis to scions of the Vanderbilt clan, you'll find that all Americans are obscenely wealthy!

Which is odd because that scion of the Vanderbilt clan, Anderson Cooper, thinks that most Americans are obscenely poor.

There is a clear case for redistributing the endowments though. Think how hard it is for the poorer colleges to compete on an equal playing field? Harvard should man up and give the cow colleges most of its cash every year.

Worse than that, they hog most of the research grant money and are the few places that can afford data.

Their profs get to do research on exclusive datasets, and then lay claim to being better researchers.

How often does a top notch researcher from a third tier institution make it to a top tier? Few, unless you claim to be a Native American and marry a Harvard prof.

I suggest a version of George W. Bush's education policy in Texas. We could call it the George W. Bush Education Equity Act. Who could object to that?

Simply put, instead of supporting the best researchers *over*all*, which is inequitable because they are all at Harvard and Yale, we should support the best researcher, singular, at each university. Every university would get an equal grant. Funded by the Feds - as well as Federal management of the Universities' collective endowments in a big central fund. Because the government knows how to spend money better than we do, right? So the best geologist at the Jim Jones Bible College would get the same grant as the best geologist at Princeton. As would the second.

I doubt that the science would be any worse and might even be better. And, anyway, equity. Who could object?

I think that dig at the Law Squaw was quite uncalled for. But I did enjoy it.

"We could call it the George W. Bush Education Equity Act. Who could object to that?"

Could we go with the "Barack H Obama No Community Organizer Left Behind Act"?

Morgan Price's comment I think is spot on, the chart is misleading and doesn't really support an otherwise strong point. I agree with the article's author that it is difficult to run at least a scientific research group in the US because it is highly competitive to get the necessary funding sources to pay for grad students (~$80k/yr incl. tuition) and equipment.

Comparing all 'research institutions' however does not reflect that: many universities support some sort of research component to their faculty's time, but it is not a central part of their strategy. They do not hire or promote faculty based on it, do not invest their endowment money in it, etc. That is not like comparing people's incomes: everyone wants a higher income, not every department wants to devote their time to research. I know Professors at what are actually relatively strong research universities where the administration is trying to *reduce* the research programs even though they are cash-positive for the institution (due to university overhead charged on grants).

So there are confounding factors in the underlying 'type' of each research university that grossly skews that chart, as the author essentially admits. I think a far better assessment would be to look at the grant dollars *per doctoral student* to look at the inequality in the resources that each doctoral candidate has, since presumably all PhD candidates have similar research objectives.

Whether it's gone up or not it's "high". Meritocracies produce inequality. I think that's the message.

Wow did you read this wrong!

It's interesting to see the equanimity with which inequality in progressive bastions -- such as academia -- is discussed. From Morgan Price we get observation that an increase in inequality may be due to new "lower-skilled" entrants into the pool. From Mesa, we adduce merit as the explanatory variable. Both valid points. Would be great if more analysts were as skeptical and non-cynical when interpreting inequality studies of the economy at large.

Wow, look how endowment inequality just *skyrocketed* ...

Oh wait, it's just another graph that starts well above zero.

Yeah, how is Bill not all over this?

If most of the variance in a time series is coming from a short span, that is worth noting.

The interesting data point here for me is that federal research support is more unequal than endowment research support. One read is that all these people boosting the endowments of the Harvards and Princetons of the world, they're not actually funding much research (which fits into a narrative that Robin Hanson might like about how this sort of funding isn't about what it claims to be about). However, another read is that the analysis might take a narrow view toward research funding... for example, it's tempting to exclude the cost of (and appropriate fraction of) the building the laboratory is housed in when computing research funding for the laboratory, but I think it should be counted.


That was interesting to me as well. I can understand it for DoD funding (and maybe a couple of other areas), where the dollars are often aimed at actually buying something useful, but much federal R&D money seems more about spreading the fun around. It's surprising to me that it's so concentrated.

Incidentally, since familiarity with the Gini coefficient is not universal, I think it's always helpful to tie it to something (hopefully) more concrete. Such as US income inequality, which has a Gini coefficient of roughly 0.5. So university research funding is MUCH more unequal that US income!

I would suspect that successful research institutions don't feel obliged to redistribute their funding to less fortunate institutions. I think the point that is interesting here is that successful academic institutions are probably deemed to have earned their support, while successful business people are not, they having generally thought to have earned their success through luck or inheritance. From the endowment and research funding data it seems universities have both high income inequality and wealth inequality, to use terminology from the current debate.

"that successful academic institutions are probably deemed to have earned their support, while successful business people are not, they having generally thought to have earned their success through luck or inheritance."


It's hard to make much of anything of the article Morgan kindly linked: Looks like a bunch of facts to me, some of which are wrong. E.g. "Historically, the world center of science has shifted several times; in the past nine decades, America has dominated." On raw figures, that's got to be off by two decades, and we know why seven decades is correct. Then adjust by population and per capita income. So what?

Let's apply Gini coefficients to "I Dream of Gini", or "Gini With the Light Brown Hair"

Makes about as much sense.

What is the Gini coefficient of Einstein, Edison, Salk, Currie, or Galileo?

Precisely. Or the Gini coefficient of Oscars won by actors and directors, or NBA championships won by basketball players.

Ok. But why does the same logic not apply to general wealth/income inequality? Are there no Einsteins or Edisons of business? (Oh wait, Edison was a businessman who raised wealth inequality by getting very rich).

The logic is simple. Ask yourself this question: Do you measure the race of meritocracy based on where you started (inherited wealth, etc.) or where you finished without those attributes. If you believe in meritocracy, you have one view; if you don't, you favor systems where the head start is preferred.

Sounds like gobbledygook. Why do 'inherited brains' get special treatment? They had an unfair head start and must be cut down to size!

You joke, but it was only a couple of weeks ago when fans of Greg Clark were advocating exactly that in this forum. It was some sort of genetic-determinism/communist-tyranny mash-up.

Of course, the (only?) reason that schools like Harvard and Yale inevitably top these lists is that they were the first established universities in the US. Why is the Ivy League is all in the Northeast -because people in the northeast are so much smarter, or because it was the first part of the country that was settled (by people who had universities)? How is that different from the endowment effect of being born rich? (I mean, it is different, but is it *meaningfully* different?)

Stanford, in this analogy, is the direct equivalent of a Silicon Valley "new money" billionaire.

and, dan, do you know how a Gini coefficient is computed. Try it on the persons above, and ask: does the Gini tell me anything other than that these persons would have an infinitely small addition to any gini coefficient.

Right- it's high. That's the point! It's weird how once people get paid for success it becomes amoral. So I guess its compensated success that's problematic for most people.

Success is autocorrelated. I think that's the point being missed here by many. In a completely "fair" system, you expect winners and losers, and you expect winning to persist. People who are supposed to vote on winners and losers (federal funders/employers) will realize that it's not a random walk.

A lot of the success in society and in universities is generated by 10% of the population. That's where the high Gini coefficients come from. The fact that one person (Einstein) doesn't affect the Gini coefficient is a silly comment.

The fact that winners generate endowments (wealth) and that increases the autocorrelation of their success is true. Why is that bad for society? It's one of the incentives for success. It is bad for losers, but is it bad for society?

Comments for this post are closed