Larry Summers is wrong. Harvard should be spending much, much more of its money.

We may earn a commission from links on this page.

Larry Summers is, by a long shot, the worst thing that ever happened to Harvard.

In just five years as Harvard’s president—the shortest tenure of any Harvard president since the Civil War—Summers managed to behave so loutishly towards star professor Cornel West that West left in a firestorm of grievance; gave a tone-deaf speech about supposedly innate differences between the sexes; paid $26.5 million to settle a government lawsuit against his friend and fellow professor Andrei Shleifer, and then claimed that “he was not sufficiently familiar with the facts of the case to comment on it”; was on the receiving end of an unprecedented vote of no confidence by the Faculty of Arts and Sciences; spent $1 billion on a hole in the ground known as the Allston Science Complex; lost another $1 billion when he was forced to unwind a bunch of idiotic speculative derivatives trades linked to that hole in the ground; and lost $1.8 billion more by investing Harvard’s cash in its endowment fund, against the advice of the fund’s own managers, in advance of the financial crisis.

There’s more, but you get the idea.

Anyway, Summers evidently thought it was a good idea to weigh in last week on one of the most fraught questions in higher education: how much of its endowment money Harvard should spend each year.


“If it makes sense for Harvard University to pay out 5% of its endowment in 1999 when the real interest rate was 4%, it’s really quite unlikely that it makes sense to pay out 5% of its endowment in 2016 when the real interest rate is zero,” said the man who came thisclose to becoming the Federal Reserve chairman in 2013, and who might therefore be expected to know a thing or two about interest rates and their effects.

Summers’s argument isn't particularly fleshed out, but he seems to be saying that if real interest rates have come down, then expected investment returns will also have probably come down. And if the endowment therefore isn't going to grow as much this year as it was expected to grow in 1999, then maybe it doesn't make sense to pay out as much this year as it made sense to pay out back then.


He is so, so wrong.

For one thing, an endowment lives in perpetuity, not just for one year’s investment returns. It funds faculties and students with reasonably fixed costs, year in and year out, and therefore its payouts should be similarly fixed. (Try telling a tenured professor that her pay is going to be half what it was last year, because the president of the university decided to cut the endowment’s payout rate.)


In some years, an endowment’s investment returns are going to be high; in other years they’re going to be negative. The trick is to set a level of payouts that can be sustained indefinitely, without worrying too much about how the fund managers perform in any given year. And a 5% payout rate is entirely sustainable. No investor has a longer-term time horizon than the Harvard endowment, and over the long term, the funds in the endowment will return substantially more than 5% per year. Indeed, over the past 20 years, the endowment has returned some 12% per year on average. Even if that number comes down over the next 20 years, 5% should be easily doable.

The result is that over the long term, the Harvard endowment can comfortably pay out 5% of its assets every year, even if you make two key assumptions: firstly, that the endowment shouldn't suffer any kind of long-term decline in its value, and secondly, that its calculations shouldn't take into account the value of any new money coming in.


But here’s the thing: both of those assumptions are insane. The current value of Harvard's endowment is roughly $40 billion. That's an all-time high, and far above the levels that allowed the university to become one of the foremost academic institutions in the world. There is no good reason for the endowment to be that enormous: after all, the size of the endowment represents unspent money, which could instead have been spent on research, or student aid, or things like a living wage for all Harvard employees. Why sock it away for a rainy day, when there are already tens of billions of dollars socked away for that very purpose?

What Summers fails to mention is that in 1999, the Harvard endowment was worth a mere $14.4 billion. If it made sense to spend 5% of $14.4 billion in 1999, then it made sense to keep $13.7 billion for the future. Adjust that for inflation, and $13.7 billion in 1999 dollars is the same as $19.6 billion today. So if Harvard wanted to keep its rainy-day fund constant from 1999 to today, then it would have to spend a whopping $18 billion this year, or 48% of the total endowment. Then, and only then, would its rainy-day fund come down to a level that was more than adequate in 1999.


Summers also fails to mention that Harvard is doing a spectacular job raising funds, and not only in big chunks like the $400 million donation from hedge fund manager John Paulson. In 2014, for instance, Harvard raised $1.16 billion in a single year, up from $792 million in 2013. Those numbers are only going to get bigger: one major philanthropist told me recently that he expects both Harvard and Stanford to raise more than $100 billion each over the next 20 years or so, as people who made their millions in science and technology decide to give back to the institutions at the cutting edges of those fields.

In other words, it’s bonkers for Harvard to base its payout calculation on the implicit assumption that it’s never going to raise another dime, when in reality Harvard’s endowment is going to be fattened by a steady stream of enormous inbound checks for the foreseeable future.


As a monetary economist and would-be Fed chair, Summers really ought to know this better than anybody. The whole point of a zero interest rate policy is to flood the economy with money and encourage people to spend it more freely on the things they love, including gifts to their alma mater. Whatever hit the Harvard endowment is taking in terms of lower expected investment returns, it’s more than making up for that in terms of ever-rising fundraising from a group of alumni who have been the big winners in the rise of global inequality.

Which means that Summers is talking nonsense when he says that Harvard’s 2016 payout rate should be lower than its 1999 payout rate. Indeed, the Harvard payout rate should not be a function of prevailing short-term interest rates at all. A smarter model would look at the cost of spending a marginal extra dollar: what future opportunities might Harvard lose, if it spent that dollar today? And it would also look at the benefits of spending that marginal extra dollar. Might it help to create new intellectual property which would turn into a cash cow for the university? Might it create more highly-valuable goodwill among a group of students who got a free ride and then became enormously wealthy later in life? What is the return on investment for every dollar that the endowment spends? It's probably negative, but not hugely so. Those spent dollars don't simply disappear: a lot of them end up coming right back into the endowment, one way or another.


At the same time, a movement is beginning to take hold whereby big university endowments might start to lose their precious tax-exempt status. A bill currently wending its way through the Connecticut legislature, for instance, would force Yale to start paying tax on its unspent investment profits, and frankly there’s no good reason it shouldn’t do exactly that. Connecticut needs money, and Yale clearly has more money than it knows how to spend. The rich should pay their fair share of taxes, and endowments like Harvard’s and Yale’s are among the richest of the rich.

Thanks to a quirk of nonprofit tax law, university endowments are not required to spend at least 5% of their value every year in the way that foundations are. But that 5% minimum figure for foundations has acted as an anchor: both foundations and endowments gravitate towards it, in the absence of any other guidance on what is a suitable number. The fact is, however, that for universities with monster endowments—Harvard, Yale, Stanford, and the like—a 5% payout rate is always going to be far too low. It was too low in 1999, and it's definitely too low today.