Felix Salmon

Larry Summers’s billion-dollar Harvard gamble

By Felix Salmon
July 24, 2009

Greg Mankiw adds some insider detail to the story of Larry Summers’s ill-fated interest-rate swap, in the form of an email from “someone knowldgeable about the financial situation at Harvard”.

The email is clearly meant to exonerate Summers, at least a little, but I’m unconvinced. Taking the three points in sequence:

1) The instrument in question was highly liquid and could be sold fully within a few days; essentially all money was lost in 2008 two years after Larry Summers left.

This is true, but misleading. When people speculate in the markets, it’s the act of putting on the position which is the point at which the gamble is made. After that point, you make money if the position rises in value, and lose money if it plunges. Interest rates could have fallen at any time after the bet was made, and Harvard would have lost the same $1 billion.

The argument about liquidity only serves to underline how speculative this bet really was. If it was a genuine long-term hedge of certain future borrowing needs, Harvard would not have needed the liquidity since the position would have been designed to sit on the university’s books for decades. On the other hand, if Harvard was intending to trade in and out of this position, then the liquidity helps, but the swap can no longer be considered a hedge at all.

Was Harvard maybe intending to keep the swaps on its books in the event that interest rates rose, while selling them if rates fell? That seems to be the implication here: that if Summers had still been around, he would have liquidated the swaps when rates fell, and thereby avoided massive losses.

Again, however, this argument doesn’t hold up to scrutiny. If Summers had wanted to buy a swap with limited downside, one which automatically unwound if rates fell to a certain level, he could easily have done so. But that’s not the instrument he bought. Instead, he bought a sophisticated financial product which left Harvard potentially on the hook for $1 billion or more — and then did nothing to address that tail risk.

2) Harvard has a system where the treasurer makes these decisions with approval of the corporation and involvement of a debt management committee on which president does not serve.

Does anybody believe that this hare-brained scheme was the idea of Harvard’s treasurer? Come on. Harvard had $1.6 billion in floating-rate debt, and it’s conceivable that the treasurer might want to swap that debt into a low fixed rate. It’s not conceivable that the treasurer would be interested in swapping nonexistent future floating-rate debt into today’s fixed rates — especially not when the hypothetical future borrowing wouldn’t even take place for as long as 20 years. This deal has Summers’s fingerprints all over it, and would never have been done had he not been president of the university.

3) Given the plan to borrow large amounts of debt in the future, doing something to lock in low rates made sense. Iif Harvard was borrowing big, there would be offsetting saving now. The big error was the failure to adjust hedge when Allston was scaled back and to take account of the risks associated with the change in the university’s credit rating.

I honestly don’t know what Mankiw’s anonymous source could be talking about when he or she refers to “offsetting saving”. Was it an egregious dereliction of fiduciary responsibility to keep the swaps on Harvard’s books even after the excuse for putting them on — the multi-billion-dollar plan to expand the university into Allston — was put on hold? Yes, of course. And you can’t blame Summers for that, since he had left Harvard by then. On the other hand, there was always a risk that Allston would be scaled back, and indeed one of the most likely reasons for scaling back Allston was that there might be a national economic crisis — exactly the sort of thing which is normally accompanied by a reduction in interest rates.

Summers was well aware of the risk of an economic crisis. Indeed, in 2004, at about the time that the swaps were put on, he gave a major address at the IMF/World Bank annual meetings about the systemic risks posed by the US current-account deficit, and warning of “a slowdown in growth that would be unacceptable in the United States and would have very severe consequences for growth globally”. But maybe because he had gone through so many other current-account crises abroad during his tenure at Treasury, he was pretty clear that he thought the big risk was that interest rates would go up, rather than go down. In response to one question from a central banker, he said:

I certainly would not want to suggest how you or any other central banker should manage your reserves, but I would point out that when you buy U.S. treasury bills, what you get is 1.75 percent, and it doesn’t really matter whether the U.S. economy grows rapidly or grows slowly. And that is, as I said, a negative interest rate in real dollar terms, and I think that’s the number that one should focus on.

Summers couldn’t have been much clearer that he was pretty convinced that interest rates in the US were going to have to rise: it seems quaint now, but back then 1.75% really did seem like an incredibly low interest rate on T-bills.

Given his analysis, and his ego, it’s pretty obvious how Summers decided to use the future Allston expansion as an excuse to engage in a massive interest-rate gamble outside the purview of the Harvard Management Company, which is the arm of Harvard with a real mandate to play the financial markets. The real reasons for the rate swaps can be found in that 2004 lecture, not in vague ideas that Harvard was sure to issue floating-rate debt at some point in the 2020s. And given those real reasons, it’s easy to see why there was no clear mandate to unwind the swaps when Allston was scaled back.

Basically, Summers took a massive gamble with Harvard’s money, and lost — big. The buck stops with him, and I look forward to Summers admitting as much sooner rather than later.

14 comments so far | RSS Comments RSS

“I look forward to Summers admitting as much sooner rather than later.”

It’s comedic touches such as that one that keep people coming back. Well done.


“It’s not conceivable that the treasurer would be interested in swapping nonexistent future floating-rate debt into today’s fixed rates ”

No, it is conceivable. It would be a pretty bad idea but it’s certainly conceivable.

Posted by dsquared | Report as abusive

Meh. I think point #2 from Mankiw’s correspondent pretty much refutes your argument, and simply asserting that the deal “has Summers’ fingerprints all over it” isn’t a real response. As it stands right now, there’s no hard evidence whatsoever that Summers was the driving force behind, or had any formal role in, the swaps at issue — it’s pure speculation. Now we learn that Harvard evidently has a well-defined process for making these kinds of investment decisions, and that Summers didn’t have any role in that process. So not only do you have no hard evidence that the forward swaps were Summers’ idea, but you also have no evidence that Harvard failed to follow its own process for making investment decisions.

Also, no one has any clue how Harvard was using these swaps, their structure, etc. We have no idea why Harvard was holding long-dated forward swaps, what the makeup of their balance sheet was at the time, or what their investment strategy was. But the LEAST likely explanation is that Larry Summers thought he knew in 2003-2004 where interest rates were going to be in 2020. Come on. It’s just not plausible that Summers thought he could predict interest rates in 2020. He’s arrogant, but he’s not a college sophomore.

Sometimes, something that everyone thought was a story for long time turns out to be nothing.

(Harvard’s Treasurer, by the way, was the head of one of the largest mutual funds in the country.)


“The big error was the failure to adjust hedge when Allston was scaled back.”

OK, I may be wrong about the timeline, but wasn’t Allston scaled back at the same time that rates fell? That is, the error wasn’t the failure to adjust the hedge; it was the failure to recognize the correlation between the level of Harvard’s endowment and interest rates. A sufficient fall in the endowment would cause the cancellation of Allston; but at the same time, the only thing which would cause a sufficient fall (aka a market crisis) would have been accompanied by a fall in interest rates. Basically, Harvard should have had more bonds in its portfolio in order to have managed this risk. Don’t know, though, how this alternate portfolio would have stacked up against Harvard’s actual portfolio (from 2004 on).

Posted by a | Report as abusive

fact of the matter is that these are the same people jumping on cnbc summer 08 saying everything was rosy

Posted by dvictr | Report as abusive

2) Harvard has a system where the treasurer makes these decisions with approval of the corporation and involvement of a debt management committee on which president does not serve.

If the treasurer ( in stead of Larry Summers) was primarily responsible, he should have been fired. Was he fired ?

Posted by RN | Report as abusive

“with approval of the corporation” – ie., the approval of Larry Summers, right? I don’t think that statement #2 really lets Summers of the hook that easily.

Posted by Nate | Report as abusive

While the wisdom of “hedging” debt that you are not yet prepared to issue is doubtful — especially with the advantages of hindsight, it seems unfair to claim that intent of Summers or the treasurer was speculative.

There’s a very important issue that is being completely ignored here. The interest rate swaps themselves undoubtedly involve repeated relatively small payments over a period of time (rather like mortgage payments). There is no evidence whatsoever that contracting to make repeated small payments was harmful to Harvard’s endowment fund.

What killed the endowment fund was the fact that these contracts — nowadays — include collateral posting which is based on the mark-to-market value of the swap. Because of these collateral requirements, the endowment was effectively required to “prepay” the whole sequence of payments. Just as if you as a mortgagor were required to deposit at your bank the full value of the loan.

In short, collateralized swaps involve extraordinarily high levels of liquidity (or collateral posting) risk that nobody’s bothering to talk about (except Richard Green http://real-estate-and-urban.blogspot.co m/2009/06/i-once-felt-good-about-having- larry.html). The real failure of Summers/treasurer was the failure to take into account the worst case collateral posting scenario.

It seems to me that collateral posting risk makes interest rate swaps unsuitable for most organizations that don’t have access to the Federal Reserve’s lending facilities.

Posted by csissoko | Report as abusive

The financial funny business in Cambridge continues apace! Harvard is sticking to its claim that the endowment shrank “only” 30% over the past year, although the cutbacks at Harvard obviously far exceed those at other such places (Princeton, for example) that also claim losses of that magnitude. Is Harvard Management cooking the books in the endowment appraisal? Impossible to tell, since Harvard does not practice true transparency.

But Harvard’s BANK LENDERS know a lot more about what is happening there than the public does. Those banks are now requiring that Harvard pay a new margin (over LIBOR) on it’s big revolving loan credit facility that is FIVE TIMES the old margin …. and accept a loan capacity $250 million less than the old one. Looks like Harvard’s banks have some seriously increased concerns about the probablility of getting their money from Harvard. Harvard is likely to accept the new terms, although this revolver is a main source of back-up liquidity (although not “same day” liquidity) and the university is widely believed to be have an ongoing liquidity issue.

Here’s a link to the student newspaper article on the matter: http://www.thecrimson.com/article.aspx?r ef=528607

Maybe Greg Mankiw should look into this credit facility mess, too. You know, before ALL of the milk is spilled?

Posted by James Tiberius Madison | Report as abusive

“This deal has Summers’s fingerprints all over it, and would never have been done had he not been president of the university.” Felix Salmon offers no proof of either claim. Clearly Salmon hates Summers as the numerous insults demonstrate. This is a hate based attack on Summers, nothing more. Obviously if Harvard has lost so much money, something went very wrong. This does not, however, prove (or disprove) that Summers is to blame.

Posted by Rob Luginbuhl | Report as abusive

If the idea was to simply try to take advantage of prevailing low interest rates (but not enter overly speculative trades), why not trade swaptions instead…

Posted by mkamdar | Report as abusive

Are you kidding? The response posted by Mankiw’s correspondent obviously knows way more about the specific situation than you, and he obviously knows more about the financial instruments used in interest rate hedging than you (there’s no fundamental reason NOT to use a hedging technique that is also liquid and there are pros and cons to any such strategy). Your belief in fingerprints aside… you are out of your league.

Posted by Regret | Report as abusive

csissiko has it right above. Also, El-Erian should have liquidated the swaps in 2007 and Mendillo should have in 2008.

Oh well, it’s just a 25% loss, likely made back quickly. As they say, “You can’t ring the bell every time.”

With billions of dollars on the line, I would try to ring it rather hard.

Posted by Nathan Herold | Report as abusive

On point (3): So long as there was borrowing for the Allston project, the interest rate position was not something that created risk– it reduced risk, by hedging. Of course, if Summers had an opinion on how interest rates would move, that would make him all the more eager to hedge to a zero net position instead of gambling the wrong way.

This relates to your point (1), but makes it backfire. The position could have been undone at any time, and so it could and should have been undone when the Allston borrowing was halted. Up to taht point, the interest rate position reduced risk; after that, it increased risk.

I wouldn’t be surprised if this was Larry SUmmers’s idea, even tho he didn’t have formal reponsibility and there is no evidence for him being involved. If so, maybe it illustrates the perils of having a smart leader introduce an innovative new policy: After he leaves, the dummies left behind can make things worse because they don’t understand the purpose of the innovation.

Posted by Eric Rasmusen | Report as abusive

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/