Jul 12 2012

A question about Bain Capital’s stellar returns under Mitt

Posted by: Nick Shaxson in: Thoughts

Update: a couple of new thoughts at the bottom.

One of the most widely cited statistics in defence of Mitt Romney comes from a Deutsche Bank prospectus looking at 68 companies that Bain Capital funds invested in between 1984 (when Romney set up Bain Capital) and 1998.

Here is the punchline from that prospectus:

“Bain Capital’s first five private equity funds achieved an average annualized rate of return of approximately 173% per annum on all realized investments through December 31, 1999 and 88% per annum on all investments through December 31, 1999.”

The 88 percent is the non-cherry picked number, and, well, wow. Take $1 million and grow it by 88 percent per year over that 15-year period, and you end up with over $12 billion. Exhibit A: this man sure is a successful businessman, and knows how to make a lot of money.

Not so fast.

Of course, there are the obvious mitigating factors that cut that number right down. First there is inflation, which was much higher in those days and would cut that to about $8 billion by 1999 in real terms. Still a huge number, though. Then there are the fees, which are special, and huge. Then there are the profit shares taken by Bain Capital. Then there is the leverage, which throws another spanner into all the calculations, at least if you are thinking about the health of the U.S. economy, rather than purely about the returns to Mitt Romney and his colleagues.

So far, so standard: plenty has been written about this, though perhaps too many have overlooked the issue of fees in particular.

But there is this other thing I came across while researching my Vanity Fair article about Romney. Bear in mind that 88 percent number, then take a look at this passage from a 2005 book by David Swensen, who manages Yale University’s $16 billion endowment fund. It’s the bit I’ve marked in bold that raised my eyebrows.

“A Yale Investments Office study provides insight into the additional return required to compensate for the risk in leveraged buyout transactions. Examination of 542 buyout deals initiated and concluded between 1987 and 1998 showed gross returns of 48 percent per annum, significantly above the 17 percent return that would have resulted from comparably timed and comparably sized investments in the S&P 500. On the surface, buyouts beat stocks by a wide margin. Adjustment for management fees and general partners’ profit participation brings the estimated buyout result to 36 percent per year, still comfortably ahead of the marketable security alternative.

Because buyout transactions by their very nature involve higher-than-market levels of leverage, the basic buyout-fund-to-marketable security comparison fails the apples-to-apples standard. To produce a fair comparison, consider the impact of applying leverage to the hypothetical public market investments. Comparably timed, comparably sized, and comparably leveraged investments in the S&P 500 produced an astonishing 86 percent annual return.”

Swensen’s 86 percent, it seems, puts Romney’s 88 percent number in a certain light.

Did I miss something here? (The reason I ask is: why has this particular question not been aired?)

Am I making an apples-to-oranges comparison? I don’t think so. The timescale is pretty similar though not identical: 1987-1998 in the study Swensen mentions, and 1984-1999 in the Deutsche Bank / Mitt Romney case. Other factors may need to be taken into account too, such as leverage ratios. But based on my conversations with people, Mitt was a huge fan of leverage, right from the beginning of Bain Capital and even several years beforehand: this would make the comparison with the Swensen number even more unflattering, as a higher leverage ratio than Swensen used in his calculations would mean still higher returns on the hypothetical leveraged portfolio for that time period.

I asked Swensen whether there was a fair comparison to be made here, but he declined to comment. And I came across this figure very late in my research, so I didn’t have the time to get proper expert comment, or to do my own calculations.

Perhaps someone else will.

Update: having thought about this a bit, I realise that the comparison I made here is probably not a perfect one, although I don’t know exactly how Swensen’s study worked. Presumably a manager would take fees for leveraging up the S&P 500 for you, and those fees would cut that 86 percent number down a fair bit. But the fees on such a simple operation would be nothing like the fees that Bain extracts, so it wouldn’t cut the Swensen number down inordinately.

3 comments so far

Alien Edouard 7th July, 2012 6.17 pm

You have a number of problems here.

First, the 88% number is for NET IRR, so after management fees, performance fees, costs, etc. but typically before taxes. But it does not mean that if you had committed $1 with Bain, it would have grown by 88% every year (note also that in the private equity world ‘committed’ is very different from ‘invested’). This is where this number is highly misleading (as is often the case with private market performance measurements). The 88% is probably an average of the returns on all realized and unrealized investments as of 1999. Since these investments were of varying size and crucially duration, the 88% disclosed figure does not help you much determine how much your initial commitment of 1$ would have grown to. For this, you would need to ask for other common private equity performance metrics: the multiple of invested capital (MOIC), or multiple of committed capital (MOCC). Neither seems to have been disclosed.

In short, the 88% is probably useless for most purposes.

Second, the S&P also brings its fair share of problems. To start with, it is already a levered index, since most if not all companies carry a certain level of debt. The leverage ratios are generally lower than in the context of leverage buyouts, but it is not unusual for S&P 500 companies to even issue non-investment grade debt. However, the biggest problem in Swensen’s study (or at least as you report it) is that it ignores the index’ survivor bias. Because the index is designed to track the performance of the largest public companies by market capitalization, it is regularly re-adjusted with weak performers substituted with stronger newcomers. This obviously significantly distorts the index’ performance upwards.

In short, you are comparing apples and oranges, but if it is any comfort you are far from being alone. I see the same errors committed regularly in both in academic and industry circles.

Nick Shaxson 7th July, 2012 3.59 pm

Thanks Edouard I entirely agree that the 88 percent number is useless in terms of estimating investor returns – though I think I reflected that in the blog. The thing about embedded leverage in the S and P 500 is a good point, as is the one about survivor bias – though the fact is that I couldn’t access Swensen’s report in the time available so I don’t know how he calculated it. It may still be that I am comparing apples with oranges, it may not. I didn’t have time to dig into this, and I wish I had.

[...] I mentioned earlier a Yale Investments Office study that noted (keep that 88 percent number in mind): “consider the impact of applying leverage to the hypothetical public market investments. Comparably timed, comparably sized, and comparably leveraged investments in the S&P 500 produced an astonishing 86 percent annual return.” [...]

Leave a comment