Making co-investment work

A new report suggests co-invests are likely to underperform. But surely this is an execution issue, not a structural issue?

If there's one truth universally acknowledged in today's private equity market, it's that LPs love co-investment. Whether as a way to reduce overall fees, mitigate the J-curve, juice returns or manage portfolio risk, most LPs seem to want more access to co-invest at the moment. GPs on the fundraising trail tell us that they're constantly being asked the question – sometimes with the implication that it would facilitate a commitment decision – while the number of dedicated co-investment funds currently in market also bears witness to its perceived attractiveness.

So it's no wonder that a report released by advisory group Altius Associates this week has raised a few eyebrows. Based on a sample set of nearly 900 realised US buyout and growth investments between 1979 and 2012, Altius concluded there was a 'substantial probability' that a co-investment portfolio of 10 assets would return an IRR of less than 0 percent. Even with a 20-company portfolio, you’d still have a decent chance of losing your money.

So what's the explanation for these findings (which are along similar lines to those of a Harvard Business School paper published last year)? Altius highlights three main problems: that LPs only get offered the bigger deals, which by definition are in a segment of the market the GP is not used to; that LPs' risk becomes too concentrated in too small a group of companies; and that GPs will try to keep the very best investments for the fund, in order to maximise their own upside.

All of these arguments seem plausible in theory. Clearly the bigger the equity cheque, the more opportunities there will be for co-investment. Equally, the largest deals have (on the whole) been more likely to underperform in recent years. So it’s not improbable that LPs looking to build their co-investment portfolio could end up with a selection of sub-optimal deals.

However, most of the people we’ve spoken to this week about the report have suggested that it doesn’t chime with their own experience.

For a start, many big LPs will tell you that their co-investment portfolios are actually performing very well. For example, AlpInvest produced a white paper last year suggesting that their co-investments have delivered an average return of more than 2.7x (with upper- and lower-quartile multiples of 3.7x and 1.0x). Similarly, ATP PEP told us recently that the 25 or so co-investments they’ve done since 2000 have returned more than 2.5x. That certainly bears comparison with primary fund returns.

And while GPs do obviously have a greater need for co-investment on bigger deals, size is by no means the only reason for them to offer it. Some do it for relationship reasons – to keep existing LPs happy or warm up potential new ones. Others do it to improve the shareholder group – perhaps to bring in some specific domain expertise, or to have a friendly local voice on their side of the table when they go outside their home market.

Equally, GPs strongly deny that they keep all the best deals for themselves. Even if you accept that managers can predict in advance which deals will be the big winners (most will tell you that’s very difficult at the best of times, and that they wouldn’t do any deal unless they thought it had that potential), it wouldn’t do their LP relationships much good if they only ever shared their worst performers.

Of course, they would say that. And the fact is that most co-investment portfolios are relatively small and new, so it’s hard to prove the point either way.

But while it may turn out to be true that there are some structural issues with co-investment, we’d be wary of any suggestion that the whole model is somehow inherently flawed. If anything, all of the above just highlights the importance of scale and selection skill: those LPs who build a sizeable, sophisticated and proactive operation will be more likely to get shown good deals, more likely to select the best performers, and more likely to have a portfolio that’s sufficiently large and diversified to mitigate concentration risk. All of which will make underperformance a lot less likely.

In other words: LPs shouldn’t avoid co-investment. But they need to do it properly.