Cashflows in a right-skewed world

Investors have long known that private equity funds usually take several years to return capital. In the past few years, distributions have been particularly slow. As a result, many investors would like to construct their portfolios with an eye to cashflow as well as to IRR. Some investors might be tempted to commit to one or two funds or areas they feel are most likely to be top performers. However, the data suggests that a portfolio with commitments to a portfolio of private equity funds will usually return capital faster than the median single fund. This cashflow portfolio effect is in addition to the better-understood impact of portfolio construction on risk and return.

Figure 1 shows the ratio of aggregate distributions to contributions for venture funds over time, as drawn from the Venture Economics (VE) database. Taken all together, the venture funds formed each year from 1986 through 1997 have more than paid back contributed capital. For example, the 1997 venture funds in the VE database have already paid back about 1.7x the cash contributed to them.

It would be tempting to conclude from Figure 1 that the typical 1997 venture fund ? the median ? has already returned capital. However, this turns out not to be the case. Figure 2 reveals the same ratio of contributed to distributed capital, but breaks it out for venture funds at the top quartile, median, and bottom quartile of IRR.

Although 1997 venture funds in aggregate have returned 1.7x contributed capital, what Figure 2 illustrates is that the median 1997 venture fund has not. In fact, the median 1997 venture fund has returned only about 0.7x contributed capital. It's interesting to note that there is not yet any vintage year for which the median venture fund has seen unusually large cumulative distributions due to the tech boom of the late 1990s.

By contrast, venture funds above the median, especially those formed in 1994 through 1997, have done very nicely indeed. Top quartile venture funds formed in 1995 and 1996 have already returned 4.5x contributed capital, and the very top funds have done much better than this.

There are two major conclusions we can draw from these observations. The first will be no surprise to anyone involved with private equity: it's better to be in the best funds than in the typical fund. This is true whether we're discussing time to return capital, IRR, or practically any other metric.

The second conclusion is probably not as well known ? namely, that a portfolio of private equity funds will usually return capital faster than the median fund in the universe that the portfolio was drawn from. This can be shown by considering the universe of all venture funds in a given vintage year as one large portfolio. For every vintage year, such a portfolio (as shown in Figure 1) returns more capital than the median fund in the universe (as shown in Figure 2). If one were to construct a venture portfolio by randomly selecting funds from the venture universe, its ratio of distributions to contributions would be expected to approach the aggregate ratio of Figure 1. For portfolios containing a small number of funds, there might be a substantial amount of variability from portfolio to portfolio; for larger portfolios, this variability would be reduced. (Reasonable management of variability can often be obtained with 15-25 distinct fund choices.) In any case, the portfolio is expected to return capital faster than the median fund.

Why does this happen? As Figure 2 reveals, cash returns in the venture universe are right-skewed. That is to say, the best funds are farther above the median than the median is above the worst funds. As a result, the best funds in a portfolio speed up the return of capital more than the worst funds slow it down.

This observation ? that a portfolio of funds will generally return capital faster than the median fund in the universe the portfolio is drawn from ? applies just as much to top-quartile venture funds as it does to the venture universe as a whole. It's easy to see that the top quartile has some funds that return capital very quickly, but no funds that return capital very slowly. In other words, the top quartile is itself rightskewed. Therefore a portfolio of top-quartile funds will usually return capital faster than the median top-quartile fund.

So far the argument has only taken into consideration venture funds. Does the same logic apply to all private equity funds? Data in Venture Economics for non-venture funds is relatively sparse, but what's there is suggestive. Figure 3 looks at how long it takes for top-quartile funds in aggregate to pay back contributed capital, for vintage years 1986-1996.

This shows that top-quartile venture funds and top-quartile non-venture funds (buyouts, distressed, and so on) have surprisingly similar payback statistics. Out of 11 vintage years in this arbitrary sample, the non-venture top quartile paid back faster in four cases, the venture top quartile paid back faster in six cases, and there was one tie (4.25 years for aggregate top-quartile venture and non-venture funds formed in 1993). And it is worth bearing in mind that this sample period includes the extraordinary years for venture funds of 1995 and 1996. As a result, the results discussed above apply just as well to non-venture funds as to venture funds, and a diversified portfolio is more likely to give the best cashflow over time than a portfolio that concentrates on one sector alone.

This analysis provides some guidance on how one might construct the best private equity portfolio, for both IRR and payback rate. First, pick strong managers of all kinds ? don't just concentrate on the strategies currently in fashion. Next, build a well-diversified portfolio of perhaps 15-25 funds from among these managers. Over the long term, we believe such a portfolio is the best bet, whether measured in terms of IRR or time to break even.

Barry Griffiths is a Vice President at Goldman Sachs and Head of Quantitative Research for the Private Equity Group (PEG), which manages approximately $11 billion in capital commitments across primary partnership, co-investments, and secondary partnership investments. He can be reached at barry.griffiths@gs.com.

This material is provided for educational purposes only. In the event any of the assumptions used in this material do not prove to be true, results are likely to vary substantially from those shown herein. Opinions expressed are current opinions as of the date appearing in this material only.