Return to search

INVESTOR BASE: Juggling J-curves

A new study suggests J-curves can actually be used to predict performance – from an early stage in a fund’s life.

J-curves are a good illustration of why LPs have such a difficult task – in that they show how long it takes for funds to prove their worth.

However, a new paper argues that J-curves have a hitherto unnoticed advantage: they can start predicting how a fund will ultimately perform from as early as its third year.

The study, by Cyril Demaria of the University of Sankt-Gallen in Switzerland, is based on aggregated performance data for US and EMEA LBO and VC funds raised between 1984 and 2001. These were split by vintage year and divided up into four categories based on returns – very high, high, medium and low/negative. For each of these categories, an average J-curve was then calculated, based on recorded cash-flows. Demaria’s theory is that because these J-curves are quite distinctive in shape, it starts becoming clear from about year three which performance category a fund is likely to end up in.

There are some important caveats around this (notably around the limitations of grouping funds by vintage year, a possible over-reliance on US data, and the potential impact of external factors like dividend recaps). Nonetheless, Demaria found that after two years, it was possible to exclude some performance outcomes; and by the end of the fifth year, the final category could be predicted with a reasonable degree of certainty.

The consequences of this are potentially significant. Theoretically, it ought to help LPs assess the level of risk in their portfolio more accurately at any given time – helping not only to inform new investment decisions but also mitigate regulatory capital demands.

It will also allow for better-informed decisions in the secondary market. As Demaria puts it: “LPs will be able to better negotiate the discounts/premia on their existing stakes, and securitise mature portfolios. The dynamics of pricing on PE’s secondary market should change significantly, should the model be validated and adopted.” 

Lastly, it should help to benchmark performance. Currently there is no one measure that – in the eyes of all market participants – is able to capture the risk-adjusted value of an illiquid fund interest, or portfolio of fund interests. And while this is no magic bullet in and of itself (given the various caveats), it would certainly be another useful tool.