I received a great suggestion (thanks Etienne, from Preqin) to consider the J-Curve created by fund IRRs (opposed to the J-Curve created by net cash flows of a single private equity fund). This could apply to an individual fund or an aggregation of funds based on vintage year, fund size, etc.
Without thinking too deeply about it, using IRRs allows you to compare the J-Curves of multiple funds on a like-for-like basis. This is because IRR is a relative measure, whereas net cash flow is an absolute measure.
The accompanying chart (thanks again Etienne) plots the median IRR for each vintage between 2000 and 2008. However, due to this blog being stuck in the 1950s and sans colour, the chart is a little difficult to interpret. But, let me give it a try via narration.
The X-axis is year of operation and the Y-axis is IRR. The longest line obviously represents the 2000 vintage, since funds of that vintage have approximately nine years of performance data. But, rather than deciphering each line, let me divert your focus to a few high-level observations.
All vintages recorded negative median IRRs in the first year of operation. Most were still negative in the second year and by the third year, most were at 10%, which is quite respectable. Other than the 2000 vintage, all other medians were above 0% by year 3 and in the 15-30% range by year 4 (and they stayed there). The J-Curve effect is quite clearly represented, but not as protracted as the net cash flow J-Curve; it accelerates quickly and then stabilises.
The most surprising takeaway for me is that the median IRR for each vintage stabilises at quite a high level (approx. 20%) after only a few years of operation. While I acknowledge that these are median IRRs (hence, not representative of the performance you would receive from a dollar invested in a single fund), I still find it an insightful statistic. The data may become skewed with new financial reporting standards (I’m thinking FAS157), but if anything, we may see a more defined J-Curve from fund IRRs.