May 5 2022

IRR is a vanity metric

I’m observing that IRR is a metric that is becoming an increasing focus in venture, replacing fund return multiple as the key metric of success. I understand the draw of IRR, and – as a fund draws to a close – there’s no question it’s an important metric. LPs have plenty of options for where to put their money and IRR is an easy way for them to compare how their money grows across various investment options and a convenient way to determine whether they believe they’ve been adequately compensated for the relative risk and varying degrees of liquidity available to them across different asset classes.

BUT (and you knew there was a but here given the post title…) it’s a poor way to compare the relative performance of funds early, and even into the mid/later, stages of a fund’s life. But that doesn’t seem to be stopping funds or LPs from touting IRR as the gold standard for evaluating relative fund performance. This is a mistake. Interim IRR is much too easily manipulated and in some cases incents behavior counter to the long-term benefit of LPs (and GPs). A few examples:

  • In an up market, IRR values quick capital deployment. Whether we’re in a full correction or not has yet to reveal itself, but for the years leading up to 2022, the market for all things venture was extremely strong. That was great for companies needing to raise increasing amounts of money and fun for investors who often saw companies receive large mark-ups relatively quickly after raising their last round. How many of those valuations will hold remains to be seen, but that kind of market clearly advantaged funds that deployed capital very quickly. Gone was the J-curve (named because often early fund returns are negative – mostly due to fees that drag down NAV prior to initial company write-ups). In markets like this, funds that deploy quickly generate markups on their full portfolio; those that show more patience suffer from an IRR perspective as a larger portion of their fund is yet to be marked up by subsequent financing rounds. IRR doesn’t track this and comparing two funds from the same vintage (fund year) without considering how quickly each deployed capital into an upmarket is a mistake.
  • Recycling hurts IRR. Recycling is how funds pay back their management fee and invest the full amount of the fund that they raise (consider a theoretical $100M fund with a 2% management fee for the 5 year investment period and a 1% fee for the tail 5 years; that fund would have to generate $15M of recycling in order to cover the fees that they charge investors). In fact, many funds have the ability to invest more than 100% of their fund size (via profits generated from investments). Traditionally LPs have viewed this as positive. The fund is investing profits, the LP is paying fees on what effectively becomes a smaller fund (in my example above, if the fund invests 110% of the fund, LPs were actually paying a 1.8% management fee). This shrinks the gross/net return spread and is good long-term for LPs (assuming in this analysis that the 10% overage invested returns the same or better than the fund as a whole). But this doesn’t help a fund’s IRR. It actually suppresses it in a similar way that investing over a somewhat longer time period does: by putting capital in the ground later in a fund’s life. Better for the overall return, I’d argue, but not for the interim IRR.
  • A lot of things change in 10 years. Most funds have a 10-year life. Really more than that, as 10 years is an increasingly outdated vestige that seems for some reason to live on in fund documents, even if it doesn’t in reality. But in any event, the life of a fund is at least 10 years,  and increasingly 15 years or more. A lot changes in that period of time. So if you’re considering IRR at year 3, 4, or 5 – especially for Seed and Series A funds – there are a lot of valuation assumptions that in hindsight will turn out to be completely wrong, but which can drive significant IRR (well, not real IRR, since they never materialize, but interim, paper IRR). For example, I was recently on an update call with a fund and did a quick calculation that over 80% of the gain they were showing came from just 2 companies. Both had raised attractive up rounds. Both were valued at greater than 50x revenue (one was, I think, 70x revenue). This isn’t a commentary on either of these companies – they both may end up being great. But there’s a lot of work for these companies (and many, many like them across the venture landscape) to grow into their paper valuations. If you abstract this across the venture industry or asset class (seed stage venture in this particular case) you run the risk of having a few outlier valuations driving what appears to be spectacular IRR. Some of those companies will be successful. Many will not. We’ll know eventually. But not now. [NOTE: After distributions, the typical capital adjusted duration of even an early stage fund is maybe 8 or 9 years; that’s probably a good benchmark for when you can have a reasonable idea of where IRR will land, but even then there are some large late returns that can change the IRR late in a fund’s life.] 

So, food for thought as our industry perhaps becomes overly fixated on a metric that is too easily manipulated (we haven’t even talked about how subjective valuations can be). I’m not saying ignore interim or early fund life IRR. Just take it with a grain of salt. Venture is a long game. And long games take a while to play out.