Vintage year benchmarking and quartiles

Key takeaways

• A vintage year assigns each fund to a single calendar year, but providers use different anchors (first close, first capital call or first investment), so the same fund can sit in two different vintages.
• Quartile breakpoints are percentiles of the peer distribution, and the result depends on the percentile formula: in the example peer set the top quartile starts at 16.0% under one common method and at 16.8% under another.
• Peer sets built from voluntary reporting can omit weaker funds, which pushes breakpoints up; adding three non-reporting funds to the example lowers the median from 11.75% to 10.2%.
• Rankings early in a fund's life rest mainly on unrealized marks and IRR timing effects, so they are unstable: eight of twelve example funds changed IRR quartile between year 3 and year 8.
• A fund can be second quartile on IRR and third on TVPI in the same peer set, so a quartile claim must name its metric.
• Treat a quartile ranking as a screen that prompts questions about vintage definition, peer set, metric, method and fund age, never as a verdict on its own.

What a vintage year is

A vintage year is the single calendar year a benchmark assigns to a fund so that it can be compared with funds that started at roughly the same time. The idea comes from wine, where the year the grapes were harvested shapes the product. For a fund, the relevant conditions are the prices at which it bought assets, which depend on when it started investing.

There is no single rule for which year that is. The three common anchors are the year of the fund's first close (when the first LPs were legally admitted), the year of its first capital call (the first cash flow from LPs), and the year of its first investment in a portfolio company. For many funds all three fall in the same year. When a fund holds its first close late in one year and calls capital early in the next, the anchors disagree.

Some managers also choose their own vintage in marketing materials, often the year of final close or of first investment. The choice is not neutral, because adjacent vintages can have very different peer returns. A fund that looks above median against one vintage may look below median against the next.

Compare within vintage and strategy only

Vintage controls for market timing. Strategy controls for the kind of risk and the shape of the cash flows. A venture fund, a buyout fund and a senior direct lending fund of the same vintage differ in expected return, in dispersion and in how quickly they return capital. Ranking a buyout fund against an all-private-equity peer group that includes venture mixes different distributions, and the quartile breakpoints will describe neither strategy well.

Finer cuts help up to a point. Geography, fund size and sector focus all affect returns, and a mid-market European buyout fund is more comparable to its direct peers than to global mega-funds. The cost of narrowing is sample size. A peer group of eight funds produces quartile breakpoints that move a great deal when a single fund is added or removed. Practitioners balance relevance against statistical stability and should say which cut they used.

Computing quartile breakpoints

A quartile breakpoint is a percentile of the peer distribution. The upper quartile breakpoint is the 75th percentile: funds at or above it are first (top) quartile. The median is the 50th percentile. The lower quartile breakpoint is the 25th percentile: funds below it are fourth (bottom) quartile. Industry usage numbers the quartiles from the top, so 'first quartile' is the best, which is the opposite of the statistical convention in which Q1 is the lowest quarter. Always check which way a report counts.

With a small peer set, a percentile rarely lands exactly on one fund, so a formula interpolates between neighboring values. This topic uses the inclusive linear interpolation method: sort the n returns from lowest to highest, compute a position p x (n - 1) counting from zero, and interpolate between the two funds on either side. It is the method behind the inclusive percentile and quartile functions in common spreadsheet software. A fund exactly at a breakpoint is placed in the higher quartile.

The peer set is twelve hypothetical 2016 vintage buyout funds, measured at the end of their eighth year on net IRR and net TVPI.

The quartile formula changes the breakpoints

Spreadsheet software and statistical packages offer several percentile formulas, and providers do not all use the same one. A common alternative is the exclusive method, which computes the position as p x (n + 1), counting from one, and so pushes the upper breakpoint further out and the lower breakpoint further in. For large peer sets the methods converge. For the small peer sets typical of a narrow vintage and strategy cut, the difference can decide a fund's quartile.

The difference matters most for a fund being evaluated against a benchmark it is not part of, which is the usual situation in due diligence. Such a fund is compared with breakpoints, and a small shift in the breakpoint can move it across the line.

Survivorship and self-reporting bias

Benchmark peer sets are built from the data a provider can obtain. Some providers collect cash flows from LPs, some from GPs, some from public filings and some from a mix. Where reporting is voluntary, a GP with a weak fund has less reason to report it, and a firm that failed to raise a successor fund may stop reporting altogether. Funds that disappear from a data set are disproportionately likely to be poor performers. The remaining peer set then overstates what a typical fund achieved, which is survivorship bias.

A related problem is selective reporting within a manager's own track record, where earlier or weaker funds are omitted, or where a fund is described as belonging to a strategy in which it compares better. Both effects push breakpoints upward, which makes a given fund's ranking look worse against the benchmark than it should, and a manager's self-selected record look better than it is.

The direction of the bias is predictable even when its size is not. The worked example shows the mechanism on the twelve-fund peer set.

Why early rankings are unstable

A fund's return in its first few years rests mostly on unrealized NAV, which is the GP's estimate, and on IRR timing effects. A single early exit can produce a very high IRR on a small amount of capital. Subscription lines can delay capital calls and shorten the apparent holding period. Conservative and aggressive valuation policies differ most in the early years, before sales test the marks. And the J-curve means that funds with heavier early fees or slower deployment look worse early regardless of what they will eventually deliver.

As the fund matures, the ranking comes to rest on cash. Early winners that were marked up may be sold for less, early laggards may produce large exits, and the IRR advantage from an early small exit fades as more capital is deployed and held for longer. The quartile a fund occupies at year 3 is therefore a weak predictor of where it finishes.

The table follows the same twelve funds at year 3 and year 8. Breakpoints were recomputed from the peer set at each date.

Different metrics, different providers

Quartile claims also depend on the return measure. In the year 8 table, Fund 7 is second quartile on IRR (12.5 percent, above the 11.75 percent median) but third quartile on TVPI (1.71x, below the 1.755x median). Fund 8 is the reverse: third quartile on IRR and second on TVPI. Fund 7 returned money faster, Fund 8 made more money in total. A manager can choose whichever metric flatters its fund, and some will. A careful reading asks for both, plus DPI and a PME.

Providers also disagree. Commercial benchmark providers draw on different sources of data, cover different sets of funds, apply different vintage and strategy classifications, treat currencies differently, and use different quartile methods and as-of dates. The same fund can therefore be top quartile against one provider's benchmark and second quartile against another's, with no one making an error. The honest presentation shows the provider, the peer group definition, the number of funds in it, the metric and the as-of date alongside the claim.

Using quartiles well

Quartiles are useful because dispersion between managers in private markets is wide, so knowing roughly where a fund sits against its true peers is valuable. They are most informative late in a fund's life, on net figures, with DPI high enough that the ranking reflects cash rather than marks. They are least informative early, on gross figures, against a loosely defined peer group.

For a manager with several prior funds, look at the pattern across all of them rather than the best one, and look at each fund's ranking over time rather than at a single date. Pair every quartile ranking with a PME, which asks a different question (did the fund beat the public alternative?) and does not depend on which other private funds reported. A fund can be top quartile in a weak vintage and still trail the index, or third quartile in a strong vintage and still beat it comfortably.