Funds of the same year share an entry market and an exit window. Two correlations, not one, show what pacing is worth.
Measure it with two correlations instead of one: funds of the same vintage move together more than funds of different vintages. On illustrative inputs, a 0.55x dispersion of net TVPI per fund, a 0.45 correlation within a vintage and 0.15 across vintages, 15 funds committed in one year have a dispersion of 0.384x; the same 15 funds spread over five vintages have 0.273x, 28.8 per cent less risk, and five funds over five vintages already beat fifteen funds in one.
Worked in full in The Private Markets Limited Partner by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Funds raised in the same year buy in the same market, borrow at the same rates and sell into the same exit window. That shared exposure is a correlation, and adding more managers in the same year does not diversify it away. The standard formula for an equal-weighted portfolio uses a single average correlation, as in how many private equity funds you need to diversify. Splitting that correlation into a within-vintage and an across-vintage figure turns pacing from a liquidity habit into a risk decision with a number on it. All inputs are illustrative.
| Input | Value |
|---|---|
| Funds in the portfolio, equal weights | 15 |
| Standard deviation of one fund's net TVPI | 0.55x |
| Correlation between funds of the same vintage | 0.45 |
| Correlation between funds of different vintages | 0.15 |
| Expected net TVPI of each fund | 1.60x |
Variance of the portfolio = σ² ÷ N² × [N + (within-vintage pairs × ρw) + (across-vintage pairs × ρb)]
With k vintages of m = N ÷ k funds: within-vintage pairs = k × m × (m − 1); across-vintage pairs = N × (N − 1) − within pairs
Excel: =0.55*SQRT((15+W*0.45+B*0.15)/15^2), with W and B the pair counts
Step 1, one vintage. All 15 funds share a vintage, so all 210 ordered pairs carry 0.45. The variance is 0.55² ÷ 225 × (15 + 210 × 0.45), and the standard deviation is 0.384x.
Step 2, five vintages of three funds. Within each vintage there are 3 × 2 = 6 ordered pairs, 30 in all; the other 180 pairs are across vintages at 0.15. The average correlation falls from 0.450 to 0.193 and the standard deviation to 0.273x.
Step 3, compare. Same funds, same managers, same expected return: 28.8 per cent less dispersion, purely from the calendar.
| Vintages | Funds per vintage | Average correlation | Std dev of TVPI | Risk reduction | P(TVPI < 1.0x) | P(TVPI < 1.3x) |
|---|---|---|---|---|---|---|
| 1 | 15 | 0.450 | 0.384x | 0.0% | 5.89% | 21.7% |
| 3 | 5 | 0.236 | 0.294x | 23.3% | 2.08% | 15.4% |
| 5 | 3 | 0.193 | 0.273x | 28.8% | 1.40% | 13.6% |
| 15 | 1 | 0.150 | 0.250x | 34.8% | 0.82% | 11.5% |
The loss probabilities use a normal approximation around a 1.60x mean, so read them as a ranking rather than a forecast. They make the point plainly: the chance of the whole portfolio returning less than its money falls from 5.89 per cent to 1.40 per cent by spreading the same commitments over five years. A single fund, for comparison, has a 13.8 per cent chance on the same inputs.
Vintage spread beats manager count. Five funds over five vintages give 0.311x; fifteen funds in one vintage give 0.384x. No number of funds in one vintage gets below 0.55 × √0.45 = 0.369x, while the floor across many vintages is 0.55 × √0.15 = 0.213x.
The within-vintage correlation is the input an investment committee should argue about, because it decides how much pacing is worth.
| Within-vintage correlation | One vintage | Five vintages | Risk reduction |
|---|---|---|---|
| 0.25 | 0.301x | 0.258x | 14.4% |
| 0.35 | 0.345x | 0.266x | 23.0% |
| 0.45 | 0.384x | 0.273x | 28.8% |
| 0.60 | 0.435x | 0.284x | 34.8% |
The more a vintage's funds move together, the more the calendar is worth, and the less the result depends on how many managers are added. Beyond five vintages the floor flattens: 0.252x with five, 0.233x with ten and 0.227x with fifteen, so most of the benefit is captured by a five-year cycle.
Counting funds and calling it diversification. A new programme that commits its whole allocation in one or two years to reach target quickly can hold twenty managers and still carry the risk of a single vintage, 0.369x at the floor here, because one entry market sets most of the outcome. The second mistake is to treat a market dislocation as a reason to stop committing. Skipping a vintage gives up a year that, in this model, carries only a 0.15 correlation with the funds already held, and leaves the programme more concentrated in the vintages it bought at higher prices. Steady pacing, the approach that also smooths the cash flows shown in when a private equity programme becomes self-funding, is a risk control in its own right.
The two correlations are estimates, and the within-vintage figure is the hard one: it rises in a crisis, which is exactly when it matters. Treat the 0.45 here as a placeholder and test the range above against your own programme's history. Nor does spreading commitments come free. A programme that paces evenly reaches its target allocation later, and the years spent underweight carry a cost of their own, which should be set against the 28.8 per cent of risk it removes.
The portfolio construction and pacing figures of the book are reproduced in the free workbook for this book, with a blank version for your own programme.
Because funds raised in the same year buy at the same prices and sell into the same exit window, so their returns are correlated. On illustrative inputs, no number of funds in one vintage takes dispersion below 0.369x of TVPI, while spreading across many vintages lowers the floor to 0.213x.
Most of the benefit comes from about five. With fifteen funds on illustrative inputs, dispersion is 0.384x in one vintage, 0.294x over three, 0.273x over five and 0.250x over fifteen. With very many funds the floor is 0.252x at five vintages and 0.227x at fifteen.
Vintages, once a handful of managers is held. On the worked inputs, five funds spread over five vintages have a dispersion of 0.311x, lower than fifteen funds committed in a single vintage at 0.384x. Adding managers within one year cannot reduce the shared vintage risk.
This article is one calculation from The Private Markets Limited Partner. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: private equity and private markets → · All 1340 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.