For the 1,438 with TVPI data, the simple average is a multiple of 1.79.
Overall, the results: 1) quantify the relationships of fund outcomes (next-stage funding, IPO, and M&A) to realized performance (IRR, TVPI, and our hybrid measure that combines the two), 2) quantify persistence effects through the relationships of prior outcomes to fund performance, 3) assess the relationships of sector-specific experience and agility to performance, 4) quantify the persistence of the firm's investment outcomes, 5) assess the separate effect of firm reputation on performance, and 6) verify that our results are robust to sources of potential bias.
Table IV contains four different models of fund performance where net IRR is the dependent variable (Models A- 1 through A-4), as well as two models in which TVPI is the dependent variable (Models B-3 and B-4) and two in which we combine IRR and TVPI into our hybrid measure (IRR:TVPI) that is more closely aligned with the true dependent variable of interest, fund net present value (NPV) over consistent investment horizons (Models C-3 and C-4).
Results are more significant for the TVPI and hybrid models.
If this were occurring, we would expect to find a different result when the fund TVPI is used as the dependent variable.
Based on Model A-3, a one standard deviation (20 percentage point) increase in IPO percentage is associated with a 6.18 percentage point increase in fund IRR and a 0.366 point increase in the TVPI multiple, whereas a one standard deviation (15.7 percentage point) increase in M&A percentage is associated with a 3.58 percentage point increase in IRR and a 0.179 point increase in TVPI.
For performance as reflected by TVPI, the contribution is usually around two-thirds as strong as IPO percentage.
Firm outcome coefficients are consistently strong economically and, except for some TVPI coefficients, always statistically significant in Table IV.
Firm-level M&A percentage also is not significantly related to TVPI, but it is to the other two performance measures.
In Model A-3, a one-investment increase in experience is associated with a 0.042 percentage point higher IRR, and a one standard deviation increase in experience (93.9 companies) is associated with a 3.96 percentage point higher IRR and a 0.122 point higher TVPI multiple.
Table V presents the results of this analysis for each of the three performance measures (IRR, TVPI, and IRR:TVPI).
(23) We apply this technique by performing simultaneous quantile regression for each decile of the IRR and TVPI distributions.