A small venture firm reports a value for every company it holds, every quarter, and most of those values are judgments. Its investors read them once a quarter. When the firm raises its next fund, prospective investors read them closely, because the current fund's interim returns are most of the track record a young firm has. The person who sets the marks is often the person raising the fund.
The operating decision is what the firm's reporting should let a careful reader check, and when; how the firm reads the interim marks it receives from the companies and funds it holds; and what it records about valuation changes around its own raise. I went looking for evidence on whether fund managers lean on their marks while fundraising, and found a study that tested it on a large set of fund records. It studies private equity funds reporting to large institutions, not a small firm reporting to its own investors, and I say where that matters.
Source
What the study measured
Gregory Brown (University of North Carolina), Oleg Gredil (Tulane University), and Steven Kaplan (University of Chicago) published "Do Private Equity Funds Manipulate Reported Returns?" in the Journal of Financial Economics (volume 132, issue 2, pages 267 to 297, 2019). The publisher would not serve me the article, so I read the authors' open working paper, NBER Working Paper 22493, dated August 2016. Every figure and page number below comes from that version. The published abstract reports the same findings, but the journal's figures may differ, and I could not check them.
The data come from Burgiss, which collects them from investors rather than managers: quarterly cash flows and reported net asset values (NAVs) for 2,071 funds, supplied by over 200 institutional investors with about $0.75 trillion committed to private equity (page 2). There are 997 buyout funds and 1,074 venture funds, with data through March 2012 (page 9). About 60 percent of the investors supplying data are pension funds and 20 percent endowments and foundations (page 8). A second database, from StepStone, confirms whether and when each manager raised its next fund (pages 8 and 9). The main regressions cover 1984 through 2011 (page 25).
The authors date each manager's fundraising by the first capital call of its next fund, and treat the three years before it as the time the firm is most likely raising (page 13). For a manager that never raised again, they place the event 13 quarters before the fund's last report (page 12). They measure each fund's reported performance against the public market. The test is simple: if a manager inflates values to raise a fund, its returns should fall back later, when companies are sold or written off (page 11).
Source
What it found
Raising a next fund was the norm: 86.5 percent of buyout funds and 83.2 percent of venture funds had a follow-on fund (page 10).
The funds that failed to raise one showed the pattern the authors were looking for. Their excess returns rose in the period when the firm would be making a final effort to raise, then reversed in the fund's last years as companies turned into cash, ending in negative excess returns (page 15, Figure 2). The authors read this as suggestive of attempts at manipulation that did not succeed (page 15). The paper draws this as a chart and does not print the size of the rise or the reversal.
The size comes from the regressions, as a bias in one quarter's reported return. Funds behind their peers tended to report higher returns the next quarter, and funds ahead tended to report lower, a pattern the authors call peer chasing; they put a typical distortion at about 2.6 percentage points of quarterly return (pages 25 and 26). For a buyout fund two standard deviations below its peers that had gone five years without raising, the model puts the upward bias in the next quarter's reported change at nearly 5 percentage points; for a similar fund two standard deviations above its peers, about 1 point; venture funds were similar (page 26). These are the authors' model estimates for a typical fund, not a measured overstatement of any one fund's value.
Investors appear not to have been fooled. A fund whose performance fell after the fundraising date, a sign that its value had been optimistic, was 11.6 percentage points less likely to have raised a next fund among buyout funds and 8.5 points among venture funds. Being in the top third of peers made a next fund 9.5 and 10.2 points more likely (page 18; Table 3, specification 2). A run-up in reported returns in the year before the event also lowered the odds, but large distributions in the same year, more than 20 percent of NAV, largely offset that penalty (pages 17 and 18; Table 3, specification 4). Cash back was the signal investors trusted.
Interim rank was mostly informative. Among funds that did raise again, 61.6 percent of buyout funds in the top third at fundraising finished in the top third, and 61.2 percent in the bottom third stayed there; for venture funds the figures were 60.3 and 55.6 percent (page 16, Table 2). About 10 percent of funds that finished in the top third had been in the bottom third during fundraising (page 16).
The best funds leaned the other way. Top-third funds tended to report downward-biased returns for the next quarter when they were one or more standard deviations above their peers, which the authors read as conservative marks, a cushion against bad luck later (page 28). Managers with two or fewer prior funds were not more aggressive; if anything, they were somewhat more conservative (page 28). The authors conclude that overstating interim returns has not been a winning strategy on average (page 5).
Synthesis
What the study does not tell a small firm
The readers in the study were large institutions. The authors cite a survey in which more than 90 percent of investors computed their own measure of a manager's past performance, and 30 percent made their own fair value assessments of reported NAVs (pages 6 and 7). A small firm's investors are often individuals, family offices, and a few institutions, and few of them rebuild a manager's marks. The finding that investors see through inflated values rests on readers who check. The authors say as much: the cost of the practice depends on whether less skilled investors misallocate capital (page 32).
The data are old. They end in 2012, and before 2009 managers had wide discretion to hold assets at cost until an explicit change (page 11). Fair value reporting under ASC 820 narrowed that, and the authors found only weak evidence that the change affected venture fund reporting, in a test confounded by the financial crisis (page 32).
The unit differs too. The study sees fund-level patterns, never a single mark. A small venture firm's marks are mostly last-round prices, so the discretion sits in timing: when a company is written down, and when a holding moves off cost. The paper names exactly that as a source of bias for venture funds (page 11).
And not every rise before a raise is manipulation. Managers may exit their best companies early to show results, or learn late that an investment has failed (pages 13 and 14). A pattern is a reason to ask, not a finding.
Synthesis
What carries across anyway
The incentive carries unchanged. Interim marks are the track record, and the person marking is often the person raising. The risk is concentrated where the study found it: a firm whose current fund trails its peers and that has gone a long time without raising. The pattern a reader can check carries too: reported value rising in the year before a raise, with no cash coming back, then falling after it.
The two behaviors the study rewarded are the ones a small firm controls. Distributions were credible where reported gains were not, and the best managers marked conservatively. A firm that marks so that a later write-down surprises no one is doing what the top funds did.
Proposal
What the firm's reporting should let a reader check
Every quarter, in the same format, I would report for each holding the current mark, the prior mark, the method (the company's last round, a recent round led by others, a write-down, or comparable companies), the date and source of the evidence, and a one-line reason for any change. At the fund level, I would show cash returned separately from value still held, so a reader can see how much of the return is realized. The firm's valuation policy goes out with the first report each year.
Two timing rules matter most. A change in valuation method is announced in a quarterly report before it is used, never first in a fundraising deck. And every figure in the deck traces to a quarterly report investors already have.
Proposal
How I would read the marks the firm receives
For each fund the firm holds, I would put the reported value beside the cash. Is the manager raising now? Did the fund's reported value rise over the last four quarters without distributions? Where does it sit against its peers? A fund that trails its peers, is raising, and shows a run-up without cash is the case the study describes, and I would ask its manager for the method behind each mark that moved. For companies, I would compare the figures a mark rests on with the audited figures when they arrive, and note the difference.
Proposal
What the firm records around its own raise
From the quarter the firm decides to raise until four quarters after the final close, I would keep a valuation log: each change in a mark, its evidence, who proposed it, and who approved it. Someone not raising the fund, the chief compliance officer or an outside reviewer, reviews every increase. The log also records write-downs the firm considered and deferred, with the reason.
After the raise, a lookback: compare each mark at the time of the raise with what came next, a realization, a new round, or a write-down. That is the firm's own version of the study's test, and it shows whether the firm's marks held.
Open question
Will the firm's own investors read any of it?
The study found that inflated marks did not work because the investors it studied checked. A small firm's investors may not. The reporting above lets them check, but it cannot make them. I do not know whether a firm should go further and ask an outside party to review its marks before a raise, at a cost that is large for a small firm. The lookback, run after each raise, is the cheapest way I know to find out whether that review would have changed anything.