Audit adjustments are the cleanest evidence an operational due diligence analyst has on valuation. Their sign says more than their size.
To test whether a GP's unaudited NAV is biased, compare it with the audited NAV each year and count the adjustments by sign. In an illustrative eight-year record, seven of eight audits cut the year-end NAV, which happens by chance only 3.5 per cent of the time. The adjustments average minus 1.52 per cent: small enough to pass as rounding one year at a time, consistent enough to say something about the valuation process.
Worked in full in Operational Due Diligence in Private Equity by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Operational due diligence rarely gets to re-perform a valuation. It can, however, compare what the manager reported before the audit with what survived it. That comparison is the cleanest evidence an analyst has, because the auditor's figure is the one the manager had to defend line by line. The test needs eight numbers and one binomial formula.
| Audit year | Manager A | Manager B |
|---|---|---|
| 1 | −1.8 | 0.9 |
| 2 | −2.4 | −1.1 |
| 3 | −0.6 | 0.3 |
| 4 | −3.1 | −0.8 |
| 5 | 0.4 | 1.4 |
| 6 | −1.2 | −0.5 |
| 7 | −2.0 | 0.6 |
| 8 | −1.5 | −0.2 |
| Mean | −1.52 | 0.07 |
Ask for these figures directly: the unaudited NAV reported to LPs at year end, and the audited NAV in the financial statements. Use the fund level, and the same fund or a consistent set of funds across years.
Under no bias, each adjustment is negative with probability 0.5.
P(at least k negative of n) = Σj=k..n C(n, j) ÷ 2n = (C(8,7) + C(8,8)) ÷ 256 = 3.5%
In Excel: =1-BINOM.DIST(k-1,n,0.5,TRUE), so =1-BINOM.DIST(6,8,0.5,TRUE) for seven or more of eight.
Manager A has seven negative adjustments out of eight. If its unaudited marks were unbiased, the chance of seven or more downward revisions would be 9 in 256, or 3.5 per cent. The test is one-sided because the hypothesis, that pre-audit marks lean high, was stated before looking at the data; tested both ways the figure is 7.0 per cent.
Manager B has four negative adjustments out of eight and a mean of 0.07 per cent. Four or more of eight happens 63.7 per cent of the time by chance: nothing to explain. Its average absolute adjustment, 0.72 per cent, is the ordinary noise of a year-end close.
| Downward adjustments | Probability |
|---|---|
| 4 or more of 8 | 63.7% |
| 5 or more of 8 | 36.3% |
| 6 or more of 8 | 14.5% |
| 7 or more of 8 | 3.5% |
| 8 of 8 | 0.4% |
Manager A's record is unlikely to be chance. Its size, a net 1.52 per cent a year on average, would not trouble a materiality threshold in any single year, and that is exactly why the test looks at the sign. A manager whose auditor trims the NAV every December has a valuation process that starts high, whether through optimistic inputs, stale comparables, or the absence of an independent challenge before the auditor arrives.
A positive result is the start of the conversation, not the conclusion. Ask which positions drove each year's adjustment, whether the same names recur, who prepares the year-end marks and who challenges them before the auditor does, and whether the valuation committee minutes show any mark being reduced on internal review. A process with genuine internal challenge produces adjustments of both signs, because the challenge catches errors in both directions before the auditor sees them.
Short records need care. With four audits, four negative of four happens 6.25 per cent of the time, so even a perfect run is weak evidence on its own. Eight years is the minimum for a confident reading; with fewer, combine funds or use quarterly comparisons against later realisations, as in backtesting valuations against exits.
The pattern has a price for anyone transacting on unaudited figures. A secondary buyer bidding 90 per cent of a 500 unaudited NAV pays 450. If the audit takes the NAV down by Manager A's average, to 492.38, the buyer has paid 91.4 per cent of audited value, a discount of 8.6 per cent rather than the 10 it negotiated. In the worst year, a 3.1 per cent cut, the effective price is 92.9 per cent.
The same overstatement flows into anything computed on unaudited NAV between reporting and audit: interim performance figures in a fundraising deck, NAV-based fees in later years, and the collateral value of a NAV facility. Ask which of these the manager calculates on pre-audit numbers.
The common mistake is reviewing adjustments one year at a time against a materiality threshold. Each of Manager A's adjustments is 3.1 per cent or less, and each would be described in the audit as immaterial. The information is in the run, not in any single figure. The second mistake is accepting the explanation for each year separately: a late-arriving portfolio company account, a comparable that moved in January. Every explanation can be true and the pattern still mean the process starts high.
Collect eight years of unaudited and audited NAVs, count the downward adjustments and compute the binomial tail: seven of eight is 3.5 per cent by chance, four of eight is 63.7 per cent. Ask about the valuation process when the sign is one-way, whatever the size. The free workbook for this case includes a case on marks that come down at year end and the effect on a buyer pricing off interim NAV.
A test of whether audit adjustments are more often negative than chance allows. Without bias, each adjustment is equally likely to be up or down, so the number of downward revisions follows a binomial distribution with a probability of one half. Seven or more downward revisions out of eight has a 3.5 per cent probability; five or more has 36.3 per cent.
Direction matters more than size. Adjustments averaging 1.52 per cent that are negative seven years out of eight show a pattern; adjustments averaging 0.72 per cent in absolute terms that go both ways, four of eight negative, do not. A single large adjustment is a question for the auditor; a one-way run is a question about the valuation process.
The bias is added to the price. A buyer bidding 90 per cent of a 500 unaudited NAV pays 450 for an interest the audit will value at 492.38, which is 91.4 per cent of audited value. In this manager's worst year, a 3.1 per cent cut, the same bid is 92.9 per cent of the audited figure.
Year-end marks that keep coming down are the subject of a companion case extending chapter 11 of Operational Due Diligence in Private Equity. 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.