Hillsan Credit Opportunities Fund reports annualised volatility of 4% with first-order autocorrelation of 0.5 in its monthly returns, because it marks illiquid loans to model. What is its likely true volatility, and what does that do to its Sharpe ratio?

Hillsan's true volatility is probably about 6.9%, not 4%, so its Sharpe ratio falls from 0.75 to about 0.43. Autocorrelation of 0.5 means each month's reported return is half the new shock and half last month's mark, so losses are spread out and look small. Unsmoothing multiplies volatility by the square root of 3, about 1.73, and divides the Sharpe ratio by the same factor.

In a liquid market, last month's return tells you almost nothing about this month's, because prices absorb news quickly. A book marked to model behaves differently. Think of a thermostat that reports the average of the last few readings: a sudden cold snap shows up as a gentle dip spread over days. When a fund's marks lag the true value, each reported return is partly this month's shock and partly last month's, which shows up as positive autocorrelation and makes volatility look lower than it is. An autocorrelation of 0.5 is a strong signature.

Model the reported return as a blend: half the true return this month and half last month's reported return. UnsmoothingReversing the lag in reported returns by assuming each reported figure blends the true return with the previous reported one, then solving for the true series; the approach is associated with Geltner and with Getmansky, Lo and Makarov. solves for the true series. With a blend weight of 0.5, true volatility equals reported volatility times the square root of (1 + 0.5) over (1 - 0.5), which is the square root of 3, about 1.73. Hillsan's 4% becomes about 6.9%.

The average return does not change: smoothing shifts gains and losses between months but does not create them. So the Sharpe ratio, excess return over volatility, falls by the same factor of 1.73: from 3 over 4, 0.75, to 3 over 6.9, about 0.43. Two other numbers move with it. Drawdowns are deeper than reported, because losses are spread across several months of marks. And the fund's correlation with equities and high-yield bonds is understated, since a lagging mark cannot move with a market that falls in a week, so its diversification value is overstated.

Compare Hillsan with liquid credit funds only after unsmoothing, and ask how the book is marked: how often, by whom, and against what evidence. A Sharpe ratio of about 0.43 for a credit fund with quarterly liquidity and model marks is ordinary, not exceptional, and the allocator should price the illiquidity rather than reward it. The limit of the method is worth saying: 0.5 is estimated from five years of monthly data and is itself noisy, and some autocorrelation can come from genuine momentum in credit. Treat the corrected numbers as the better estimate, not the truth.