Do hedge funds rely on off-the-beaten-track (illiquid) positions to fuel performance? In his April 2011 paper entitled “Hedge-Fund Performance and Liquidity Risk”, Ronnie Sadka investigates aggregate market liquidity as a predictor of hedge fund performance. His calculates liquidity based on trade-by-trade price impact estimated monthly for individual stocks and aggregated by averaging. Using net monthly returns for a broad sample of live and dead hedge funds during 1994 thorugh 2009 and contemporaneous trade-by-trade stock prices for liquidity calculations, he finds that:
- Returns for eight of 11 hedge fund category indexes (Convertible Arbitrage, Emerging Markets, Equity Market Neutral, Event Driven, Fixed Income Arbitrage, Fund of Funds, Long/Short Equity, and Multi-Strategy) exhibit significantly positive relationships with liquidity.
- Hedge funds with high liquidity risk subsequently outperform those with low risk by an average of about 6.5% annually over the entire sample period, but they underperform during liquidity crises.
- Results are independent of share restrictions, suggesting frequent imbalances between the liquidity funds offers investors and the liquidity of fund positions.
- Among commonly used stock return indicators, price-impact liquidity correlates most strongly with VIX (-0.27).
In summary, evidence indicates that liquidity risk accounts for a substantial part of hedge fund returns, which somewhat resemble those from selling out-of-the-money put options.
Results suggest the possibility flow-driven implosion during liquidity crises of hedge funds with high liquidity risk and mismatched liberal investor withdrawal policies.
Cautions regarding findings include:
- Tick-by-tick stock price-impact liquidity calculation is impractical for many investors.
- Obstacles to hedge fund switching, such as withdrawal restrictions and tax implications, may limit exploitability of findings.
See also “Hedge Fund Outperformance: Skill or Liquidity Risk?” for a similar study with similar conclusions based on data through 2006.
