Conditional Information Ratios and the Information Curve

Conditional information ratios (CIR) measure lower partial performance of a portfolio conditional to the distribution of returns on the benchmark. The Information Curve plotted by a serial CIR provides an overall view of performance especially under downside scenarios. The CIR approach is closely related to the theory of conditional stochastic dominance and is thus distributional and modeling free. Statistical inference for CIR performance evaluation is developed and is free from the assumption of normality. Empirical evidence shows that although many US equity mutual funds suffer from the 2005-2010 downtrends, some funds outperform the market and provide good downside protection. Additionally, this article suggests Conditional Sharpe Ratios are useful for measuring hedge fund performance.