Recent research in mathematical methods for finance suggests that time series for financial data should be studied with non-stationary models and with structural changes that include both jumps and heteroskedasticity (with jumps in variance). It has been

This paper examines the sensitivity of various measures of portfolio performance to the choice of the benchmark model using the asset-pricing kernel methodology. It derives the appropriate estimation frameworks that are suitable to perform evaluations of fixed-weight and dynamic portfolio strategies. Various asset-pricing kernel-based benchmark models are tested using a comprehensive sample of Canadian equity mutual funds over the period, 1989–1999. The performance statistics and inferences are sensitive to the choice of the kernel-based benchmark model and to the choice of liquidity as an alternative sorting variable for forming the passive benchmark portfolios. However, they are robust to the removal of ex post index mimickers and somewhat less robust to the presence of nonlinearities in the structure of the pricing kernel. Since conditioning has a more pronounced impact on absolute than on relative performance inferences, this supports the common industry practice of comparing the performance of funds against each other.