Reading 43: Evaluating Portfol....rmance－LOS r

The null hypothesis states that the expected value-added return is zero. We are testing the managers ability to generate positive expected value added returns. There is no reason why skewness in the distribution should exist and thus a Chisquared distribution is inappropriate in this case. We want a consistent process to ensure that the distribution of value added returns about their mean is constant. We do indeed assume that value-added returns are independent and normally distributed.

In order to determine statistical significance or otherwise, confidence intervals will need to be constructed using the standard deviation of the returns, this is not observable from plotting the manager's value added returns alone and would need to be separately calculated. All the other conclusions are correct.

In portfolio management, quality control charts are used to determine if a manager has statistically significant excess returns. The managers returns versus a benchmark are plotted on a graph where time is on the x-axis and value-added (excess) return is plotted on the y-axis. A confidence interval is formed around the x-axis of zero. If the managers returns plot outside the confidence interval, we conclude that the manager has generated statistically significant excess returns.