Outliers - NumXL Articles

In this paper, we’ll examine a claim by a portfolio manager (let’s call him trader B) about his ability to generate statistically significant Alpha. Alpha is the excess-return compensation for the risk borne, and thus commonly used to assess active managers’ performances. Using plain summary statistics and empirical distribution plots (e.g. Histogram, QQ-Plot), we identified a single data-point with relatively high-return, and once excluded from the sample, the so-called alpha simply vanished.

The case deals with the predictability of the daily closing EUR/USD exchange rate, given its prior history. First, we constructed a trading strategy where we buy EUR at market open, and close the position (selling EUR) at the market close and calculate the daily return time series. After thorough statistical analysis, we found those daily returns behave much like a white-noise. No predictability power.