21 – M4 L3a 171 Turnover As Proxy For Real World Constraints V2

Since liquidity and transaction costs are dependent upon market conditions at the time of a trade, it’s difficult to simulate actual transaction costs when evaluating an Alpha factor. So a useful proxy for these real-world constraints is to measure turnover. Generally, factors that are updated quarterly such as fundamental factors, have lower turnover compared to factors that are updated daily such as those based on price and volume data. Turnover measures what fraction of the portfolio’s total value gets traded in a time period. For example, let’s say a portfolio is valued at $100 million. On a single day, let’s imagine one million dollars are brought and another one million dollars are sold. So, the turnover in this case is two million dollars divided by 100 million or two percent. Since we’re out the Alpha research stage and not dealing with an actual portfolio, turnover can be defined in terms of the change in portfolio weights. We calculate turnover by taking the difference between weights for each stock between yesterday and today, then we sum up the absolute values of the weights for all stocks in the portfolio. Recall that when constructing an Alpha factor, we’re likely to convert raw factor values into ranks. So, there is another way in which we can measure turnover called the Factor Rank auto-correlation. We’ll see this next

%d 블로거가 이것을 좋아합니다: