All, I have trouble understanding this sentence from the text- Volume 3 page 158 answer to eoc # 13
The factors commonly used in the factor-based approach generally have low correlations with the market and with each other. This results from the fact that the factors typically represent what is referred to as a zero (dollar) investment or self-financing investment, in which the underperforming attribute is sold short to finance an offsetting long position in the better-performing attribute. Constructing factors in this manner removes most market exposure from the factors (because of the offsetting short and long positions); as a result, the factors generally have low correlations with the market and with one another. Also, the factors commonly used in the factor-based approach are typically similar to the fundamental or structural factors used in multifactor models.
I didn’t go back and look at the question but it sounds like Beta hedging? Simplistically Like… buy one bank stock, short another bank stock… by doing that you’re basically removing the systemic risk and left with the difference in specific risk of each bank therefore low systemic risk and low correlation with each other.
Consider constructing “inflation” factor by long on treasury bond and shorting the inflation-indexed bond (delivers real rates)
Using the above example you have essentially isolated inflation and eliminated market risk in the process; If you weight these securities by their market risk or beta, then long 1 and short the other is essentially your zero-dollar investment
Consider how the 2 factors now show low correlation between each other and also show low correlation between these and market risk individually
This way, you can construct a portfolio with essentially each of what you want without the market noise
Appreciate your answers. I understand how to isolate the factors and what that does. I am struggling with correlation concept. They say this portfolio will have low correlation with the market and each other. So can you elaborate on this for me? Thanks in advance.
That is just historical correlation.
The performance of small caps (size factor) seems to have not much to do with how the momentum factor performs or other factors.
I guess you can find time periods for which the statement is wrong. But even if you have correlation then (randomly) it is not automatically causation.
Correlation with market removed = Beta low or close to zero.
Constructing a long-short portfolio in order to isolate the factors removes the Beta => Hence, correlation with market low. Factors portfolios constructed that way have low beta => low correlation with market.
Removing Beta = low or close to zero systematic risk = only retain idiosyncratic risks.
Factors portfolios only retain their own idiosyncratic risks; which are not necessarily related with each other => hence, low correlation with each other.