For investors looking for excess premium, it is relatively well known that size doe snpt matter when it comes to market cap as
large cap stocks over the long term.
However, a common recommendation within asset management is to remove the size factor as it offers relatively weak performance over the short term, according to Scientific Beta’s recent white paper.
The study, named Does the Size Factor Still Have its Place in Multi-Factor Portfolio, analyses the performance of the size effect in comparison with other factors.
The scrutiny of the size factor is nothing new. A paper by Banz in 1981 shows returns to be quite underwhelming between 1926 and 1976. However, there was a spike in the data used from 1976 to 2018.
Many years after the release of the study, the data used in Banz’s paper had its shortcomings and was not as “clean” in the 1980s as today’s standards allow.
In contrast, a recent study by S&P Dow Jones Indices supports investors interest in small cap stock. The index provider also found the most profitable small cap stocks offered greater returns.
Scientific Beta says there is a conflict between results documented in asset price research and provider claims about the weakness of the size effect. There also remain inconsistencies in providers’ conclusions of the size factor.
Perspective is to be one of the reasons for the inconsistencies on size factor. Smart beta providers usually asses and compare the factors individually when investors are actually more interested in seeing how the factor contributes to investment outcomes when used alongside other factors within a portfolio.
Scientific Beta calculate the unexplained returns for each factor. The results show that excluding the size factor increases the proportion of unexplained returns from around 60% to 80%. Excluding the momentum and high profitability factor also leads to the deterioration of the model’s performance, but with slightly less magnitude than the size factor. There is no evidence from the test which says excluding the size factor improves performance.
Looking at size premium, Scientific Beta found the factor’s performance to be economically and statistically significant over long-term US history when accounting for the presence of commonly used equity factors.
Size’s returns, when it is unrelated to exposures to other factors, is significantly positive and is similar in magnitude to the premia of other factors such as value and high profitability.
Size also offers diversification given its low correlation with other factors. Not only would it keep a role in a multi-factor portfolio if its premium was significantly lower, it also diversifies macroeconomic risks, in particular, interest rate risk. The size factor receives substantial allocations in a mean-variance efficient portfolio which includes bonds.
To conclude, Scientific Beta says size does matters and supports the factor as it is not an obvious candidate to remove from a factor menu.
Looking at its performance individually isn’t sufficient or beneficial for making decisions on its inclusion in a portfolio which would include other factors.