Sizing Up The Size Premium by by Gerstein Fisher
Since Rolf Banz published his groundbreaking paper that identified the so-called “small stock effect” in 1981, the investment community has acknowledged the existence of a return premium afforded to smaller-capitalization stocks over their larger counterparts. Banz’s study demonstrated that between 1926 and 1980, the smallest quintile of the stocks on the New York Stock Exchange outperformed both large stocks and the overall market.
Between 1961 and 1980, the size premium earned an annualized return of a full 5%. From the time the Banz paper was published through the end of 2014, however, that premium shrank to closer to 1%. So is it possible that the size premium is a result of data mining or biases in the historical data, but does not hold in reality? This paper explores this question, as well as some of the key theoretical arguments as to why a size premium should exist – and persist into the future. From a practical standpoint, we touch on considerations surrounding what we think is the most efficient way to access the size premium.
Looking over the period from 1926 through the end of 2014 – nearly a century of data – there has been an almost monotonic relationship between firm size and return, where smaller-capitalization stocks have earned higher returns. (See Exhibit 1).
When we narrow the time frame for analysis to the past 20 years, the pattern is not as smooth, as can be seen in Exhibit 2. The smallest decile of US stocks still outperformed, but with more “noise” in between the extremes. This result is consistent with what we would expect when analyzing a shorter time period in which the data can be more idiosyncratic.
Exhibit 3 examines the last 10 years of data. This chart reveals what we think is the shorter-term risk inherent in factor investing, particularly when investments are focused on a single factor. As seen in Exhibit 3, the smallest-cap stocks outperformed only the very largest stocks over this period.
Some researchers have sought to expose the size premium as more of a myth than a verifiable and lasting phenomenon. Arguments against the existence of a true size premium include: the inaccuracy of returns; the possibility that the size premium is the result of data mining; the fact that small cap long-only indices ignore trading costs; and the fact that the size premium is much weaker after 1981.
While we acknowledge the validity of some of these points and we do agree that the size premium is weak, it is still positive and we believe it still makes sense for portfolios to be tilted toward small-cap stocks in spite of these arguments.
Let us first explore the theoretical argument for why the size premium exists. Kalesnik and Beck (2014) explain that if a large company splits into two smaller companies, size theory proponents would argue that expected returns increase, while the authors claim nothing has really changed.
See full PDF below.