Gallery:Step 8|Step 8: Riskese

The Five Qualifications of a Risk Premium

Gallery:Step 8|Step 8: Riskese

How can we determine whether or not we have identified an exploitable risk premium—a source of higher expected returns that we can implement in our portfolios? To understand the concept of a risk premium, imagine that the only available investment is Treasury Bills, widely considered to be a risk-free asset. Note that we are using the term risk-free somewhat loosely, for even though the risk of nominally losing money is pretty much non-existent,  there is still a high risk of losing purchasing power due to a failure to keep up with inflation. At today’s yields, that risk appears to be a virtual certainty at least for the short-term. Now suppose a risky asset class such as equities becomes available. The risk premium of this asset class is the amount of additional expected return investors would require to move to it from the risk-free asset class. Academic researchers have identified several risk and return premiums such as market (stocks have a higher expected return than T-bills), size (small caps beat large caps), relative price (value beats growth), and profitability (high profitability companies beat low profitability companies). The purpose of this article is to define the criteria for deciding when we have indeed found a dimension of higher expected return. Throughout the article, we will use the size premium as an example.

The first question we ask is whether or not it is sensible. Can we intuitively understand why investors should expect a higher return for taking on this risk? For small companies vs. large companies, answering this question is quite easy, as any reasonable investor would agree that smaller companies tend to be riskier than larger companies. The relationship between risk and company size is apparent in the credit ratings assigned to bonds issued by different companies, according to this research report in which the author explains the relationship: “Larger companies tend to have higher credit ratings. Empirically, size metrics offer the strongest statistical correlation with credit ratings—reflecting important qualitative factors such as geographic and product market diversification, competitive position, bargaining power, market share and brand stature.” If smaller companies have a higher cost of capital for their debt, then it stands to reason that they would have it for their equity as well. Recall that a company’s cost of capital is paid to its investors as returns.

The second question is whether the premium is persistent across time periods. If a premium is the result of just a few concentrated years, then it may be more of an anomaly and not the basis of a sound investment strategy. The size premium was first identified by Rolf Banz in 1981, using data going back to 1927 from the Center for Research in Security Prices. A fair question to ask is what has happened with the size premium in the years after it was identified—the “out-of-sample period.” As the bar chart below shows, although it has not been as strong as it was in the earlier years, it has persisted.

One possible explanation for the diminishment of the size premium is the very low return received by extreme small cap growth stocks. These companies tend to have low, zero, or negative profitability, so they are priced based on anticipated cash flows in the distant future and thus tend to be of a highly speculative nature. The funds that IFA advises for our clients exclude companies from this corner of the market.

The third criterion is whether the premium is pervasive across markets. If it is only observed in one or a few financial markets, then it may again be chalked up to an anomaly. IFA constructed its own version of Fama and French’s SmB (small minus big) premium for international developed and emerging markets. The bar chart below shows that the premium pervades markets around the world. We found that the results are statistically significant at a 95% confidence level (t-stat in excess of two).

The fourth test we perform is to see whether it is robust to alternative specifications. For the small premium, we have arbitrarily drawn the line between the stocks that make up the top 90% of market capitalization (i.e. the large and mid-cap stocks) and the stocks that make up the bottom 10% (small caps). Although it may seem counterintuitive, there are actually many more names included in the latter category. If the small premium were attributable to only a few names in one part of the small category (perhaps occurring in the 7th to 10th percentile), then it would not hold up if we specified small caps in a different way. We tested this out by comparing micro cap (the bottom 5% of the market) to the rest of the market. The chart below shows that when micro cap is substituted for small cap, the size premium increases, which is exactly what we would expect.

The final question we ask is whether it is cost-effective to capture in well-diversified portfolios. This very objection has been raised many times in connection with small cap companies because they are less liquid than large cap companies and thus costlier to trade. If the trading costs were to overcome the anticipated gain from exposure to the premium, then there would be no point in attempting it. To address this important issue, Dimensional Fund Advisors (DFA) developed their patient and opportunistic approach to trading. Specifically, whenever possible, they play the role of a liquidity provider to the market rather than a liquidity seeker. For example, if an active manager is holding a small cap stock that he now believes is “overvalued”, DFA may offer him a price that is superior to what he would achieve on the open market (where he would end up moving the price) yet is lower than the most recent bid price. What allows DFA to be indifferent among particular stocks is its willingness to view securities with similar valuation characteristics as substitutable. In an internal study1, DFA recently estimated that their improvement in trading costs for small cap stocks ranged from 0.21% to 0.45%, depending on the designated market and the methodology used by a liquidity-demanding trader.

Putting all of this together, the chart below shows how three dimensions of expected returns (size, relative price, and profitability) have performed in the global equity markets.

Combining multiple dimensions into a single portfolio is not a simple task. For example, loading up on profitability automatically causes a decrease in exposure to small cap and value stocks. The good news is that a portfolio that is already tilted towards small cap and value can reduce its tracking error relative to the overall market by incorporating direct profitability and still keep its higher expected return. In this article, we discussed the potential impact of incorporating direct profitability into the IFA Index Portfolios. If you would like to learn more about how IFA builds global portfolios to capture the various dimensions of risk and return, please give us a call at 888-643-3133.

 


1Wiley, Ryan J. and Dave Twardowski, “Global Trading Advantages of Flexible Equity Portfolios”, Dimensional Research, April 2014.