A Wealth Creation Journal

Category: Academic

Small cap value works, especially when controlling for quality

Recently I read an excellent Cliff Asness (AQR founder) paper Size Matters, if you Control your Junk.  It is very comprehensive research on the size effect. Highly recommended.

First popularized by Fama & French, the size effect says small companies outperform large by a few percentage points per annum.

After Fama & French, the size effect got a lot of criticism from new empirical research however for not being statistically significant or being the result of data mining. The main pain points that make the size effect appear less statistically significant – than for example value or momentum – are basically:

  • small caps do not outperform consistently
    • over time (in the ’80-’00 period they underperformed)
    • globally
    • over certain stock characteristics (value vs growth, positive vs negative momentum)
  • small caps outperformance seem to be concentrated in
    • January
    • illiquid stocks
    • the most extreme size deciles (smallest companies and largest companies): the size effect does not exhibit “monotonicity”. In other words, the best performing decile might be the smallest companies’ decile, but the remaining deciles do not have monotonically decreasing returns toward the last decile

The authors then surgically show that all these criticisms are trumped by the size effect if you control for “quality” (defined by high profit margin, low leverage, high sales growth, good quality of earnings). In other words, small caps have not significantly outperformed globally, over time, over certain stock characteristics, in Feb to Dec, etc. because stocks in the smaller company universe tend to be more “junky”.

If you buy small companies that are on average as high quality as large companies, you actually get size outperformance that is consistent over all the above metrics (time, geographically, value and growth, liquidity, all year) and  the effect becomes monotonically decreasing with the size decile.

Now that we can rest assured the size effect is significant and robust over all these variables when we make an apples-to-apples comparison with large stocks in terms of quality, the cherry on top is that the value effect much larger in small caps. This great paper shows how much.


Who are the Value and Growth investors?

This paper “Who are the Value and Growth investors” is a great read because it investigates who are the suspects that create the value effect.

[..] we relate the value tilt to household characteristics. Value investors are substantially older, are more likely to be female, have higher financial and real estate wealth, and have lower leverage, income risk, and human capital than the average growth investor. By contrast, men, entrepreneurs, and educated investors are more likely to invest in growth stocks. These baseline patterns are evident in both stock and mutual fund holdings. The explanatory power of socioeconomic characteristics is highest for households that invest directly in at least five companies, a wealthy subgroup that owns the bulk of aggregate equity and may therefore have the greatest influence on prices.

Note from the statistics table that three large explanatory variables are sex, education level “human capital” and immigrant. 

Women tend to be value investors, while men and immigrants buy glamour

The academic top growth or “glamour” stock decile historically shows a histogram of returns with more fat tails (see books by Haugen previously linked on this blog). Thus, glamour stocks look more like lottery tickets. I presume this contributes to why women tend to shun glamour, as they have been shown to be more risk averse. This speculation is corroborated by the immigration dummy: immigrants buy more glamour: are they more risk seeking than locals?

Higher education correlates with glamour investing

Maybe we are close to the peak usage of the word moatThe word sounds smart, but it means a million different things to different people. That makes it hard to scientifically test. I am of course not arguing that investing in companies with moats is stupid, as the grandfather of the word Buffett has done phenomenally well. Rather, I am questioning the ability of the majority of moat seekers to outperform, especially when everyone is looking for moats [Bloomberg: Moat is the latest Silicon Valley jargon].

The study shows that smarter people tend to buy high multiple stocks. High multiple stocks are related to companies with explosive/high growth rates and/or high (incremental) returns on capital. Although it was shown empirically that some “quality” factors (such as high and stable historical return on capital) outperform without the use of valuation metrics (Remark 1), the magnitude is marginal versus the value effect.

Hence, for the average smart person, seeking great companies with moats might be a dangerous game when the crowd is paying high multiples.

Indeed, what I think is going on here is an example of Kahneman’s insider view problem. The more people are experts in a field, e.g. investing or technology, the more they are drawn to go against the base rates in their field of expertise. It is ironic that knowledgeable investors might well be aware of the value effect but still find high moat investing to be more intellectually challenging.

A practical advice for investment professionals and myself that seek to do quality investing could be that one should very slowly, and with a lot of self-contempt, evolve toward high moat investing. This might well be what Warren Buffett did over his long life.

A primary focus on staying within one’s circle of competence, a secondary focus on expanding that circle slowly.


Value investing is for old folks

The not often discussed fact that value investing bears lower duration risk (to borrow a term from bond investors) than growth stocks is cited to explain why older people – who will soon draw money for their pension – own more value stocks.

value loading age.PNG

I think this is a great explanation.

Remark 1 

Buying high & stable return on capital businesses solely by looking into the rear-view mirror outperforms the market. This means that part of high moat investing can be tested and works. I call this the quantitative moat strategy. There are three caveats though:

  • Quantitative moat strategy return component: the historical out-performance of this naive rear-view mirror strategy is marginal versus e.g. the value effect
  • Qualitative moat strategy return component: most returns from buying high moat companies come from outwitting the market in the pool of easily identifiable high ROIC companies. In other words, by having a variant view on magnitude and/or longevity of anomalous ROIC before it reverts back to the mean (see e.g. the book Accounting for Value). This is where moat investing becomes more art than science.
  • return detractor: the out-performance of the moat strategy can be washed away, and then some, if one is buying high multiple stocks

Remark 2

I am not sure of the paper’s last conclusion. Based on the observation that employees in more cyclical industries buy more growth stocks, the authors assume owning value is more correlated to key macroeconomic risk factors (as the die-hard efficient market believers do).

The central tenet of the rational approach is that the value premium is compensation for forms of systematic risk.


To the best of our knowledge, our paper is the first to provide direct evidence of hedging demand of any kind in the risky portfolios of households. It also lends support to the link between the value premium and income risk, which has been the subject of a vast asset pricing literature.17 In his Presidential Address to the American Finance Association, Cochrane (2011) develops the following interpretation of the value factor: “If a mass of investors has jobs or businesses that will be hurt especially hard by a recession, they avoid stocks that fall more than average in a recession.” Our results confirm Cochrane’s prediction.

I think the paper by Lakonishok & Shleifer (1994)  shows compelling contrary empirical evidence to this cited theory. Lakonishok looks at value versus growth relative performance in the worst economic times (i.e. recessions) to conclude that value is safer in an environment where safety is at a premium.

If we look into Cochrane’s (2011) paper that was cited, we find – beyond an amazing amount of theory – no empirical data for value versus growth in recessions. The only data that is cited in this paper is across asset classes, not within equities. Cochrane cites Fama, French (1996):

Consider Fama and French’s (1996) story for value. The average investor is worried that value stocks will fall at the same time his or her human capital falls. But then some investors (“steelworkers”) will be more worried than average, and should short value despite the premium; others (“tech nerds”) will have human capital correlated with growth stocks and buy lots of value, effectively selling insurance. A two-factor model implies a three-fund theorem, and a three-dimensional multifactor efficient frontier as shown in Figure 17. It is not easy for an investor to figure out how much of three funds to hold.

When we then look at the F&F (1996) paper, we find that it addresses the Lakonishok (1994) “LSV” paper without any data (see page 24 to 26 “The distress premium is irrational”).

Fama responds that the argument that value stocks outperform growth stocks in general recessions is not sufficient to disprove that value is riskier:


In my opinion the burden of hard evidence is on F&F as they contend that value is necessarily more risky than growth. Rather, their answer to LSV as to why value is riskier than growth rests on the non-quantified concept of “relative investor distress”. In other words, the theory that investors care more about downside risk at exactly the time when a diversified value basket under-performs.

Despite the finding that value outperforms in a recession when everyone most likely cares more about downside risk (except perhaps efficient market intelligentsia), F&F respond that this data is not sufficient, and that there might be a multitude of offsetting times that employees/investors in various industries care more about downside risk when value under-performs. It seems that the deus ex machina for F&F is that we just can’t measure subjective distress as this argument is not elaborated on. No empirical data was shown formulating their answer.

I find it quite ironic that the the individual industry distress argument is coming from proponents of CAPM, the theory that stipulates that diversifiable risk should earn no premium. If an investor that works in the cyclical steel industry shuns value stocks in steel companies, wouldn’t it suffice to own a diversified basket of value stocks of different industries to diversify employment risk?

All the best,



Illiquidity as a return driver

In my post on Horizon Kinetics’ Bregman, I described several factors that deter index funds. Today I look at a factor that deters not only index funds, but many institutional and speculators in general.

An interesting academic paper on this factor is the famous Ibbotson paper Liquidity as an investment style.

Key takeaways

  1. Liquidity is a stronger factor for returns than size (see below by comparing columns).
  2. Another interesting finding is that the value effect is stronger in illiquid companies, so strong that it overshadows the size effect (see below and compare most liquid smallest companies with most liquid largest companies). In other words, the market is less efficient in the universe of illiquid stocks
    • Fama would not agree with me as he argues value stocks are more risky in general and that risk is more nuanced than simple volatility (for example, we do not know when the prospensity for risk is lowest), but this is debunked in the famous Lakonishok et al. ’94 paper Contrarian investment, Extrapolation and Risk and an updated 2009 study. The papers show that not only is value less volatile than glamour in general, it outperforms glamour in the bad states of the world, i.e. recessions, where risk aversion is most probably higher due to job losses and the powerful animal spirit fear.

TC comments

  1. Many institutions are simply too large and cannot support research into these illiquid companies from a cost-benefit perspective, as they would drive up the price to build an economically meaningful position
  2. Total friction costs for an investor are driven by liquidity and total trading volume. Ideally, an investor in illiquid stocks should
    1. be small,
    2. have a low turnover portfolio

The other benefits of low turnover tie into one of my favorite passages in the book I am currently reading, Capital Returns by Edward Chancellor. Low turnover* allows the investor to minimize the amount of decision making. Having to make a lot of decisions disproportionately increases errors by increasing time pressure. The pressure of highly frequent decision making is a disincentive to long term thinking about hard questions that matter. Guy Spier talks about the benefits of reducing the amount of decision making too in his book The education of a value investor.

As such, illiquidity forces the investor to think twice before investing in a stock.

The right spirit to invest in illiquid stocks is to worry about being right.

In my investing experience, illiquidity has been on balance a positive for me. Having said that, most of my investing has been in the time frame of the 2009 – 2017 bull market.

*Note that I use low turnover and not long term investing. In my view, low turnover is not a necessary condition to get the label “long term investor” and is not to be worshipped as an end in itself. Long term is merely a succession of short terms, and rational fundamental-driven investors should be willing to act short term if Mr. Market provides the opportunity to close a position at a margin of safety that is inadequate.









© 2022 Pembridgecap

Theme by Anders NorenUp ↑