Steve Reading wrote: ↑Fri Sep 04, 2020 11:59 pm
I'll just ask the Prof. himself:
I should clarify, in case there is any confusion: I worked as a professor for a few years after my PhD, but left that position several years ago. So I'm no longer a professor, except to my kids who sometimes have to listen to my lectures.
Steve Reading wrote: ↑Fri Sep 04, 2020 11:59 pm
1) Ok long story short, I don't see a good, theoretical reason for why there should be a small cap premium. I get the value, quality and investment factors. I agree that, all else equal, the higher the B/P, the higher the profitability as a function of assets, and the less reinvestment a company needs to do, then technically the higher the returns. That's what the Miller and Modigliani equation tells us.
But size, I don't see why it should have an effect. I think smaller companies tend to be more distressed, with lower credit ratings. Smaller companies tend to load more on beta. Smaller companies tend to be more illiquid. But given you control for distress via fundamental factors, control for the extra beta, and control for the illiquidity premia, is there a good, theoretical or economic reason why the market cap itself should affect the cost of capital of a firm?
For reference, a couple of my earlier posts:
viewtopic.php?p=5475359#p5475359
viewtopic.php?p=5475511#p5475511
2) If anything, if you believe the market is mean-variance efficient, one should expect a large cap premium as the large caps would need to offer some incentive for the added concentration in the average investor portfolio. Since Fama believes the market is efficient, then I assume he's probably even further perplexed by a small cap premium. Am I correct in my thinking here?
Thanks man, your insight is always highly appreciated.
The only factor that emerges easily from basic economics is the market correlation (beta) factor. The rest are all complicated, and whether they are better explained by risk or mispricing is debatable. To me, the mispricing story is more believable.
But first, why basic economics tells us that market correlation should work: Assets that are highly correlated to the market should be cheap (have higher expected returns) because they are paying out in good times, when everything else is doing well. So they are not as valuable to us as the assets that pay out when times are tough. This explains why stocks are cheaper (have higher expected returns) than bonds, and why call options on a market index have higher expected returns than put options on a market index.
Once market beta is taken into account, it's not obvious why size, book to market, profitability, quality, momentum, and anything else should matter any more than the name of the company or whether the CEO plays golf. Note that it's critical to control for market beta when evaluating a story. Often the reasons given for why value companies deserves a higher expected return is that these companies are under duress and if the economy tanks, then they will do badly. But that story is not controlling for market beta. Controlling for market beta means there is a premium over and above what is merited by how the company will do when the economy tanks. We would have to explain why even when they pay out in tough times, people don't seem to want to hold it and demand a premium for holding it. That is a much harder story to tell.
My take is that factor premiums are due to mispricing. We know there are bubbles--from tulips to bitcoin to tech stocks. People get too excited about certain things at certain times. Run-ups in prices draw more people in driven not by fundamentals but by recent price increases. Dramatic bubbles might be infrequent. But some chronic low-grade overexcitement might be more prevalent, and will result in some stocks being chronically overpriced and some chronically underpriced. If so, we might expect that big, popular, growing firms will be overpriced and small, unpopular, embattled firms will be underpriced.
There is another thing that I find really telling--factorheads are almost always tilted towards small and value, and rarely towards large and growth. If this was all really about risk, then there is no reason why factorheads wouldn't be evenly distributed across the factors. So why aren't they? The fact that they are all tilted in one direction suggests that they seem to think it's mispricing, not risk. Who digs deeply into Fama-French, believes that the factor premiums exist, and then decides that they should tilt towards large growth and not small value? [Edit to clarify: I'm not saying I would expect the same person to tilt towards both small-value and large-growth. Different people should tilt to different things. The question is why are factorheads as a group tilted almost exclusively towards small-value. Where are the risk averse factorheads who read this literature, believe it to be empirically sound, and then decide to tilt towards large-growth?]