Friday, January 13, 2017

"Can Investors Profit Using Academic Research?" (II)

I used that headline once before, for this in 2014:
Here's an oddball example. In 1977 Roger Ibbotson and Rex Sinquefield published a very famous paper, Stocks, Bonds, Bills, and Inflation: The Past (1926–1976) and the Future (1977–2000), which tackled the equity premium puzzle and was right but for the wrong reasons.

The paper's equity risk premium and therefore its estimated returns were too high but the paper got a lot of people into equities and for the 18 years 1982-2000 the major U.S. indices returned 18% per year in price appreciation. Ta da!

Mr Ibbotson sold his company to Morningstar, put the proceeds into his Zebra Asset Management and has a comfy chair at the Yale School of Management! Profit baby, profit....
After my flippant intro the post went on to get serious with some meta-analysis of investment research but I still chuckle when I think about how things worked out for Professor Ibbotson.

Anyhoo... here's the latest iteration of the question, this time from Morningstar, January 3:

Academic Research Doesn't Always Make Grade in Real World
Jeremy Glaser: For Morningstar, I'm Jeremy Glaser. There has been no shortage of academic contributions to the investing world finding ways for investors to maximize their returns. I'm here with Alex Bryan, he is our director of passive research here in North America, to look at some of these academic effects that maybe don't translate quite so well into the real world.
Alex, thanks for joining me.
Alex Bryan: Thank you for having me.

Glaser: Let's start with Fama-French Value Factor. This is one of the most famous. Can you talk just a little bit about what this research was and why investors haven't quite seen the full effects that Fama and French saw?
Bryan: Sure. So, Fama and French back in 1992 formally documented this effect where stocks that are trading at low valuations, specifically low price-to-book, have tended to give you higher returns than stocks that were trading at higher multiples. Now, practitioners have known this for quite some time. It was just discovered in the academic world about 25 years ago.
But anyway, the way that they constructed this factor is they ranked all U.S. stocks on the New York Stock Exchange, the NASDAQ, and the NYSE based on their price-to-book ratios and they sorted the stocks that traded at the lowest price-to-book into the value bucket. The highest 30% of stocks in the market went into the growth bucket. And then they tracked the performance of the cheap stocks, the value stocks relative to the growth stocks over the next year, and then they repeated that sorting process.

The way that they've constructed this portfolio is very different than the way that most actual investment strategists go about it, because they went by the number of stocks in the market. So, they are targeting the 30% of stocks that are trading at the lowest valuations, whereas most indexes, they are not looking for a certain number of stocks. They are looking to provide a certain coverage of the market. So most value and growth indexes try to cover about half of the market capitalization out there. The Fama-French factor gives you this bit of an imbalance where you're really going to a more extreme section of the market with the value bucket than you would with a traditional value index. So, it's very different from that way that most funds are actually being constructed.

Glaser: If it's being constructed differently, how big is that performance gap then?
Bryan: It's substantial. So, there's a few reasons for that. One, like I mentioned, the value factors that they constructed are a bit more extreme in its tilt than most actual index funds. But they also gave a lot--they gave equal weight to small-cap stocks the same weight that they have given to large-cap stocks. So, historically, the value effect has worked the best among the smallest stocks. And in the way that they constructed their portfolio, they give the same weighting to small-cap value stocks as they do to large-cap value stocks. So, if you look at the last 50 years of performance or so that academic factor returned about 3.7% annualized. Now, value and growth indexes don't go back that far. But from what we have for the Russell 1000 Value and Growth, that goes back to 1978, the 1000 Value Index outperformed its growth counterpart by about 1.1 percentage points annualized. It's quite a big difference from that 3.7% outperformance that you saw from the academic factors. So, there's a bit of slippage that you can see.

Glaser: Let's look at profitability, another academic or another factor that academics care for quite a bit. What have we seen in terms of the gap for that factor?
Bryan: So, this is newer one that was actually just discovered, so to speak, in the academic literature within the last decade. And it basically has documented this effect where stocks that are more profitable--either if you look at gross profitability, margins, or if you use a different measure of profitability--the more profitable stocks have tended to outperform their less profitable counterparts, particularly during market downturns. There's a couple of different explanations as to why that is. But this effect appears to be about as robust as the value effect that we've known about for quite some time. So, that's pretty interesting research and it jibes with the way that a lot of practitioners had thought about investing for quite some time....MORE
HT: Abnormal Returns