20 for Twenty from AQR
AQR, a Connecticut investment management firm famous for its factor investing offering with 195bn AUM (Assets Under Management) has published a selection of 20 papers that they view as fundamental to investing. The book can be downloaded for free here.
I review below two of the articles that I read today. This is a thought-provoking book.
The great divide: EMH vs Behavioral
Nobel Prize-winning Economist G Fama. has been credited with the next evolution of the Efficient Market Hypothesis (EMH) by introducing more factors than the single market factor introduced by Sharpe in the '60s, to 'explain' the alpha.
Factors reported by Fama were high vs low price/book value, small vs large capitalization stocks, and later momentum.
The alpha associated with each factor portfolio is then called as a 'factor risk premium' as this is the name of excess return generated by taking a certain risk. The approach is purely statistical is not accompanied by an evaluation of the actual risk associated with taking these positions.
Cliff Asness makes the point that EMH cannot be asserted without some security pricing model. For instance, CAPM which states that the expected return on a security is proportional to the beta of a security is false.
He then makes the cogent point that to be an EMH supporter, one needs to believe that EMH with CAPM is false because CAPM is too simplistic and this model does not take into account a risk whose rational expectation would explain prices.
In contrast, someone who asserts that risk premia are driven by irrational psychological factors belongs to the behavioral school.
G Fama follows the EMH in that he 'explains' that value stocks are riskier because of the observation a positive factor premium for this factor. The common opinion amongst value investors since B Graham is that even safer high-quality stocks can be priced lower because the business seems boring to the investor.
MPT (Modern Portfolio Theory) does lead to the definition of risk factor premia, and these risk premia price are fixed by the market. It seems very reasonable to harvest these factors as Ben Graham set himself up to do. I do not understand C Asness deference to the EMH as harvesting these premia implies that one does not believe that they are the result of rational expectations.
100% Equity vs 60-40 Portfolio
In this article, Cliff Asness criticizes the proposal to invest 100% in equity. First, he explains that large allocation in equity has only become the norm in the '60s and that portfolios were invested in bonds before.
He uses US history since 1930 to make the point that a 60-40 portfolio (60% equity, 40% bond) has lower return but a higher Sharpe ratio than an equity portfolio (in fact, the 60-40 is the optimal mix given US past history). He then advises for people to invest in a 60-40 portfolio with monthly rebalancing and lever it if they wish to enhance returns.
His argument is about bonds being there for a longer time does not take into account that for most of the period described. Money was on a gold standard, and that bonds, therefore, were guaranteed a positive real return. As strong inflation in the 20th century ended this, it took much time for investors to review their policy concerning assets vulnerable to the new risk of inflation.
I find the idea of using historical returns and volatility from past bond moves in the US not very compelling as yield can never go far below 0, whereas they are unlimited with fiat money.
A bond investor is picking pennies in front of a steamroller in a fiat money system, he needs to trust the government not to increase the money supply to a level that would cause inflation. He has fixed upside and unlimited downside.
The statement about the 60-40 portfolio having a higher return is mathematically correct for the US. But a bond is an asset whose risk is not properly captured by volatility: when Asness suggests a levered 60-40 strategy levered at x1.5 rebalanced monthly based on 80 years of backtesting, it should be noted that this strategy would have bankrupted the investor in Austria, Germany, France, Japan during the same period. Bond investors were all wiped out but some equity investors survived.
Bonds pay less than equity because they look safer, but they are not safer over the long run.
The macro case for buying low yielding G3 government bonds is much weaker now. You bear the risk of inflation in exchange while there is no prospect of return at all (we call this the return free risk).
There was a strong macro case in 1992 to buy bonds when real rates were at a record high as they had both the potential to complement other asset strategy return while bringing some predictability to the return.
In the current interest rate regime, it seems much safer to invest 60% in stock and 40% in less liquid income-producing real estate, which is less correlated to stocks than REITS and is also an inflation hedge, but that's a question of how you define your investment universe.
Value and Momentum
In this article, Asness et al. study the excess return of value and cross-sectional momentum strategies across several markets: US, Japan, UK, Germany as well as momentum across other asset classes (bonds, futures, and currency).
Setup involves measuring value using price/book value and momentum using 12month to 1month return. There are ways to improve these measures, but the goal of the paper is to paint broad strokes and understand the correlation of these factors across regions and between themselves.
The findings are that
- value and cross-sectional momentum deliver excess return compared to their reference equity market
- both are anti-correlated, so that a 50/50 strategy has much lower risk and better performance
- momentum is positively correlated to measures of the availability of leverage (TED spread, repo, swap-T bill spread) whereas value is negatively correlated to it.
- correlation of the two strategies to economic indicators such as GDP growth is non-existent.
Book by Asness on Amazon: Expected Returns: An Investor's Guide to Harvesting Market Rewards