Momentum works - for all sizes and all geographies

 

Asness (2014) quote on comparing momentum and value within large cap stocks: “Putting it starkly: in-sample, out-of-sample, calculated in Greenwich Connecticut, Chicago, Boston, Palo Alto, Santa Monica, Austin, or in the library with a candlestick, wherever or however you want to look, along any dimension, those who make the claim that momentum fails for large caps, yet being supporters of value investing, are not simply mistaken, they have it backward.”  

Charts and quotes taken from "Momentum Works Everywhere" from EAM Investors 

The folks form EAM performed one of the greatest comparisons of major factors I have ever seen. They look at momentum, value, growth, quality, and the market and find that momentum is the clear winner across time, geography, and size. It will perform better than any other major factor. Of course, momentum will not do well in every sub-period and may not do well in all markets, but the evidence is very strong as measured by the t-stats and by excess return. Follow momentum even if it is a very simple strategy and simple look-back period. You will do well with this as an adjunct to any strategy. 

The appraisal ratio - another way of looking at risk

The appraisal ratio is the alpha of an investment strategy divided by the unsystematic risk of the strategy relative to the benchmark. Take the investment return and run a linear regression against a benchmark like SPX. The regression will generate a beta, an alpha, and a residual error which is the unsystematic risk not associated with alpha or beta. The ratio then just looks at how much alpha is generated versus the unsystematic risk. A higher appraisal rate means that the manager can generate more return versus risk unassociated with beta. It is the manger's skill relative to the risk taken adjusted for beta.

The appraisal ratio can be compared with the information ratio which is the strategy return minus a benchmark (excess return) versus the standard deviation of the tracking error.  

Both measures attempt to measure sill versus a benchmark excess or tracking risk measure.

What makes trading "bubbles " so difficult?

A recurring theme has been looking at bubbles in financial markets, or to be more precise, significant moves in price that are greater or less than normal. The simple view is that if there is a significant move over a given period, it will not continue. Hence, fade the large move under the view that prices will mean-revert. 

However, when is a move too big? Look at the NVDA for the stock market. An extreme was hit well before the beginning of 2024. Look at cocoa futures prices over the last 3-6 months. There would have been good reason to sell or sell short these two high profile markets. 

The paper  "Bubbles for Fama"  shows that even if we place a high return number as a threshold there are still strong reasons for prices to go higher. It is not a given that there will be mean reversion. The probability of a revision is high and there are factors that will increase the likelihood of a decline but calling a bubble top is not easy and should be done with caution.

See:

More on Bubbles - trends last longer than expected

Bubble Time - Perhaps if we look at history

An interesting chart on the history of rolling returns asks the question - Are return extremes associated with the business cycle or bubbles?  At the end or just after a recession, there usually is a nice jump in returns. This is a reversal of the past declines. This return behavior makes sense. Prices are driven down in a recession and risk averse investors need a higher return. 

The more interesting question is the driver behind high returns not associated with recessions in the business cycle. What causes these infrequent extreme?  There is the suggestion of a bubble. This term is thrown around, yet it may not have any substantive meaning.  Can high market returns just mean-revert? Returns post a 20% increase are likely to be lower, but that does not mean losses or a bubble.