Fund Managers Have Some Valid Reasons To Avoid Momentum

Includes: EEM, QQQ, SPY, TLT
by: Michael Harris


Momentum strategies require a transition from passive to active management.

With momentum strategies there is possible loss of investment discipline.

Momentum suffers from data-snooping bias.

Momentum, relative, absolute or dual, is essentially a timing strategy that is used for the purpose of achieving better risk-adjusted returns in the longer-term as compared to passive allocation strategies or even buying and holding.

Below is a backtest of a dual momentum strategy with two assets, S&P 500 Total Return and cash, and a 12-month timing period, since 1989.

Created with Portfolio Visualizer Click to enlarge

It is clear that risk-adjusted returns of this dual momentum strategy are superior when compared to those of an equal weight portfolio (50% in S&P 500 Total Return and 50% in cash) or to those of a passive investment in S&P 500. Specifically, the annualized return of the dual momentum strategy (blue line) outperforms a passive investment in S&P 500 total return (yellow line) by 160 basis points and drawdown is lower by a factor of 3.

The above results illustrate the potential of timing models, especially when combined with relative momentum. However, this is a trivial example and most investors prefer a certain degree of diversification. In addition, the improved risk-adjusted performance of the above trivial strategy can be attributed to trend-following, which can be achieved by a wide variety of simpler strategies, for example moving average crossovers.

Below I list three reasons why investors neglect momentum:

Reason #1: Momentum strategies require a transition from passive to active management

This transition is not trivial and actually requires that a fund manager is also a trader. Going from passive allocation to timing models requires different systems and operating structure. In an era of constant bashing of active management, some fund managers decide that the transition is risky for their business.

Reason #2: With momentum strategies there is possible loss of investment discipline

Timing models require trading discipline. The most difficult task of trend-followers is adhering to strategy rules. This is in contrast to passive allocation schemes that offer inherent discipline because they only require rebalancing. Loss of discipline can cause friction in a fund management firm due to different opinions of managers about whether or not to adhere to strategy rules and signals.

Those of us who have actually used timing strategies can understand the impact of loss of discipline and the friction in can create. In reality, using timing strategies without a mechanism to enforce discipline slowly leads to random decisions and losses. Most fund managers know the risks involved but researchers do not have actual experience with the dangers involved in transitioning from passive to active management.

Managing the savings of people is a job that requires high level of professionalism and respect for the customer. Those who wonder why momentum is neglected should try to answer the following question: If you were given today $1B to manage, would you choose a passive allocation scheme or a timing method? Most fund managers choose the passive allocation scheme because they understand the risks of trading timing models. This decision is not because they do not understand momentum. Actually, momentum is a trivial timing strategy.

Reason #3: Momentum suffers from data-snooping bias

This is a very serious objection against using momentum and also other technical strategies despite the convincing backtests offered by some researchers even if they include robustness and out-of-sample tests. Note that if a strategy is optimized, robustness tests are unlikely to fail. Also, note that out-of-sample tests make sense only in the case of a single independent hypothesis. As soon as one mixes and matches assets to produce a desired result based on backtested performance on already used data, out-of-sample tests lose their significance. It is known that if one tries many strategies on historical data, a few of them may outperform in out-of-sample testing by luck alone. Let us look at some examples of dual momentum strategies below.

The first strategy is for SPDR S&P 500 Trust ETF (NYSEARCA:SPY) and the iShares 20+ Year Treasury Bond (NYSEARCA:TLT) and with 12 months timing period. Below are the backtest results:

Created with Portfolio Visualizer Click to enlarge

It may be seen that the dual momentum strategy (blue line) underperforms the equal weight portfolio in SPY and TLT. The annualized return of dual momentum is 300 basis points lower and maximum drawdown is higher by nearly 9%.

Next, EEM is added in an effort to provide exposure to emerging markets. However, as soon that is done, data-snooping is introduced. Below are the results:

Created with Portfolio Vizualizer Click to enlarge

It may be seen that although the dual momentum strategy outperforms equal weight, there is a correction in equity (blue line) in 2015. The return for 2015 was -8.5%. However, this is not the main problem with this attempt to improve the asset mix in an effort to obtain superior performance. Actually, the outperformance was possible due to conditions in emerging markets (NYSEARCA:EEM) that may never occur again, or better said, the risks of never occurring again are high.

Specifically, in 2005 EEM was up more than 55% and in 2009 the return was close to 72%. However, last year emerging markets crashed. Therefore, a fund manager employing this strategy in 2015 paid the price of data-snooping bias.

But why EEM and not QQQ? Below is the backtest for SPY, QQQ and TLT dual momentum with a 12-month timing period:

Created with Portfolio Visualizer Click to enlarge

In this case, the equal weight portfolio generated 360 more basis points of annualized return with just 7% more drawdown and it outperformed dual momentum.

One may find many backtests where dual momentum works well and many where it does not. This is actually the point, and the risk involved. If your research shows a specific asset mix where dual momentum worked well, I do not care about any out-of-sample and robustness tests unless you can prove that there was no data-snooping involved. Since providing such proof is highly unlikely, I can understand why most fund managers neglect momentum. Besides, momentum becomes a crowded trade when its signals align with strong uptrends and are influenced by passive investment decisions.

In the era of Big Data and machine learning, it is difficult to know which strategy represents a unique, independent hypothesis, or it is the result of data-snooping and p-hacking. Thus, many fund managers hesitate in adopting popular strategies that are based on trivial rules and fully disclosed in books, articles and blogs. They may be wrong but I do not blame them for their decision in adhering to passive allocation. Momentum is part of technical analysis and many traders know that this type of analysis has contributed to a massive wealth-redistribution in recent history.

Note: Charts created with Portfolio Visualizer.

Original article

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.