A recent paper by Cam Harvey and his co-authors on the statistical validity of many so-called finance anomalies has attracted a great deal of attention in quantitative finance crowd. It seemed worthwhile to discuss the implications of the paper with one of our favorite quants, Wes Gray of Alpha Architect and manager of the ValueShares US Quantitative Value ETF ($QVAL) and the ValueShares International Quantitative Value ETF ($IVAL). Our questions appear in bold and Wes’ answers follow. You should also check out part one of our Q&A.
AR: One of the big issues with momentum is high turnover and for taxable investors this can reduce returns on an after-tax basis. Are individual investors limited to putting their momentum-based strategies in non-taxable accounts?
WG: Listen, I always like to say that taxes are more important than alpha. The largest fee for any US-based investor is the 23.8 percent (20 percent Long-Term Gains + 3.8% Healthcare Tax) performance fee assessed on gains every year (if you are a short-term trader or a New York or a California resident it can be 50 percent or higher!) If you aren’t minimizing Uncle Sam’s performance fee, you aren’t going to be successful. A great example of the ultimate tax-minimizer is Warren Buffett. The guy has some of the largest tax deferrals on the planet and is avoiding the payment of these deferred liabilities by giving the money away to the Gates Foundation. His lifetime tax-rate will certainly set a new record. And good on him–he figured it out a long time ago that tax-deferral is more important than alpha.
But back to momentum, specifically. I think Cliff Asness and the crew over at AQR do a great job highlighting that momentum investing is equally tax-efficient as value, primarily because momentum systems let winners ride and cut losers short. But, as the authors highlight, there is certainly some tax-leakage because of the intense turnover of momentum strategies. Unfortunately, the turnover is required for momentum to work effectively. Our own research on the topic shows very clearly that momentum portfolios should be rebalanced frequently. In general, the more frequently you rebalance, the higher the returns, in expectation, you can achieve. Of course there comes a point of diminishing returns where transaction and other frictional costs will start to overwhelm the benefits of increasing the frequency of rebalance, but in general momentum requires a high rebalance rate, and also concentrated positions, to get the best exposure to the anomaly. We think with low transaction costs and a portfolio of 30 to 40 liquid stocks, you can realistically rebalance monthly, which leads to the best momentum returns, in expectation. This is one of the big drawbacks to many of the momentum funds out there, i.e., they only rebalance every 4-6 months, and hold sometimes hundreds of stocks. This is clearly not the best way to access the anomaly.
On the tax question, non-taxable accounts are obviously ideal for momentum strategies, since capital gains taxes aren’t applicable. However, for taxable accounts, the only sensible vehicle for investing in momentum is an ETF structure, or perhaps a private placement insurance wrapper (but this is complex). ETFs structures can minimize tax burdens via the creation/redemption process, which is unique to the investment vehicle. We think for taxable accounts, a momentum-based mutual fund doesn’t make a lot of sense. And if the momentum mutual fund contains several hundred securities and rebalances infrequently, it makes even less sense. Now you’re simply paying active fees for a closet-index product that is tax-efficient. Yuck.
AR: Changing gears, one of the more pleasant developments of late has been the flowering of relatively straightforward tactical asset models. Meb Faber’s “Ivy Portfolio” approach, Gary Antonacci’s “Dual Momentum” and your “Robust Asset Allocation.” For instance, Meb’s paper has been downloaded a great deal, but do you think many investors use these models in real-time?
WG: The emergence of transparent, simple, and effective asset allocation tools is a development that is good for investors and bad for those elements in the financial services industry that are slow to adapt. The emergence of the so-called “DIY Financial Advisor” is here to stay. But the DIY Financial Advisor movement is a small element embedded in a larger theme: the disintermediation of the entire financial services industry. We see disintermediation everywhere: robo-advisors, ETFs, crowd-sourced venture capital, crowd-sourced real-estate, online factoring, peer-to-peer lending, and so forth. We also see a wave of free and engaging research being published on the blogosphere every day. Your daily blog curation puts education and information at the fingertips at all market participants–something that was unavailable 10 years ago.
An increasingly educated investor base matters. Some investors are starting to figure out that there’s no “magic,” or “secret sauce,” to whiz-bang mean-variance methods or other complicated asset allocation strategies recommended by many high price and/or conflicted investment advisors. I think Meb’s new book, Global Asset Allocation, does a great job of highlighting how similar global diversified portfolios perform. It’s a sort of a democratization of asset allocation that is occurring, and that’s a good thing for after-fee returns and for investors that put in the effort to get informed.
With respect to Meb’s paper, Gary’s book, Dual Momentum, and our system, I assume you are referring to the work on technical analysis to time markets. The examination of so-called “technical analysis” rules to time markets has been around well before Meb and I were even born (although, Gary may have actually been around J). What made Meb and Gary’s work so popular is the ease with which they were able to communicate these rules and show how they reduced large drawdowns. We have our own opinions on how to use these rules. We essentially decided that we liked both Meb and Gary’s ideas. My team and I have been doing intense research on market-timing for 5+ years. You name it, we’ve probably reverse engineered it and tried to improve it. With respect to technical rules, we’ve researched all of them, and we’ve slowly come to the conclusion that simple trend-following and/or time-series momentum rules are effective ways to risk-manage a portfolio, at least historically. Hat-tip to Meb and Gary for showing us the way!
Now, back to your broader point on whether or not these rules actually are used and/or help investors. One thing I’ve noticed in financial markets is that just because a rule “works,” doesn’t mean investors actually follow it and/or make effective use of the rule. Value and momentum stock selection techniques are a perfect example. With respect to technical analysis, there is actually a funny paper by Tom Etheber, Andreas Hackethal, and Steffen Meyer, who study account-level data on individual investors in Germany. After examining 2.7 million transactions across 35,000 investors, they find that a substantial portion of individual investors do follow technical analysis rules. However, they also find that these investors lose, on average, to basic buy-and-hold investors because these “chartist” investors trade too much and are under diversified. I can’t prove it, but I will venture to guess that these “chartist” didn’t actually follow the rules. They probably followed the model’s advice when they liked it, but disregarded the model’s advice when their gut told them otherwise.
Among certain sophisticated groups, the use of technical rules is well established. On one hand, when I speak to my friends who manage endowments, pensions or large family offices, I get blank stares when it comes to technical market timing rules. They think I’m crazy. However, when I speak to my PhD buddies who work at large asset management firms, they are well aware of the benefits of technical trading rules, but few actually implement them in practice.
Which begs the question: Why would someone acknowledge the benefit of a potential trading rule, but not follow it?
It turns out that just because you are aware of a reasonable investment approach, doesn’t mean you will put it in to practice. Incentives matter and work in mysterious ways. Strategies that work aren’t comfortable and often force an investor to do things that put them in a position to lose their job. And as Shleifer and Vishny show in their 1997 Journal of Finance paper, “The Limits of Arbitrage,” asset managers are afraid of strategies that can deviate from standard benchmarks because their investors will pull their capital following poor short-term relative performance. Taking Meb’s simple moving average rule as an example, if you are sitting on cash because a technical rule takes you out of the market, while the market proceeds to rip 10 percent, you’re going to get fired by your investors. It’s that simple. So what you end up with is a marketplace full of buy-and-hold closet-indexers posing as active managers “adding value.” Think “expensive Vanguard.” Few asset managers can be truly active, because their fear of poor short-term relative performance, and the associated bankruptcy of the firm when all assets leave the building, is simply too high. Unfortunately, this is the real-world marketplace we know and love. Fortunately, I’m too naïve and idealistic to care about my job security. A blessing and a curse in the asset management business.