This is part two of our discussion. For best results we suggest you start with part one.
Long-time readers of Abnormal Returns are by now familiar with the work of Wesley Gray of Alpha Architect. He is the co-author of Quantitative Value: A Practitioner’s Guide to Automating Intelligent Investment and Eliminating Behavioral Errors. Last year we published both an excerpt and a Q&A related to the book. In addition posts by his firm dissecting recent academic research are linked to frequently on this site.
What you may not know is that Wes’ firm last week launched its first ETF, the ValueShares U.S. Quantitative Value ETF ($QVAL). You can read more about the fund in its Fact Sheet. With that in mind Wes was kind enough to answer some questions about the strategy behind the new fund and launching a new active equity ETF amidst recent market volatility. Our questions appear in bold, Wes’ answers follow.
The backtests for the quantitative value strategy show relatively high tracking error. Just how patient are investors going to need to be to follow this strategy?
Probably very patient, since we expect plenty of short-term underperformance, even though we are confident in the long run. Honestly, the biggest challenge we have is identifying the right investor. If an investor underperforms for a few years and asks, “hey, is value investing dead?” then we have not done our jobs at educating them on the strategy. If an investor is going to have success with our strategies they need to be disciplined and committed to a long-horizon holding period. If you have underperformed and therefore have stopped believing in value investing, there is a pretty good chance this is exactly the wrong time to abandon the strategy. By design, we believe our strategies work because much of the institutional capital can’t handle the tracking error embedded in our investment processes. They say, “Look, this strategy has underperformed recently, and so it is clearly is not working.” That’s a very natural reaction, and as I mentioned earlier, this is one reason why value investing keeps working! We also try to avoid other, similar institutional behavioral pitfalls: we don’t neutralize sectors, we don’t stop-loss or stop-gain positions, we don’t window dress, we don’t follow Sell-Side recommendations–we simply buy what our model believes to be the cheapest, highest quality stocks in the market. Period. Over 5-year horizons we believe the risk-adjusted performance can be favorable, but in a given year, anything can happen.
To really understand why a long horizon is necessary to exploit behavioral anomalies, your readers, who tend to be more on the “geeky” side, should read “The Limits of Arbitrage” by Shleifer and Vishny. I make all my advanced investment management students read this paper. This academic article highlights the tension between investment managers and investors in managed funds in a rigorous way. The thing I like about this research piece is that is summarizes the key reasons why asset markets are rarely perfectly efficient, but also highlights why an inefficient market doesn’t necessarily imply there are easy profit opportunities for smart investors.
Your research shows some benefit to more frequently rebalancing this strategy. However that would potentially detract from the tax benefit of holding positions in excess of a year. How do you think about the trade-off between taxes and re-balancing?
In general, most “anomaly” strategies have improved performance with higher frequency re-balancing. This makes sense intuitively, since you are making the best available statistical bets more frequently, instead of less frequently. At the margin this can make a material difference to returns. Of course, as you alluded to, with higher trading activity comes higher trading costs and higher taxes. These are quantifiable. And while transaction costs are a pain, taxes really suck. I am not sure about the direction of the markets, I am bullish on tax rates, meaning in general I think they are going up. Being bullish on tax rates also means I am especially keen on identifying ways to minimize tax burdens. Less frequent re-balancing is one way to minimize taxes, but as I mentioned, the trade-off is lower expected performance. So what can you do? How do you deal with this conundrum?
There are techniques such as tax-harvesting, strategic re-balancing, and one can always hire a banker to devise a structured product or swap, but all of these involve a lot of brain damage, and are often not very effective. Another potential solution to the tax/re-balance trade-off is the ETF vehicle. Despite that ETFs offer very effective solution to this tax issue we’re discussing, we believe that there are many investors who are not familiar with the tax features of ETFs. I invite all your readers to google “tax-efficiency of ETFs” and read the various research pieces on the tax benefits of the ETF structure. If I’m a family office and there is a vehicle that allows me to engage in an actively rebalanced strategy and simultaneously minimize my tax burdens…well, I’m all ears. Or at least I should be, if I’m a family office investor thinking strategically about taxes within the context of my portfolio strategies. It’s amazing what 50% marginal tax rates can do to incentives.
Your strategies are managed according to a multi-factor quantitative strategy. Anyone who uses a quantitative strategy knows it gets tweaked over time due to performance, liquidity and new insights. What is the plan to modify the model over time?
We are incredibly reluctant to change our models. Why not? Our approach is based on well-established approaches to investing and academic finance. This is not some complex day trading system based on mathematical equations or esoteric market patterns. To quote a famous data mining example, we are not using Bangladeshi butter production to predict the stock market. The mission of our model is to buy the cheapest, highest-quality value stocks, in other words, a systematic approach to a value-investment philosophy. So we are primarily value investors, and secondarily we are users of the most effective quantitative tools we can identify that are consistent with this approach. Our model’s mission is NOT to create amazing backtesting results.
While our core quantitative philosophy will not change in any material way (at least until my team votes me off the island), our approach to portfolio construction might adjust due to liquidity issues. For example, we might need to move from a pure equal-weight construction to a strategy that uses liquidity buckets or a market-weight construction.
The academic literature is awash in findings in favor of a momentum effect. Are you only focused on value, or do you look at momentum as well? Also, why not try to layer on a momentum screen to enhance the returns from the strategy outlined in your book?
We are about 12 months into a massive research and development effort to identify the best way to implement a momentum strategy at the stock-selection level. Jack Vogel, my partner and former research assistant at Drexel for 4 years, is steering that effort and finding some fascinating things. We’ll be publishing a lot of our findings via our blog and traditional academic channels. Stay tuned.
We’ve thought about, examined, and tested the idea of mixing momentum into value, but frankly, momentum doesn’t seem to reliably enhance the strategy outlined in our book, Quantitative Value. And even if momentum did enhance risk-adjusted performance in expectation, I’m not sure we’d want to jumble the two approaches because it can cause ex-post assessment issues for our clients. How do you decompose the exposures are you getting with a jumbled strategy? You don’t really know.
It’s funny, we see this issue sometimes when we sit on the other side of the investment table, and perform due diligence on various strategies and managers. I like reviewing clear, transparent, systematic strategies–they are much easier to review and understand. When a manager starts mixing and matching “anomalies” it is hard to tell what is driving returns. For example, if we have a value/momentum model and the system generates 20%, what drove the returns? Value? Momentum? Timing between value and momentum? Sure, one can use some statistical tools to try and tease that answer out, but it is much easier to ascertain when a strategy is “clean.” So, again, our thought is we should offer these strategies independently, on a “pure-play” basis, and it is up to the investor as to how he wants to use them or combine them. We try and follow Einstein’s sentiment when it comes to strategy assessment and implementation, “Keep things as simple as possible, but no simpler.”
Charley Ellis has been quite outspoken about the high costs of active management. You have also talked about the hidden high costs of so-called smart beta funds. How should we think about the costs of active strategies?
Charley Ellis works with Wealthfront, which promotes a mean-variance-analysis approach to tactical asset allocation. Research shows, pretty convincingly, that a simple 1/n equal-weight strategy works better. So isn’t mean-variance-analysis and the whole efficient frontier bit, which Wealthfront implements, a form of active management? I think arguably it is. Does Dr. Ellis have some magic insight on the appropriate covariance estimates across asset classes? I think it’s hard to argue that he does, which is kind of the point of all this research I’m referring to. So at some level, Ellis and Co. are injecting what they would tell you is their expertise and wisdom into an investment process. How is this not active? In other words, some self-proclaimed “passive” managers may be more active than they would have you believe. So to be clear up front, our view is that Charley Ellis is an active manager. But hey, everyone is trying to sell something, and people trust Jack Bogle (as they should), and so the “passive” management philosophy sells well, even if some variations on it are perhaps not as pure as Mr. Bogle’s vision. We should remember that everyone wants to earn a paycheck.
Now all that said, the arguments for passive management are persuasive because active management usually not only costs too much but also often does not perform–a double whammy. The pain gets worse when you roll in the hidden costs behind the scenes–soft dollars, distribution costs, tax-drag, cash-drag, and so forth. It’s time for another option in the market and it’s the reason why we get up in the morning. There’s a better way. We don’t claim to be the cheapest investment solution on the planet, but do strive to be the best value-proposition in the market. In other words, when one weighs the all-in costs of what we provide against the effectiveness of our approach and all-in benefits, we think our investment offerings are compelling and can compete very favorably with passive solutions on a cost/benefits basis.
Thanks to Wesley Gray of Alpha Architect for his time in answering our questions.