By Samuel Lee
A version of this article was published in the April 2014 issue of Morningstar ETFInvestor. Download a complimentary copy here.
Wesley Gray is the finance professor turned quant behind the upcoming ValueShares U.S. Large Cap exchange-traded fund. He cowrote with Tobias Carlisle Quantitative Value, a lucid, entertaining, yet meaty guide to the construction of an automated stock-picking strategy. Gray's upcoming exchange-traded fund will implement a version of the strategy laid out in the book.
Aside from his qualifications, including a Ph.D. in finance from the University of Chicago and a professorship at Drexel University, Gray's credibility stems from the good work he's done on quantitative investing, tackling interesting topics like valuation-based market-timing (it's hard) and the efficacy of various dividend stock strategies (total shareholder payout is a better signal than dividends alone). I've followed Gray's career for a few years, watching his firm evolve from a shoestring operation to a serious, albeit scrappy, money manager. His blog is one of my must-reads.
Because Gray is both voluble and transparent about his thinking, anyone is able to conduct due diligence on his processes and capabilities in-depth. Most funds - ETFs, closed-end funds, open-end funds - provide nearly content-free marketing materials on their processes to the public, making real due diligence hard, if not impossible, with the materials on hand. I've studied a lot of the content from Gray's firm, Empiritrage (now Alpha Architect), and am impressed with its candor and comprehensiveness.
What follows is a lightly edited interview conducted via email.
Samuel Lee: Let's start with a softball--Why launch an exchange-traded fund? The ETF business is brutally competitive. A viable fund is said to require $100 million or so in assets.
Wesley Gray: Mutual funds and hedge funds belong in the dustbin of business history. The ETF structure is transparent, lower-cost, and much more tax-efficient. From an investor's standpoint, the decision is a slam dunk. From a fund manager's standpoint, the decision is much tougher, since there is less opaqueness, less stickiness, less ability to use the traditional distribution channels, and so on. In other words, less profitable for the "croupiers." But we don't have a problem with that. One of our firm's core values is to be consumer-friendly, so we are going "all-in" on ETFs, and we think we have a unique product that is highly differentiated from the competition. We think there really isn't a true active manager in the ETF space today--you either find passive or closet-passive, such as so-called smart beta. We are explicitly avoiding indexing and quasi-indexing approaches. We want to be a genuine high-tracking error, index-irreverent, high expected value-add active asset manager, exploiting mispricing caused by a combination of behavioral bias and limited arbitrage. The key differentiation between us and traditional active managers is we want to deliver affordable active management, as opposed to overly expensive active management, which is the status quo among mutual funds and hedge funds. I want to make my Ph.D. dissertation advisor--Eugene Fama--proud to support active management by making it affordable. Although convincing him we can provide "alpha" will be a challenge.
We see two challenges in the ETF business: garnering the initial capital and distribution. For initial capital, we have largely solved this problem: We currently work with several large family offices and various wealthy individuals who have committed an initial $50 million to our first ETF, so we think we can be break-even right out of the gate. For the distribution problem, we have decided to be "disruptive" and avoid the traditional Wall Street man channel, which adds expense for the investor. Instead, we are developing a direct-to-consumer marketing channel. This approach will liberate the investor from the costs imposed by the middleman and also decrease overall costs.
Lee: What do you intend to charge?
Gray: Right now we are targeting the 50-75 basis point range, and we may bump that a bit for international strategies where operating costs are higher. A bit more expensive than "smart beta," but to be painfully clear--we aren't closet-passive benchmark-huggers. We are true active management. We will be holding 40-50 stocks and building out true active ETFs. Unlike other ETF providers, we will be hyperfocused, with a business plan that should never extend beyond a stable of 10 funds. As a point of reference, iShares has nearly 300 ETFs. We've researched the space, and in cases where the competition offers these same genuine active products, they offer them for twice our price, often due to bloated infrastructure, and in tax-inefficient vehicles, such as mutual funds. We are able to offer a lower cost because we have extraordinarily lean operations (blame my U.S. Marine Corps background for that), solid relationships with core ETF service providers, and, again, we are minimizing our distribution costs by avoiding the Wall Street sales engine and going direct to the consumer. We like the overall value proposition for the investor.
Lee: Your first proposed ETF, Empowered Funds Quantitative Value, will have to reveal all its holdings daily, exposing it to copycats and front-runners. Moreover, you estimate the strategy's capacity to be $1 billion, but you won't be able to close the ETF. It's a first-class problem, but does that keep you up at night?
Gray: Before getting into the questions, I wanted to clarify that the ETF business entity is Empowered Funds, LLC, but the name of our ETF business is ValueShares. So the first fund will be the ValueShares US Large Cap ETF. ValueShares will leverage our years of research and development to give investors access to the value premium in the most effective way we know possible.
Regarding these excellent questions about copycats, front-running, and capacity, yes, we have thought hard about these, and I'll explain some structural advantages we possess that allow me to get to sleep at night.
An advantage we possess is that we only offer ETF strategies based on deep, liquid stock markets. We can do this because our strategies are highly robust, and so a move to ultraliquid names doesn't create a large drop-off in returns. This liquidity advantage is twofold. First, on front-running, even if someone could reverse-engineer our algorithm and identify our rebalance trades, can they really cause a front-running problem in Microsoft (NASDAQ:MSFT)? Not really. Second, due to our basket liquidity, with small tweaks, we can increase the capacity to $10 billion.
As far as copycats, they are fine by us, since there are challenges to any would-be copycat. The beauty of the ETF vehicle is the tax-efficiency. So sure, someone can simply replicate the names in our basket and make sure they adjust after we rebalance each quarter (and one-offs throughout the year), but the churning will absolutely kill their tax bill. Whereas we'll be operating within an ETF construct, and so by working with in-kind redemption rules, we can minimize capital gains distributions. The tax costs to a copycat would quickly swamp any benefit of saving 50-75 basis points. Plus, at 50-75 basis points, do most people really want to deal with the day-to-day pain of managing a portfolio? Possibly, but we are comfortable with this possibility based on our own experience managing money and offering people 100% transparency on everything we do. The reality is investors quickly realize being a copycat is simply not worth the brain damage when costs are very low, although if we charged 2/20 the story might be different.
Lee: How similar will the fund be to the strategy you and Tobias laid out in Quantitative Value? What are the biggest differences and why?
Gray: Very similar, but we've added another layer to exploit an additional area of behavioral bias not discussed in the book. Step 1 is to eliminate firms that pose a risk of potential permanent loss of capital. Step 2 is to identify the cheapest set of securities, where people are most likely to suffer from representative bias. In Step 3, we identify the high-quality firms among the cheap stock bin, thus exploiting limited attention to fundamentals. Finally, Step 4 is where we have added a new wrinkle: We identify the set of highest-quality cheap firms, from steps 1-3, that Wall Street dislikes the most. We use sell-side estimate data to create a measure that ranks the distaste Wall Street has for the securities we want to buy. The more the Street hates our stocks, the more we like it. We include this element in our system because we want to exploit availability bias, which is a bias where humans overweight information that is more available and in their face. We originally contemplated creating a "CNBC hatred" measure, where we would tabulate negative comments mentioned by CBNC commentators on the stocks we wanted to own, but doing this analysis in a systematic way was not plausible. In the end, we want to buy cheap, high-quality stocks that Wall Street hates--what makes others feel a little sick makes us feel good. It's the uber-contrarian approach, by design.
Lee: What percentage of your liquid net worth is invested in the Quantitative Value strategy? Are your friends and family members invested in the strategy?
Gray: We are currently managing a little over $155 million across 30 clients--almost all are close friends and family. I am a huge fan of the idea of eating your own cooking. My entire net worth, that I'm able to control anyway, is invested in our different investing programs. The vast majority of my U.S. equity allocation is dedicated to the quantitative value approach. Many of my friends and family are also invested. It's not for everyone, though. The nature of our strategies, and their reliance on the exploitation of behavioral bias, requires that our investors invest some time in education. They should understand not only how, but why these strategies work. Even if I had a rich uncle willing to dump millions on our firm, if he wasn't willing to learn the financial, statistical, behavioral, and academic foundations of our process, which collectively explain it, we wouldn't want to manage his money. We've been at this for a couple of years now in separately managed accounts with investors who have taken a lot of time to understand our process. Not everyone has that patience or willingness to learn, and if they aren't willing to get educated, it may not be in their best interests to invest with us because they'll pull capital at the exact wrong times.
Lee: Your firm tests a lot of signals. How do you guard against too much data-mining?
Gray: We are highly paranoid about overfitting and data-mining. To avoid getting into big trouble, we focus on well-established elements of the value investing philosophy. We believe in concepts like buying cheap, margin of safety, financial strength, and returns on capital, which have stood the test of time. We also believe there are some ideas that have been extensively explored in academia over many decades, so that we can safely say they are not fads or the result of in-sample cherry-picking. The challenge with many of these is that the concepts can be subtle, or not widely known or accepted, or can appear complex. We try to avoid unnecessary complexity. We always strive to keep things as simple as possible, but no simpler. We find it can be very hard to make things simple. We also spend a lot of time on robustness studies and understanding why a strategy works. For example, we start off by identifying behavioral bias. Next, we try and understand how that bias will manifest itself in stock prices. Next, we try and understand why other investors aren't taking advantage of the bias, or in academic jargon, we look for "limits to arbitrage." Finally, we conduct a barrage of out-of-sample tests and try and distill our models down to the simplest components that effectively capture the effect.
Lee: In August 2007 a big long-short quant equity fund quickly unwound its positions and set off a chain reaction of margin calls and unwinds that led to big (albeit temporary) losses for many quant equity funds. Do you worry about quant funds herding into the same stocks?
Gray: We get that question a lot, and we understand why people might have that concern, but we'd like to think we're a bit different from the herd of quant funds you describe. First, we like to think of ourselves as primarily fundamental investors that use quant tools extensively, and not strictly quants. Fine-tuning factor models and walking around the office doing differential equations on the board really isn't our business nor our passion.
Second, we do not use leverage. So while we are fearful of crowding and fire-sale situations, we believe we can minimize the pain of this problem by staying long and never being in a position where we would be unwound due to a margin call. The reality is that there is no way to prevent a fire sale in any market or any strategy. The only thing one can control is making stupid mistakes during a fire sale. If we purchase a bunch of cheap, high-quality stocks that Wall Street hates and they magically get cheaper and cheaper because of a short-term fire sale, we won't be forced to do anything. It's possible we could even benefit from a fire sale--for example, if we rebalanced during a fire sale and bought even better bargains. We follow our model, minimize transaction costs, and hold for the long term. Mark-to-market performance can be painful, but we aren't going to change our investment process to "time" fire sales--it's impossible.
Lee: How do you decide whether a signal is broken?
Gray: One must always weigh the benefits of judiciously tweaking a model relative to the benefits of maintaining model discipline. Our inclination is to be disciplined to the model, but we also understand that committing investing suicide by following a model no matter what can also obviously have its pitfalls. Thus, we submit our strategies to an ongoing three-tiered review process. The first element is "circuit breaker." We have a great understanding of the historical performance of all our models. If a live performance event occurs that drifts toward the extreme of our historical model performance, we investigate and try to understand what happened.
The second element is "mechanism failure." This is a rare bird, but here is the concept: Let's say you have a strategy that clones 13Fs for large hedge fund managers. Prior to Reg FD (regulation on fair disclosure), you might expect this to be a great strategy because the cloned hedge funds are getting all the "inside baseball" before everyone else. However, after Reg FD, this ability to get better scoop decreases dramatically and the 13F cloning strategy stops working. This would be an example of a mechanism failure.
The third element is "tenure review." Every five years we do a full review of a strategy's live performance. The concept is that after five years, we have enough data to make an informed statistical judgment as to whether a strategy is working out of sample or needs to be tossed.
Lee: What are your thoughts on the following factors: value, size, momentum, quality, and betting against beta?
Gray: Value.--This is a legitimate factor, but it requires a long-term view and the ability to stomach volatility and tracking error.
Momentum.--This is also a legitimate factor, and it too requires a long-term view and ability to stomach volatility and tracking error.
Size.--Overall, we think the size effect is questionable, after accounting for liquidity and transaction costs. In our view, many early studies in particular did not take account of these considerations in a sufficiently robust and sophisticated way.
Quality.--As a stand-alone factor, we think it is not great. As an integrated factor, however, that helps separate the winners from the losers among the cheapest stocks in the universe, so we think it adds value.
Low Beta.--We think this factor is interesting but lacks robustness--definitely not a top-shelf factor. I can't prove it, but low-beta sure feels a lot like value.
Disclosure: Morningstar, Inc. licenses its indexes to institutions for a variety of reasons, including the creation of investment products and the benchmarking of existing products. When licensing indexes for the creation or benchmarking of investment products, Morningstar receives fees that are mainly based on fund assets under management. As of Sept. 30, 2012, AlphaPro Management, BlackRock Asset Management, First Asset, First Trust, Invesco, Merrill Lynch, Northern Trust, Nuveen, and Van Eck license one or more Morningstar indexes for this purpose. These investment products are not sponsored, issued, marketed, or sold by Morningstar. Morningstar does not make any representation regarding the advisability of investing in any investment product based on or benchmarked against a Morningstar index.