Q: What is the history of the company and the fund?
The TCW Total Return Bond Fund is an intermediate category bond fund that had undergone significant management change a little over six years ago.
In late 2009, The TCW Group, Inc. purchased Metropolitan West Asset Management LLC, a smaller investment management company here in Los Angeles, of which I had been a part since 2001. A team of us, which formed over the past 20 years since Met West began in 1996, was immediately installed at TCW to oversee all of its U.S. dollar fixed-income assets under management, including this fund. We do not have one individual managing this or any other fund; we have always used a team approach to fixed-income management.
The team consists of generalists and specialists. The generalists act as a board of directors for the Fixed Income Group, while the specialists, such as myself, handle day-to-day management responsibilities for specific sectors of the market. I focus on securitized products and co-head TCW’s Securitized Products Division that oversees about $75 billion in fixed-income assets, including the vast majority of assets in this fund.
This fund, which started back in 1993, has a current net asset value of about $9.6 billion.
The benchmark is the Barclays Capital Aggregate Bond Index, a widely used multi-sector benchmark. The fund, however, has always been a mortgage-centric fund, and in fact, we see it more as securitized product-centric, because that’s a slightly broader term that includes not just mortgages, both residential and commercial, but also asset-backed securities and, within mortgage securities, both agency and non-agency securities.
Most intermediate term bond funds tend to be multi-sector and include, for example, corporate bonds and a few other categories in which this fund does not participate.
Q: What are the underlying principles of your investment philosophy?
We have a few philosophical tenets. One is that the fixed-income markets are generally mean-reverting—over time, they move in cycles, and certain sectors get out of whack for various reasons before they return to more fundamental value.
The second is that assets can move away from fundamental value, often driven by technical forces, for temporary but sometimes prolonged periods. That’s happening right now with experimental monetary policies around the globe along with a significant increase in regulation and capital requirements in the financial services industry.
The third, which is where bottom-up analysis enters the process, is that persistent fixed-income market inefficiencies can be exploited by doing extensive security level research, keeping an ear to the ground, and being regularly involved in the marketplace.
Q: Can you cite some examples of fixed-income market inefficiencies?
There are tens of thousands of often complex, varying assets, particularly in the mortgage and securitized products market, that can be driven by investors who are often incented to make non-economic decisions. Sometimes there are segmentation and duration issues, and sometimes investors are driven by the quality or rating restrictions of a particular security type that leave other securities fundamentally undervalued. We might purchase securities that are not bid up by those non-economic factors, or, if overpriced, sell into them.
Another inefficiency is how federal and local governments interact, e.g., intervening in the capital markets’ pricing process. In a perfect world, in a free market, a security’s price would reveal key reliable information, like supply and demand for credit and duration, but monetary or fiscal policies, or government regulations, can drive asset prices away from their normal free-market levels. Since the Great Recession, regulation such as the Dodd-Frank Act has driven liquidity lower, while monetary policy has driven asset values higher, out of an equilibrium state.
Q: What is your investment strategy?
We have two core missions. The first is to seek long-term total return for investors, which occurs over a full cycle of interest rates.
Our second mission is to beat the benchmark and produce risk-adjusted excess returns over a full cycle of interest rates. That said, the benchmark is multi-sector, and includes corporates, government agency debt, Treasuries, and more, whereas this fund is mortgage-centric, so we must be alert to potential additional tracking error from time to time.
We want to ensure the alpha, the risk-adjusted return, that we achieve compensates for that tracking error over the full cycle, and in the six years we’ve been managing the fund, we have achieved that goal
In 2015, the fund produced a total return of 108 basis points while the benchmark’s return was 55 basis points, or 0.55%. The past three years, annualized, the fund returned 2.83% while the index only achieved 1.44% annualized return, a yearly excess of 138 basis points. Now, for the past five years, we produced 5.12% return annualized while the index managed only 3.25%, giving us a 1.8% annualized excess return.
If we look at the five-year information ratio, the ratio of excess returns to the volatility of those excess returns, it is 0.96. Generally, anything over 0.5 is considered to be good.
Q: How does your investment process work?
We use a disciplined, dollar-cost methodology in our approach to buying and selling assets. Rather than try to calculate the bottom or top in asset prices at any point in time, we watch to see when assets diverge from their fundamental valuations. If they become cheaper, we dollar-cost-average them in. If they become pricier, we dollar-cost-average them out.
Such fundamental valuation requires specific value-driven research, so we have a large team covering the sectors, with most analyzing, researching, and trading securitized products. We believe persistent inefficiencies in the securitized products market can be exploited through disciplined research and bottom-up issue selection. We focus on issue selection as a primary driver of alpha over a full cycle versus predicting interest rates over short or intermediate periods, as we think interest rates offer the least potential for alpha.
Ours is a diversified process. Even though this is securitized product-centric, we diversify among the asset types and even on specific assets. Since cash flows among securities within the asset class can have a wide range of alternative outcomes, no one asset dominates, and often they represent small percentages of the overall net asset value.
Q: What is your research process and how do you look for opportunities?
One research-intensive area for us is the non-agency mortgage-backed securities market, those not backed by Fannie Mae, Freddie Mac, or Ginnie Mae, but instead securities collateralized by loan pools, with sometimes thousands of loans underlying each issue. We have a proprietary method to identify what underlies these complex assets, and to model cash flows from these securities over a wide range of scenarios
For example, on a monthly basis we collect data, directly from the trustees or servicers, on the vast majority of the non-agency loan market, 30-plus million outstanding mortgages. Getting accurate and timely loan level analytics is crucial to predicting borrower behavior in current market environment. This information tells us what the borrowers’ credit and propensity to pay on their mortgages look like today.
This non-agency mortgage-backed security area is ripe with alpha because the assets became distressed in 2008 and 2009, during the Great Recession. We began investing in these securities at discounted levels. We have been able to add value through insight into this complex asset class by making significant capital investment in infrastructure and human resources, which we started doing well before the housing crisis occurred.
We also use a proprietary system to run the potential cash flows in each security over a wide range of possibilities or distributions over time, at all different prices, and highlight those prices that look at proprietary statistics, including current loan-to-value ratios that we estimate on our own independent methodology.
We get third-party information as well, such as the borrower’s credit to gauge the likelihood of that borrower to pay his/her mortgage in the future. And, on a macro basis, we consider home prices in the various geographic regions in which these loans exist, the current unemployment conditions and forecasts, foreclosure rates, and current payment rates on those loans. We provide our traders immediate access to information on almost any CUSIP that comes to the market for bid, so if somebody is looking to sell it, we can quickly calculate our own valuation estimate.
Generally, when an investor wants to sell a group of assets—known as BWICs, “bid wanted in competition”—we are not the only investors seeing those assets. We enter details on each day’s BWICs into our database so that when a BWIC comes to market, our traders can estimate relative value based on the most up-to-date information. For example, we can look at the percentage of loans that have been current for the last 12 months, without delinquencies, which can tell you about the borrower base, or see the true loan-to-value ratio of underlying loans use our proprietary methods. We have several dozen different ways to slice the information.
Q: What is your portfolio construction process?
There are three aspects to our construction process: diversification, which I mentioned earlier, risk control, and consistent implementation.
For example, when analyzing what assets might produce a significant amount of alpha over a full interest rate cycle, we look for a flat risk profile. We want assets that, over a wide distribution of outcomes of interest rate and home prices and other macro economic factors, show no large return swings. If we can dollar-cost-average those kinds of flat risk profile assets over a full interest rate cycle, we’re more likely to produce alpha.
Consistent implementation is dollar cost averaging performed in a diversified way. We might find an attractively priced asset, but no matter how relative value we think it provides currently, we won’t excessively overweight it in the portfolio. We obviously try to be right more often than wrong in asset selection and valuation, but it is possible to be right and wrong and still produce alpha over a full cycle if investing is done in a diversified way.
We also invest in U.S. Treasuries from time to time, mostly for two reasons. One is to help with duration and partial duration management relative to the benchmark, and the other is to provide some convexity to the portfolio. Mortgage securities generally have negative convexity—this is similar to selling options, where average lives tend to shorten as rates fall and get longer as rates rise because of different prepayment expectations over different interest rate cycles. Treasuries help normalize and improve that convexity profile.
In addition to agency and non-agency mortgages and U.S. Treasuries, we have commercial mortgage-backed securities as well as asset-backed securities that aren’t typically mortgage related. These five groups of securities can constitute anywhere from 90% to 100% of the fund’s assets.
Q: How do you define and manage risk?
Our information ratio is one way to assess whether we’re producing true alpha. The higher the information ratio, the more likely it is alpha—risk-adjusted excess return as opposed to simply excess return with excess risk. We look at the standard deviation of that excess return, a traditional quantitative way, but there are qualitative methods we use as well.
What we learned from the most recent significant financial crisis is the need to continue our disciplined process of investing, and how having a diversified portfolio and a disciplined process, for both bottom-up research as well as dollar-cost averaging investments, are critical to producing risk-adjusted returns.
We do not set a minimum position size. Depending on whether the fund is growing or shrinking (it has been growing for the past 5+ years), older positions that still represent good value remain in the portfolio and become smaller over time—some become so small as to be immaterial to the fund’s performance. Still, we do not sell such assets just because they are immaterial. If you add up enough of those insignificant immaterial assets, there could be some materiality to it.
The fund can have a large number of positions, but after the first 125 or so positions, they drop to relatively small allocations. If these small allocations continue to shrink over time, it equates to an increasing number of portfolio positions that won’t individually impact the fund’s performance, but will assist returns in aggregate.
