Fri, 13 Apr 2012
Legg Mason's Randy Befumo points to how staying clear of significant industry weightings allows his fund to control for volatility.
Bridget Hughes: Hi, I'm Bridget Hughes. I'm an analyst here at Morningstar. I'm here today with Randy Befumo, who is the director of research at Legg Mason Capital Management and is also one of the portfolio managers on the Legg Mason Capital Management Disciplined Equity Research fund.
Thanks for coming in, Randy.
Randy Befumo: Thanks for having me.
Hughes: So this has been a mutual fund now for about 18 months, but it's a mandate that the firm has had for about three years. So, please talk about what this research fund is, what it's supposed to do, and what it's supposed to be for people?
Befumo: What's different with Disciplined Equity Research, or what call internally the Research Fund, from the rest of the Capital Management funds is that it's team-driven. So we have six broad-sector teams within research. Each of the teams is allocated a portion of the portfolio based on their team's percentage of the index which is the Russell 1000, and then they have a high degree of autonomy in deciding what to do with the money, what number of positions to put it in, weightings, and so on.
The only rules effectively for the Research Fund are that it's something on which we've actually done research. We've applied our investment process and come up with some valuation in a sense of where our edge is and where we differ from the market. And so, to the degree that those hurdles are met, each team basically decides independently what it wants to do. So the portfolio as a whole really reflects what we believe are the best ideas of research, but it also does it in a way where there are not significant sector bets and there are not significant industry bets. We believe that a lot of excess volatility can come from industry bets. All in, the Research Fund is a lower-volatility version of the same core Capital Management process that you see in all of Legg Mason's other products.
Hughes: Let's take a little bit more about that research process at Legg Mason Capital Management and maybe how you think it's different from some of the other analyst-run or best-ideas portfolios that are out there?
Befumo: I would say that there are sort of two key points of differentiation in what we're doing at Capital Management at the research level. We're an evaluation-driven shop like there are a lot of evaluation-driven shops. We're long-term. There are a lot of long-term shops. The Research Fund, like all of our other products, runs with very high active share relative to most other products.
Where we differ, though, are I think two key things at the stock analysis level. The first is, we very much have a probabilistic framework for how we think about valuation. When you read about what an analyst thinks the stock is worth, it's typically a single-point estimate. We believe, however, the future is not a single point, but rather it's a distribution of outcomes. And we actually do a fair amount of work to try to figure out what distribution is implicit in any stock that we value.
So, we have a core value that we call our central tendency value. It's kind of the middle of the distribution that we tend to work with, but we also have high and low points that we think are also reasonably likely and give us some sense of the shape of that distribution. Is it a normal distribution? Is it logistic? I could bore you with statistic terms for half an hour, but the basic idea is not just collapsing our view on what something is worth into a single point that you can be right or wrong on, but rather understanding sort of the shape of the future and what the stock might be worth in a variety of scenarios, and being able to apply that as part of portfolio-construction process.
The second thing that I think is a little different, not a lot, but a little different is that we're very empirical in terms of figuring out what exactly is our edge. So, the market is incredibly efficient; only about 45% of stocks outperform every year. A really good investment process maybe gets you to a 55% or 60% hit rate. In a very stochastic environment like that, you've really got to narrow in on what do you believe is different from this very efficient, very challenging, and very tough market and make it: 1) empirical--it's got to be something that exists in the world; 2) measurable--it's got to be something you can put numbers behind; and 3) falsifiable--it's got to be something on which you could actually end up being wrong.
We codify this in what we call our investment cases. Every investment case has two to five variant perceptions, and we also do some work on what we call premortem or "what would make me wrong." The idea there is just figuring out in advance if things are going bad on the stocks--if we're actually going to lose money on it--what would we be seeing along the way? This idea is just to make ourselves a little bit more flexible and a little more disciplined in saying, "Well, we think that the stock is well below intrinsic value and the investment case is actually working out. So, this is something we still have high confidence in," versus us saying, "This company has faced a couple of setbacks. Everything we believed when we bought it has turned out to be wrong, but golly, it's 4 times earnings. How can we sell it here?"
The discipline of having that investment case coupled with a dynamic sense of valuation of a variety of scenarios we believe yields better outcomes over time. That's sort of the core Capital Management stock-analytic approach, and that's really what's intrinsic to the fund.
Hughes: One of the things I think that sometimes goes hand-in-hand with some of the funds that don't take sector bets is that there is some sort of tracking-error constraint which doesn't seem to fit with Legg Mason Capital Management's high-conviction approach. Do you control for benchmark risk?
Befumo: Perversely, we actually seek to increase it by limiting the number names in the fund. So, there is an absolute limit which depends on the allocation of different sleeves, but it's effectively somewhere between 100 to 110 names, which the maximum number of names. The fund typically has 80 to 90 names. It's often running with active share between 70% and 80% which shows up as high active share and almost all ways of thinking about it.
So it's a very concentrated fund. However, the way we believe we can control for volatility is by simply not taking big sector bets. So, our quantitative portfolio manger Arturo Rodriguez has done a lot of what we believe is path-breaking work on sort of sector bets and how they contribute to the overall volatility, particularly tracking error. His specific finding is he believes about half of the tracking error in a product comes from sort of sector variation. So we believe we can run the core Legg Mason Capital Management process but actually end up with, call it, between 3% and 6% tracking error just by not being as extreme on the sectors which is a byproduct to how the fund is constructed. We've got to have some way of divvying up the money among the team, and the sector teams are basically how our research functions. So, it's really the best way to do it.
Hughes: Great. Well, thank you very much.
Befumo: Thank you.