Thu, 18 Oct 2012
When buying into nontraditional assets, investors should monitor the risks of high correlations to equities, backward-looking data, and manager performance, says Vanguard's Fran Kinniry.
Christine Benz: Hi. I'm Christine Benz for Morningstar.com. Ever since the bear market, investor interest in alternatives has increased dramatically. On a recent visit to Vanguard headquarters, I sat down with Fran Kinniry, a principal in Vanguard's investment strategy group, to discuss some recent research on the asset class.
Fran, thank you so much for being here.
Francis Kinniry: Thank you, Christine.
Benz: During the past decade or so investors have become much interested in adding nontraditional asset classes to their portfolios, and certainly some data support doing so. But you note that there are some practical hurdles to adding nontraditional asset classes. First, let's talk about the types of assets that people typically think of when they are thinking about that other portion of their portfolio.
Kinniry: Typically what we see as investors, the other part, would be beyond stocks, bonds and cash. They really come in two forms. One would be, what we would call betas, like different parts of the capital markets, and the other would be alpha, which would active managers either through hedge funds and private equity. They are very different, but that's what we've mostly seen, is trying to move out of the traditional asset classes of stocks, bonds, and cash and into some of these more exotic beta or alpha mandates. That's been a common theme, which I don't think is going to let up because of really the low-return environment.
Benz: Right. One thing that you found in analyzing some of these alternative asset classes is that oftentimes asset-allocation models are the most enthusiastic in their recommendations of alternative asset classes after they've already had a good performance runup.
Kinniry: Yeah, that's right. I mean with a traditional model portfolio, sometimes the more sophisticated you get, the more dangerous it can be. With the explosion in computing power and the speed at which computing power is, we can all go back to our desks and find a 20- or 30-year asset class that has outperformed on a risk-adjusted basis. And so, a lot of these model portfolios are formed through a technique known as mean variance optimization. You're trying to either increase the return with the same risk or lower the risk with the same return. The problem with all of that is, no matter how you do all of that, you're using backward-looking data.
So, you're using returns, risks, and correlations, and a lot of these asset classes may only have 10, 20, 30 years of history. And I want to remind everyone that we've been in one long super cycle of disinflation. So, we've had year-over-year lower interest rates and year-over-year lower inflation since 1977. So a lot of these models have not even seen different cycles. You'll see that no matter how you try to govern these, you're going to load on those asset classes that while looking great on a risk-adjusted basis they may be selling at some extreme valuations relative to their starting point.
Benz: Your data would also point to certain asset classes, even though they look good over a longer period of time, during periods of real market duress where the stock portfolio is down 35% or more, they haven't necessarily held up that well. What did you find on that front?
Kinniry: Again, back to the computing power, a lot of people will put into their models or how they build a portfolio, they will use average correlations or use median correlations. And so with the typical diversifiers you see of real estate investment trusts, commodities, hedge funds, and private equity, you will see that they have lower correlations than the stock market, but have returns like the stock market. So, you would think, "Well, that adds to my portfolio." But I would just throw out there for every investor to think about, if you have 30%, 40%, or 50% in equities, what you really want is to have low correlations or negative correlations with your primary risk asset; equities are the primary risk asset in most portfolios.
So, you don't want to look at average correlations, where commodities and REITs and emerging-markets assets and hedge funds may look well, but [you want to look at how] those asset classes perform when your primary risk asset, equities, is having a tough time. And so, within [our recent] paper and some of our research, we wanted to see how did these asset classes perform when stocks were in their bottom decile or bottom-10% events. In there, almost all of these nontraditional diversifiers had negative returns, when stocks were in their bottom 10%. The only asset class that actually had positive returns were investment-grade corporate bonds, Treasury bonds, and high-grade municipal bonds. So, the classic traditional diversifier of fixed income I worry about today because a lot of people are trying to think about "Let me move out of fixed income with low yields, and I can go to some of these other nontraditional diversifiers."
Benz: And you think there are big risks there. Another risk that you highlight is the manager risk. Anytime you're using a product to provide you with alpha, there is a manager selection risk and overall manager risk. Let's talk about that.
Kinniry: Back to the betas, we can tell you if you buy a REIT index or a commodity index, you're going to get that index. When you go to active strategies, which would be hedge funds, private equity, or even traditional active, you're really just picking one little manager out of a very large distribution. So, how you even model that becomes very complicated. Really, the bottom line is if you can get the top-quartile manger, you would want to go all in, and if you're getting anything below the top-quartile manager, you wouldn't want any part of it. And that's a tough game because especially in the nontraditional hedge funds, private equity, and venture capital, it's not a democratic access issue. A lot of the large endowments and foundations that have been doing this for a very long time have unique access, unique relationships to that area, and even in the traditional long-only mutual funds space, we have not seen persistence. Morningstar has done a lot of work. We've done a lot of work that just because you had good alpha or excess return over a one-, three-, five-, or 10-year period doesn't really give you much to say it's going to exist in the next three-, five-, and 10-year period.
Benz: So, there are obviously a lot of reasons to proceed with caution in this area once moving beyond that traditional stock and bond portfolio. In your research paper, you highlight what you call some best practices for people attempting to make good decisions in this area. Let's talk about some of those.
Kinniry: I think the best practice, first and foremost, is to really fix what your strategic allocation should be. How much return do you need? And understand that return and risk are related. The more return you need or want, chances are the higher risk you're going to have. There are not a lot of free lunches out there, where you're going to have higher return and keep risk the same. So, really concentrate on how much equity you want and how much risk-free or lower-risk assets like fixed income you want, and rebalance to that. And in this low-return world, really fixate on costs because a lot of these nontraditionals have higher costs, and if we're in a bond market of 2% to 3% [returns] and a stock market of 6% to 8% [returns], you can pick your number, I think most people are kind of comfortable with a balanced portfolio being in the low single digits, 4% to 8%.
Benz: In terms of the return that you could expect?
Kinniry: Exactly, the returns you can expect. Then really thinking about managing your costs, which you know them in advance and they're highly controllable, is going to be a real secret to success in the future.
Benz: Well, thank you so much, Fran, for sharing this research.
Kinniry: Thank you, Christine. I appreciate it.