Laurence B. Siegel, director of investment policy research at the Ford Foundation, has great investment instincts. But he knows too much to follow them blindly, preferring instead to temper instinct with disciplines developed over a lifetime of experience in delving deeply into the best investment research available. Especially the quantitative stuff. A good bit of which Larry, back when he served as managing director of Ibbotson Associates, which he helped to establish in 1979, did much to put on the map.
But Larry isn't a head-in-the-clouds intellectual, either – something he can prove equally by quoting early rap song lyrics or by pointing to his service as chairman of the investment committee of the Trust for Civil Society in Central and Eastern Europe. He is, in short, precisely the sort of fellow you'd expect to keep his head – and maintain a long-term perspective – even amid volatile markets. Which is exactly what he's doing. -KMW
The markets seem to have gone wacky, Larry. Is this part of a process that means value-oriented strategies had their day, and growth stocks are about to revive? You did, after all, pen (along with John Alexander) that very well-timed piece on "The Future of Value Investing" that appeared in Invesco's first quarter 2000 report – when almost everyone believed value stocks were death.
Value strategies have performed so well since March 2000 that their expected return, relative to the overall market, is much more modest than it was back then. But I still think value stocks have more attractive valuations, and more "visibility" in the sense of earnings that are believable and predictable. And value is a better strategy overall, although a well-constructed portfolio has growth as well as value components.
What do you mean, value is a better strategy?
Value stocks are already priced to include the possibility of a disaster. If the disaster doesn't occur – and there will be many more non-disasters than disasters – you eventually get a capital gain as the market figures out the true worth of the company's assets. So the active value manager only has to figure out which companies are most likely to collapse. Growth stocks are much better companies. However, they are typically priced for high rates of long-term growth that are only rarely realized. If there is an earnings disappointment – and for most growth stocks, there eventually will be – the price will fall. To beat a growth benchmark, the manager has to figure out which of many companies are going to have the competitive edge that enables them to become the next Microsoft or Wal-Mart. That's too hard.
Let's talk a bit about the "science" of money management. The problem is that many portfolio managers seem to have forgotten the essence of investment risk: Loss of capital, or purchasing power. You're a quant –
I've been trying to shake that designation, but obviously not very successfully. My first reaction is that while volatility (standard deviation) isn't exactly coextensive with risk, it's a pretty good proxy for risk. You have to measure risk somehow. Moreover, there is no clearly better measure. Downside deviation, shortfall risk, and other approaches to measuring risk have been tried but they aren't very revealing. The whole process of bringing science to bear on the question of deciding how to invest has been, by and large, tremendously positive for everyone. And you can't do that unless you adopt some measurement techniques, including the ad hoc definition of risk as volatility, which makes the calculations doable. But of course, the science isn't perfect. There are unusual circumstances that crop up from time to time in the market, which make that wisdom look foolish for a while. An example of that is the Internet bubble.
Exactly. I suspect that there's a crucial disconnect between trying to make a quantifiable science out of something as inherently unstable as a market – which after all reflects collective human emotions. Fear and greed don't always fit into neat equations.
That's right. During the height of the Internet bubble, a number of companies with essentially no sales, earnings or assets rose to huge capitalizations. That meant that the market-cap weighted indexes, which are not only the basis for all index funds but the basis for the performance measurement of active managers, had a lot of weight in these companies. If one accepts portfolio theory, the market portfolio is, by construction, mean variance efficient. But we knew it wasn't. It's not efficient to own a bunch of stocks that you're pretty sure are going to crash. So in the spring of 2000, it was a pretty easy call to say that something was going to beat this overstuffed S&P 500 or Russell 1000, which had large weights in a few tech stocks with little or no assets. It just wasn't clear what-it could have been bonds, it could have been value stocks, it could have been international stocks.
An "easy" call? Then why didn't more investors get out at the top?
It must have been easy or I would not have been able to write that article with John Alexander. I don't do hard calls! Kidding aside, at the time, it is always hard to go against the grain and say, "I am going to sell," when other people are getting rich by buying. That is the tracking error problem: you are judged as taking "risk" if your holdings are materially different from those of the benchmark. But some of these technology or Internet stocks were barely stocks at all. They were lottery tickets. Of course, some investors expected these companies to generate earnings or cash flows, eventually. The fundamental concept of investing is discounted cash flow. But if you projected the cash flows that you expected the companies to generate, then used a reasonable discount rate to reduce those future cash flows to a present value, and compared that result with the price, most of them were still lousy buys. But it seemed that at the time only value investors were doing that math because the growth investors knew that if they did the math, they couldn't justify the prices of the stock! Sure, there was a growth rate you could put in that would make the stock fairly priced or even under-priced –
All you had to be willing to do was discount the hereafter.
Right. These were growth rates that had never been achieved by anybody. So was it an easy call? Yes and no. Let me tell you a brief parable. You've read our April 2000 paper, in which I was pretty clear that value would beat growth. But what did I do individually? Nothing. I already had a value fund in my portfolio and I basically kept the growth fund where it was, as a hedge against the possibility that I might be wrong. After all, I had been wrong for while. I had started to get nervous about growth stocks a full year before then. When you actually build portfolios, you realize how little you know with certainty. So the possibility that the other guy might be right begins to dominate in your psychology and you become a chicken. I believe that professionals in our field-plan sponsors and money managers-are about the worst personal investors you can imagine. That is because they are always saying, "What if I am wrong?" So they have difficulty acting according to their own advice-which is typically good advice.
Paralysis by analysis. But aren't there are other things about the way portfolio theory has come to dominate the business that explains so many managers' failure to act when the handwriting was on the wall?
Yes, very much so. Portfolio theory says that the cap-weighted portfolio is mean variance efficient. So you are taking "risk," relative to that portfolio, by accepting tracking error, by deviating from the market. If I had followed my own advice and put 100 percent of my U.S. equities into a value fund, I would have subjected myself to a 20 percent or 30 percent a year tracking error to the S&P 500 benchmark, and I could have been wrong. Nobody wants to be "wrong and alone."
But the S&P 500 is anything but a good proxy for the market, much less an efficient one. It's an extremely actively managed portfolio – and one that became tremendously overweighted in techs.
The S&P 500 is not that bad a proxy for the market. The correlation of the S&P 500 with the Russell 1000, which is just the top 1,000 stocks with no active selection, is 0.996. The problem was with the market, not the index used to measure the market.
Only if you're willing to accept a pretty narrow definition of "the market" as the top 1,000 stocks, still cap-weighted.
The top 1,000 stocks comprise more than 80 percent of the total market, so you get a high correlation with the S&P no matter how many stocks you add. You would have to define the market as the equal-weighted market, or use some other measure that ignores market cap, to get a rate of return substantially different from the S&P. Equal-weighted indices are interesting to study, but if any serious amount of money went into that strategy, there would be a shortage of the smaller-cap stocks and managers would not be able to construct the portfolios.
Isn't "tracking error" of the positive variety precisely what active portfolio managers are supposed to add to a portfolio-incremental performance?
A leading active manager recently said the same thing – that tracking error was supposed to be as large as possible, as long as it was positive. That is very cute, but can he do it? While there are always a few active managers who are flying high, the track record of the active management community in-the-large is not good. Plan sponsors and their consultants had better measure the quantitative characteristics of their managers to avoid getting snookered – paying high active fees for index-like performance or worse. And tracking error is one of those characteristics. Of course, for those few active managers with real skill, tracking error feels like a shackle, since they are often expected to buy stocks they know are bad investments (and prevented from stepping outside their boxes, to buy good ones). The problem is that plan sponsors or asset owners don't know which managers have skill, so they put the shackles on all the managers.
That's an institutional cop out if I ever heard one. If the plan sponsor isn't willing to do the research and then accept responsibility for making a decision (that, yes, might be wrong), then he might as well mechanically split his resources between an index and Treasuries and retire.
You are not far from right. Plan sponsors will go to the ends of the earth to avoid taking responsibility for a decision. I am currently writing an article with Barton Waring, of Barclays Global Investors, that says that sponsors should make numerical estimates of the alphas of each of their managers, for use in an optimizer. If the sponsor doesn't think he or she can make such an estimate, then the sponsor should index instead.
But is trying to minimize the odds of being wrong and alone what an institutional money manager is really paid to do? Isn't he or she hired to preserve the purchasing power of the clients' capital over time, and to earn a return on it, consistent with an agreed level of risk?
That is called the principal-agent conflict. Preserving purchasing power and earning a return on capital is good for the client. Minimizing the risk of being wrong and alone is good for the manager. There are a variety of techniques to align the interests of the two parties. One that is popular in the hedge fund world is to have the manager's own assets invested heavily or exclusively in the fund. I think that is very risky. Who would want to hire a manager who ignores the most basic principles of diversification? So there is no good single answer. Thoughtful use of performance measurement techniques is another answer – and that is what I am advocating.
But it was pretty obvious in the spring of 2000 that growth stocks were going to have some problems. You've said so yourself. Lots of highly paid professionals most likely saw the same thing – but had their hands tied. I've heard any number of growth managers complain that if they had gone outside of their box, they would have been fired. Which essentially meant they believed their clients preferred losing capital to allowing them to make informed stock selections, or even to raise cash.
Yes, once in a generation the division of labor between plan sponsors and managers that I just described will result in throwing a lot of money away. In the spring of 2000, it would have been better to let our growth managers buy value stocks, bonds, cash, real estate, or anything but large-cap U.S. growth stocks. However, that's not the way that a fund should be run 95 percent of the time, so that's not the way we're going to run it.
There is a real conflict, a real logical disconnect, in what you are saying –
There sure is. And that is one reason why hedge funds have suddenly become so popular. They take back at least part of the asset and style allocation decision – and promise that they'll always make a positive rate of return.
If you believe that, there's a bridge to Brooklyn I'd like to sell you. But they are at least aiming at making absolute returns. Doing relatively well (but still losing capital) doesn't cut it in a bear market.
But bad or mediocre managers can start hedge funds, too. Few of the 6,000 hedge funds now in existence will behave like Soros or Robertson. There aren't that many geniuses in the world. In the end, they are going to perform like the long-only managers that they were just a few years ago.
Complete with fully developed herding instincts.
They are going to act like people because that is what they are. Every time I am tempted to stand up and say, "Look, it is okay to take benchmark risk as long as you are avoiding real,' or absolute risk," I am reminded how hard it is to do that. There are a few people who will beat their benchmarks over a long time period, but it is very hard to figure out in advance who they are. So it is very important to have benchmarks, and to measure managers against them, rather than letting them sell you on the idea that you should just let them do whatever they want. Absolute return investing, in particular, is a way that agents try to convince principals to not subject them to the discipline of performance measurement and to permit them to do whatever they feel like at the time. But I think that benchmarking, performance measurement, and other quantitative analysis probably keeps more arrogant egotists from ruining portfolios than it keeps misunderstood geniuses from contributing to them.
The way things have been going, portfolio managers might even be chastised for not holding cash someday –
Yes. Investors expect their managers to be fully "invested," as if cash is not an investment. Many value stocks are now fully priced; but some are still bargains, especially after this most recent downturn, which has been quite severe. Yet the alternative is not growth stocks – it is very hard to figure out the fundamental values of many of them – but to buy bonds or other asset classes instead. But I would caution that putting your money to work in a well-diversified, risk-controlled and well-monitored alternative investment program is extremely labor-intensive. I don't know if the people who are late to the party in investing in hedge funds appreciate that.
Maybe what has really changed is that relative performance won't cut it in this new investment world. A world without a rising tide.
Yes, then hedge funds really do have an advantage over long-only equity managers. However, you don't have to choose between long-only equities and hedge funds. There are always bonds, TIPS (inflation-indexed Treasury bonds), cash, real estate, and other assets – even commodities, which tend to perform well when other assets are doing poorly. There is always a way to make money.
Thanks, Larry.
Kathryn M. Welling is the editor and publi-sher of welling@weeden, an independent research service of Weeden & Co. L.P., Greenwich, Conn. http://welling.weedenco.com
