The Policy Portfolio revisited

by Peter L. Bernstein, President, Peter L. Bernstein, Inc.

A few years ago I gave a talk that I thought was interesting and controversial. I was making a point—and did I ever! It got everyone all agitated. I ended with the observation that we have to learn to live without the crutch of policy portfolios. I’m going to qualify that conclusion now because I’ve learned a little bit in the intervening few years, although I think that the basic ideas are still there.

At the core, what you really have to think about is the long run because the policy portfolio is not what you’re going to have today, but what you’ll have over an extended period of time. It’s based on long-run expectations and considerations of risk. So, we really have to think about the long run—how well we understand it and how we use it. How much of the future can we foresee, and what do we do with the answer to that question? I’m going to begin with a brief description of what I mean by a policy portfolio, then try to develop a sense of how we should approach the fundamental asset allocation problem.

First, a policy portfolio—what is it? And why is it?

What is it? It’s flexible—I use the word “crutch” again, by which investors indicate that they have an ordinary portfolio built on some kind of underlying view about risk and return and covariances among asset classes. It’s a broad statement of the investment committee’s view of the relative long-term attractiveness and risk in the variable asset classes. And long term, as I’ve said, is the operative word. A policy portfolio is an explicit statement that an institution has some sense of how much risk it wants to take, but it’s also a reflection of its beliefs about the long-run future.

Flying blind without a policy portfolio

Now, why do we have these things? Why do we go through this exercise? Policy portfolios provide a set of guidelines around which all subsidiary decisions can revolve. Without a policy portfolio, investment committees have the feeling that they’re flying blind. The policy portfolio reflects viewpoints about what the future holds and what kinds of risks the portfolio can take.

It serves another important purpose, and this is one of the reasons I’m eating a little bit of crow about the original presentation. The policy portfolio provides a benchmark against which the active bets can be managed. If we just took the policy portfolio and held it, how would we have done? And how have we actually done, because over time it’s impossible not to deviate from the policy portfolio.

Today, investors are generally employing three different approaches to managing the actual portfolio within the structure of a policy portfolio. One is periodic rebalancing to the policy portfolio allocations. This is what David Swenson of Yale does with zeal. He rebalances every day so that he holds the policy portfolio almost constant. A second approach is to allow wide swings around the policy portfolio, but with a view to ultimately rebalancing back to the policy allocation. The third approach is more ad hoc, namely to sit down and revise the policy portfolio when changes to the fundamental environment justify those revisions. All three, however, involve periodic rebalancing, allowing swings and then deciding when to revise. We start from the viewpoint that the policy portfolio is a longterm statement that should not be tinkered with in the short run.

But what if the long run is an uncertain guideline? Then how do we manage asset allocation? How do we benchmark our managers if we say we’re really scared of the long run and we’re going to have to do something differently? I don’t know whether this can be done, but I think it’s worth asking ourselves this question and thinking about what we would do if we didn’t have a long-run view expressed in a policy portfolio.

What goes up will come down

Essentially, policy portfolios express unwavering faith in regression to the mean—that what goes up will come down, that things will fluctuate, but around a basic setting…or, more hopefully, that trend lines matter and after considering the risks involved we want to have most of our money in assets with trend lines that slant upward. For example—and this is the leading example—in the long run, equities will provide the highest real rate of return even though they have short-term volatility.

Now, why do we think this about equities? Simply because equity returns must dominate, or the system will fail. When you can expect to earn more on a bond, which is a contract with the borrower, than you can expect from holding equity, where nothing is promised, risk-taking will come to a halt, business activity will dry up and the economy will turn into dust. In other words, the whole system will come to an end if the expected return on bonds is higher than the expected return on stocks. Because if you expect to get more by being a lender than by being an investor, then our whole economic system is in big trouble.

I’m not arguing that the long run has no relevance. But I had made the following assertions, and I think it’s within these assertions that we have to look at this matter: Visibility is never what we think it is. It’s very, very short. Uncertainty is a constant rather than a variable. It’s not more or less, it’s always there. We do not know what the future holds. And, as a result of that, the long run has been a frail reed to lean on.

Consider just a few facts about this. Jeremy Siegel made his reputation because of 7 percent. He found that over the long run, 20 years or more, with extraordinary consistency equities have provided real return of 7 percent. When you break that figure down, it turns out that real price appreciation contributed only 2.1 percent to that 7 percent. In other words, 5 percent is missing and that came from dividends; so that the real price return was only about 2 percent. But we got 7, and we needed a decent yield on dividends and reinvestment of dividends to get there. So, more than half—substantially more than half—has come from that source.

Scant dividend yields

The average dividend yield over Siegel’s long run was 4 percent. Today, the dividend yield is less than 2 percent. So, in order to achieve 7 percent real return over the next 20 years, we’re going to need 5 percent growth in earnings or dividends, which is very hard to visualize over the long run. We certainly can’t count on buybacks—they’re too uncertain. Just in the last 12 months buybacks among S&P 500 stocks have totaled over $400 billion, which is like taking Exxon and saying that it was bought out by a private equity company and disappeared from the market. When you do something like that, you’re changing the character of the market and, therefore, long-run structures will differ.

It’s also important to realize that if you look at long-term growth in earnings and dividends in the U.S., the real growth rate is less than the growth rate in GDP. So, you can’t walk around with big numbers in your head. It just isn’t automatically there. You can argue that valuations will increase. You can say that today stocks are not really that expensive but that over the long run they will get more expensive. Do you really want to make that bet?

Furthermore, in trying to figure out the long run—looking at the past and drawing trend lines and taking averages and standard deviations and so forth—I think we forget what the nature of history is all about. It’s not a pool of inconsequential happenstances, one after another, or individual episodes or a series of accidents. Each episode is equally instructive and is a consequence of the preceding episode. Markets have long memories.

So, volatility matters. Volatility matters. You can say long run, but in the meantime you’re spending money. And the current return that you’re getting is not enough to cover it. So, this is always a consideration in managing the principal in one’s portfolio. The portfolio can absorb the volatility over the long run when you hold an asset with a long-run total return high enough to compensate for the short-run volatility. But is there such an asset? This line of reasoning is what led me to recommend a more opportunistic and less structured approach than the policy portfolio approach.

Uttering the dirty words

But if you think about what I just said, I’m talking about market timing. And those are dirty words. But why are they dirty words? The old joke went like this: If somebody says you can’t deviate from the policy portfolio because it would be market timing, just tell them that the asset liability mix has changed.

The big risk is that market timing is very difficult. I’m going to get back to that. But the big risk in market timing—and this is what has given me some second thoughts—is what Ron Kahn and Richard Grinold refer to as breadth in their book, Active Portfolio Management. They define breadth as “the number of times we can use our skill. If our skill level is the same, then it is arguably better to be able to forecast returns on 1,000 stocks than on 100 stocks.” If you’re picking stocks, you’re making a lot of different decisions, a lot of different choices. Some will be wrong, but some will be right. In market timing there are really relatively few moves that you make. Over the course of your investment lifetime, it might be a dozen. If you’re wrong, you’re likely to be wrong big. It’s not a diversified kind of bet. You don’t have enough opportunities to diversify it.

So, there’s a genuine problem with market timing and it’s the only problem I can see—that it is hard and you don’t get enough opportunities to diversify your bets. But if market timing isn’t easy, try this: How easy is it to manage portfolios when market fluctuations drive actual allocations away from policy allocations? What do you do then? Those decisions are hard, too. How easy is it to decide when to rebalance? Should you do it like David Swenson? Should you do it every six months? Should you do it when the deviation away from the policy portfolio is x percent? These are important questions to which there are no obvious answers.

How easy is it to decide when to make changes to the long-run policy allocation and what changes to make? That’s hard, too. Unless you’re Rip Van Winkle, this job has always been a challenge. And while I don’t say market timing is easy, no decisions in this business are easy.

The policy portfolio does, as I suggested earlier, have one very important utility, and that is to function as a benchmark against which to measure active bets. And, it does focus attention on the big picture. But that doesn’t mean you should have a policy portfolio in cement. It should be in glue, so that you can move it around as you want. I’m merely suggesting that you should face up to new challenges, which may be no more difficult than the old ones, and in some ways may be more fun than the old ones.

Three thoughts on portfolio decisions

I offer three ways to think about this. Perhaps as more people follow this line of thinking additional approaches will develop. The first is something that at one point when we first started quant investing was very popular and that is tactical asset allocation, which is a kind of disciplined market timing strategy. It isn’t as popular as it used to be, but I think it deserves a fresh look in an environment where longerrun expected returns on stocks and bonds are not easy to put our arms around.

Second, we can take another step into this more opportunistic environment by moving away from the dreadful process of putting managers into cubbyholes. This is just another outcome of a policy portfolio. We want this kind of manager and that kind of manager—small cap, large cap—you know what I mean —along with a stifling obsession with tracking error. I think
today there is some movement away from cubbyholing managers, and it’s certainly true in nontraditional assets and absolute return.

Third, experiments with nontraditional assets and absolute return in recent years are a strong, positive conceptual development. As these assets do not perform like traditional assets, they require rethinking the roles of the traditional assets in the portfolio.

Overall, I don’t think anybody is going to go back to the office and say, “We’re going to dump our policy portfolio.” But I’d like to provoke you to think about what it is that you’re really doing. And, I’d like to challenge investment committees and investment officers to welcome the opportunity to move away from the beaten path, see what other kinds of solutions, what other kinds of structures, there are where flexibility, creativity and out-of-the-box thinking offer more rewards than doing the same thing forever.

The policy portfolio: Commonfund's point of view

by Lyn Hutton, Chief Investment Officer, Commonfund

Peter Bernstein is, and has been over the last 50 years, one of the most thoughtful and provocative commentators on asset allocation and portfolio strategy for long-term investors. His accompanying article on policy portfolios is worthy of more than one reading. Commonfund agrees with many of the points he articulates in this article, but we have a different take on others.

Bernstein writes: “A policy portfolio is an explicit statement that we have some sense of how much risk we want to take…it is also a reflection of our beliefs about the long-run future.” We wholeheartedly agree with the first statement but question the second: Inasmuch as the future is unknowable, we would ask whether a policy portfolio can or should express a long-term view of the future expected direction of the capital markets. Investors should not confuse the role of a policy portfolio with active management.

We start from the view that there are four basic principles, each of equal importance, for long-term investors:

In our view, and considering the principles articulated above, a policy portfolio is the truest expression of an institution’s unique and specific tolerance for taking the risk necessary to meet its objectives over its planning horizon. A policy portfolio assumes that diversification matters—that it is possible to diversify away some amount of investment risk—and that an equity bias is more likely to preserve purchasing power than is holding wasting assets.

A policy portfolio assumes no active management risk or excess returns from active management. A policy portfolio is not designed to “beat the market” or the peer group—it is devoid of any excess returns from active management of the whole portfolio as well as from active management within asset classes. It does not “express a view” regarding the direction of the capital markets. Rather, a policy portfolio tries to reconcile how much risk investors are willing to take as a matter of policy—their risk tolerances—with the fact that the future is unknowable. Consequently, a policy portfolio assumes that capital markets in which institutions invest are in equilibrium. But as Bernstein has written, “the markets are seldom, if ever, in equilibrium.” Consequently, it is the role of active management to reflect the strategic outlook by asking, Where am I going to be compensated for taking investment risks and are those risks consistent with my long-term policy targets?

As the old adage goes, “A person can still drown in an average of two feet of water,” meaning that even those with a perpetual investment horizon are still concerned with near-term outcomes—three, five or 10 years. Active management deals with this near term. It asks and tries to answer the questions: Is the current “price of risk” or discount of possible future “known” outcomes imbedded in asset prices? Are these prices reasonable and, on the whole, is this price of risk consistent with my policy? Are active management strategies within asset classes consistent with my long-term policy targets, for example, owning an aggressive, small cap, growth-oriented, high equity beta portfolio instead of a large cap, low beta, yield-tilt conservative one while having a significant investment in venture capital partnerships? How much leverage of the absolute return strategies in my portfolio is within my policy range? Given current conditions, have I diversified away as much risk as possible, or does my institution need more or less in the way of inflation/deflation hedging strategies? And, have I fully exploited the long-term opportunities available to me given my long-term time horizon?

Without the framework of the institution’s policy portfolio, active management is blind. Active management requires that we rebalance to the risk tolerances of the policy portfolio. The policy portfolio is a guidepost—not a crutch and not a view of future expectations—allowing active management to make those strategic and tactical shifts between and within sectors that reflect one’s current outlook based on what is known while remaining consistent with the investor’s longterm tolerance for risk.


feedback: mmeditor@cfund.org