The discussion is cordial and ends with a friendly wager, but John Hussman and Gus Sauter find little common ground.
Valley Forge, Pa., and Ellicott City, Md., may be small towns, but they are the breeding grounds of some big investment ideas. Valley Forge is home to Vanguard, the client-owned firm that manages $1.4 trillion in more than 100 U.S. mutual funds. Ellicott City is where John Hussman, a former economics professor, and his firm, Hussman Econometrics Advisors, have operated since 1989. With their combined assets of $7 billion, Hussman's two funds--Hussman Strategic Growth
While Vanguard has branched out into alternative strategies like long-short and complex asset-allocation vehicles such as its Managed Payout Funds in recent years, it is still regarded as a bastion of traditional, sober, long-term investing. Meanwhile, Hussman's research, which he often publishes on his Web site, frequently challenges conventional wisdom. We thought that getting Vanguard CIO Gus Sauter, who has been with Vanguard for more than 20 years and oversees all of its internally managed funds, and Hussman together to talk shop would result in an interesting Morningstar Conversation. We were right. It was one of the liveliest interactions we've had.
Their exchange on Nov. 3 ranged widely over alternative asset classes, global asset allocation, inflation, and current market valuations. The two differed sharply on many topical subjects, such gold's suitability as a long-term investment, the prospects for higher inflation, and the relevance of the concept of equity duration to asset- allocation strategies, including target-date retirement funds.
Equity duration is an academic idea that has been around since the early 1980s, but it has descended from the ivory tower in recent years and influenced investors, including Hussman. Duration, a measurement usually associated with bonds, measures interest-rate sensitivity and weighted average time to receipt of the bond's cash flows. In bonds, the higher the yield to maturity, the shorter the duration will be.
It turns out that you can also calculate the duration of the stock market. In general, as the market's (dividend) yield rises, its duration decreases, and vice versa. Hussman uses this measurement, among other things, to argue that the glide paths of target-date funds should be dynamic even for an investor with no particular view about market conditions or future returns. Vanguard, like most other target-date fund providers, uses static glide paths.
The conversation has been edited for clarity and length.
Ryan Leggio: Let's start off talking about alternative asset classes. Gus, what types of alternatives do you believe are most appropriate for mainstream investors?
Gus Sauter: Commodities, REITs, and Treasury Inflation-Protected Securities are alternative investment classes that make sense in many people's portfolio allocations.
At the same time, we're looking at other types of alternatives as well, primarily on the beta side of the equation. We've looked at the hedge fund industry, and while we believe that hedge funds offer more beta than people realize, we still don't think that it's the right strategy for us. Other strategies we're looking at are still in the research phase, but we like alternatives in general.
RL: Do you include gold in commodities, or is it a separate asset class?
GS: Gold is buried in that commodity bucket. Gold, honestly, is a foreign-currency play, particularly the way it's performing today. Gold is not responding to a fear of inflation; it's responding to a weak dollar and has rallied because of that. So, gold is certainly not something we would recommend as a long-term investment.
We have stayed away from foreign-currency plays because, in the long run, we don't expect there to be any return to foreign currencies. We think it's a net-zero-sum game; the only way to benefit from it would be to try to time it. We just don't have any confidence in our ability to time it.
I should also point out that there's a difference between investing in spot commodities and commodity futures. The rationale behind investing in commodity futures is supported by the theory of normalized backwardation, and that's our thought process behind commodity futures as being a good diversifier with reasonable returns in a portfolio. If you go back 50 years, you'll see that a broad basket of commodity futures has earned an appealing rate of return, somewhat in line with long-term equity returns with similar levels of risk. That is fully in keeping with the theory of normalized backwardation. One example is that from the early 1980s to the mid-1990s, energy prices were down something like 5% on an annualized basis, but an investment in energy futures returned about 7% per year.
RL: John, what are your thoughts regarding alternative asset classes?
John Hussman: For the typical investor, I don't believe that a standard commodity allocation, a standard TIPS allocation, or a standard allocation to a lot of passive classes is a good idea. If you look at precious metals and precious-metals shares over the long run, they tend to have a negative beta. You would expect that the long-term return on those kinds of assets would look a lot like the long-term return on anything that has some insurance properties, which is that they provide relatively low sustained returns. So, I don't think investors do themselves a lot of good by having as a passive asset class a fixed allocation to precious metals or to a wide range of alternatives.
On the other hand, in one of our funds we do maintain periodically an allocation to precious-metals shares, TIPS, foreign currencies, and so forth, and those allocations are specifically driven by valuation and market conditions. For example, you will find fairly consistently throughout history that precious metals do best when there's downward pressure on real interest rates, particularly when precious-metals shares are favorably valued relative to bullion prices. So, appropriate signals to hold some allocation to precious metals are downward pressure on nominal rates, upward pressure on inflation, generally weak economic conditions, and a relatively low ratio of gold stock prices to the bullion.
The primary objective of a financial advisor and a long-term investor should be to align a portfolio with the likely liabilities that the investor will incur over his or her lifetime. For somebody who anticipates a significant commodity demand over their lifetime, a fixed allocation to commodities may make sense. For the broad class of investors, a group of stocks, bonds, and cash is a reasonable portfolio. In an increasingly globalized environment, you could add foreign-stock mutual funds and foreign currencies to some extent, and as a means of managing fluctuations in purchasing power, possibly precious metals and TIPS. Once you go beyond that group, I think it gets a little more esoteric for the average investor.
GS: In trying to create an asset allocation for an investor, part of it is related to future liabilities or expenditures, but that's an argument for international investing, as opposed to domestic investing. The bigger picture that you're trying to solve is producing attractive returns with low volatility.
GS: That's where commodities come into play. You don't buy commodities because people are going to have a demand for commodities in the future. You buy commodities because you think they're going to provide you an attractive rate of return, albeit with a lot of volatility.
JH: Yes, although if you look at the long-term returns of commodities and their volatility, you have to rely on an extraordinarily negative correlation to make a fixed allocation to commodities create a reasonable increase in the return/risk profile of a portfolio. Commodities can be useful, but I would view them as more opportunistic than as a fixed allocation.
I agree with Gus that what we're observing right now in gold--at least as of the time of this conversation, when gold is trading at $1,086 per ounce--is primarily a reflection of a weak dollar. But I'm not so sure that I would partition that as not being a fear of inflation. It certainly is not any near-term fear of inflation. We still have a very weak economy, and we still have some credit concerns that are creating willingness to accept an extremely large issuance of government liabilities.
But over the longer term, that extremely large issuance of Treasury debt, the monetary base, and general government liabilities that we're seeing as part of this bailout effort are likely to not only depress the foreign currency value of the U.S. dollar but result in pressure on the purchasing power of the dollar, which is another way of saying inflation.
The Inflation Debate
GS: I am not at all concerned about inflation--not near-, medium-, or even long-term. Where we get a good indication of the marketplace's analysis of inflation is by looking at TIPS. Right now, TIPS are projecting about a 2% inflation rate out for a decade, so there isn't a fear of inflation in the market.
I would point out that the creation of debt by the government does not lead to inflation; it is only the monetization of that debt. I could be proven wrong, but I don't think the Fed is going to let that happen. The Fed has two goals: to maintain moderate price increases and to maintain full employment. If the Fed lets inflation get away from it, it can't fulfill either one of those requirements.
JH: Gus, I don't agree. If you look at the relationship between monetary base growth, for example, and inflation, you actually observe a very poor correlation between the two, even if you take it over four-year periods. Where you observe a strong correlation is between government spending and inflation. The reason is that it does not matter how a government funds its spending, whether it's through monetary creation or through Treasury issuance, because Treasuries and base money compete almost perfectly in the portfolios of individuals. If you have too much of them, their marginal value goes down.
What we see is a fairly good correlation between four-year growth in U.S. government spending and four-year rates of inflation that is much stronger than you observe for the monetary base.
The other issue about monetization is, what does a doubling of the monetary base over a period of a year reflect other than monetization, unless the Federal Reserve is actually able to unwind the enormous portfolio of mortgage-backed securities that it's taken on at the prices that it took them on at? I'm not so sure that we're likely to observe a successful unwinding of that portfolio, because from what we observe in terms of defaults, the rate of defaults on mortgages continues to remain high and is likely to accelerate with a new reset cycle in Alt-A and option ARMs.
We've got a Treasury that's issued an enormous amount of debt, and we've got a Federal Reserve that's issued an enormous amount of monetary base. I'm not convinced that those dollars are simply going to go away and be reabsorbed, because that would require that the assets that government agencies and the Federal Reserve took on are actually worth every dollar that they paid for them.
RL: But what about Gus' point? According to the break-even rate--the difference between nominal Treasury yields and TIPS yields--the market is not expecting inflation.
JH: If nominal yields are low because of other factors, particularly risk aversion and the strong demand for default-free securities, you may mistake low yields for low inflation expectations, where what you are really seeing is a flight to safety and, in particular, a flight to safety from default. Unless investors bid TIPS prices up to the point where their real yields are deeply negative, which forces them to accept poor returns in the short run in order to protect against inflation in the long run, inflation expectations can't be properly reflected in the yield difference between Treasurys and TIPS here.
If you look at the relationship between historical TIPS inferences and actual rates of inflation, I'm not convinced that there's nearly enough variability in that inflation-expectations component to account for inflation rates that we actually observe in practice.
Duration of Equities
RL: John, you've written a lot about how advisors and investors should factor duration into their asset allocation. Please describe the concept.
JH: One of the things we think about in terms of setting up portfolios is that the overall duration of a portfolio--the effective life of the assets--should roughly be matched to the period over which those assets will be exhausted. I'll give you an example that I observed once where this was not the case.
In the late 1990s, there was an institute for disabilities that was being built, and they had an endowment that had been created to provide for the construction. The construction was going to be done over a period of two years, yet the bulk of the portfolio that was funding the construction was in common stocks. I was on the board but not on the investment committee at the time, and I was absolutely out of my mind trying to have them change that allocation. Unfortunately, the endowment, over the course of building, wasn't available, and they had to scramble to raise new funds during the 2000-2002 bear market.
So, one of the things that is extremely important for investors, particularly as they approach retirement, is to think about appropriately matching the character of the assets that they hold with the horizon over which they'll need the money.
With bonds, we call that duration. Duration is effectively the average date over which you get your present value back. For example, the maturity of a 10-year Treasury bond may very well be 10 years, but if you look at the fact that you're getting some amount of coupon payments every year in advance, the actual duration of that bond is closer to seven years. Depending on the stream of payments that you're getting as you're holding that asset, the larger the payments, the shorter the effective life of that investment is. For bonds, you can figure that out. In fact, you can look it up on a Bloomberg terminal.
For stocks, an equivalent calculation can be made. It's difficult to do this with individual stocks, because payouts can change over the life of a stock. But the effective duration of common stocks taken as a group is very close to the price/dividend multiple. For example, from the postwar years until the early to mid-1990s, when things started getting very overvalued, the dividend yield on stocks averaged around 4%. A dividend yield of 4% gives you a price/dividend multiple of about 25, which means that if you had 100% of your money in stocks, your overall horizon is effectively about 25 years. If you had a horizon of 25 years, you didn't have to worry much about where stocks would go in the interim. You could be comfortable that your terminal value wasn't going to depend enormously on the ups and downs of the stock market over your holding period.
On the other hand, as we got into the high levels of valuation up to the 2000 peak, the dividend yield was about 1.25%, which meant that the duration of stocks was about 80 years. In other words, if you were drinking from a sippy cup at the time, it may have made sense to have all of your assets in stocks and still have some confidence that it wouldn't matter where stocks went over the lifetime. But if you had a horizon of less than that, it mattered enormously. People found that out the hard way, not only in 2002, but in the decline that we've recently had.
You want to match the overall duration of your assets to the lifetime over which you expect to use the money.
RL: Gus, do you think that John's analysis that the overall allocation of a portfolio may have to change dramatically with the underlying valuation of the market is appropriate for advisors and investors?
GS: In concept, yes, absolutely. We could perhaps come up at different endpoints using different methodologies, but the idea that as your time horizon shortens you need to adjust your portfolio is absolutely recognized by all investment advisors.
When we measure the duration of a bond, what we're really seeing is the volatility relative to changes in interest rates. The best example of using duration is for something that's truly liability-driven, like a pension plan, because there we know that the liabilities are driven by changes in interest rates; therefore, you want to make your investments ... very sensitive to changes in interest rates as well. A lot of corporations are moving to liability-driven investing, and it makes all the sense in the world.
Applying that to the individual investors is more difficult, because your liabilities aren't just changing with interest rates. It's the fact that you can't afford the same level of volatility as your time horizon is shortening. So an asset allocation should be derived based on the circumstances of the individual, and it should be adjusted as those circumstances change.
An Algebra Concept
RL: Does moving the glide path of target- date funds with duration make sense? Right now, the glide path of a 2055 fund would be exactly the same if equities had a 40 P/E ratio or a 10 P/E ratio.
GS: I am not familiar with any work that says that the volatility of equities goes up as dividend yield goes down. That's really the important thing that investors need to solve. They need to moderate volatility of their portfolio.
JH: But duration is, in fact, strictly a mathematical measure of volatility. It's the elasticity of the security price to changes in the underlying rate of return. Let's put it this way. If you have a 100-basis-point increase in a 2% dividend yield, by definition you've lost a third of your value, because you've gone from a 2% dividend yield to a 3% dividend yield. On the other hand, if you go from a dividend yield of 4% to a dividend yield of 5%, you've lost a much smaller percentage of your assets. It's just a strictly mathematical relationship. You take the old dividend yield divided by the new dividend yield and subtract one.
GS: Well, that holds with bonds.
JH: No, it's a mathematical fact for stocks, too. If the dividend yield on the S&P 500 goes from the current 2% to 3%, an investor, barring massive dividend growth, is going to lose about a third of his money. It's not a bond concept; it's an algebra concept.
GS: You're assuming that the payout ratio is constant, and in fact .
JH: No, I haven't said anything about earnings.
GS: But you can easily get to a different dividend yield just by changing the payout ratio without changing .
JH: That's why I said that I'm assuming that we won't have a substantial variation in dividends, which we're really not observing in terms of the relationship between dividends and normalized earnings. That payout ratio is not highly variable at all.
GS: Payout ratio has gone down, and stock buyback has gone up.
JH: But stock buybacks are a way of reducing the float as a result of corporate insiders handing their earnings over to themselves and not wanting it to appear in the float. We can argue about the notion that buybacks are a distribution to shareholders, which I would strongly argue against. But the fact is that you're not seeing an enormous amount of retained earnings less buybacks, which is actually what investors are going to get outside of the dividend yield. If you were, you would see book values expanding much faster than you've observed over the past decade or two.
GS: You don't see that high a correlation between dividend yield and price/earnings ratio, and you also haven't seen that high of a correlation between dividend yield and volatility of equities.
JH: The reason you don't observe it in the price/earnings ratio is because earnings year to year have enormous amounts of noise. You have to normalize these things. For example, right now we're looking at a trailing 12-month figure for S&P 500 earnings of something around $7. That's certainly not anything close to the normalized level of earnings that would be expected to be sustained over the long run. But it certainly messes with the price/earnings ratios in ways that make the correlation with dividend yields disappear. If, on the other hand, you measure earnings yields on a normalized basis and compare them to dividends yields on a normalized basis, you actually get extremely strong correlations.
GS: You see a secular decline in dividend yields, and that's a phenomenon of the past 15 years.
JH: But the accompanying phenomenon of the past 15 years is that stocks have delivered bupkis. The reason we have not achieved any meaningful return in stocks is precisely because valuations have been enormously rich, and low dividend yields and high-normalized price/earnings ratios have been an extremely good signal about the prospective long-term returns on stocks. You could have observed that back in the late 1990s, when on the basis of any reasonable historical expectation for long-term earnings growth, and any reasonable terminal price/earnings ratio, stocks were priced to deliver returns well under the risk-free Treasury bill rate, which, in fact, they have.
GS: That was a time period dependent in the latter part of the 1990s, but it wasn't true in the early part of the 1990s, and it wasn't true in 2007 either.
JH: It wasn't true in the early part of the 1990s, though, because we were about to embark on a decade that took stocks from a well-below-average multiple to a price-to-normalized-earnings multiple of about 34, where previously every bull market had topped out at about 19 or 20 times normalized earnings. It wasn't true in the decade of the 1990s, but that was because the decade of the 1990s was the precursor for horrifyingly poor returns and experiences for the average investor.
GS: Well, it doesn't hold up with 2007-- that year was a horrifying experience, and the P/E ratios were about 16, 17 times .
JH: Not on the basis of normalized earnings. The problem with 2007 was that profit margins in 2007 were about 50% above the historical norm, so they made observed price/earnings ratios look quite attractive. But then, look what happened. What we observed in 2007 was a relatively rich earnings multiple applied to profit margins that were 50% outside the norm.
So, it is true that we didn't have the kind of extremes that we saw in 2000, but that was driven by the aberration in profit margins, which have been quickly normalizing during the recent downturn and, without the benefit of resumed leverage, are unlikely to go back to their 2007 peak.
GS: The issue is whether or not this leads to extra volatility in the stock market. We haven't seen a significant change in volatility in the stock market as yields have gotten to this low level. We went through historically low periods of volatility in the mid-1960s as yields collapsed and through historically low periods of volatility before this particular collapse. We've had extremely high periods of volatility in the 1970s at the same time dividend yields were high.
JH: In the 1970s, yes, we had extremely high volatility and dividend yields were high. Why were dividend yields high? Because stock prices had collapsed and because dividend yields were previously extraordinarily low.
So, yes, by the end of that decline, it is very correct that we saw high dividend yields and high volatility. But that high volatility was generated by the fact that stocks lost half of their value in 1972-74.
GS: The market actually recovered in 1975. It was as high as it was in 1968. The demise of stocks in the 1970s is greatly exaggerated. The market went virtually straight up--204%--from the end of 1974 through 1980.
JH: And then by 1982, it had settled back.
Now, by the way, between the 1974 crash and 1982, when dividend yields on the S&P 500 shot up to about 6.7%, it is very true--that by the end of 1982, you had an enormously good long-term buying opportunity for stocks. Even if you were to observe no change in stock valuations over the future, if you were to just have dividends growth at the same long- term rate as earnings, which is about 6.25%, you were priced to achieve long-term returns near 13%.
So, yes, by the time that you got extremely high dividend yields, you had gone through an enormous amount of pain, and that was amply rewarded over the next two decades. But to say that low dividend yields don't mean trouble ultimately is counter to the historical evidence.
GS: No, I'm not trying to predict returns. I'm talking about volatility that you experience.
JH: Low dividend yields have to be associated with higher volatility over time--at least if you measure the amount of loss that you end up achieving. Downside volatility in particular is horrific after periods of rich valuations, which are also characterized by low dividend yields. If you look at the historical record, it's very hard to controvert that.
GS: We have low dividend yields right now.
JH: Which I'm worried about.
GS: And I feel very comfortable with.
RL: Where do the both of you see relative valuations right now?
GS: A lot of people have asked, what's the equity risk premium looking forward? Is it zero? Is it negative? Is it small? Or is it the historic norm, with the historic norm being in the 5.5% to 6% range? I would say that, on average, the equity risk premium is at historic norms all the time. So, I think that we're looking at average rates of return going forward, and that's based on the concept that we're rewarded for investing in stocks because of the risk inherent of investing in stocks. If we weren't going to be rewarded for that, we'd sell stocks, and we'd sell them down to a price that made them attractive again. In fact, that's what happened from the end of 2007 to the beginning of 2009.
I believe we're going into a slow-growth U-shaped recovery, but I think we can still get a reasonable rate of return from stocks, because the price was driven down to a point that makes the future returns reasonable. So, we are at a policy benchmark weight in equities in our Managed Payout Funds, thinking that we'll get historical rates of return over the intermediate to longer term.
RL: Are you saying that the other asset classes in the funds are at some type of reasonable valuation range, including long-term Treasuries and TIPS?
GS: If there's a bubble, it is in Treasuries. There has been a tremendous flight to quality, and having taken on credit over the past year was a good investment move. There's still some advantage to taking credit risk above and beyond Treasuries. We do think that the spreads between Treasuries and credit will continue to narrow, but they will by Treasuries backing up--not in the next month or three months, but over the next year or two.
RL: John, what is your thinking?
JH: On the stock side, our view as of the present, which is about 1,041 on the S&P 500, is that the S&P 500 is priced to deliver total return over the next decade somewhere in the region of 6.5%.
The way that we get our longer-term projections of total returns on stocks is fairly straightforward. If you look at S&P 500 earnings over the long run, you'll find that they have historically behaved in a very well-defined growth channel of just over 6% annually. You can draw that on logarithmic paper as far back as you care to look. Then you look at where earnings are, anticipate that over the next decade you'll get to about midchannel earnings, and then apply a historically reasonable range of P/E ratios to those terminal earnings, anywhere between the extreme low of seven, which is what we saw in 1974 and 1982, to a multiple of 20, which is where most historical bull markets have run into a great deal of trouble. You look at about the midrange of those estimates, and you're at about 6.5% long-term returns.
In terms of the bond markets, we have cross currents there, because I am convinced that the enormous amount of issuance in government liabilities that we've seen over the past year, and anticipate seeing for some time, will eventually be accompanied by a reduction in the willingness to hold those assets. But that reduction in willingness is likely only to occur after we've seen some amount of stabilization in credit conditions. Our view is that credit conditions are still deteriorating, and we're about to get a second wave of mortgage defaults based on the reset schedule that's been in place for years. Near term, we're not likely to observe sustained upward pressure on bond yields. In fact, we may get additional flights to safety like we saw in 2008. But over a longer horizon beyond the next four to five years, that's when we're likely to observe more difficulty with inflation.
GS: We've got to have a $10 bet. The difference between your 6.5% return on equities versus my 10%. I say it's going to be greater than 8.25% 10 years from now; you say less than 8.25%.
JH: If we mark those expectations against 1,041.52, which is where the S&P is as we speak, you have a bet. Rather than $10, let's wager lunch.
GS: All right.
JH: And Gus, I really enjoyed this. I hope you don't take any of my disagreements personally.
GS: Nor mine with you.
Ryan Leggio is a mutual fund analyst with Morningstar.