Last year the return on the S&P 500 was ~30%, quite the performance. However, that's in the past and we need to look forward. In driving up the market we have bloated valuations that have historically led to poor long term expected returns. I consider myself a long-term investor and I am not a slave to the daily market gyrations. Prudent, long-term investors plan to capture long term risk premia via a number of persistent factors that have historically delivered significant excess risk-adjusted returns. However, from current valuations, future returns don't look very promising for long-term investors. So what's an investor to do?
I am going to limit myself to two asset classes: equities and fixed income. They are generally considered competing assets and are important to all portfolio construction strategies.
Lets begin with equities. Since 1881, the market has traded around 16.53x trailing ten-year earnings. It currently stands at 25.33 or about 50% overvalued. This is startling for long term investors.
It is clear that the market is very expensive and is closing in on the 2007 valuation.
However, from the early 90s to the present, valuations have been persistently higher than their long-term averages. Has something changed? Can current valuations be explained or even justified? Yes to the former, maybe to the latter. We will get to that later.
In the long-term, returns are driven by valuation and Federal Reserve policy. In the short-term, markets are driven by sentiment, both rational and irrational. There are a number of valuation metrics that have had strong forecasting power for long-term returns. These include:
- Tobin's Q
- Market Capitalization to GDP ratio
- Price-to-Peak earnings using peak trailing 10 year earnings
- Price residuals
Here is the reliability of such valuation metrics:
It is clear 6 out of the 7 are highly explanatory and significant in the statistical sense (boasting very high t-statistics). Also, ten-year returns are negatively exposed to these valuation metrics. In other words, the higher the valuation, the lower the expected returns.
Also, in defense of the CAPE, the real 10-year earnings for the S&P 500 are above trend. This contradicts the claim that the measure is distorted from the disastrous collapse in earnings during the financial crisis.
Using these four indicators we can come up with an approximation of forward ten-year returns. We have decided to use these four indicators to construct a forecast based on long-term data as well as short-term data. Pundits claim that using data from 1881 is no longer relevant because markets have changed and have no bearing on future returns. So we will accommodate both sides of that argument. Utilizing the results from our time-series regressions, both short and long-term, our forecasted 10-year annualized total returns are as follows:
The average is a 3.74% annualized nominal return for the next ten years. Using the Survey of Professional Forecasters, who are projecting a 2.3% inflation rate over the next ten years, the real return over the next decade becomes 1.44%. Not very promising if your expecting equities to deliver historical premiums.
Before I sound too much like a valuation bear, I also want to present work done by Philosophical Economics.
From 2003 until 2013, the annualized nominal 10-year total return on the S&P 500 was 7.3%, which is certainly strong given the second worst financial crisis in history. So remember, while I think returns going forward will be less than 7.3%, that doesn't mean investors can't make money in the market.
Also, earlier I had mentioned Federal Reserve policy. As everyone knows, they are highly accommodative. They are tapering but that is not the same as tightening. Overnight rates are still 0%. With inflation below their target, high unemployment, and a poor demographic outlook it's hard to believe that when the Fed decides to raise rates they will do so significantly. This seems to coincide with the Fed model, which accepts a lower earnings yield, as interest rates remain low. This is only true historically when controlled for investor's perception of risk. Below we will discuss this in detail.
Lets take a look at fixed income.
The best valuation metric for long-term bond yields (using the U.S. 10-year Bond) has been current real yields.
The 10-year real bond yield currently stands at 1.2%, about half the historical average. It is evident that bonds are also expensive.
Bonds are expensive and stocks are expensive. However, as investors, we need to make choices and seek the most promising long-term investments. So which is it, stocks or bonds? Lets take a look at another model that has been a reliable predictor of the relative returns between stocks and bonds.
We can see again, bonds are expensive and stocks are expensive. However, relative to each other they are about average. Again, stocks have demanded a risk premium because they are inherently riskier. So over long periods of time stocks are expected to outperform bonds. The current spread is 2.8%, which implies that stocks are going to outperform bonds over the next ten years by about 2.8%. This doesn't neglect the fact that expected returns for equities going forward are below average, it just means that they are going to do better than bonds. Now this is the current real yield, in order for us to compare our forecast to 10-year bond yields we must adjust for future inflation. The current yield on the 10-year is 2.5% and our forecasted inflation is 2.3% becomes a .2% annualized return. This is consistent with GMO's long-term fixed income forecast. With our forecasted 1.44% real total return on the S&P 500 and .3% forecasted real return on bonds, taken at face value, equities look favorable relative to fixed income.
Additionally, with some adjustments, this model can be used to estimate long-term returns for a traditional 60/40 portfolio (as constructed by Cliff Asness and Antti Ilmanen).
Again, expected returns are exceptionally low.
In contrast to my prediction, pundits will cite the Fed Model. The Fed model attempts to justify high market valuations given the secular decrease in interest rates.
What they generally show you is the period from 1960 to the present. Looking at the period prior to that, it is evident that no relationship exists. Proponents will claim that if the earnings yield equals the 10-year yield then stocks are fairly priced and bonds are fairly priced. If stocks are yielding less than bonds then they are expensive and vice versa. However, it would be more accurate to say that stocks relative to bonds are cheap or expensive. For instance, currently stocks and bonds are both expensive, but relative to each other they are about average. I am in the Cliff Asness camp that finds it irrational to place a valuation on equities or fixed income using the Fed model. It leaves out investors perceptions of risk. I have replicated Asness' study "Fight the Fed Model". His data runs up to 1998 whereas I include data up until the present.
Through time-series regression, using the period from 1929-present, the 10-year yield explains about 2% of the variation in the earnings yield with a t-statistic of 5.9 and a correlation of .17. The Fed Model does poorly when looking at the entire time period.
Now lets adjust for risk perception. I will accept Asness' trailing 20-year volatility measure to gauge investor's perception of risk. This is measured for both stocks and bonds.
After adjusting for risk, the correlation becomes .62 and the R2 becomes .4.
The results are highly significant and logically make sense. As equities become more volatile, the earnings yield moves higher (or a lower P/E). As bonds become more volatile, the earnings yield moves lower. Controlling for risk, we can see that the relationship between stocks and bonds holds. Again, this is after we have controlled for investor's perception of risk.
For a visual, here is how trailing 20-year volatility has looked over history:
So am I justifying the high valuation multiples? Now I am. I am also answering the question "why are investors pushing valuations so high on equities, lowering their expected returns?". Bonds have become quite volatile and it appears investors are willing to accept lower volatility in equities for below average expected returns.
So what does this all mean?
I have forecasted stocks to outperform bonds and comparing real CAPE yields to real 10-year yields appears to predict the same. Also, the volatility of bonds has exceeded that of stocks so adjusting for investor's perception of risk the Fed Model holds. Therefore, low rates imply a lower earnings yield. So in that case I am justifying current market valuations. Not simply due to earnings yield versus 10-year yields, as many mistakenly do.
In terms of strategy I think it's important to look at the views of two highly respected individuals: Warren Buffett and Cliff Asness.
Warren Buffett on bonds:
"In terms of…bonds, some day they will sell the yield a whole lot more than they're yielding now…I don't know…when it'll happen…the question is always when…but you could have interest rates very significantly different than what they are now, in some reasonable period in the future…I like owning stocks. I do not like owning bonds now. There could be conditions under which…we would own bonds, but they're conditions far different than what exist now…You shouldn't be 40% in bonds…I would have productive assets. I would favor those enormously over fixed dollar investments now, and I think it's silly to have some ratio like 30 or 40 or 50% in bonds. They're terrible investments now…I bought bonds back in…the early 80s…We made a lot of money and we bought zero coupon bonds…The price of everything determines its attractiveness…You've got some guy buying $85 billion a month…and that will change at some point. And when it changes, people could lose a lot of money if they're in long-term bonds."
Cliff Asness on diversification:
[What is] one investment for a long-term diversified portfolio? What would you have us all own some of?
"I am going to be very counterintuitive, and this is not a forecast. I am actually picking an expensive investment, one more expensive than stocks. I'm going to pick the bond market. And I don't want anyone to listen to your show and go 'god, Asness's forecast - he loves bonds!' Having said that, people understate the power of diversification.
Take the 1970s. The 1970s was a disastrous period for bonds (the decade). A portfolio that took equal risk in stocks, bonds, and commodities, something like we might prefer, did better than all stocks, and it had way more bonds."
"In the 70s?!"
"In the 70s. One thing: commodities were strong, which helped, which in an inflationary period, …when you're really going to see your bond disaster, is not a certainty, but is not a bad bet, I think.
Second, the power of diversification is that strong…having three different horses even if one doesn't work. You know, if you commit yourself to diversification…the glass is half empty way to view it is you're always in the worst thing, but the glass is half full way is you're always in the best thing. Turned out to be commodities that decade.
If you look over the long term, we think balanced risk [can make sense], even when rates mildly rise. We look at periods from the 40s to the 80s. You know your interest rate[s], they mildly rose, and then they shot up in the 80s, and they've been coming down ever since until very recently. Even over that rising period, having a relatively equal amount of risk in bonds, not even dollars, worked better than traditional approaches. So this is not a short-term forecast. I am certainly not sitting here telling your viewers, 'here is an undervalued asset - bonds."
"But don't abandon bonds, and always have a portion in your portfolio."
"Exactly. And I love it because it's counterintuitive, and it gives me a chance to beat to death what is almost always our most important theme: diversification beats timing."
I like the risk premiums that can be generated over the long-term with various investment styles but I also believe in allocating risk. Being heavily invested in equities at these valuations is a poor allocation of risk. Also understand that risk premiums have become very narrow at these valuations and that:
"No asset (or strategy) is so good that you should invest irrespective of the price paid"
In sum, don't abandon equities altogether because of valuation alone. It is a strong predictor of future returns but there are a number of other dynamics at work. Expect lower than normal long-term returns and allocate risk and assets appropriately to protect yourself from the imminent downturn in equities over the next ten years.
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.