To Bond or Not To Bond
A Valuation Approach to Asset Allocation Decisions
As of this writing, first drafted on April 22 2024, the Shiller PE sits at 33.27. Many analysts and investment managers will tell you to fear this number. In his latest memo, Jeremy Grantham says that today’s price-to-earnings metrics sit in the top 1% of modern history, sounding the alarm for U.S. equity bubble territory.
Well, the U.S. is really enjoying itself if you go by stock prices. A Shiller P/E of 34 (as of March 1st) is in the top 1% of history. Total profits (as a percent of almost anything) are at near-record levels as well. Remember, if margins and multiples are both at record levels at the same time, it really is double counting and double jeopardy – for waiting somewhere in the future is another July 1982 or March 2009 with simultaneous record low multiples and badly depressed margins.
I don’t think it’s quite so simple; it might not be appropriate to look at a single asset class in a vacuum the way that many in the investment community do. Is a 30+ PE high? Objectively, it sounds pretty frothy. If bonds were yielding 10%, I’d almost certainly say that bonds were more attractive. If they were yielding sub-2% like much of the post GFC decade, it might not be as straight forward. At a Shiller PE in the low 30’s, we have a very conservative 3% earnings yield (remember, Shiller averages the past 10 years of earnings) before even accounting for earnings growth. One might conclude that stocks have the slight edge in this case.
The point is, we can’t look at a single valuation metric and make an informed decision. We have to consider valuations of equities against the universe of other asset types.
With this post, my aim is to take a more holistic look at valuations - particularly valuation spreads - and see if we can’t make investment decisions based on our findings.
A Simple Visualization
A great place to start with this analysis, and the place that I started when I first began exploring this topic, is a quick visualization plotting stock yields vs bond yields. By doing so, we can start to form a picture on where we are in respect with history.
It’s important to point out what inferences we might try to gage from this chart.
First, intuition tells us that high earnings yields and high bond yields (as defined by the 10-year treasury, in this case) would lead to high forward equity and bond returns, respectively. So the further right on the plot we are, the higher the future equity returns might be. Likewise, the higher (vertically) the point is, the higher the bond returns should be.
With further inspection, the right most points correspond with the years surrounding the late 1910’s and early 1920’s; leading into what has been monikered the roaring 20’s.
1982 is also highlighted on this plot; which was the kickoff to one of the strongest bond and bull markets in history.
These are in-line with our expectations: high returns happen when yields are high. Duh. Don’t worry, there’s more.
More generally, the further up and to the right we are on the aforementioned graph, the better we can expect forward returns to be for a diversified portfolio.
It’s apt to point out that 2022 was basically the inverse of 1982, having the lowest bond & stock yield combination in the modern era. In fact, the post-GFC era was essentially hugging the lower bounds of both stock and bond yields compared to the pre-GFC era.
We can also start to see a shadow of how bonds and stocks might be related. Perhaps when bonds are yielding higher than stocks, stock returns suffer in relation to bonds. We see that the year 2000 (the dotcom bubble top) had equity earnings yields just over 2% (the lowest in history) while treasuries were yielding nearly 7%. We all know how that turned out.
On that note, one might hypothesize that the spread between stock earnings yields and bond yields might be a predictor on how portfolios perform over time. More on that later.
Historical Equity-Bond Spreads
Let’s first define what the Equity-Bond Spread is:
Equity-Bond Spread = (1 / CAPE) - (10 Year Treasury Yield)
Again, the implication is that the higher the equity-bond spread (simply referred to as “spread” moving forward) the more attractive equities are in comparison to bonds (i.e., equity earnings yield of 10% looks more attractive than a 3% bond yield, the spread being 10% - 3% = 7%)
The figure below shows us the historical distributions of equity-bond spreads. Also noted, that today’s valuations lie in the left side of the distribution.
Excess Returns
The goal of this study is to see if we can find some indication on whether the spread between stock and bond yields is predictive of future returns.
The easiest way to accomplish this is to compare a stock heavy portfolio to a bond heavy portfolio. One might argue between something super stock heavy like a 90/10 (stock / bond) vs 60/40. But let’s first look at complete opposites of the spectrum: 90/10 vs 10/90.
We’ll define “excess return” as follows:
Excess Return = (10 year annualized return of 90 / 10 portfolio) - (10 year annualized return of a 10 / 90 portfolio)
As an example, in the year 1990, a 90/10 portfolio had a 10 year annualized return of 13.6% while a 10/90 portfolio had a 10 year annualized return of 5.3%, giving an excess return of 8.3%.
Also, in the year 1990, the Shiller PE was 17.05 giving a equity earnings yield of 5.87%. The 10 year treasury yield was 8.21% at that time. This gives a spread of -2.34%.
The point for 1990 is shown on the plot below at (-2.34% , 8.3%).
The red arrow denotes where we are in 2024.
The big takeaway from this plot is that 1) stocks outperform bonds almost always and 2) there is a decent correlation between the equity-bond spread and excess returns. When stocks yield much higher than bonds, stock heavy portfolios tend to do better, in comparison, vs when the spread is low or negative.
But we already knew that stocks typically perform better than bonds. The better assessment might be when to overweight stocks compared to a more traditional portfolio. Or, better yet, when to take the foot off the gas on a stock heavy portfolio. So let’s do the same exercise, this time comparing a 90/10 to a more traditional 60/40.
I’ve left the original 10/90 comparison on the plot for the visualization. As expected, the excess returns, across the board, are less pronounced because we’re comparing a stock heavy portfolio to a slightly less stock heavy portfolio. But the conclusion is clear. The spread does appear to have an impact on excess returns. In negative spread environments, we’re not paid nearly as much for the extra risk as when spreads are positive and wide. In highly positive spread environments, excess returns can be in the range of 3 - 5%. Which, we all know, can be very impactful over the long-run.
Understanding Valuation Drivers
For bonds, valuation is pretty easy: an investor can purchase a bond for a given yield-to-maturity (although returns on bonds aren’t quite as simple).
For equities, we should examine the components of the discounted cash flow model.
In the long run, a PE ratio might be estimated as follows (this is the terminal value equation):
Above, “g” is the long run earnings growth rate, and “d” is the discount rate. In the case of price-to-earnings, “d” will be the cost of equity. I won’t cover these more in depth here because this is a very simplified look, but cost of equity is essentially a measure of risk or the required expected return for the asset.
From this, we can actually glean a lot of useful information.
If the security is considered very safe (ie low risk), the discount rate “d” will be low (since the required rate of return is typically lower for a safe asset). A low discount rate in the equation above will lead to a higher PE ratio.
Conversely, a risky security will have a high discount rate, which will lead to a lower PE.
A high long run growth rate, “g”, will increase the numerator and decrease the denominator, leading to a higher PE for a given discount rate.
From these three ideas, we see that risk and growth are comingled in valuations. Something that’s low risk and has low earnings growth might actually have the same high PE valuation as something that’s high risk and high earnings growth. But the expected return will actually be higher for the high risk security.
This all just to say that while PE ratios are related to forward expected returns, they don’t tell the full story. This is an important caveat to the next section.
Current Valuations By Asset Classes
The following data was pulled from Vanguards Website.
VOO = S&P 500 BND = Bond Index
VEA = Developed International VNQ = REITs
VWO = Emerging Markets
This chart isn’t meant to be used to decide what asset mixture to make your portfolio. Instead, it’s meant to be used, qualitatively, as a starting point to see what asset mixes might make sense to hold.
Typically, in terms of valuations, the further up and to the right (high starting yield + high earnings growth) on this graph indicates higher predicted forward returns.
But there are trade offs. Namely, this doesn’t account, directly, for risk. Bonds (BND) is considered ‘risk-free’, but it doesn’t offer any potential for earnings (or coupon) growth. Developed international (VEA) looks attractive compared to the S&P 500 (VOO) on a starting yield basis, but it has offered less earnings growth, and comes with extra baggage in terms of geopolitical risk. But high risk does typically mean higher potential returns. The same goes for Emerging Markets (VWO), but to an even greater extent.
Does History Have to Look Like the Past
Something else to consider, especially when looking back at the first couple of sections, is “does today have to look like the past?” Do current market environment have stocks overvalued, or is it that historic valuations had stocks inordinately undervalued?
Maybe stocks aren’t as risky as we first thought. Especially in the U.S., the largest companies might not carry a ton of risk at all. In that sense, maybe it was the early days of modern capitalism that were inefficient, and we’re now getting to a more balanced regime in terms of valuations, where risk-free bonds yield in the 3-5% range, and slightly riskier stocks return in the 5-7% range. In this case, the current spread environment would make sense, where starting yields are much closer, and the earnings growth potential of stocks makes up the difference in forward expected returns. But this would be all the more reason to hold a diversified portfolio. Why hold only stocks, when stocks and bonds will give a similar range of outcomes.
Stocks also offer other advantages over bonds. Namely inflation protection. If inflation spikes, bonds an investor is currently holding will not only lose value due to rising interest rates, but the purchasing power of the dollars tied to those bonds will decline over time. Stocks are somewhat more resilient in that revenues and earnings (assuming steady margins) will rise with inflation. In this sense, stocks are actually less risky than bonds or cash.
Inflation also affects the spread in another way. The CAPE ratio uses inflation adjusted earnings from the past. What this means is that in a high inflation environment, the CAPE ratio comes down without any correction in price. We saw this in 2022 where the CAPE fell nearly 30% while the S&P 500 only fell 18%. Due to this phenomena, in a high inflation environment, the metrics used above can correct themselves even while equity prices are climbing.
Another potential issue with this study is that accounting standards have changed over time. Earnings today may not be comparable to earnings of the past. I haven’t explored these potential differences here, but it might be prudent to do so if you were to use this study for actionable advice.
Conclusions
Are we in a Bubble?
To give Jeremy Grantham a rebuttal (although, I’m sure he’s not asking for one). No, I don’t think we’re in an outright bubble. U.S. markets might be frothy, and forward returns will probably be lower for U.S. stocks, but we’ve seen in the data above that 10 year returns have been fine given any market spread and valuation. Would I be surprised if we had another bear market? No. But I’d be just as un-surprised if we average 6-8% equity returns for the next decade.
Asset Allocations
To me, when presented with the data above, it doesn’t seem likely that we’ll be rewarded for holding an overweight U.S. equity portfolio. While equities should continue to outperform bonds for the next ten years, if today’s environment rhymes with history, holding an underweight stock portfolio won’t cost us much in terms of returns. But it may come with the added benefit of lower volatility and overall risk. An underweight portfolio also still has some potential to outperform. That all seems like a good trade-off.
In addition, international (both developed and emerging) markets have relatively enticing valuations and return prospects. While there’s no guarantee that either will outperform U.S. equities, they may offer uncorrelated returns that also won’t drag too much on the overall portfolio.
In general, given the current valuation environment, a balanced portfolio might be the best path forward for risk adjusted returns.
Citations
Shiller PE and Treasury Yield Data:
https://www.multpl.com/shiller-pe
Historical Return Data:
https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/histretSP.html