6.2 Expected future returns and risks
In Article 6.1, the story was based exclusively on the history of returns and risks. You have probably seen the disclaimer on financial websites and statements that “past performance is no guarantee of future results”. While we learn history in part to avoid repeating past mistakes, our decisions must also consider whether any particular historical mistakes are even relevant to our current situation. After all, no historical period is exactly like right now.
Expected future returns
The returns picture is not pretty for the near future. You can find all sorts of predictions of expected future returns based on various factors, calculations, and models, but unfortunately, most of them point to a rate of return for both stocks and bonds in the next few years that is below historical averages. The “Bogleheads” website puts together a few of these estimates and describes some of the differences between the estimates, which can be important. Additional forecasts are also available on the internet from McKinsey & Company, Christopher Brightman for Investment Wealth Management Associates, Norbert Keimling at StarCapital Research, and Wes Gray at Alpha Architect. This table provides a simplified overview of these studies.
Somewhat confusingly, two of these estimates are inflation adjusted, while the others are not. On a non-adjusted basis, stocks are predicted to return 4 to 6% moving forward instead of the historical 10% returns. And bonds will return 3% at best, given Bogle’s estimate assumes taking on some slight additional credit and interest rate risk as compared to the overall bond market. Bernstein’s inflation adjusted estimate of negative 1% for bonds is important. That means, corrected for inflation, money invested in bonds would be expected to actually decrease in spending power over the next 10 years. However, the relationship between expected future stock and bond returns is still remarkably similar to historical estimates. Stocks are still expected to provide returns that are around twice that of bonds.
The above summary focuses on one overall central tendency estimate (like an average). However, some of these studies examine the probabilities of various return outcomes. Vanguard conducted a study looking at the probabilities of various future stock returns using a price earnings projection as shown in these graphs.
The graph on the right is for non-inflation adjusted (nominal) returns and the left adjusts for inflation (real returns). The StarCaptial Research study also provides some probabilities around potential future stock returns as shown in this graph.
Both these studies highlight that predicting the future is a notoriously uncertain activity. Vanguard notes that, “The average is in fact a poor description of the norm…about two-thirds of 10-year periods had realized returns that deviated from a 5% band around the best-fit line. In other words, for a majority of history, the forecast was wrong by a meaningful margin.” Charlie Bilello at Pension Partners compared past 7-year forecasts by GMO (an investment management firm) to actual returns and found large differences between predicted and actual results. Clearly, there is a broad range of reasonable future outcomes around any of these expected annual return estimates. Using the Vanguard example, there is an about 80% percent chance that the future 10-year stock returns will be somewhere between 0 to 16%, which is a pretty darn wide range. Nonetheless, as best we can tell from these studies, the most likely non-adjusted return outcome is in the 4 to 6% range for stocks, as opposed to the historical average of 10%.
Expected future risks
Rick Ferri’s 30 year forecast also estimates future standard deviations for stocks that are right in line with historical data (in the 20% range). That study forecasts standard deviations for bonds in the 7 to 9% range, which is a little higher than historical data. As noted above, standard deviation by itself is only the first step in estimating the true risk of permanent loss. Nonetheless, it provides a consistent point of comparison that we should not expect stocks and bonds to be substantially more or less volatile in the near future, which eliminates one potential variable. Just like with predicting future returns, predicting future risks for any given period is a highly uncertain exercise, and these future risk estimates are just central tendencies. We should expect such risk forecasts to be wrong as often as the return forecasts. But they still represent the most likely estimated outcome based on the assumptions of the model used.
The future of bonds
Because bonds are particularly sensitive to interest rates and currently interest rates are at historically low levels, many have suggested that bonds now have a return/risk profile that is substantially different from the past. Let’s examine whether there are any more fundamental concerns with bond investing right now beyond simple return/risk metrics. First we have to understand how bonds produce positive returns.
Where do bond returns come from? – As I already mentioned, the expected future return on bonds is likely to be minimal at best, with the central tendency estimate at perhaps 2% before inflation, and zero or less after inflation. Why are bond returns expected to be so bad? Most of the return from bonds comes from the interest income paid, or the bond’s yield (as opposed to changes in bond prices), as illustrated in this chart from NewFound Research.
The interest income paid by a bond is determined mostly by prevailing interest rates for loans in general. No one is going to buy a 10-year bond that yields only 1%, if instead they could loan that money out elsewhere for a similar period and get 3% interest. Bond issuers must keep pace with the prevailing interest rates if they expect people to purchase those bonds.
History of interest rates – What does the history of interest rates tell us about the current and future environment for bonds? This graph, also from Newfound Research, shows 10-year U.S. treasury bond rates (which reflect interest rates in general) since 1875.
As the orange circles highlight, today’s interest rates are at truly historic lows. The 10 year bond rate was at 2.4% as of November 2017. Merrill Lynch conducted an analysis that reached back to the invention of money itself 5000 years ago. As shown in the chart below, the current interest rates are nearly unique across all history. Because of the current low interest rates, and because most of the return from bonds comes from the bond yield, the expected returns from bonds are also historically low.
Short term U.S. bond rates (6-month to 2-year bonds) were in the 1% to 1.7% range as of November 2017. At the same time, some countries in Europe and Japan were issuing short and even intermediate term bonds with negative interest rates to stimulate their economies. A negative rate paradoxically means that the bond issuer gets paid income from the bond purchaser. However, this is not an investment “opportunity” that many will want to pursue for obvious reasons. So, sustained and substantial negative interest rates are not likely, and if they were, it would mean that such bonds only have value if rates continue to move even further into negative territory. Clearly, this would be a questionable long-term investing approach for the individual investor. Although U.S. 10-year bond rates have recovered by close to a percentage point from historic lows around July 2016, they are still extremely low as compared to just about any other time in history.
Future of bond rates – It would be logical to assume that, because interest rates have no place to go but up, bonds will soon recover to their former historical rates of return. The problem is almost no one is predicting a rapid rise in interest rates. The U.S. Federal Reserve System (the Fed) sets national base interest rates and has indicated a policy intent to gradually raise interest rates in the coming years. The Fed is currently projecting (as of November 2017) that interest rates will be at about 3% by 2020.
The Fed has also been clear that the pace of interest rate increases is highly dependent on the health of the economy. If the economy starts to grow very fast and inflation starts to pick up, then interest rates may need to rise higher or more quickly. But almost no one is predicting a fast growing economy or a substantial increase in inflation soon. As a result, Robert Shiller, an often prescient Nobel economist, has warned of the “New Normal”, which is continued historically low interest rates and yields for some time to come.
Let’s assume that the Fed policy forecast is correct and about a percent or two will be added to existing 10-year bond yields in the next few years. That would about double the current (November 2017) yield for a 10-year bond to around 4.5%, but that is still a historically below average yield for such bonds. And due to prevailing economic conditions, this could be the fastest likely rate of increase for bond yields. So, all expectations are that bond returns will continue to be very low, although perhaps somewhat better than they are now. Like any prediction, there is considerable uncertainty around this estimate, but again, it appears to be the most likely scenario.
Future of bond prices – There is another problem with the current low bond yields. When bond yields go up, bond prices go down. This relationship is explained more here. Even though most of the return from bonds comes in the form of interest income, decreasing bond prices still take a bite out of those returns. As a general rule-of-thumb, for every percent increase in interest rates, there will be a 1% decrease in the bond price for each year of bond duration. So, if interest rates rise by 1%, the price of a 10-year bond will go down by about 10%. The Wall Street Journal presented a bond price calculator using a more precise algorithm. That calculator provides the following price declines for the 10-year US bond, which yields about 2.4% as of the November 2017:
- 1% yield increase = 8% price decrease
- 2% yield increase = 16% price decrease
- 3% yield increase = 23% price decrease
As noted above, a 2% yield increase in the next five years appears plausible. The net effect of somewhat higher yields and falling bond prices is pretty dismal. For example, if you purchase a 10-year bond fund, and the yield of that fund goes from about 2 to 4% in the next 5 years and the price goes down 16% in those same 5 years, your net annualized return over 10 years will be only about 1.7%. (This estimate assumes no reinvestment of interest payments and that neither the yield nor the price of the fund changes further in years 6 through 10.) This example is offered as only one reasonably plausible scenario for the next 10 years. If you adjust this annualized return for an equally plausible 2% rate of annual inflation, you are actually losing spending power (negative real returns).
My one example here is consistent with other’s observations about the range of possible future outcomes for bonds. Cullen Roche examined the “worst case” for bonds moving forward and noted that “…interest rates rose from 2% to 15% from 1940-1980 and that the 10-year U.S. government bond generated an average annual return of 2.85%.” However, like my example, the average annual inflation over this same period was 4.5%. Again, the inflation adjusted bond returns were negative in this period. We can start to see why Bernstein predicts a -1% return on bonds over the next 10 years as discussed in the “Expected future returns” subsection above.
These examples are also consistent with a study conducted by Research Affiliates. As shown in this graph using the 10-year U.S. government bond, historically there has been a nearly one-to-one relationship between the starting bond yield and the subsequent total bond returns in the next 10 years.
So, even allowing for the difficulties of predicting future interest rate changes, history suggests that low bond yields today are likely to provide low returns in the future. Applying today’s (November 2017) 10-year bond yield of about 2.4% to the above graph, we should expect 10-year returns on such a bond fund to be about the same.
Ideas for mitigating bond investing – There is a pretty constant stream of media articles nowadays pointing out ways to partially avoid some of the current problems with bonds. Let’s review some of the more popular suggestions one at a time.
- If you buy actual bonds, not bond funds, and hold them to maturity, you get back the face value on the bond with no price decrease. However, if you need to invest now, that strategy just gets you back to the historically very low return provided by the bond yield only (no potential price increases). For example, if you buy an actual 10-year U.S. treasury bond (not a bond fund) as of November 2017, you are locking in an approximate 2.5% annual return for the next 10 years on that money. Given inflation is expected to be around 2%, you are essentially guaranteeing that your bond purchase is barely maintaining spending power over the next 10 years. That’s a pretty long time frame to virtually guarantee no real return.
- Another way to mitigate bond price declines is to favor shorter term bonds, because longer term bonds suffer greater price declines when interest rates rise. Using our rule-of-thumb from above, when interest rates rise by 1%, the price of a 2-year bond will decline by about 2%, but the price of a 10-year bond will decline by about 10%. Although the rule-of-thumb starts to break down a bit with long duration bonds, price declines in the 20% to 25% range are possible for 30-year bonds when interest rates rise just 1%. These steep price declines sound more like the risks typically associated with stock market crises. Top bond investors like Bill Gross have been shouting warnings for years against holding long duration bonds. Cullen Roche and others suggest that the sweet spot for bond duration is an intermediate duration around the 7-year mark. This provides a balance point of somewhat higher yields with somewhat lower price sensitivity. However, as shown in the above examples using 10-year bonds (a close neighbor to a 7-year bond), intermediate duration bonds will most likely give you about a 2.5% annual return for the next 10 years regardless of whether you hold a bond fund or an actual bond for this period. And less likely scenarios won’t do much better. A rapid increase in interest rates, like that seen in 1940 to 1980 example above, would likely only get you up to about a 3% annual return. Likewise, continued low and flat rates will keep returns down in the 2% range for bonds with durations less than 20 years.
- Another mitigation strategy is to buy a series of bonds with increasing durations (like 2, 4, 6, 8, and 10 year bonds) in something called a bond ladder. As the shorter term bonds mature, and presumably interest rates have gone up, you can apply the short-term cash to longer term bonds at higher interest rates. I could go into a huge tangent here about bond ladders. Suffice it to say that if you don’t expect interest rates to rise very rapidly, and even if you hold actual bonds to maturity (not a bond fund), you are still fiddling with returns in the few percent range. For example, as of November 2017 the rates on 2 to 10-year government bonds range from only about 1.5% to 2.5%. So, unless interest rates really take off in the next few years, your bond ladder is still unlikely to outpace inflation. If you are interested, here are a few choice thoughts on bond ladders in various interest rate environments.
- One final mitigation strategy often mentioned involves a specific type of bond called Treasury Inflation Protected Securities (TIPS), which provides yields that are expressly augmented based on inflation. (Because TIPS are linked to the time dependent factor of inflation, I will come back to this option in Article 8.3, which is about investing over time.) However, Rick Ferri points out in “All About Asset Allocation” that TIPS are not the free lunch they might initially appear to be. He says, “Before you get too excited about the inflation-adjustment features of TIPS, you should understand that all bonds already have an inflation forecast built into their expected return.” For example, as of November 2017, the traditional 10-year government bond is yielding about 2.4% and the 10-year TIPS yield is about 0.5%, with the difference being investors’ collective inflation expectation (1.9%). You’ll note that this investor-expected inflation rate is about the same as the 2% government forecast I noted previously. So, TIPS might provide a return nearly the same as the traditional 10-year bond. But that depends on where inflation actually heads in the next few years. TIPS can be a good hedge against unexpected rapid rises in inflation, which have occurred periodically in the past. But as already discussed, most people are not expecting rapid inflation soon.
While we can quibble about details that might moderately boost bond returns, all of these approaches likely result in real returns that hover around zero. Further, some of these bond scenarios are both more complex to implement and not significantly different from simply holding cash. I define cash to include relatively short-term and guaranteed bank certificates of deposit, which currently pay about 1 to 2% annually across a 1 to 5 year maturity range (assuming a reasonably large chunk of money is being invested). And even some vanilla bank saving accounts will give you a 1.5% annual return as of November 2017.
Timing the bond market? – As a bit of a preview, Article 8.5 discusses that attempting to “time the market” by buying and selling stocks as the market goes up and down is not a particularly mindful approach. So, isn’t avoiding bonds right now essentially a form of bond market timing? My short answer is, no. I think there is a fundamental difference between stocks and the current bond situation. Even when stocks get into bubble territory (extremely high valuations) as was seen in 2000 and to some extent in 2007, there is no practical reason why stock prices can’t go even higher. There can always be “greater fools” who are willing to buy the same stock at ever more inflated prices. And it is very hard to determine when stock prices are so inflated that the bubble is about to burst. The practical floor in bond rates is fundamentally different given there are very few fools who will want to invest in zero or negative yields for the long term. While there is no ceiling in stocks, there is clearly a floor in bond yields, and objectively and historically speaking, we are still very close to that floor.
Bond dangers – Many have been pointing out that bonds are actually more dangerous in a historically low rate environment than normally portrayed:
- Burton Malkiel, author of Random Walk Down Wall Street, said he thinks bonds (and particularly bond funds) “may be one of the riskiest plays you can make”
- Ben Carlson pointed out that the unusually low bond yield environment may actually cause an increase in bond volatility in the future
- Brian Portnoy wrote an article on the painful lessons looming for those who are piling into bond index funds in the last few years
- And Bill Gross has mentioned multiple times that bond returns in the future will look nothing like they have in the last 40 years.
It’s also worth pointing out that some knowledgeable investors, like Charles Ellis, have for many years questioned the value of the lower returns of bonds in a long-term portfolio. Consequently, a mindful perspective tells us we should be skeptical about the standard claims that a diversified portfolio needs to contain a balance of stocks and bonds, or at least we should be skeptical that bonds need to represent a substantial proportion of that balance at all times. I come back to this point in more detail in Article 7.3.
The future of stocks
Like bonds, future stock returns are also expected to be below historical averages as I discussed above. Also like bonds, whenever there is any news of potential interest rate increases, almost every media outlet publishes an article with a headline that says something like: “Stocks decline on speculation of Fed rates increase”. So, it might be reasonable to assume that stocks are also going to suffer greatly as interest rates continue to rise.
As previously noted, the Fed is projecting a base interest rate up to about 3% by 2020. While history may not be a perfect mirror for the future, it’s worth looking at the history of stock returns and interest rate movements as shown in this graph from JP Morgan Asset Management.
Despite all the news articles, history shows that when the yield of the 10-year government bond is below 5%, stock returns tend to be positively correlated with increasing interest rates. While there will be short-term gyrations of stock prices when interest rates are raised, evidence suggests that there won’t be the systematic stock price declines that we expect with bonds.
Outside of interest rate rises, is there anything else special about our current situation that would likely impact the future of stocks? There are many international, economic, government policy, technology, and business factors (among a few) that could impact stock returns and risks in any given period. Nonetheless, it’s worth examining some of the most common measures used to judge future stock performance, which are exactly the types of measures that go into the future expected stock return models and calculations that I reviewed above.
One popular measure of future stock performance is to look at the current price earnings (P/E) ratios. Here is one typical analysis from JP Morgan Asset Management. This is simply taking the stock price and dividing it by the annual earnings per share.
The forward-looking P/E for the S&P 500 is about 16 as of December, 2015 when this chart was prepared. (“Forward looking” means that the earnings estimates are based on analysts’ projections of stock earnings for the next year. The other common P/E measurement is to look at previous years’ earnings. Both measures can be useful at times.) In the past, the five-year annualized return for forward P/E values in this range was around 10% to -2%, with the best fit line being around 9%.
Forward P/Es can be tricky, because they are predictions themselves, and as we have seen, most predictions are highly uncertain. Here is another estimate of future stock returns from StarCapital Research that is based on the past rolling ten years of P/E ratios, known as the Shiller CAPE ratio.
As of November 2017 the S&P 500 CAPE ratio is at about 31, which is historically quite high. So using the yellow dots (S&P 500) in the graph above, we can see it’s unlikely that the next 10-15 years of stock returns (inflation adjusted or “real” returns in this case) will be any higher than about 4%, and it’s more probable they will be lower. As discussed above, these are only broad and highly uncertain estimates, and StarCapital Research estimated an overall uncertainty range from 11% to -1.5% real returns using the CAPE ratio from December 2015.
Attempting to predict stock returns in the next year or even few years is essentially an impossible exercise because stocks are so volatile and can react suddenly to myriad random events from around the world. The one thing we know for sure is that stocks will go down as well as up, although the overall trajectory will likely be upward over sufficiently long periods of time. This graph from JP Morgan Asset Management is a good reminder to expect the unexpected with stocks.
Even though average annual returns for stocks have been around 10% for the last century, this graph shows you can expect significant within-year declines to occur even in years that have very good year-end returns. The future of stocks is very likely to be the same roller coaster it has always been. One year that’s very good can be followed by another year that’s very bad, and every conceivable mixture of good and bad years will also occur. To sum up, although it’s pretty clear we should expect lower than historical average returns for stocks, there is little evidence for a strong downward force on stock returns due to expected interest rate increases that is anything like the bond situation.
Our conclusions about “what to invest in” based on a mindful assessment of historical and expected future returns and risks in Articles 6.1 and 6.2 include:
- Even given a wide range of potential future outcomes, it appears very likely that stocks will always provide substantially better returns than bonds.
- For the near future, stocks are the only option that has a reasonable prospect for keeping ahead of inflation and increasing the purchasing power of your money.
- Stocks may be less “risky” than commonly portrayed, particularly if they fit within your investing time horizons. (More on time horizons in Article 8.)
- History suggests that we live in a unique time with regards to bonds, which makes them more “risky” than commonly portrayed.
A mindful perspective also demands humility. These conclusions are based on history, which will never be repeated exactly. Likewise, predicting the future is notoriously difficult, particularly based on economic and finance models that are imperfect at best. It’s extremely important to not rely on any of these conclusions too heavily. Humility suggests that we cannot entirely ignore the traditional role of bonds in investment portfolios. Particularly if interest rates rise faster than expected in the next few years, many of the dangers of bonds will diminish, and bonds could once again provide a reasonable balance of risk and returns. The appropriate mix of stocks and bonds (and cash) is discussed more in Article 7 on diversification.