A foundational principle of IvyVest is that for most investors, there is a "95-5" rule in investing. Put simply: 95% of your results will come from what, to at least most investors, is 5% of your efforts. The 5% that the data proves really matters: it's not whether you buy IBM or Microsoft, but what your overall asset allocation is – how much of your money goes into US stocks, European stocks, Asian stocks, bonds, etc. We try to guide that "95%" in as data-driven of a way as possible and leave it to you to decide what, if anything, to do with the other 5.
That put forth, when new data comes out that purports to inform the best way to allocate assets for long-term investments, we pay attention. All this to say that we will warn you in advance that this month’s newsletter is going to be a little bit wonkish, but it will bring you up to date on some recent economic research into this most important area of investing. We are going to review a working paper that was recently published by the National Bureau of Economic Research. The full reference is given at the end of this newsletter, but the title is “The Rate of Return on Everything, 1870 – 2015”.
You might ask what this has to do with how you should invest today, and perhaps that would be fair. But the historical record, particularly over such a long time period, is certainly a key data point to inform asset allocation strategies. History tells us about the “risk premium” (return over a risk-free rate like a savings account interest rate) that investors have historically demanded to invest in risky assets like stocks and bonds. When looking to make an educated guess about what is likely to happen in the future, at some level, all we have to go on is history, and what analyzing that history tells us about how the world works.
While the title of the paper we analyze today is perhaps a bit overarching, the paper looks at the returns on bills (short term government securities), bonds (long term government securities, typically with a ten year maturity), equities (stocks), and residential housing over the period from 1870 to 2015.
Let’s discuss the most topical subject first: bonds. Interest rates in the US have recently risen, breaking a multi-year trend line and leading many to predict that the “bond bubble” is finally breaking. What insights does history have to offer about where interest rates stand today relative to their extremes? The authors calculated real (inflation adjusted) bond and bill interest rates (i.e., they subtracted the rate of inflation from the interest rate). They defined a global rate of return by averaging the rates in each country, weighted by its gross domestic product (GDP). (They also stated that using equal weights instead of GDP weights would not change the picture significantly). They presented the results in the form of decadal averages, so the data point for any time is the average of the five preceding years and the five subsequent years. This smooths the data without eliminating major trends. With these provisos, the computed global real rate of return on bills and bonds is shown in figure 1 for the time from 1870 to 2010.
There are a couple interesting features in this chart. The most striking feature is that there were considerable periods when global real returns were not just low, but actually negative. In fact, based on the data in figure 1, our current low interest rates are not far out of line with the long-term average. Many people, ourselves included, have said that current interest rates are historically low. The reason for this opinion is shown in figure 2, which was adopted from reference 2. It shows the average government bond yield for the G7 countries for the period from 1864 to 2014. The rates shown in Figure 2 are nominal (not inflation adjusted). Based on figure 2, present interest rates are far below average, and therefore one might expect that a reversion to the mean is likely bring higher future rates. However, the picture in figure 1, with inflation adjusted rates, is very different and suggests that current real rates are already close to the long term mean. We should note that the periods of negative interest rates shown in figure 1 were times of war (World Wars I and II) or times when inflation was very high (the late 70s). When we look at periods without major wars or high inflation, current real rates are quite low compared to historical rates, even in figure 1. Therefore, we continue to believe that some reversion to higher rates is likely. Nevertheless, figure 1 casts some doubt on that conclusion.
Figure 1: Global (16 countries) real (inflation adjusted) returns on government bonds and bills. The returns are plotted as decadal rolling averages. The data was taken from reference 1.
Figure 2: Average ten-year government bond yields (not inflation adjusted) for the G7 countries. The data was taken from reference 2.
Figure 1 is the global average of the 16 countries. For comparison, figure 3 presents similar data for just the U.S. In this case the data was taken from reference 3, and the values are 7 year rolling averages (the value for any year is an average of that year plus the three preceding years and the three subsequent years). The U.S. data is quite like the global data (but of course, the U.S. data was part of the global data). To further investigate the correlation between interest rates in the 16 countries, the NBER authors considered them in pairs and calculated a correlation coefficient for each pair. Then the correlation coefficients were averaged and plotted in figure 4. If you are not familiar with correlation coefficients, it suffices to say that a coefficient of 1 means that data tracks perfectly; a coefficient of 0 means that there is no relation between the two data sets; and a negative coefficient means that one data set goes down when the other goes up. Figure 4 shows that the real bond rates in the countries were quite well correlated at times (they moved in the same direction together), but at other times the correlation was much worse. However, it was always positive, meaning there was always a tendency for the rates to move in the same direction.
Figure 3: Real returns for U.S. one-year and ten-year bonds. The data was taken from reference 3.
Figure 4: Average correlation coefficient of government bond rates for the 16 countries considered in pairs. The data was taken from reference 1.
The comparison of bond and bill rates (the rates that government pays on its debt) to the rate of GDP growth is also of interest. If the GDP grows at a rate that is greater than the bond and bill rates, then a government can reduce its total debt burden (as a percentage of GDP) without diverting tax revenues to debt service. In Figure 5, the average of the bond and bill rates is compared to GDP growth rates for the 16 countries. As in figure 1, this is done using decadal averages. Over most of the time span, the GDP growth rate has exceeded the average real interest rate. That is also true now. The current 10-year treasury rate is about 2.6%. Inflation in the past year has been about 2.1%, giving a real interest rate of about 0.5%. Real GDP growth for the year ending at the end of September 2017, was 2.4%, putting real GDP growth well above the real interest rate.
Figure 5: Comparison of global (16 countries) government interest rates with GDP growth. The interest rate is the average of the bond and bill rate. The data was taken from reference 1.
Now let’s turn to the “risky” assets -- housing and equities. Here, the paper breaks new ground by reporting a very long history of housing returns that include the true returns of owning a home — not just price appreciation. Just as the the return on equities is the sum of dividends and price growth, likewise, the return on housing is the combination of “implied rents” and price growth, where “implied rent” is a measure of how much you would have had to pay to rent a house of similar quality to the one that you own. Of course, rents must be adjusted for maintenance costs and depreciation, and the NBER authors attempted to do this. In reference 4, Shiller showed that housing prices have grown just a little faster than inflation. This new NBER study adds in the rent component and shows that housing gives a very good return. Figure 6 shows the results for both housing and equities. Housing returns have been nearly equal to equity returns, but housing returns have been much less volatile.
For both housing and equities, the returns are far above government bond returns. Table 1 shows the average real returns for the U.S. and for the 16-country global average for three different time spans. Real equity returns were strongly positive for all the time spans. The U.S. equity returns were slightly better than the global returns for the two longer periods, but for the post 1980 period, the unweighted global average was slightly better than the U.S. average. Although U.S. equities outperformed the global average in the longer time spans, Finland was the top performer in all three time spans. The U.S. was the second-best performer in the full-time span going back to 1870, but in the 1980 to 2015 period, 12 of the 16 countries did better than the U.S. This supports the notion that foreign equities can add diversity to a portfolio without significantly reducing returns. For the GDP weighted global average, housing outperformed equities for the longest time span. For the more recent time spans, equity outperformed housing, but not by a whole lot. Looking at just the U.S. data, equities outperformed housing in all time spans, but only in the most recent time span was the difference appreciable. Housing gives returns that are close to equities but with much less volatility.
Figure 6: Housing and equity global returns for the 16 countries. The data was taken from reference 1.
Table 1: Global and U.S. Equity and Housing Returns (Inflation Adjusted)
|Returns, %1980 - 2015|
|1870 – 2015||1950 – 2015|
|Global Average, equal weights||6.60||7.25||8.25||7.46||10.68||6.42|
|Global Average, GDP weights||7.04||6.69||8.13||6.34||8.98||5.39|
To better understand the diversity added by foreign stocks, it is helpful to look at the correlation coefficients between the performance of equities in the various countries. A high correlation coefficient between equities in two countries indicates that the two are moving together and there is little gain in holding both, but lower correlation coefficients indicate that holding both can reduce risk. Figure 7 shows the average of all the pairwise correlation coefficients between the 16 countries. The average correlation was relatively low over most of the time span, but it has recently gotten much higher. The more recent results are reflective of a global economy and indicate that foreign stocks are not providing as much diversity as they once did.
Figure 7: Correlation between equity performance in the 16 countries. Correlation coefficients were computed on a pairwise basis for all pairings of the 16 counties, and the average of these correlation coefficients is plotted. The data was taken from reference 1.
Another interesting issue is how equities and housing do in inflationary times. Figure 8 shows the correlation between equity performance and inflation and housing performance and inflation. Equities seem to show a slight negative correlation with inflation, at least recently, but housing is positively correlated with inflation. The positive correlation between housing and inflation is somewhat surprising because we are dealing with real return; the positive correlation means that housing prices or rents must be moving up faster than inflation. If you expect inflation, it’s nice to own a house. Stocks probably adjust to inflation over the long term, but in the short-term inflation creates high interest rates which hurt the performance of stocks.
Figure 8. Correlation of equity returns and housing returns with inflation. The data was taken from reference 1.
Before we leave the discussion of housing, we should note that most people now buy homes using mortgages. This means that their investment in housing is leveraged, which increases both the return and the volatility. The authors discuss this issue and conclude that it doesn’t have a big effect on the global picture, but we suspect that the real returns from housing may be higher than what is given in this paper. A final issue may be of interest to the most wonkish of you. Thomas Piketty (reference 5) has argued that the normal performance of capitalist economies leads to growing inequality. His explanation for this is that the returns on wealth (a combination of equities, bonds, and housing) exceeds the growth rate of the GDP. You can see this easily if you picture the 10% nominal return of stocks next to a 5 or 6% nominal GDP growth — if the rich stay invested then historically at least, the rich have gotten richer in the sense that their wealth has grown as a share of the overall economy. Wealth accumulates with those who already have it.
This is an issue that requires much more discussion that we can give here, but the NBER study supports the basic premise. The authors looked at the relative components of wealth in the 16 countries and determined a distribution of stocks, bonds and housing that constituted wealth. Then, they computed the rate of return on that distribution of assets. The results, presented as the return on wealth minus the GDP growth rate, is shown in figure 9. Apart from the war years (WWI and WWII), the yield on wealth has been well above the rate of growth of GDP over the entire 145-year period. This agrees with Piketty, but it leaves us somewhat perplexed as to how the system can be stable in the long run. However, taxes have not been included in the return of wealth calculation, and they may resolve the puzzle. The practical implication of this is that future equity returns may be well above the GDP growth rate.
So what conclusion can IvyVest asset-allocators draw from the paper? Overall, a positive one — it supports the idea that over the long-run investors are rewarded for owning stocks, it supports the idea of foreign diversification, and it adds some additional historical context to illustrate that the current era of low interest-rates might not be as historically unprecedented as many have imagined. Finally, if you have not yet purchased a house, the paper adds to the weight of evidence indicating that you should perhaps consider this one of your highest priority investments.
Figure 9: The return on wealth minus the GDP growth rate. Taken as a whole, this NBER paper reinforces our belief that investors should maintain a strong commitment to equities, both domestic and foreign. It is probably no surprise to anyone that owning a house also looks like a good long term investment.
- Òscar Jordà, Katharina Knoll, Dmitry Kuvshinov, Moritz Schularick, Alan M. Taylor; “The Return on Everything, 1870 to 2015”, National Bureau of Economic Research Working Paper 24112, December, 2017.
- Jim Reid, Carol Nicol, Nick Burns, Sukanto Chanda; “Long Term Asset Return Study: The Next Financial Crisis”, Deutsche Bank Markets Research, September, 2017.
- Robert Shiller, online data at econ.yale.edu.
- Robert Shiller, “Irrational Exuberance”, Princeton University Press, 2000. 5.Thomas Piketty; “Capital in the Twenty-First Century”, Harvard University Press, 2014
Get our next article delivered to your inbox.
Sign up below and be the first to know about our freshest data-driven thinking on the markets, and investing. We will send you no more than one email a week. This is free.
Ready to start putting this into action?
Take a free two-week trial to IvyVest premium -- our premium subscription service. You'll get access to our rules-based dynamic asset allocation model, tools that will show you exactly what you need to buy in your own discount brokerage account (and when to re-balance) to implement it for yourself, and an insightful monthly newsletter that will keep you on abreast of the most important things going on in the markets. There is no credit card required. Get Started Now!