Another look at the long-term trend in real rates

Some time ago, I wrote a post about the fascinating data collected by Paul Schmelzing who showed that the real rate on long-term government bonds is not stationary over time but follows a very slow downward trend over the centuries. Now, Paul has teamed up with Ken Rogoff and Barbara Rossi to examine what could explain this trend.

Most importantly, using real interest rates from 1300 to today, they clearly show that real interest rates are not stable over time. They follow a statistically and economically significant downtrend that puts hits negative real rates right about now. In other words, they show that even though real rates can deviate from this trend for some time (the time to return to the long-term trend is 1 to 10 years), we shouldn’t get our hopes up that real rates are going to normalise to levels anywhere near the levels seen before the financial crisis of 2008. Instead, current positive real rates in the United States are probably just about what we can expect in the coming decade or possibly too high already.

That’s bad news for savers, and it leads to the question of possible breaks in that long-term downtrend. This is what the new study by Paul Schmelzing and his collaborators has tried to identify. They showed that there are essentially two possible trend breaks over the last seven hundred years. One came with the Black Death in the middle of the 14th century when the balance between capital and labour shifted very much in favour of labour and against capital for a generation or more. The second possible breakpoint was the triple default of France, Spain and the Netherlands in 1557 which shook confidence in the ability of empires to repay their debt.

Most notably, there is little evidence of a trend break with the introduction of the Bank of England in 1694, the Fed in 1913 or central banks anywhere else. Similarly, there is no indication of a trend break in real rates after the Great Inflation of the 1970s and the introduction of today’s monetary policy tools.

So, if central banks and monetary policy did not change the downtrend in real rates, what does? We don’t really know the correct answer. But the new paper sheds light on two commonly cited drivers of real rates. First, economic theory postulates that higher economic growth should be linked to higher real interest rates. But a look at the chart below already shows that if anything, the opposite is true. Higher growth tends to coincide with lower real rates, not vice versa. A more sophisticated statistical analysis confirms that there is no long-term link between growth and real rates, at least not the one economic textbooks predict.

There is no link between economic growth and real rates

Source: Rogoff et al. (2022)

The other driver for real rates that has commonly been proposed are demographics and population growth. But I have written before that demographic changes are unable to explain the shift in the natural rate of interest on the short end of the yield curve. The new research now shows that at the long end of the yield curve there is also no link between demographics and real rates.

There is no link between demographics and real rates

Source: Rogoff et al. (2022)

In other words, the declining trend in real rates is pervasive over the last 700 years and real interest rates tend to return to this declining trend over time. Neither the introduction of central banks, the Two World Wars in the 20th century nor the invention of more and more sophisticated monetary policy tools had any material impact on changing this long-term trend. And neither economic growth nor demographics seem to drive this trend.

It is fabulous! Here we finally have an empirical observation that shows us that we do not know what drives real interest rates and that should make us rethink all our theories of interest rates, monetary policy or even economic growth. It is an invitation for a macro revolution. Alas, so many people will try to keep this revolution from happening even though there is more and more empirical evidence against existing theories.