Interest rates have declined from over 5% in the late 1960s to under 0.5% in 2020 (I’m using long-term treasury bond rates as an estimate of pure risk-free rates).
There’s clearly something unusual about savings rates, and it’s causing a semi-stable decline in interest rates.
In my review of Shut Out, I guessed that demographic effects of the boomer generation explain 25% the housing boom, via high savings rates and the resulting low interest rates. I’m now revising that estimate down to 10%.
I now see a similar but bigger effect from a different set of demographic changes. I changed my mind due to the paper by Mian, Straub, and Sufi titled What Explains the Decline in r*? Rising Income Inequality Versus Demographic Shifts (think or r* as an idealized version of interest rates).
Wealth affects savings rates more than age does.
The higher saving rate of the top 10% together with the large shift in income to the top 10% combined to generate a significant increase in savings entering the financial system from high income households. Overall, we estimate that between 3 and 3.5 percentage points more of national income were saved by the top 10% from 1995 to 2019 compared to the period prior to the 1980s. This represents 30 to 40% of total private saving in the U.S. economy from 1995 to 2019. The rise in saving by high income households is likely a powerful force putting downward pressure on r?.
(It’s a bit confusing that the most commonly cited measure of savings shows a significant decline in the saving rate from 1975 to 2005. That measure seems to exclude capital gains, whereas Mian, Straub, and Sufi use an apparently more useful measure of savings that’s based on changes in wealth.)
What does this mean for investors?
It represents a fairly large shift in my long-term outlook.
For starters, it suggests that the current low interest rates and gigantic government deficits might be almost as sustainable as current inequality levels are. I.e. maybe they only need to revert 20% of the way back to their late 20th century mean, which would be a much smaller reversion than I previously thought would be needed.
The effects of boomer generation retirement will likely cause some increase in interest rates over the next decade, but I now expect that effect to be fairly modest.
There’s a saying “The four most expensive words in investing are: ‘This time it’s different.'”. So I’m nervous about investing based on expectations that inflation-adjusted interest rates will remain near zero.
Yet it’s also a good idea to take most market forecasts seriously. Bond markets are somewhat less prone to irrational exuberance than are the smaller, more volatile parts of the stock market that have experienced clearer bubbles. Bond markets are clearly forecasting low interest rates for as far as they can see.
One widespread effect of lower interest rates is increased demand for housing. Housing is more affordable than it was when rates were higher, so people are buying more of it. I’m pretty bullish on homebuilding stocks, and on a number of stocks in related businesses.
Some of the stock market’s higher price/earnings and price/sales ratios are due to there being more wealth available to invest. There’s also some irrational exuberance causing some companies to be overpriced – it’s hard to have as big of a stock market rally as we’ve had without producing some excesses (GameStop? Rivian? Tesla?).
Stock market prices look overvalued by most traditional measures. But if those measures had been designed to change in response to large, lasting changes in interest rates, they’d likely say that current stock market levels are likely to be about as stable as current low interest rates.
Low interest rates cause earnings 5-10 years in the future to be more important than usual, so investors are taking a longer-term outlook than usual. But beware that investors have, in that process, become overconfident in their predictions about which companies will do well 5-10 years from now.
Let’s compare companies that needed a lot of long-term investment to survive in 2000 with similar companies in 2021.
Webvan seems to be the clearest example of such a company in 2000. Webvan did lots of things right. It sure looks like it could have succeeded if it had raised a few $billion more money. At the peak of the dot.com boom, that looked like a reasonable gamble. But they bet the company on finding that money. A year later, it shut down due mainly to the shortage of investors who believed it would raise enough money to reach the economies of scale that it needed.
In 2015-2020, Lyft took a similar strategy, with very different results. It was able to weather a pandemic that caused its revenues to drop 35%, and now seems able to survive if it gets no more investment.
Tesla was a bit less clearly dependent on massive investment, but its investors tolerated a decade of losses that reached about $1billion per year before it became profitable.
Aurora Innovation looks like an even clearer example of a company that will need to spend several billion before getting anywhere near to profitability. It seems to have most of what it needs.
These patterns are spurring a moderately important increase in innovation in areas such as transportation that had previously been stagnating as part of the Great Stagnation.
It’s weird to see a large overlap between people who support large government deficits and the people who talk as if income inequality needs to be significantly reduced from current levels, now that I expect substantial reductions in inequality would force large reductions in deficits.
But it’s not as weird as it sounds, if we take into account the lack of realistic plans to make more than minor dents in inequality.
The only force that has been proven to dramatically reduce income inequality is a major war such as WWI or WWII. The U.S. and China might be on a collision course for a major war over Taiwan. I can’t quite see how that will approach being WWIII, but I also can’t see what will cause either side to compromise or back down.
There are probably other ways that income inequality could be reduced enough to have a major impact on interest rates, but I don’t see any that look politically feasible.
P.S. Another interesting divergence:
in any given birth cohort, the households that are saving more today relative to previous birth cohorts are those at the top of the income distribution. Households in the bottom 90% are actually saving less relative to previous birth cohorts. The Euler equation logic suggests that these patterns are due to the fact that households in the top 10% of recent birth cohorts expect lower income growth relative to the top 10% of previous birth cohorts, whereas households in the bottom 90% of recent cohorts expect higher income growth relative to previous cohorts.