Inequality, Interest Rates, Aging, and the Role of Central Banks

Bad things lead to more bad things. But (some) bad things can be fixed.

Here are three facts about the recent history of the world:

  • Income has become more concentrated within societies. Both the shares of household income going to the top 1% and corporate profit shares have increased dramatically since the 1970s.

  • Inflation-adjusted interest rates are much lower now across much of the world than they were in the early 1980s. That’s contributed to a surge in asset values relative to incomes—and likely means that future investment returns will be much lower than in the past.

  • The human population has aged rapidly, with the global median age rising by about ten years since the early 1980s. The share of the population aged 65 and older in the high-income countries and China, which together account for the vast majority of global economic output, has already jumped from less than 7% in the 1970s to 13% today—and the United Nations projects that share will rise to 25% by 2060.

New research suggests that these facts are directly connected: the drop in interest rates has been caused by changes in population structure and by shifts in the distribution of income.1 Lower expected returns on investment are just another consequence of deeper forces holding back economic dynamism.


Two recent papers provide complementary explanations of the mechanisms at work. Adrien Auclert, Hannes Malmberg, Frédéric Martenet, and Matthew Rognlie focused on the demographic angle, while Atif Mian, Ludwig Straub, and Amir Sufi focused on income inequality.2 Before digging into their specific arguments, a brief refresher on the basics.

What are interest rates, anyway?

For most of history, humanity faced a harsh tradeoff: we could produce goods and services that could be enjoyed today, or we could invest in the production of capital goods and research that would hopefully support higher living standards in the future. But we couldn’t do both. Until relatively recently, simply securing enough food to prevent starvation was a major challenge.

That constraint isn’t nearly as binding now as it was when we lived as subsistence farmers. In fact, our big problem for the past few decades has been a glut of capacity that we’ve refused to put to work. We’ve been living below our means and have been actively discouraging new investment. Nevertheless, we still retain many of the social institutions that we invented to endure the long era of scarcity.

One of those social institutions is “interest”: a reward that society pays to those who voluntarily sacrifice spending power today in exchange for the promise of more spending power later. That sacrifice frees up resources so that others can spend more than they earn now. In exchange, the borrowers agree to either spend less later (if they are borrowing to buy consumer goods and services) or to produce more later (if they are borrowing to invest in expanding society’s productive capacity). Whether this is good or bad—and therefore deserving of reward or punishment—depends on each society’s circumstances.3

In theory, interest rates are supposed to encourage or discourage spending according to each society’s needs at different points in time. If there is too much demand for goods and services relative to current production, high interest rates can potentially help by making it more expensive to borrow and by raising the rewards for those who abstain from spending.4

But if nobody is consuming or investing, then it’s pointless to get people to spend even less. Thus inflation-adjusted interest rates tend to be higher when the economy is growing rapidly and lower (or negative) during downturns and periods of slow growth. From this perspective, the relentless decline in interest rates is a signal of serious social problems.

The search for “neutral”

In most economies, baseline interest rates are “set” by central banks. Private investors and traders then set the borrowing costs for businesses, households, and governments on top of those baselines according to duration, credit risk, and other factors.5

Central banks can impose any level of (local currency) borrowing costs on the economy that they want, which is why some say that “interest rates are a policy variable.” But central banks generally avoid exercising that power arbitrarily—and for good reason. Interest rates that are too high or too low relative to what makes sense can lead to undesirable consequences ranging from mass unemployment to currency collapse.6 That’s why most central banks instead try to meet the economy and financial system where it is.

However, central bankers have an incredibly hard time estimating the “neutral” level of interest rates because there are so many countervailing forces at play. Here are just a few factors that aren’t being discussed in the recent batch of papers, but that clearly matter:

  • Governments’ willingness to spend, borrow, and tax

  • The state of consumer and “macroprudential” financial regulation

  • Perceptions of crisis risk and the need to self-insure against those risks by nonfinancial corporations, pension plans, and foreign reserve managers

  • Globalization and changes in openness to cross-border trade and finance

  • The rate of technological innovation, which in turn is affected by research productivity and by businesses’ success at commercialization and development

  • Attitudes towards debt, consumption, entrepreneurship, etc. by ordinary people and business executives

No model can capture all of the relevant mechanisms, which is probably why every model that’s produced estimates of “neutral” (or “natural”) interest rates generates extremely strange results at least some of the time. In practice, the best that central bankers can hope for is to get reasonably close with trial and error. As Richard Clarida, the Federal Reserve’s Vice Chairman, put it at the end of 2018, the “key parameters that describe the long-run destination of the economy are unknown”—and the only solution is to constantly update your estimates as new data come in.


Among other things, Clarida was referring to the relentless decline in estimates of the “neutral” rate over the past few decades. In fact, Fed officials’ estimate of “neutral” dropped another 0.5 percentage point just between Clarida’s 2018 speech and the end of 2019. The Fed’s new framework and the European Central Bank’s updated monetary policy strategy were both responses to this phenomenon.7

But even though it’s more or less impossible to know the level of “neutral” at any point in time, it’s easier to track changes in “neutral” over time. And that’s where the new research is particularly helpful.

Demographics and interest rates

Auclert et al argue that population aging—and slowing population growth—is partly responsible for the global drop in interest rates because slower population growth reduces investment. There is less reason to reward those who put off spending when there are fewer people trying to build factories, houses, or other types of capital.

This effect should only get bigger if the United Nations’ forecasts pan out:

There will be no great demographic reversal: through the twenty-first century, population aging will continue to push down global rates of return, with our central estimate being -123bp, and push up global wealth-to-GDP, with our central estimate being a 10% increase, or 47pp in levels.

In the 1960s, total population growth in the major global economies (the “high-income countries” plus China) averaged almost 2% a year. That slowed to just 1.2% a year by the 1980s, 0.9% a year by the 1990s, 0.6% a year by the 2000s, and just 0.4% by the eve of the pandemic. The combined population of these economies is projected to shrink starting in the 2030s, eventually falling nearly 20% from the projected 2030 peak by the end of the century.

Put another way, the number of children aged 0-14 in these economies fell from a peak of more than 600 million in the mid-1970s to about 465 million now. The number of children is projected to plunge almost 30% from current levels to just 335 million by 2100.

That pushes down interest rates, according to Auclert et al, because fewer people means there is less need to provide for the desires of future generations. This effect outweighs the fact that older people have much lower saving rates than everyone else. An aging society might produce less, but demand falls even further and faster. The process began in the 1980s and could continue for decades to come.

That’s consistent with what I noted almost six years ago when writing about Japan. There, population aging in the 1990s and 2000s pushed the household saving rate to zero during a period of sustained government budget deficits—yet interest rates went down. The reason was that households are only one piece of the broader economy. In Japan’s case, the decline in business investment and the rise in corporate profitability (which in turn was partly attributable to lower pay for workers) were more than enough to offset what was happening in the rest of the economy.

Inequality and interest rates

Mian, Straub, and Sufi, in a paper presented at the Federal Reserve Bank of Kansas City’s Jackson Hole Economic Symposium, focus on how changes in the income distribution affect saving rates, borrowing, and consumer spending.

The key insight is that the ultra-rich are different from you and me: they have much higher saving rates regardless of their age. No matter how expensive your tastes, there’s a limit to how much you can consume, which means any income above that threshold has to get saved. The ultra-rich therefore spend relatively small shares of their income on goods and services that directly provide jobs and incomes to others, instead accumulating stocks, bonds, art, trophy real estate, and other assets.

The ultra-rich need no encouragement to refrain from buying goods and services, so any increase in income concentration should put downward pressure on interest rates. Another way to look at it is that an increase in income concentration boosts the demand for financial assets, which should push up prices and push down yields.


But that effect can be offset by the fact that rising income concentration pushes everyone else to borrow more to finance their own consumption or investment, either directly or through the government budget deficit. The supply of financial assets can rise at the same that demand is also rising. The net impact on interest rates is therefore a function of how the extra borrowing (asset sales) relates to the extra lending (asset demand).8

While there are several ways this can play out in the short-term, rising income inequality leads to some combination of higher indebtedness and weaker consumer spending.9 To simplify only slightly, in a highly-unequal global economy, the ultra-rich lend everyone else the money they need to buy the goods and services sold by the businesses the ultra-rich own. Their profits depend on the ongoing growth of these circular financial flows.

As Mian et al put it, the increase in saving by America’s high earners was offset by everyone else saving less.

But that’s inherently unstable. Consumers getting paid less and less relative to national income can’t keep spending more and more unless they also borrow more and more. And the only way consumers can keep taking on more and more debt relative to income is if interest rates keep falling. That dynamic also makes new investments in productive capacity less attractive: indebted consumers with minimal income growth aren’t going to be great for sales growth. All of that should push interest rates down.

Thus the real estate, junk bond, and consumer debt bubbles of the 1980s were followed by a big decline in interest rates in the early 1990s, the EM and corporate debt bubbles of the 1990s were followed by a big decline in interest rates in the early 2000s (and the inflation of new EM and real estate debt bubbles), and the excesses of the 2000s were then followed by the collision of interest rates into the “effective lower bound.”10

All of this has happened before.

Here’s how Marriner Eccles put it in his testimony to the U.S. Senate Finance Committee in February 1933:

The debt structure has obtained its present astronomical proportions due to an unbalanced distribution of wealth production as measured in buying power during our years of prosperity…The time came when we seemed to reach a point of saturation in the credit structure where, generally speaking, additional credit was no longer available, with the result that debtors were forced to curtail their consumption in an effort to create a margin to apply on the reduction of debts.

This naturally reduced the demand for goods of all kinds, bringing about what appeared to be overproduction, but what in reality was underconsumption measured in terms of the real world and not the money world…There must be a more equitable distribution of wealth production in order to keep purchasing power in a more even balance with production.

The Great Depression didn’t really end until wartime mobilization caused a surge in incomes and production that wiped out old debts, leveled the wealth distribution, and gave people confidence in the future. The end of the war also kicked off a baby boom after a long drought of births. Not coincidentally, interest rates marched up for decades until the early 1980s.

Since then, we’ve endured a great reversal in the trends underlying the postwar prosperity, capped off by a global financial crisis that bore many similarities to the Depression. In fact, even though the 2007-10 downturn was shallower, the subsequent recovery was so much weaker that the net effect is that GDP per American grew less in 2007-2019 than in 1929-1940.

But there are reasons to hope that things can improve.

The U.S. government’s response to the pandemic—and the attendant impact on household balance sheets—in some ways resembles the wartime mobilization of the 1940s, albeit at a much smaller scale. And while population aging will be tough to reverse, the trends in inequality that have retarded growth and have pushed down interest rates were choices that can be changed.

Let’s get to work.


There is also the argument that central banks are causing the increase in inequality by lowering interest rates. Lower rates lift asset prices and create a bias in favor of those who can borrow the most to buy the riskiest collateral: leveraged buyout shops, hedge funds, and real estate investors. Ordinary savers ostensibly pay the price in the form of lower returns on liquid assets.

There’s something to this mechanism, but I think it’s more accurate to say that “the political and economic forces that put central banks in charge of responding to downturns and that require consistent declines in interest rates” increase inequality. If rising income concentration is itself one of those forces—and I think it is—then higher inequality is self-reinforcing with a normal central bank.


While Mian et al frame their research as being partly opposed to Auclert et al, my reading of the two papers is that they are complementary. Auclert et al looked at the impact of aging on investment, while Mian et al looked at the impact of rising inequality on consumer spending and saving. All of those things affect interest rates.


For more on this, I would recommend reading Trade Wars Are Class Wars—just released in paperback with a new preface!—especially chapter 3.


There are some who argue that high interest rates are harmful in this situation because they discourage both current demand as well as investment in additional capacity that could help meet that demand.


Some central banks instead manage the exchange rate and let interest rates move accordingly. Ecuador, El Salvador, and Panama all use the U.S. dollar as their currency and therefore have no direct control over domestic interest rates. Hong Kong’s Monetary Authority maintains a strict peg between the Hong Kong dollar and the U.S. dollar, with interest rates following the Federal Reserve. The Monetary Authority of Singapore manages the level of the Singapore dollar against a basket of foreign currencies, with interest rates often lower than in the U.S.


Big gaps between government policy and whatever makes the most sense for the economy as a whole can show up in a variety of ways, including credit rationing and black-market exchange rates. See, for example, the Argentinian mortgage market.


The downward trend in rates arguably goes back a lot further, although not from the perspective of anyone currently alive. One study from the Bank of England recently found that interest rates have been falling at least since the Renaissance.


It gets even more complicated if you look at how these forces interact in individual countries within the larger context of the global economy, as Michael Pettis and I did in Trade Wars Are Class Wars.


In theory, an increase in income concentration could lead to more financial asset sales by businesses, rather than consumers, which could support additional capex. That could make sense in poorer countries where resources are scarcer. But squeezing ordinary people to finance additional investments in productive capacity makes no sense in an advanced economy. Who would buy the extra goods and services once they’re available? In practice you would end up with higher indebtedness as the result of the bad investment.


David Levy memorably called this the “bubble or nothing” economy.