Can Putin Survive?
The Lessons of the Soviet Collapse
The past five years of U.S. economic policy have been noisy, as the Trump administration and its allies in Congress pursued a controversial agenda: a trade war with China, a push to repeal the Affordable Care Act, tax cuts that mostly benefited the well-off, and so on. Behind this sound and fury, however, lies a story of quieter but deeper economic changes that will have far-reaching implications. That story revolves around four interconnected developments: the fall in the natural rate of interest, the remarkable decline in the price of renewable energy, the stubborn persistence of inflation below the U.S. Federal Reserve’s target of two percent, and the stunningly fast collapse and then partial rebound of the economy during the COVID-19 crisis.
These changes do not necessarily call into question any fundamental principles of traditional economic theory. In fact, in many cases, they confirm the value and validity of certain core concepts. They were largely unexpected, however: taken together, they require economists to rethink some key parts of their models. They also open new dimensions in old debates about taxes and spending. And what is perhaps most consequential, they present new opportunities for policymakers when it comes to the fight against climate change.
The natural rate of interest, or r*, is the real interest rate—that is, the actual current interest rate minus expected inflation—that would prevail in an economy enjoying full employment without any government intervention. Savers and investors use r* to project interest rates over the long run; for example, the expected long-term value of the yield on ten-year U.S. Treasury securities is r* plus the expected long-term rate of inflation. Monetary authorities can hold real interest rates above or below the natural rate for months or years—but not forever, so r* is the benchmark against which monetary policy is considered tight or loose.
Although r* is a revealing metric, it is also a problematic one. The economy, after all, is never at full employment without government intervention; the Fed always has its hand on the wheel. And because expected inflation is an estimate, the real rate of interest is impossible to measure directly at any given moment. So gauging r* is a tricky proposition.
Yet there are still several ways to estimate r* and to observe its steady decline. The simplest one is to select a relatively long period of time in the past, identify the average interest rate on a low-risk asset during that period, and then subtract actual inflation during that period. Take, for example, 30-year U.S. Treasury bonds. The return on those bonds, minus the rate of inflation, has fallen consistently over the course of the past four decades, from an average of 5.9 percent in the 1980s to 4.7 percent in the 1990s, 2.8 percent in the first decade of this century, and 1.6 percent in the 2010s. More sophisticated estimates of r* draw on statistical models. Although no single method is determinative, they broadly converge on the finding that r* has fallen by approximately 1.5 to 2.5 percentage points over the past two decades, and not only in the United States but also in developed countries around the world.
Because this decline occurred over decades, it cannot be attributed to a single business cycle or to monetary policy; explanations must lie elsewhere. Economists disagree about which factors have played the biggest role. Some point to the aging of workforces in developed economies. In the United States, this has occurred as baby boomers have approached retirement age. As they have, their savings have risen, which has had the effect of pushing interest rates down. Others argue that the high savings rate in China has put downward pressure on interest rates around the world. And although the evidence is mixed, some economists cite factors such as the slowing growth rate of productivity (which has dragged down consumption) and the rise in income inequality (which causes the overall savings rate to rise, since the rich command an ever-larger share of income and tend to save a greater portion of their income than the poor do).
In some parts of the United States, clean energy sources are now cheaper than dirty ones.
Whatever its causes, the decline of r* has profound implications. One is the increased capacity of developed economies to take on additional sovereign debt. The conventional wisdom—dating to the 1960s and earlier, when r* was much higher than it is today—holds that carrying too high a level of debt is unsustainable for governments because interest payments compound and overwhelm the state’s capacity to tax, leading to a debt crisis. What constitutes “too high” has never been made clear, but the historical experience of foreign debt crises, periods of hyperinflation, and defaults provides compelling evidence that some threshold did in fact exist.
The logic of fiscal prudence is turned on its head, however, in a world of low r*. If the interest rate paid by the government is less than the growth rate of the economy, then debt taken on today can be rolled over in perpetuity—and the burden posed by the interest on that debt becomes a vanishingly small fraction of GDP. Governments can therefore finance current spending with essentially no change in future tax rates. Naturally, this “free money” argument is appealing to politicians across the U.S. political spectrum: those on the right have used it to rationalize making permanent the 2017 income tax cuts; those on the left have used it to justify massive spending proposals under the so-called Green New Deal.
To some economists, including myself, such arguments are unsettling: after all, r* is less than the growth rate of the economy until suddenly it isn’t. To the extent that the decline in r* is driven by an aging population, that demographic transition will taper off. For example, in the United States, by 2030, all the baby boomers will be at least 65 years old. Additional government borrowing will, all else being equal, place upward pressure on interest rates and on r*. The lesson from previous fiscal crises is that when markets start to doubt the ability of a government to meet its debt obligations, things can sour quickly, and no one wants to set the stage for a Chinese bailout of an overindebted U.S. government. Experts in public finance have long maintained that Americans should not burden their children with a debt-to-GDP ratio of 100 percent. Although that threshold now seems too low, economists have not yet reached a consensus on what it should be.
The evidence for a long-term decline in r* adds a new dimension to arguments about spending and taxes. Those arguments are age-old, and the role of interest rates is a familiar part of the debate. More novel, however, is the question of how the decline of r* might affect an issue that has come to the forefront of U.S. politics only in more recent years: climate change.
Policies that reduce emissions of greenhouse gases entail taking actions today that governments would not take, at least not with any urgency, absent the risk of future harm from climate change: in other words, they involve incurring costs today to enjoy benefits at some point down the road. To know whether those future benefits will outweigh the present costs, one must first assign dollar values to both and then place those monetized values, which are realized at different dates, on the same footing. This is done by converting future costs and benefits to current-year dollars using an interest rate, which in this context is called a “discount rate.” For example, at a five percent annual discount rate, $1 today is equivalent to $1.05 next year. Such a comparison of the present values of costs and benefits is more than merely sound practice: it is required by law for some federal regulations and by presidential order for others. Making such a comparison requires choosing a discount rate, which for long-term societal costs (such as damage from climate change) is conventionally taken to be the long-term real rate of interest—that is, r*.
In the case of climate policy, the so-called social cost of carbon (SCC) is the present value of the harm done by emitting one metric ton of carbon dioxide. The U.S. government has published estimates of the SCC since 2010. At the end of the Obama administration, the federal government estimated the SCC to be $51, a value it computed by setting the discount rate to three percent. The Trump administration revised this down to $7 by considering only domestic climate damages, and not global ones, and then further lowered it to just $1 by setting the discount rate to seven percent. The choice of where to set the discount rate matters a great deal: at three percent, it would make economic sense to pay $97 today to avert $1,000 in damages in the year 2100, whereas at seven percent, it would make sense to pay only $5. In this light, the Trump administration’s calculations encourage the conclusion that under current U.S. policy, climate change will inflict tremendous global damage—but that it would still be cost-effective just to let future generations deal with the problem.
The three percent figure used by the Obama administration comes from a remarkable official document released by the Office of Management and Budget known as “Circular A-4,” which was issued in 2003 and which provides detailed, thoughtful guidance to federal agencies on how to conduct cost-benefit analyses. “Circular A-4” arrived at the three percent figure by taking the 30-year average rate of interest on ten-year U.S. Treasury bonds and subtracting the rate of inflation in the consumer price index (CPI). Repeating that calculation today provides a dramatic restatement of the decline in r*: over the past 30 years, the yield on ten-year Treasuries has averaged 4.3 percent, and CPI inflation has averaged 2.3 percent, putting r* at 2.0 percent. If this calculation is performed over the past 20 years, the resulting r* is 1.1 percent—substantially lower than the Obama-era estimate of three percent.
Despite the Trump administration’s suspect abandonment of the three percent rate, that number continues to be used as a reference point in some tax rules and policy proposals. But the decline in r* implies that the three percent discount rate is too high, and by using that too-high factor, economists are underestimating the SCC. If r* is not three percent but rather two percent, then the SCC is not $51 but actually $125, and it would make economic sense to pay $209 today in order to prevent $1,000 in damages in the year 2100. In other words, in a world of low r*, many climate policies that might once have appeared inordinately expensive start looking more affordable. A principle of regulatory policy is that costs and benefits should be calculated using the best available science. “Circular A-4” is now nearly two decades old. U.S. President Joe Biden has rightly issued an executive order to update it so that it reflects new economic understanding.
Climate change policy has been framed as a tradeoff between the benefits of reducing carbon dioxide emissions and the costs of those reductions. Economists have encouraged this cost-benefit mindset at many levels, perhaps most prominently by advocating a carbon tax that would make it more costly to pollute. The implicit assumption in that thinking is that polluting will naturally be cheaper than not polluting. But over the past five years, a remarkable development occurred: the cost of green technology fell sharply, and in some parts of the United States, clean energy sources are now cheaper than dirty ones.
From 2014 to 2019, the cost per kilowatt of solar panels fell by around 50 percent. In many parts of the country, building a new wind or solar farm costs less than running an existing coal plant or building a new natural gas plant. The U.S. Energy Information Administration projects that wind and solar farms will soon account for more than three-quarters of newly installed power plant capacity. These installations were supported by federal tax credits and by incentives at the state level. Still, the main driver of their falling costs was not such subsidies but advances in technology and companies simply “learning by doing” in the production and installation of renewable power facilities. Similar cost reductions occurred in electric vehicles. The main driver of the price of these vehicles is the cost of the lithium-ion batteries, which fell, on average, by more than 87 percent between 2010 and 2019. By 2024, in many parts of the United States, an electric vehicle with a 250-mile range will reach price parity with a comparable conventional vehicle.
These dramatic price declines in two key parts of the U.S. energy system are shifting the economics of climate policy: instead of making it more expensive to pollute, policymakers are looking for ways to make it cheaper to be clean. This puts some economists outside their comfort zones: there are no randomized controlled experiments that can tease out the precise causes of these price declines. That said, the available evidence suggests a potent role for policy in driving down costs. Solar energy prices declined because of high demand for solar panels even at very high prices, thanks in part to government spending through programs such as Germany’s Energiewende and the California Solar Initiative, which committed the German and California governments to purchase solar technology even when it was expensive. In the case of declining onshore wind energy prices in the United States, it seems that one important contributor was learning by doing; companies simply have gotten better at installing wind farms—the development of which, it should be noted, was aided by public subsidies. Similar trends are shaping the market for offshore wind energy, which exists thanks only to massive early government commitments, notably by Denmark, to make major purchases.
Of course, there are plenty of examples of such policies not panning out. One study has suggested that although certain initial small grants of up to $150,000 that the U.S. Department of Energy awarded between 1983 and 2013 helped some firms get a start, the second round of funding, with grants up to $1 million, had little effect on those firms’ future business prospects. Perennial funding favorites, such as small nuclear reactors and research on nuclear fusion, have made little federally funded progress (although private investment has recently spurred a rash of exciting fusion projects). The federal grant-making process is also conservative, with high political penalties for failure. If a program aimed at developing high-risk technologies doesn’t include failures, however, then it isn’t taking enough risks. Meanwhile, prioritizing the improvement of technology need not mean abandoning the idea of a carbon tax that would make it more expensive to pollute: policymakers can make future clean technology cheaper while also making it more costly to pollute today.
The drop in green energy prices might have many causes. But explaining it is a relatively easy task compared with accounting for a more fundamental change relating to prices, one that confounds conventional economic thinking: the astonishing stability of low inflation. Economists’ main theory of the rate of price inflation is the so-called Phillips curve, named for the economist A. W. Phillips, who introduced the concept in the 1950s. In its original form, the Phillips curve showed an inverse relationship between wage growth and the unemployment rate. Modern versions, applied to prices, link the current rate of inflation to both expected future inflation and a measure of economic slack, such as the gap between the rate of unemployment and its full-employment value—that is, a measure of the value of unused resources in the economy, such as people who cannot find a job.
The past five years have not been kind to this cornerstone of macroeconomics. From the 1960s through the early 1990s, the U.S. Phillips curve was remarkably stable, with an increase in slack reliably correlating to a reduction in the rate of inflation. Since 2000, however, this correlation has dropped nearly to zero. For example, during the economic expansion of 2018 and 2019, the U.S. unemployment rate stabilized at around 3.5 percent, but the rate of inflation failed to rise. In fact, in 2019, after subtracting out the influence of food and energy prices, inflation fell to just 1.6 percent—its lowest annual rate since 2015, when the unemployment rate stood at 5.3 percent.
In a marked break from the experience of past economic downturns, the trend of inflation becoming less and less sensitive to economic conditions continued into the COVID-19 recession. During the 1990 recession, for each percentage point increase in the unemployment rate, the core rate of inflation in the United States fell by 0.8 percentage points. In the 2000 recession, that figure was 0.4 percentage points. During the recession that followed the 2008 financial crisis, it was 0.3 percentage points. And in the current pandemic recession, it has been less than 0.1 percentage points.
The past five years have not been kind to the Phillips curve.
It is worth noting that the rate of inflation has been unusually hard to measure during the COVID-19 crisis because of shifts in demand, which implies that consumption bundles are changing (toward, say, home office supplies and away from hotel stays) more rapidly than the assumptions that are built into the inflation measures. Moreover, the pandemic has produced both a negative supply shock and a negative demand shock, which are, respectively, inflationary and disinflationary in standard Phillips curve models. Still, the stability of the rate of inflation through periods of historically low unemployment and then through periods of historically high unemployment remains puzzling.
There are plenty of possible explanations for this flattening of the Phillips curve. One is that the prices for many goods and, increasingly, the level of many wages are now set internationally and thus are less sensitive to domestic economic conditions. This is consistent with the fact that the prices of goods and services that are produced and consumed locally—such as housing rentals, restaurant meals, and hotel rooms—have tended to fall during recent downturns. Another explanation is that the apparent insensitivity of prices to economic conditions reflects the Fed’s success in stabilizing prices. But that hypothesis cannot account for why the Fed, the European Central Bank, and the Bank of Japan have not been able to get the rate of inflation up to two percent despite their clear desire to do so.
The persistently low rate of inflation, combined with the decline of r*, has made it harder for central banks to react in customary ways to sharp economic downturns. In the recessions of 1990, 2000, and 2008, the Fed reduced short-term interest rates by an average of roughly five percentage points. During the first three weeks of March 2020, as the COVID-19 crisis began in the United States, the Fed had far less wiggle room, since the core federal funds rate was already at just 1.6 percent. The Fed brought this rate down to essentially zero, but even that did not provide remotely close to the level of support the economy needed. So, as it did following the financial crisis, the Fed purchased long-term assets in order to stabilize asset markets and keep interest rates low, making it easier for companies to borrow and preventing a public health crisis from cascading into a financial crisis.
One reason the Fed had the confidence to go all in on long-term asset purchases is that, as the pandemic took hold, the central bank was just wrapping up a years-long review of its monetary policy framework. The Fed had undertaken the review partly in response to the decline in r*, which had led Fed economists to conclude that it was highly probable that the federal funds rate would be stuck at zero for extended periods. The review was both evolutionary and revolutionary. It was evolutionary in that the changes it codified, such as calling for long-term asset purchases when necessary and encouraging a willingness to tolerate persistent excursions of inflation over two percent, were consistent with years of research and were broadly understood and accepted by markets. But the review was revolutionary in that it happened at all. The process was transparent and systematic, relied on the best available science, and received input from the broader community of experts and the general public. Such transparency and public discussion stand in stark contrast to the secrecy and opacity that have historically characterized Fed decision-making. As a result of this process, the Fed has become a stronger institution.
But the Fed could have gone even further. The central bank’s new willingness to tolerate extended periods of inflation exceeding two percent has introduced a window for experimentation, which could increase comfort with raising the inflation target. To its credit, the Fed, through its review, has now created a means for publicly discussing this charged issue in a rational and scientific way.
As the pandemic rages on, it is too soon to know precisely what lessons it will eventually yield. But one is already clear: macroeconomic dynamics can change far more rapidly than economists have traditionally assumed.
Prior to the COVID-19 crisis, the largest monthly increase in the U.S. unemployment rate since 1950 was one percentage point, which occurred in 1953. In April 2020, the rate increased by 10.4 percentage points, then fell by 4.6 percentage points over the following three months. The speed of this spike and retreat was unprecedented; these were changes immensely larger than what recessions typically produce. At a casual level, this might seem unremarkable: after all, schools closed across the country and many businesses shut down. But at the level of economic modeling, the speed of these changes was a dramatic departure and has led to an especially wide range of projections about the recovery. One view is that the post-vaccine recovery will be rapid, as pent-up demand is released and extra savings are spent on vacations, restaurants, and other long-delayed services. A second view is that because of widespread business closures, workers will not have jobs to return to and that, after an initial fast recovery, the usual slow business-cycle dynamics will take over.
Macroeconomic dynamics can change far more rapidly than economists have traditionally assumed.
Although such dramatic changes call for a rethinking of some basic concepts, much of economic theory has fared quite well over the past five years. Take, for example, the concept of externalities. Public interest in doing something about climate change has sharply increased, from consumers choosing “green” options, to corporations purchasing carbon offsets for their employees’ travel, to a growing preference among investors for funds that pursue social and environmental goals alongside profits, to the Congressional Budget Office including climate damages in its long-term GDP forecasts. But carbon remains unpriced, and so the carbon externality persists. And although U.S. emissions of carbon dioxide have fallen because of the shift from coal to natural gas and renewables (and most recently because of the pandemic recession), the rate of the reduction has not been nearly fast enough. Because carbon is unpriced, carbon pollution will remain a problem that markets left alone will not solve.
The COVID-19 crisis has presented another example of externalities. Wearing a mask reduces your risk of contracting the virus, so that benefit of mask wearing is internalized. But it also provides a benefit to others by protecting them from you if you are infected, and because that benefit does not accrue to you directly, that benefit is not internalized. This is a classic externality: one person’s decision not to wear a mask affects the welfare of others. Economic theory suggests multiple ways for officials address this, such as making it costly not to wear a mask by fining those who refuse, making it pay to wear a mask by requiring it in places of business, and reducing the potential social costs of wearing a mask by casting it as a patriotic duty or as an act of compassion. Oddly, and tragically, policymakers have rarely pursued such solutions—often because they have denied that the contagion externality exists in the first place, an eerie echo of the way that many of the same policymakers deny the existence of climate externalities.
Another principle of traditional economic theory that has fared well is the importance of well-functioning institutions as the basis for a well-functioning economy. COVID-19 has revealed that U.S. public health institutions are not up to the task of responding to a pandemic. The institutional breakdown resulted from a combination of chronic underfunding and a presidential administration instinctively averse to science and expertise. Economists have invested heavily in ensuring the intellectual integrity and independence of the Federal Reserve, which has operated admirably and effectively in the crisis. They would do the country a service by turning their attention to the job of making other institutions just as resilient.