U.S. Housing as a Global Safe Asset: Evidence from China Shocks

"...In 1999, Mr Bernanke had chided the Bank of Japan for failing to rekindle Japanese growth after a bubble burst, despite reducing interest rates to zero. Yet in the 15 years after he christened the saving glut, finance ministries and central bankers around the world became familiar with the struggle to maintain steady growth in the context of zealous saving. Since 2000 alone, the value of global wealth held by households, firms and governments has roughly tripled, from $160trn to $510trn, or from about 460% of global gdp to 610%, according to McKinsey Global Institute, a think-tank…The contribution of growth in reserves to savings was most pronounced around the time Mr Bernanke sounded his warning. From 1998 to 2008, official foreign-exchange reserves jumped from 5.2% of global gdp to 11.5%, powered by a steady rise in oil prices and reserve accumulation by China. During this period, reserve growth probablydominated other sources of saving; research by Francis Warnock and Veronica Cacdac Warnock of the University of Virginia suggests that reserve-accumulation in the year to May 2005 alone reduced the yield on ten-year Treasury bonds by 0.8 percentage points.Reserve growth paused during the global financial crisis, then resumed in the years after, reaching a peak of 15.2% of global gdp in 2013(see chart 2).... Work by Lukasz Rachel, of the London School of Economics, and Larry Summers, of Harvard University, reckons that over the past half century, rising government debt across rich economies pushed up interest rates by about 1.5 percentage points.This effect was more than balanced out by other factors in the past, but might not be in the decades ahead....Increased inequality accounts for about 0.6 percentage points of the decline in rich-world interest rates since the 1970s, say Messrs Rachel and Summers...High-rolling households have not been alone in stockpiling savings. For decades, corporations have been hoarding money as well, retaining a large share of their hefty net profits. According to Peter Chen, of the Analysis Group, an economic consultancy, and Brent Neiman, of the University of Chicago, and Loukas Karabarbounis, of the University of Minnesota, annual global corporate saving rose from less than 10% of world gdp to nearly 15% between 1980 and 2015. The corporate sector has been acting as a net lender to the global economy, rather than as a net borrower from it...Corporate saving, in contrast, rose relatively slowly before 2000, then much faster thereafter, as firms salted away cash from increased profits. In America, for instance, corporate profits have hovered above 10% of gdp for most of the period since 2006, after never rising above 8% over the prior quarter century....In a recent paper examining the effects of demographic change on saving,Etienne Gagnon, Benjamin Johannsen and David López-Salido of the Federal Reserve Board suggest that ageing in America may account for about one percentage point of the drop in interest rates since the 1980s.(Other recent work finds still larger effects, of as much as three percentage points.) If past is prologue, rates seem sure to remain low. Barring a surge in procreation, or the embrace of a dystopian “Logan’s Run” approach to the aged, the world’s population will continue to get older. The share of global population over the age of 50 rose from 15% in the 1950s to 25% today, say Adrien Auclert and Frédéric Martenet, of Stanford University, Hannes Malmberg, of the University of Minnesota, and Matthew Rognlie, of Northwestern University. It is expected to rise to 40% by 2100(see chart 3). That may well turn out to be an underestimate, if recent fertility trends are anything to go by. In 2021, India’s birth rate declined to just 2.0 children per woman—below the rate at which births and deaths are in rough balance. Indeed, a growing number of emerging markets have flipped to the slow population growth common in rich countries. Recent research by Matthew Delventhal of Claremont McKenna College, Jesús Fernández-Villaverde of the University of Pennsylvania and Nezih Guner of the Universitat Autònoma de Barcelona concludes that such transitions—the switch from high mortality and fertility rates to low ones which accompanies economic development—are happening faster over time. The transition took a half century or more 100 years ago, but now tends to be compressed into just two or three decades. Some 80 countries have completed this transition, and in virtually all the rest it is under way....Recent work by Noëmie Lisack, of the Banque du France, Rana Sajedi, of the Bank of England, and Gregory Thwaites, of the University of Nottingham, estimates that this habit of leaving behind savings will by mid-century depress interest rates by nearly half a percentage point relative to current levels. With neither inequality nor the level of reserves showing signs of sustained fall, the ineluctable force of demography should continue to drive savings growth. The world, in other words, may come to look ever more like Japan. There, the median age is 48, more than a quarter of the population is over 65, and the yield on a 30-year government bond is a cool 0.8%, despite a government debt load of 259% of gdp. A generation ago, Mr Bernanke reckoned that Japan’s lacklustre growth and subterranean rates of inflation and interest were the consequence of “self-induced paralysis” by the central bank. Today, such realities seem more like the dull fate of a world with more savings than it quite knows what to do with...."

Economist Staff, "Why the world is saving too much money for its own good," The Economist, February 5, 2022, https://www.economist.com/briefing/2022/02/05/why-the-world-is-saving-too-much-money-for-its-own-good

Why the world is saving too much money for its own good

In 2005, ben bernanke, then a member of the Federal Reserve’s Board of Governors, wondered at a tide of money washing over American shores—and worried about its consequences. Grasping in a speech for a way to describe the phenomenon, he coined a phrase. “Over the past decade,” he noted, “a combination of diverse forces has created a significant increase in the global supply of saving—a global saving glut.” Savers of all sorts—from older Americans preparing for retirement to oil-exporting countries accumulating sovereign-wealth funds—were shoving more money into stocks and bonds than could be put to use by those looking to invest in plants and equipment.

Now the pandemic has shaken pillars of the global economy in ways which could fundamentally alter saving patterns. Tight labour markets are shifting money to workers who are eager to spend, contributing to the highest inflation in a generation. Central banks which had found themselves unable to push interest rates down enough to keep inflation from falling below their targets are beginning to push rates up to keep inflation from soaring. Yet while new enthusiasm for government borrowing or a retreat from globalisation could help to drain savings and establish a new normal, it seems more probable that Mr Bernanke’s glut will persist, thanks to old habits and old people, who are a growing share of the world’s population.

The rising reservoir of global savings, most of which is held in bank deposits, bonds, corporate equity and property, has been fed by three main tributaries: governments hungry for foreign-exchange reserves, penny-pinching households and firms, and workers nearing retirement age. It was the first flow, saving by governments, which preoccupied Mr Bernanke. Governments’ accumulation of foreign-exchange reserves adds to saving in two ways. Resource-exporting economies save part of the windfall earned from their exports and plough it into stocks and bonds. Some of these piles are held as official reserves; the Russian government has reserves, excluding gold, valued at $460bn, while Saudi Arabia’s are worth $440bn. Windfalls have also been shifted into sovereign-wealth funds; that of Abu Dhabi is worth almost $700bn, while Norway’s is valued at more than $1.3trn.

Other economies pile up foreign-exchange reserves as they intervene in markets to reduce the value of their currencies, to boost exports or to build up a hoard of safe assets which can be drawn upon in times of financial stress. In effect, these interventions squeeze consumption in their home economies, reducing spending relative to production and thus contributing to current-account surpluses which must be absorbed by the rest of the global economy. Reserves held by South Korea, Singapore, Taiwan and India have grown into the hundreds of billions. No country has engaged in such practices to more disruptive global effect than China, which holds some $3.2trn in foreign-exchange reserves.

Reserves plateaued thereafter (and indeed fell slightly as a share of gdp), and a decline in the years ahead cannot be ruled out. A protracted period of post-pandemic financial stress could force some emerging economies to deplete their reserves. If economic strains and geopolitical tensions force Russia and China to draw down their hoards, that might place upward pressure on interest rates.

Yet it is also possible that the pandemic will lead to a new surge in reserves. Yes, a shift to zero-carbon energy may eventually doom fossil-fuel windfalls, but the transition might well mean high prices for oil and gas, since new production is likely to stagnate. Meanwhile, the pandemic and its aftermath will probably reinvigorate the appeal of defensive foreign-exchange reserves. During the financial panic of March 2020 and again in recent months, as straitened global conditions squeezed emerging markets, the economies which weathered stresses best were those with ample foreign-exchange reserves. This lesson has already been put to use. By the autumn of 2021, reserves were roughly $1trn higher than they were before covid-19.

The effects of reserve accumulation could also be offset by increased government borrowing. Government debt loads, already high pre-pandemic, have exploded over the past two years; in 2020 alone, public debt as a share of gdp surged by nearly 20 percentage points across advanced economies, to 123%, and nearly ten points across emerging economies, to 63%. Work by Lukasz Rachel, of the London School of Economics, and Larry Summers, of Harvard University, reckons that over the past half century, rising government debt across rich economies pushed up interest rates by about 1.5 percentage points. This effect was more than balanced out by other factors in the past, but might not be in the decades ahead.

A first-class problem

A second stream of saving has flowed from the households and firms which have done best over the past few decades. Since the 1970s, inequality has risen across many economies. Wealthier households have a higher propensity to save, so this shift in the distribution of income contributed to the saving glut, according to work by Atif Mian, of Princeton University, Ludwig Straub, of Harvard University, and Amir Sufi, of the University of Chicago. From 1983 to 2019, the share of American income going to the top 10% of the income distribution rose by 15 percentage points, they reckon. Because of this “saving glut of the rich”, average annual saving by the top 1% of American earners alone has outstripped annual average net domestic investment since 2000. Increased inequality accounts for about 0.6 percentage points of the decline in rich-world interest rates since the 1970s, say Messrs Rachel and Summers.

High-rolling households have not been alone in stockpiling savings. For decades, corporations have been hoarding money as well, retaining a large share of their hefty net profits. According to Peter Chen, of the Analysis Group, an economic consultancy, and Brent Neiman, of the University of Chicago, and Loukas Karabarbounis, of the University of Minnesota, annual global corporate saving rose from less than 10% of world gdp to nearly 15% between 1980 and 2015. The corporate sector has been acting as a net lender to the global economy, rather than as a net borrower from it.

As with reserve accumulation, the relative importance of such factors has waxed and waned. Income inequality rose sharply from around 1980 to 2000. In the years since, it has levelled off in some economies, like Britain’s, and increased at a slower pace in others, like America’s. Corporate saving, in contrast, rose relatively slowly before 2000, then much faster thereafter, as firms salted away cash from increased profits. In America, for instance, corporate profits have hovered above 10% of gdp for most of the period since 2006, after never rising above 8% over the prior quarter century.

Income inequality and corporate profitability cannot be forecast with any certainty. Both reflect the interaction of myriad forces, from the balance of corporate and labour power, to the state of technological progress and productivity growth, to government tax and regulatory policy. It is possible that the trends of the past half century might be upended by the pandemic and its aftermath. Over the past 18 months, tight labour markets helped push wages upward and strengthened workers’ leverage in bargaining with their employers. Slower growth in the labour forces of ageing societies could help to preserve these gains, and perhaps enable a resurgence by organised labour. Firms—especially big and profitable technology ones—are in the cross-hairs of regulators looking to boost competition. Better times for workers should also squeeze profits, in addition to reducing inequality.

A retreat from globalisation could amplify these trends. It would increase the earning power of the working masses in rich countries, while hitting the profits of multinational firms and the higher incomes of their white-collar workers. On the other hand, substantial reversals in inequality are relatively rare in recent economic history. The great compression in incomes that occurred from the 1910s into the post-war decades occurred as fortunes were hammered by the Depression and liquidated to fund wars, as taxes on the rich soared well above current levels. For now, such upheavals seem unlikely.

What is more, as Mr Mian and co-authors write, the effects of inequality on saving can feed on themselves. As high saving by the rich pushes down interest rates, they argue, poorer households increase their borrowing to sustain their consumption. But as debt piles up, they find themselves forced to reduce spending to pay back loans. Their debt payments, furthermore, represent a transfer of more money to rich households whose purchases of assets (like mortgage-backed securities) effectively finances the borrowing of the non-rich. The trap which results—of perpetually high inequality, low interest rates, and high debt among poorer households—could prove difficult to escape, sustaining the savings of the rich as a potent macroeconomic force.

Far more certain is the third great river of savings, whose flow, which has grown in importance, might well swamp other post-pandemic changes in behaviour. The world is not getting any younger, and in coming decades the savings of the old stand to apply relentless pressure on the macroeconomy. Across time and countries, household saving follows a reliable pattern. When workers are young, they save little or even take on debt. Their savings rise through their 30s and 40s before peaking a decade or so before retirement. As populations have grown older over the past half century, in the rich world especially, the share of workers in their prime saving years has risen as well, leading to ever more money in nest-eggs and ever lower yields on the assets therein.

That may well turn out to be an underestimate, if recent fertility trends are anything to go by. In 2021, India’s birth rate declined to just 2.0 children per woman—below the rate at which births and deaths are in rough balance. Indeed, a growing number of emerging markets have flipped to the slow population growth common in rich countries. Recent research by Matthew Delventhal of Claremont McKenna College, Jesús Fernández-Villaverde of the University of Pennsylvania and Nezih Guner of the Universitat Autònoma de Barcelona concludes that such transitions—the switch from high mortality and fertility rates to low ones which accompanies economic development—are happening faster over time. The transition took a half century or more 100 years ago, but now tends to be compressed into just two or three decades. Some 80 countries have completed this transition, and in virtually all the rest it is under way.

What is more, the pandemic further depressed birth rates in many countries. China’s birth rate touched a record low in 2021, potentially bringing forward the era of declining Chinese population. America experienced a baby bust too, which in combination with falling immigration depressed the population growth rate to just 0.1% in 2021—the smallest annual increase on records going back to 1900. The end of the pandemic could bring a rebound in birth rates. But there is no mistaking the broader trend: the world is greying, fast.

Innumerable shades of grey

Will the effect of ageing on saving necessarily remain the same in future as it was over the past half century? In an influential book, Charles Goodhart, of the London School of Economics, and Manoj Pradhan, of Talking Heads Macroeconomics, a research firm, argue that the greying of the population will depress interest rates only up to a certain point, after which there will be a “great demographic reversal”. Their view rests in part on the observation that while workers on the verge of retirement save heavily, those already retired begin to spend down their stores of stocks and bonds. An increase in the share of the population above retirement age, then, could mean that the proportion of workers in their high-saving years will peak and then decline, dragging down saving and pushing up interest rates.

A great demographic reversal seems intuitive, particularly in places like America where an oversize cohort—the baby-boomers—is easing into retirement. But other economists say there are reasons to expect ageing to continue to depress interest rates. They note, for example, that it is the age profile of a population as a whole which matters. Even as more people retire, the age of the typical working person will continue to rise toward those prime saving years. There are boomers aplenty, but the median age in America is still just 38. Another reason is that, in the emerging world, a larger share of workers have their prime saving years still ahead of them. The median age in India is only 28, for instance. So long as financial markets remain reasonably integrated around the world, higher saving anywhere helps to depress interest rates everywhere.

Perhaps most important, people in retirement do not tend to spend everything. Rather, for a number of motives—to avoid outliving their savings, or to provide for heirs, among others—they tend to maintain large stocks of wealth well into retirement. In Britain, for instance, as of 2018, people 80 or older held more wealth than those aged 45 to 49. Recent work by Noëmie Lisack, of the Banque du France, Rana Sajedi, of the Bank of England, and Gregory Thwaites, of the University of Nottingham, estimates that this habit of leaving behind savings will by mid-century depress interest rates by nearly half a percentage point relative to current levels. With neither inequality nor the level of reserves showing signs of sustained fall, the ineluctable force of demography should continue to drive savings growth.

The world, in other words, may come to look ever more like Japan. There, the median age is 48, more than a quarter of the population is over 65, and the yield on a 30-year government bond is a cool 0.8%, despite a government debt load of 259% of gdp. A generation ago, Mr Bernanke reckoned that Japan’s lacklustre growth and subterranean rates of inflation and interest were the consequence of “self-induced paralysis” by the central bank. Today, such realities seem more like the dull fate of a world with more savings than it quite knows what to do with.

The FRBNY's Dean Parker and Moritz Schularick find that term spread is a useful indicator of financial crisis both domestically and overseas The effect is grounded in an uptick at the short end of the yield curve likely driven by higher risk taking by financial intermediaries in the run up to a crisis, "...It is clear that theterm spread is significantly lower in the years before a financial crisis, both in the international sample and in the U.S. case. When all countries are considered, the term spread is approximately one percentage point lower on average in the two years before the onset of a crisis. For the United States specifically, this effect is almost doubled in the year before an onset, reaching 2 percentage points lower than usual on average.For a concrete example, in the run-up to 2007-08 financial crisis in the United States, long-term rates declined despite increasingly tight monetary policy, giving rise to Alan Greenspan’s famous description of the development as a “conundrum.” As a result, the term spread fell by 3 percentage points at the time. Similarly, though not as widely discussed contemporarily, the term spread declined a total of 3.3 percentage points in the United States between 1924 and 1928 in the lead-up to the 1929 crash and Great Depression. In the run-up to normal recessions, however, we observe only limited movement.... There are various possible explanations for the predictive ability that depend on whether the decrease of the term spread is driven by the short-term rate rising or the long-term rate falling. As seen in the panel chart below, when taken in isolation, rising short-term rates seem to play the dominant role for the narrowing of the term spread. This is because the long-term rate does not seem to behave differently in the run-up to a crisis. Such an increase in short-term funding rates relative to the yield on long-term investments is indicative of declining interest margins and higher risk taking by financial intermediaries. Our findings hence echo the evidence for rising financial risk taking in financial markets before financial crises that is reflected both in quantities of credit growth and a depressed price of risk..."

Dean Parker and Moritz Schularick, "The Term Spread as a Predictor of Financial Instability," Liberty Street Economics, November 24, 2021,
https://libertystreeteconomics.newyorkfed.org/2021/11/the-term-spread-as-a-predictor-of-financial-instability/

The Term Spread as a Predictor of Financial Instability

The term spread is the difference between interest rates on short- and long-dated government securities. It is often referred to as a predictor of the business cycle. In particular, inversions of the yield curve—a negative term spread—are considered an early warning sign. Such inversions typically receive a lot of attention in policy debates when they occur. In this post, we point to another property of the term spread, namely its predictive ability for financial crisis events, both internationally and in historical U.S. data. We study the predictive power of the term spread for financial instability events in the United States and internationally over the past 150 years.

The Term Spread as a Crisis Predictor

The behavior and slope of the yield curve are longstanding tools for predicting the onset of economic recessions within the next year, or future employment growth. Consideration of the yield curve as a predictor for recessions rose to prominence with Estrella and Mishkin’s (1996) paper. Rudebusch and Williams (2008) also highlighted the power of the term spread in predicting aggregate economic downturns. While there are many academic studies and other articles on the efficacy of the term spread and its inversion for predicting the onset of normal recessions, little is known about its forecasting performance for financial crises. Our study aims to close this gap.

In normal times, a bond with a longer maturity fetches a higher rate on the market due to increased duration risk. However, in the lead up to an economic downturn, various factors can alter the shape of the yield curve. Changes in risk attitudes and search for yield, central bank policy actions, and international demand for safe assets can all influence the term spread and drive the yield for short-term debt closer to or even above the yield for long-term debt. Are such changes in the term spread predictive of a higher risk of financial instability?

The Behavior of the Term Spread around Financial Crises

For our analysis, we use the Macrohistory Database compiled by Jordà, Schularick, and Taylor (2017). The data cover eighteen advanced economies around the world from 1870 to 2017 and contains country specific short- and long-term rates. For the United States specifically, these are the three-month Treasury bill and ten-year Treasury bond rates, respectively. The main benefit of the Macrohistory Database is that it enables examining financial crises in a variety of economic and temporal contexts. Our international sample consists of eight-three crises out of 2,304 total observations, or approximately four percent of the sample. The United States has six years marked as crisis years over the past 128 years: 1873, 1893, 1907, 1930, 1984, and 2007. A financial crisis is defined by Jordà, Schularick, and Taylor as a period of banking distress characterized by major bank failures, exceptional losses in the banking sector, and/or significant government intervention. Briefly, the 1873 and 1893 crises were bank runs precipitated by railroad failures, 1907 was brought on by the collapse of Knickerbocker Trust, 1984 marks the beginning of the Savings & Loan crisis, and 1930 and 2007 saw the Great Depression and Global Financial Crisis, respectively. We also corroborate our results using a different definition of financial instability, namely large declines in the market value of bank equity (Baron, Verner, and Xiong 2020).

The chart below plots the term spreads in the five years before and after a financial crisis or recession. We test, using regression analysis, whether the level of the term spread is above or below normal levels in these windows. The blue dots and bars in the chart represent point estimates and confidence intervals, respectively. They can be read as the average levels of the term spread in the lead-up to and aftermath of an event compared to periods where no event occurred. Positive values imply a higher level of the term spread than usual and negative values imply a lower level than usual.

Using the Term Spread as a Forecasting Tool

With the knowledge that something abnormal occurs in the term spread in the years preceding a financial crisis, the natural next step is to examine its efficacy as a predictive tool and how it functions when used as part of a larger model including other known predictors of crises. Schularick and Taylor (2012) show that real bank-loan growth is a strong predictor of impending crisis and that model serves as the base for our later models. As for definitions of financial crises, we use both the Jordà-Schularick-Taylor narrative and the bank equity declines from Baron, Verner, and Xiong (2020).

We run logistic regressions to predict the two types of crises and compare models using real bank-loan growth and term spread as predictors. We also test the inclusion of an inverted yield curve (a negative term spread). The results are reported in the table below with each column reporting a separate regression.

Our main finding is that, for both crisis definitions, including the term spread alongside standard predictors yields a significant improvement in forecasting power. A smaller term spread increases the risk of a financial crisis. Using the inverted yield curve as a predictor results in a slightly worse predictive ability of the model.

There are various possible explanations for the predictive ability that depend on whether the decrease of the term spread is driven by the short-term rate rising or the long-term rate falling. As seen in the panel chart below, when taken in isolation, rising short-term rates seem to play the dominant role for the narrowing of the term spread. This is because the long-term rate does not seem to behave differently in the run-up to a crisis. Such an increase in short-term funding rates relative to the yield on long-term investments is indicative of declining interest margins and higher risk taking by financial intermediaries. Our findings hence echo the evidence for rising financial risk taking in financial markets before financial crises that is reflected both in quantities of credit growth and a depressed price of risk (Krishnamurty and Muir 2017; Baron and Xiong 2017).

The term spread is often used as an early warning indicator for recessions. In this post, we showed that it performs well as a predictor of financial crises, both internationally and in the United States alone. There is a significant benefit to including the term spread as a predictor for two separate crisis definitions. We identified that this effect is driven by the short end of the yield curve rising and offered a potential explanation based in higher risk taking by financial intermediaries in the time before a crisis.

Tim Lee on how he misunderstood the so called housing "bubble""...What if the big problem in the early 2000s wasn’t an excess of houses but a shortage of them?...The idea that the 2000s housing boom reflected a housing shortage probably sounds wrong to you. After all, “everyone knows” that the mid-2000s saw an unprecedented building boom. Here’s a chart of single-family housing starts between 1974 and 2006: This chart shows a clear upward trend from 1980 through 2005, with housing starts reaching an all-time record in 2005. However, single-family houses aren’t the only types of home! People also live in multi-family apartment buildings and in manufactured homes. When you factor in those additional housing types, things look different: Here I’ve stacked up the three types of housing to show the total number of homes constructed each year. Over the 30 years between 1975 and 2005, there was a consistent pattern where homebuilding peaked around 2 million units during each economic boom. Total housing production in 2005 was actually slightly below the all-time high of almost 2.3 million units reached in 1978. And even this chart makes the 2005 housing boom look more impressive than it actually was, because it doesn’t factor in population growth. A better way to judge the pace of homebuilding is to compare the growth of homes to the growth of people: Here I took the top line from the previous chart and divided it by the total number of housing units to get a growth rate of homes in red. I compare this to the growth rate of people in blue. As you can see, it used to be common for the housing stock to grow more than 2 percent in a year. But after the growth rate of homes fell below 2 percent in 1987, it never reached that level again. New homes in 2005 accounted for 1.8 percent of housing inventory, an 18-year high, but far below the peak construction rates of the late 1970s and early 1980s. And this slowdown in housing construction can’t be explained by a slowdown in population growth. The nation’s population was growing at around the same rate—just under 1 percent—in the mid-1970s and the mid-2000s...."

Tim Lee, "The 2000s housing bubble was greatly exaggerated," Full Stack Economics, November 5, 2021, https://fullstackeconomics.com/the-2000s-housing-bubble-was-greatly-exaggerated/

The 2000s housing bubble was greatly exaggerated

In 2006, the New York Times published a chart that I consider the defining image of the housing crisis. Based on data from the economist Robert Shiller, it showed how American home prices have changed over more than a century.

For a few years after 2006, that forecast seemed to come true. Housing prices fell rapidly between 2007 and 2012. But then something surprising happened.

Today you can download the same data set from Robert Shiller’s website, updated to the present day. And it shows that housing prices are now above the supposedly unsustainable levels of 2006. And that’s after adjusting for inflation.

But that leads to a question that at first glance might seem crazy: what if those same explanations largely explain the housing boom that peaked in 2006? What if the big problem in the early 2000s wasn’t an excess of houses but a shortage of them?

An alternative view of the housing crisis

That’s the thesis of Shut Out, a 2019 book by Kevin Erdmann, a scholar at the Mercatus Center. Erdmann’s book attracted little notice at the time it was published. But his thesis has gotten more plausible as housing prices have zoomed upwards over the last two years.

Erdmann argues that policymakers misdiagnosed the causes of the housing boom, and that led to catastrophic policy errors. In particular, because the Federal Reserve thought housing was overvalued in 2007, it didn’t cut rates fast enough in response to the housing crash. That helped turn what might have been only a mild, industry-specific downturn into a severe, economy-wide recession. And that recession, in turn, made the housing crisis bigger than it needed to be, since many previously solvent homeowners lost their jobs or saw their mortgages go under water.

This view is shared by Gregor Schubert, a UCLA professor who recently earned his economics PhD studying the housing crisis at Harvard.

“To me, the strongest piece of evidence in favor of the recession causing the housing crisis rather than the housing crisis causing the recession is the fact that prices snapped back up, in similar geographic configuration, after the economy has recovered,” Schubert told me.

Schubert points out that the places with the highest 2006 prices, like Los Angeles, did relatively well during the Great Recession and “took off like a rocket ship afterwards.” In contrast, cities that didn’t experience much of a bubble, like Atlanta, suffered big home price declines in the crash. That, he said, suggests that the housing crash was driven by broader macroeconomic factors more than an oversupply of homes in any particular part of the country.

If Erdmann and Schubert are right, we’re still living with the consequences of misdiagnosing the housing boom as a speculative bubble. After the crash, housing construction fell to its lowest level in decades, and remained depressed for several years. That under-production contributed to the housing shortages that now plague much of the country.

There was never a national housing glut

The idea that the 2000s housing boom reflected a housing shortage probably sounds wrong to you. After all, “everyone knows” that the mid-2000s saw an unprecedented building boom. Here’s a chart of single-family housing starts between 1974 and 2006:

And even this chart makes the 2005 housing boom look more impressive than it actually was, because it doesn’t factor in population growth. A better way to judge the pace of homebuilding is to compare the growth of homes to the growth of people:

As you can see, it used to be common for the housing stock to grow more than 2 percent in a year. But after the growth rate of homes fell below 2 percent in 1987, it never reached that level again. New homes in 2005 accounted for 1.8 percent of housing inventory, an 18-year high, but far below the peak construction rates of the late 1970s and early 1980s.

And this slowdown in housing construction can’t be explained by a slowdown in population growth. The nation’s population was growing at around the same rate—just under 1 percent—in the mid-1970s and the mid-2000s.

Metros with limited housing supply had higher prices

Of course, even if there wasn’t a national housing glut, there still might have been oversupply in particular cities. So let’s get more granular.

When people think of the housing boom, they think of two things: a boom in home prices and a boom in building construction. You can find examples of both things happening in the early 2000s. But to a large extent, they happened in different places.

Superstar cities like San Francisco, Los Angeles, New York, and Boston saw home prices soar to record levels, but these cities were building homes at rates far below the national average. Strict regulations prevented the construction of new homes, so rising demand drove big increases in home prices.

On the other hand, cities like Atlanta, Dallas, and Minneapolis had strong economies and significant housing demand in the 2000s, but they coped with it by building a lot of housing. So prices rose only modestly.

If every American city fell into one of these two categories, it might have been obvious that what was happening in the 2000s was not a “housing bubble” but simply an object lesson in the importance of adequate housing supply. Cities with ample housing supply maintained moderate home prices. Cities that didn’t, didn’t.

What made the 2000s confusing was that there was a third category of cities that had rapid housing growth and rapidly rising home prices. These included Phoenix, Las Vegas, Miami, Tampa, and California’s Inland Empire. If you wanted to make the case that the US was in the throes of an irrational housing bubble, these were the cities to focus on.

In Phoenix, for example, home prices doubled between 2001 and 2006, even though the Phoenix metro area was building homes at a record pace. New housing permits in the Phoenix area rose from 41,000 in 2001 to 63,000 in 2004—far more than the 37,000 housing permits in the much larger Los Angeles area. Yet home prices in Phoenix still rose by 17 percent between those three years—and then another 54 percent between 2004 and 2006.

The combination of rapid building and soaring prices in cities like Phoenix led many people to conclude that this was an irrational and unsustainable housing bubble. You couldn’t say the high prices were the result of limited housing supply because Phoenix was building homes at a brisk pace.

But Erdmann has a different interpretation—one that’s backed up by Schubert’s empirical work: the housing booms in these cities were largely driven by underbuilding in the superstar cities.

The spillover hypothesis

In the 1990s and early 2000s, a lot of college-educated people moved to Silicon Valley for high-paying jobs at technology companies. Housing was scarce, but the new workers made enough money to outbid locals for market-rate apartments. As rents and home prices rose, middle- and working-class locals found it harder and harder to afford housing. Every year, thousands of them fled the San Francisco Bay Area for more affordable housing elsewhere.

These lower-earning workers didn’t disperse randomly across the country. Many moved to Phoenix, which was also getting a lot of inbound migrants from Los Angeles. The overall scale of California-to-Arizona migration was massive. In 2005, for example, Arizona welcomed about 90,000 new residents from California—1.5 percent of Arizona’s population.

Many of these people weren’t wealthy enough to afford a home in Los Angeles or San Francisco, but they weren’t necessarily poor either. Many could comfortably afford to buy a house in a less expensive housing market like Phoenix.

Erdmann has lived in Phoenix since the 1990s, so he got to see this process first-hand. “You can go around and talk to people at neighborhood parties here and a good portion of them will say ‘we were in San Francisco and couldn't afford it any more,’” he told me.

A similar story applies to most of the other cities we think of as typifying the housing bubble. Around 50,000 people moved from California to Nevada in 2005, with many settling in booming Las Vegas. Other Angelenos were moving to the Inland Empire, which the Census Bureau considers a separate metropolitan area from Los Angeles but is only about an hour’s drive away.

On the East Coast, hundreds of thousands of New Yorkers and Bostonians moved to Florida in the early 2000s, driving housing booms in Orlando, Tampa, and Miami.

In effect, superstar cities were outsourcing their housing development problems. In the early 2000s, the Phoenix area was expanding its housing stock quickly to accommodate the flood of California refugees. But there’s an inherent limit to how quickly a region can build new houses. So even with the high rate of housing production in Phoenix, prices still rose.

In a empirical paper published earlier this year, Schubert found a causal connection between inter-city migration and home price spillovers. The more people who move between two cities, the more home prices will “spill over” from one city to the other. When San Francisco housing gets more expensive, people move to Phoenix and Phoenix housing gets more expensive too.

The conventional view holds that rising home prices in Phoenix and Las Vegas were driven by speculators hoping to find a “greater fool” who would pay an even higher price—or by gullible consumers, likely financed by subprime loans, who didn’t realize they were getting a bad deal. But the migration story suggests another possibility: middle-class homebuyers were coming from Los Angeles or San Francisco and simply saw Phoenix homes as a bargain.

I ran this theory by Mark Calabria, a Cato Institute scholar who ran the Federal Housing Finance Agency under President Trump.

“Migration trends very clearly show growth into Phoenix from places like California,” Calabria acknowledged. “A lot of the supply response in Phoenix was demand by people not being able to live in San Francisco. But I would still argue that we overbuilt in Phoenix at the time.”

Rents weren’t rising

Dean Baker is a left-leaning economist at the Center for Economic and Policy Research. He’s known as one of the first experts to “call” the housing bubble. Way back in 2002, he wrote an article bluntly concluding that “there is a housing bubble” that “must eventually come to an end.” Two years later, he put his money where his mouth was, selling his condo in Washington DC and renting a home while he waited for the bubble to pop. It turned out to be a smart move, as he was able to buy back into the market near its low point in 2009.

Baker has continued to defend the bubble thesis, most recently in a 2018 paper. So I thought he’d be the perfect person to poke holes in Erdmann’s arguments.

As you can see above, almost 11 percent of housing units were vacant in 2000. That was near the highest level in decades, and the vacancy rate rose even more during the early 2000s. If the nation were suffering from an acute housing shortage during this period, we should have expected vacancy rates to move in the other direction.

This is a good point as far as it goes. But I find it striking that vacancies never really decline during the 1980 to 2010 period. If vacancies had been low in the 1990s and then shot up during the 2000s, that would be strong evidence that there was an over-supply of housing. But the gradual rise makes me wonder whether there were longer-term forces pushing up the vacancy rate.

I’m not sure what that might be. Maybe people were buying more vacation homes. Maybe people were moving away from rust-belt cities and leaving vacant homes behind. Maybe the Census Bureau was getting better at counting vacant properties. Without knowing where these vacancies occurred and why, it’s hard to know if they are bubble-related or not.

Baker also points out that home prices rose much faster than rents during the early 2000s. The chart below compares inflation-adjusted rents to inflation-adjusted home prices—with both indexed to be equal in 2000.

This is a great point, and it convinces me that the 2000s housing boom wasn’t entirely driven by housing shortages in a few coastal cities. At the same time, I don’t think it comes close to proving that the run-up in prices in the early 2000s was entirely a speculative bubble.

Much of the divergence between home prices and rents during the early 2000s can be explained by falling interest rates. Suppose a family can afford to pay $1,000 per month on a mortgage. In 2000, they would have gotten a mortgage at around 8 percent; that would have enabled them to borrow around $135,000. By 2004, interest rates had fallen to 5.8 percent, so they could have borrowed $170,000 with the same $1,000 monthly payment.

So falling interest rates enabled families to buy roughly 25 percent more house in 2004 than they could buy in 2000 with the same monthly payment. That means falling interest rates can explain almost all of the 28 percent increase in the ratio of home prices and rents between 2000 and 2004.

If you had taken a snapshot of the housing market in early 2004, things would have looked pretty normal. Home prices in San Francisco and Los Angeles were soaring because of local housing shortages. In most other places, prices had risen modestly thanks to falling interest rates. There were a few places where prices seemed to be going crazy—especially Las Vegas and Miami. But this wasn’t the first time in American history that a few local housing markets had dramatic boom-and-bust cycles.

But the last two years of the housing boom are harder to explain. Between January 2004 and January 2006, national home prices rose an additional 20 percent, adjusted for inflation, at a time when rents were trending down and mortgage rates were trending up. Home prices got especially crazy in Las Vegas (up another 48 percent, inflation adjusted, in two years), Phoenix (up 64 percent), and Miami (up 52 percent).

So I think it would go too far to say there was no housing bubble at all during the 2000s. Things clearly got frothy near the end. The high prices of 2005 were probably not sustainable. Some correction was probably inevitable, at least in a few cities.

However, people wrongly concluded that the entire national housing boom of the 2000s was a speculative bubble. They came to believe that prices were far above levels that could be justified by fundamentals, and that there were millions of excess homes.

The Fed botched the housing downturn

This mistake had profound consequences because the perceived size of the housing bubble influenced decision-making by the Federal Reserve. The Fed started raising its benchmark interest rate in 2004, reaching a peak of 5.25 percent in mid-2006. Part of the Fed’s goal was to raise mortgage rates and thereby cool a housing market it viewed as overheated.

The policy worked—too well, as it turned out. New housing starts peaked in early 2006 and began to fall. By June 2007, new home starts were at their lowest level in a decade, and home prices were falling.

At this point, the Fed faced a crucial choice: when and how quickly to cut interest rates. Cutting earlier and more aggressively would have provided a cushion for the housing market by pushing down mortgage rates and luring new buyers into the market. An early rate cut also would have sent a signal that the Fed wasn’t going let the bottom fall out of the housing market. That would have boosted the confidence of homebuilders and might have limited job losses in the construction sector. This would have been a no-brainer if the Fed had believed that the housing boom had been mostly healthy with a bit of froth at the end.

But things looked different if you believed that a massive housing bubble had bequeathed the country with millions of extra housing units that needed to be “worked off.” In that scenario, aggressive rate cutting would have merely been delaying an inevitable reckoning. Even worse, it might have allowed the bubble to get even bigger, leading to an even bigger crisis a few years down the line. In this view, keeping rates high at 5.25 percent was tough but necessary medicine.

Unfortunately, the Fed took this latter approach. The central bank held rates at 5.25 percent at its June 2007 meeting and maintained the same rate in August. New housing starts plunged another 20 percent between June and September, and the decline of home prices accelerated.

Meanwhile, the employment situation was deteriorating, if subtly at first. The summer of 2007 had a couple months of negative job growth, though this was not apparent until later data revisions. Even in the good months, job growth was slower than population growth. By the time the Fed cut rates in September, it was too late to prevent the onset of the Great Recession in December 2007.

Today, many people believe that the severity of the 2007 housing crisis was made inevitable by the size of the housing bubble. But Erdmann and Schubert’s analysis convinced me that that’s not true.

A large portion of the housing boom between 2000 and 2006 was driven by the fundamentals. If there was over-building in 2004 and 2005, it was modest in scale and limited to a handful of metropolitan areas. If homes were overpriced at the end of 2005, it was probably by 10 or 15 percent nationally, not 30 percent.

If the Fed had understood this at the time and acted accordingly, it could have averted a lot of human misery. Home prices would not have fallen so much, and fewer people would have lost their jobs. That, in turn, would have limited the losses of banks that bet on the mortgage market, and might have prevented the 2008 financial crisis.

Other countries had big booms without big busts

You’re probably still skeptical of the claim that the post-bubble crash didn’t need to be so severe. So I’ll close with this chart, which compares US housing prices to some of our peer countries:

But starting in 2006, our housing market diverged from theirs. US housing prices fell by more than 30 percent. Home prices in the UK, France, and Canada zoomed upward for another year or two, then fell by 18, 8, and 7 percent, respectively. Prices then rebounded more quickly in all three countries.

It’s hard to look at this chart and maintain that the depths of the 2008 housing crash was foreordained. Canada, France, and the UK have economies similar to our own, and they all experienced housing booms like ours. But none of them suffered crashes that were anywhere close to ours. Perhaps with better policies, we could have had a smaller, gentler housing bust too.

His bottom line, “…So I think it would go too far to say there was no housing bubble at all during the 2000s. Things clearly got frothy near the end. The high prices of 2005 were probably not sustainable. Some correction was probably inevitable, at least in a few cities. However, people wrongly concluded that the entire national housing boom of the 2000s was a speculative bubble.They came to believe that prices were far above levels that could be justified by fundamentals, and that there were millions of excess homes….”

“…The model also reproduces the empirical pattern shown in Table 4 that higher price dividend growth during the boom forecasts higher future dividend growth and not large relative price declines. For this exercise, we collapse the model time series by year and regress the log changes in dividends and price between model years corresponding to 2006 and 2019 on the log change in the price-dividend ratio during the boom. An additional log point of growth in the price-dividend ratio in the boom predicts additional 0.41 log point of dividend growth in the bust-rebound (compared to 0.20 in column (3) of Table 4) and additional 0.03 log point of price growth (compared to 0.03 in column (6) of Table 4)….echo the result from Table 4 that price rent growth in the boom not associated with long-run fundamentals strongly predicts subsequent price decline. However, they also suggest that the role of investors was mostly or wholly orthogonal to the role of fundamentals and less important to explaining the entirety of the boom or the full 1997-2019 period, which are the focus of our paper…”

New NBER argues that there was no housing bubble, national real house prices rose 80% between 1997 and 2006, lost two-thirds of their gain by 2012, but then rose again. The episode isn't best characterized as a boom-bust, but a boom-bust-rebound, "...We reevaluate the 2000s housing cycle from the perspective of 2020.1National real house prices grew steadily between 2012 and 2019, with the largest price growth in the same areas that had the largest booms between 1997 and 2006 and busts between 2006 and 2012. As a result, the areas that had the largest booms also had higher long-run price growth over the entire 1997-2019 period. With “2020 hindsight,” the 2000s housing cycle is not a boom-bust but rather a boom-bust-rebound...."

"...We argue that this pattern reflects a larger role for fundamentals than previously thought. In a first step, we use a standard spatial equilibrium framework to motivate several determinants of house prices. We find that these explain cross-city variation in long-run house price growth in reduced-form and structural supply regressions as well as the amplitude of the boom-bust-rebound across cities and severity of the foreclosure crisis. In a second step, we introduce a “neo-Kindlbergererian” model of a fundamentally rooted house price cycle in which belief over-reaction amplifies the boom and a foreclosure spiral exacerbates the bust, and discipline the model with our empirical moments. The estimated model generates the boom-bust-rebound from a single fundamental shock and quantitatively matches the cross-city patterns...."

Their housing cycle dynamic, "...House prices clear a Walrasian market, with demand emanating from potential entrants and supply coming from the the construction of new homes and foreclosures. An endogenous boom-bust-rebound cycle occurs in response to a single change in the city’s fundamental, an increase in the growth rate of the income and amenities or “dividend” from living in the city...."

Evidence: "... Figure 1 shows the national Case-Shiller house price index, deflated using the GDP price index. After a period of zero real growth, the series begins to rise in the late 1990s, peaks in 2006Q2, reaches a local trough in 2012Q1, and then grows again through the end of our sample in 2019. In what follows, we measure the boom as house price growth between 1997 and 2006, the bust as price growth between 2006 and 2012, and the rebound as price growth between 2012 and 2019.We start the boom in 1997 because very few cities have booms that start before that year..."

"... Here, we revisit the increase in price-rent ratios with the benefit of 2020 hindsight and show that the component associated with long-run fundamentals predicts subsequent rent growth and not future price decline, consistent with our interpretation of this component as fundamentally-based. We start by characterizing the behavior of rents. Panel (a) of Figure 5 plots the growth rate of real rents in each CBSA between the 2000 Census and the 2018 American Community Survey (ACS) against the long-run fundamental, again measured as the second stage fitted value from column (3) of Table 2. Areas with higher fundamentals experienced larger rent growth over the BBR, with the relationship especially strong for larger CBSAs. Panel (b) shows the timing of rent growth using BLS CPI rent data for the 22 CBSAs with annual data since 1987, grouped into population-weighted quartiles of the long-run fundamental. Rents rise fastest in areas with the highest fundamentals, and this growth appears to represent a break from the pre-boom period. Column (1) of Table 4 shows that CBSA-level price-rent increases in the boom correlate positively with the long-run fundamental. We measure the log growth in the price-rent ratio using 2000 Census and 2006 ACS mean rent and the same house price data as above. The bivariate relationship has an R2 of 0.29. Columns (2) and (5) report the correlation of price-rent growth in the boom with subsequent rent and price growth over the 2006-18 period. This user-cost decomposition (Poterba, 1984) closely resembles the Campbell and Shiller (1988b) exercise of decomposing variation in the price-dividend ratio of a stock into future dividend growth and returns with rents replacing dividends as the cash-flow measure.15 Larger price-rent growth during the boom forecasts both higher subsequent rent growth and future relative price declines. Columns (3) and (6) restrict the variation in the price-rent ratio in the boom to the part associated with the long-run fundamental. Specifically, these columns report regressions of subsequent rent and price growth on the fitted value of the growth of the price-rent ratio from column (1). Strikingly, the rise in price-rent ratios associated with long-run fundamental growth predicts even faster subsequent rent growth than in column (2) and no subsequent price decline, validating our labeling of this component as a fundamental. Columns (4) and (7) show that the part of price-rent growth not explained by long run fundamentals predicts no subsequent rent growth and large subsequent price declines. These columns make clear that our empirical evidence admits the possibility of aspects of the housing boom not associated with long-run fundamentals; in fact, the part of price rent increases not correlated with long-run fundamentals looks very bubble-like ex post…"

Greg Obenshain and Eldar Safarov report on a Andrei Shleifer paper (attached) we previously highlighted that argues financial crisis are relatively predictable, "...The authors define periods when three-year debt growth has been in the top 20% of historical increases and three-year asset price growth has been in the top third of historical increases as the “Red Zone,” shown in the upper right-hand corner of the figure below... The probability of experiencing a financial crisis within one year of entering the Red Zone is over 13%, compared to a 4% probability in the entire dataset. Remarkably, the probability of entering a financial crisis within three years of entering the Business Red Zone is 45% and an astounding 69% in the rare instances when both the household and business Red Zones are breached. The authors note that crises do not immediately follow high debt and asset price growth, suggesting they take time to develop.... And most interestingly, we have just entered a Red Zone as of the end of 2020.This is driven by the rise of debt-to-GDP and the equity market rally. Even if debt-to-GDP is calculated using the pre-COVID GDP level, this result holds. So, if crises are by-products of credit and asset price growth cycles, as the authors’ evidence strongly suggests, then the US may have just entered the danger zone. This does not mean that we need to worry about the sky falling just yet. The authors show clearly that crises take time to develop, and only materialize 36% of the time after a Red Zone has been entered...."

Are Financial Crises Predictable?

The consensus wisdom among academics and practitioners alike is that crises are inherently unpredictable. But Harvard’s famed behavioral economics group has recently produced a new paper suggesting the consensus wisdom on predicting crises might be wrong—at least on a time horizon of three years or longer. Instead, they argue that rapid expansions of credit can drive asset price booms, which are followed by credit busts.

Building off earlier academic research, the authors combine data from 42 countries between 1950 and 2016 to study financial crises in the context of business and household credit growth as well as equity and home price growth. They define crises as declines in bank equities by greater than 30% or banking panics, which capture events such as the Great Recession (2007) and the Savings and Loan Crisis (1990), but not the bust of the dot-com bubble. In line with the theories of Hyman Minsky and Charles Kindleberger, the authors find that financial crises tend to follow periods that feature both elevated debt and asset price growth. The authors define periods when three-year debt growth has been in the top 20% of historical increases and three-year asset price growth has been in the top third of historical increases as the “Red Zone,” shown in the upper right-hand corner of the figure below.

The table below shows the probability of entering a financial crisis once the Red Zone has been breached, broken down by the type of Red Zone event: business, household or both.

As a leading indicator, the Red Zone would appear to be effective. 64% of crises were preceded by either the business or household Red Zone warning within the prior three years. When the authors expand their three-year debt growth to include the top 40% of historical increases and the three-year price growth to be in the top two-thirds of historical increases, the Yellow Zone precedes 82% of crises within the prior three years. But like most leading indicators, their presence does not necessarily result in the event. Red Zones and Yellow Zones were followed by crises only 36% and 20% of the time. To give a sense of what this data looks like, we show the Red and Yellow Zone indicators for the US housing sector along with the real housing price index.

Ideally, we’d show the same chart for US businesses debt and asset price growth. However, there were no business Red Zones in the United States in the authors’ dataset. But while the authors needed to study 42 countries across 60 years so that they could get statistically significant results, we are not similarly constrained. We can use their insights to see what happens if we use US data alone. Specifically, rather than global cut-off values for Red and Yellow Zones, we can use only the US data to define the Yellow and Red Zones. The results are shown below.

Credit and asset booms are what predict busts. In part, the authors argue, because investors tend to extrapolate these booms into the future, taking more risks than they should. Investors feel an inevitable desire to “reach for yield” in low-yielding environments, spurring credit growth that drives asset growth until the inevitable reckoning when some portion of the assets and some portion of the debt go bad.