It Wasn't Household Debt That Caused the Great Recession

It was how that debt was disproportionately distributed to America’s most economically fragile communities.


Why are nearly 10 million people still out of work today? Was it because in September 2008, the U.S. government failed to bail out the insolvent investment bank Lehmann Brothers? Was it because the two U.S. housing finance giants Fannie Mae and Freddie Mac guaranteed too many mortgages securitized by Lehman and other Wall Street firms to low-income borrowers in the run up to the housing and financial crises? Or does blame rest with the Federal Reserve’s too-easy-money policies in the wake of the brief dotcom recession in the early 2000s?

Princeton University professor Atif Mian and University of Chicago Booth School of Business professor Amir Sufi pin the blame squarely on policymakers, but not for any of these three reasons, all of which are variously popular with policymakers on different sides of the political divide in Washington. Instead, in their just-released book, House of Debt, they argue that the Great Recession was the result of a sharp fall-off in consumption due to the unevenly accumulated household debt in the first six years of the 21st century. In that period, mortgage-credit grew more than twice as fast in neighborhoods with low credit scores than in neighborhoods with high credit scores, a marked departure from the experience of previous decades. When the housing bubble popped, the economic consequences were sharply magnified by the way debt was distributed across households and communities.

How did this happen? Why did lenders suddenly shower less-creditworthy borrowers with trillions of dollars of credit? Mian and Sufi demonstrate this was enabled by the securitization of home mortgages by investment banks that did not seek federal guarantees from Fannie and Freddie—so called private-label securities, made possible by financial deregulation and the glut of cash in world markets in the wake of the Asian financial crisis of the late 1990s. That private-label mortgage-backed securities were at the core of the housing meltdown is no longer in doubt, but what Mian and Sufi bring to the debate is how an unequal distribution of debt magnified the economic risks—based on their path-breaking microeconomic research—and a new framework for considering who is to blame among policymakers for the still reverberating debacle.

Unfortunately, the two authors don’t provide answers for why so many households took on so much debt, but they do paint a cautionary tale. This is a critically important contribution to the policy debate now raging over what Congress and the Obama administration should do in the way of reforms to the housing-finance industry. And, it’s important to our understanding of whether and how inequality affects economic growth and stability. What they demonstrate is that as the U.S. housing bubble burst and home prices began to fall in late 2006, the unequal distribution of debt amplified the decline in consumer spending and the consequence was an economic disaster. Mian and Sufi’s research leads them to conclude that the crisis was avoidable if only economists had used the right framework to see what was happening around them at the time.

“Economic disasters are man-made,” they write in the opening pages, “and the right framework can help us understand how to prevent them.” By the end of the book, the reader cannot but be left appalled at the sheer enormity of the policy failures. It’s not just that 7.4 million workers lost their job during the years of the Great Recession of 2007-2009 but also that the employment crisis continues to this day. While jobs are no longer being shed at the rate of 20,000 a day, the share of the U.S. population with a job fell to a low of 58.2 percent in November 2010 from a high of 63.4 percent in December 2006, but has only increased by a fraction of a percentage since then, hitting just 58.9 percent in April 2014.

Missing the housing bubble was a massive failure on the part of policymakers. As a result, our new normal is one where there are nearly 10 million fewer people at work. This book's contribution helps us understand the important mechanisms through which this occurred.

* * *

I watched the housing and financial crises unfold from my perch as staff for the U.S. Congressional Joint Economic Committee. By the time Lehman Brothers failed, the mantra on Capitol Hill had been articulated by former Treasury Secretary Lawrence Summers, who said that any recovery package had to be “timely, targeted, and temporary.” But the stimulus that emerged was not specifically targeted at homeowners in foreclosure. If Mian and Sufi are correct, the biggest failure was—and continues to be—leaving families struggling with mortgages they cannot afford because of the fall in home prices.

The federal government has provided assistance to a paltry 940,000 struggling homeowners through the Homeowners Assistance Mortgage Program, in a nation where 5 million homes have been foreclosed on. This lack of help hasn’t just hurt those homeowners. Also caught in the downdraft are now destroyed neighborhoods, ruined communities, and thwarted lives of far too many. Protecting banks does not necessarily make the economy strong.

So, how did we get here? That’s the focus of House of Debt. Mian and Sufi spent the past decade compiling and analyzing microeconomic data to test theories about how the macroeconomy works. They conclude that inequality in wealth and debt combined with greater availability of credit to marginal borrowers are a toxic macro-economic combination. They call this the “levered losses” view, arguing that severe recessions occur when “asset prices collapse and households sharply pull back on spending,” even with “no obvious destruction of productive capacity occurs."

Their story starts with an accumulation of debt—lots of it. After the Asian financial crisis in 1997, investors were looking for safe havens to park their money. What they wanted were AAA-rated bonds. What they got were mortgage-backed securities that were rated AAA but turned out to be junk. As we all now know—but most of us didn’t know at the time—Wall Street firms in the early 2000s began slicing and dicing and then reassembling mortgage debt into more and more exotic and risky mortgage-backed securities in ways that made them look risk-free.

But, it wasn’t just that there was more securitization. It was that loans made to riskier borrowers were more likely to be securitized. This both drove the housing bubble and made the consequences of it popping all the worse. Mian and Sufi point out that between 2002 and 2005, the growth in mortgage credit and household incomes became negatively correlated, that is, credit expanded in areas where incomes were declining. This makes no sense: How can you pay back a loan if your income is falling? They point to academic research by Yuliya Demyanyk and Otto Van Hemert showing the profound consequences: By 2006, loans had become so disconnected from prudent business practices that “an unusually large fraction of subprime mortgages originated in 2006 and 2007 [became] delinquent or in foreclosure only months later.

As these foreclosures began to pile up, affected households cut back sharply on spending. Thus, the catalyst for Great Recession had begun two years before the dramatic demise of Lehman Brothers. In the second quarter of 2006, the collapse in consumption started with residential investment, which fell by a 17 percent annual rate. Non-residential investment didn’t begin to fall until late in 2008, but by then households had already pared back spending sharply.

This fallout from the collapse of the housing bubble was amplified by the unequal distribution of net wealth. What Mian and Sufi find is that counties with the largest decline in total net worth—were the ones that cut back most on spending when house prices declined. As housing prices began falling in 2006, in counties where net worth had declined most, consumption fell by almost 20 percent, compared to only five percent for the entire U.S. economy. In contrast, even through 2008, counties that avoided the collapse in net worth saw almost no decline in spending. If debt had been more equally distributed then the decline in consumption would have been less dramatic and the recession would have been less devastating.

Further, they point out that you cannot have a foreclosure crisis—or its associated sharp fall-off in demand—without debt and the way that debt grew during the early 2000s exacerbated the potential for a foreclosure crisis. Mian and Sufi find that about half of the rise in mortgage debt was among people who lived in their homes, not new purchasers. People took out home equity lines of credit and used the cash for home improvements, funds for their kids' college tuition, or other types of consumption. Once the crisis was in motion, about four-in-10 mortgage defaults were among home-equity borrowers. Thus, the foreclosure crisis was not due to people reaching to buy homes, but to borrowing against their primary asset. Had they not ramped up borrowing, falling home prices would not have affected consumption or led to record-high foreclosures.

Finally, all this subprime mortgage debt that had been structured into AAA-rated mortgage-backed securities created financial instruments in which no single investor has the incentive or legal right to restructure the loan, especially for loans to low-net-worth borrowers. This led to a situation that dramatically reduced the capacity of homeowners to get relief in form or informal backruptcy and increased foreclosures. Foreclosures reduce prices more so than principal reductions and thus amplified the decline in home prices and the loss in wealth.

* * *

Given the troubling rise in economic inequality over the past four decades, this research could not be more timely. It’s not just the questions they are asking and the results they are finding, but also the methods they are using. Mian and Sufi are part of a new generation of economists who examine detailed microeconomic data and analysis to understand the macroeconomy, giving us a deeper understanding of how inequality affects economic growth and stability. They have done this by using detailed, microeconomic data at the county and zip-code level to examine debt and consumption patterns.

Mian and Sufi’s research shows that the marginal propensity to consume—an economics term that describes the amount of spending done after receiving an additional dollar—out of housing wealth depends not just on the value of the asset but also the debt burden, settling a near-century-old economic debate between two of the most prominent economists of the 20th century.

In The General Theory of Employment, Interests, and Money, University of Cambridge economist John Maynard Keynes argued in 1936 that the distribution of income mattered for the stability of the macroeconomy. Increased spending, be it from consumers, government, greater exports, or investment, will multiply as it works its way through the economy. If additional income goes into the hands of those with a high marginal propensity to consume then the multiplier for consumption demand will be relatively larger. But if additional income goes into the hands of those with a lower marginal propensity to consume then the multiplier on consumption demand will be relatively weaker.

Two decades later, University of Chicago economist Milton Friedman hypothesized that although rich households appear to consume less, they have a pretty clear sense of what their standard of living will be on average year after year and they adjust their savings to keep themselves at that level. In good years, when they get an income bonus, they will save a more while in bad years, they won’t save as much—or will borrow—to maintain that average standard of living.

Yet neither Keynes nor Friedman had access to the kinds of data now at the fingertips of Mian and Sufi. Thus the Keynes-Friedman debate was theoretical, not grounded in empirical reality. Now, Mian and Sufi provide a definite “yes” to the question of whether we could have prevented the Great Recession—and the conclusion isn’t pretty. They argue that policymakers could have seriously mitigated the damage, pointing out that debt forgiveness would have been much more effective that the policies implemented because it would have targeted households with the largest marginal propensity to consume. This is a failure on a massive scale and more economists need to follow the lead of Mian and Sufi and look deep into the data to understand what we got wrong.

Mian and Sufi’s argument hinges on the conclusion that it was the supply of credit that drove the bubble and the heightened debt burdens, rather than increased demand from consumers. They discuss some reasons why people may have wanted to borrow more, such as the idea that people who expected higher incomes were borrowing constrained, but come down on the side that people were just acting irrationally—given that the massive increase in borrows during the credit boom was among borrowers with declining incomes, not those with rising incomes. From this, they conclude that “whatever the reason, however, consumers who were offered more money by lenders took it.”

Were people behaving irrationally? And, what (really) does that mean? The late 1990s saw the strongest labor market in decades. The typical male earner saw his annual earnings finally grow, after over a decade-and-a-half of inflation-adjusted declines; women’s employment rates hit an all-time high of 58 percent; and the typical family income grew by an average annual rate of just under 2 percent. The middle was (finally) back, so it may have been the case that people were optimistic that the recession of 2001 would not just be short and shallow, but that the recovery would look like the late 1990s.

But looking closely at the data reveals another pattern. One thing that did not happen during the recession of the early 2000s was a rise in government borrowing. The cash seeking a safe haven from the Asian financial crisis had to go somewhere, but the federal government wasn’t in the mood to borrow. So those dollars flowed willingly into the mortgage-backed securities being peddled as AAA-rated bonds. And the greater the demand, the more Wall Street packaged up their dodgy securities containing more and more subprime loans extended to those least able to afford credit.

The last few decades of the 2oth century also saw a number of marked changes for families. Women increased their labor supply steadily from the 1960s through the high employment years of the early 1990s. By the late 1990s, however, that long-term rise in employment rates had stalled. The United States went from being an economy that had one of the largest shares of women in the labor force to number 18 among 35 developed-economy member nations of the Organisation for Economic Co-operation and Development.

A variety of reasons have been presented for the sudden end in the growth of women’s employment beginning in the first decade of the 21st century and continuing today. The mainstream media play up the idea that women are “opting out” of careers in favor of motherhood, a story line developed in large part by journalists who either live in more wealthy neighborhoods and thus see this happening or write for publications that cater mostly to these wealthy neighborhoods. Yet empirical research finds that women, like men, found it harder to find and keep jobs due to the lackluster economic recovery after the recession of 2001. As families sought to cope with the slow-job growth economy in the 2000s and a labor market that still does not provide the kinds of supports and protections working parents need, many turned to increasingly-readily-available credit as a way to cope.

Now, of course, such easy credit is no longer available. Neither is a robust jobs market. It may be true that it doesn’t matter why people took on more debt prior to the Great Recession, as Mian and Sufi contend, but today the lessons learned then are critically important. The story that emerged in the early days of the Great Recession was that too many people borrowed too much to afford fancy houses. That’s not what Mian and Sufi’s data show. They show that the boom in debt occurred among borrowers that couldn’t qualify for a government-backed mortgage. That the private sector sought them out in the tens of millions to offer loans they were demonstrably unable to repay—without worry because these lenders very quickly diced up those loans and sold them to supposedly savvy institutional investors—created the housing bubble that exploded into the twin crises that led to the Great Recession.

This activity produced a bubble—one that anyone could see and one that policymakers blithely passed off as either non-existent or unimportant—to the detriment of our entire economy. Subsequent reforms to our financial system give policymakers more tools to police housing finance, yet the continuing over-reliance on debt and a lack of good jobs leaves families at risk and exposes our economy to the whipsaw of another debt-fueled credit bubble. Mian and Sufi deserve credit of another kind for detailing how ensnared the American Dream is in this tangled web of debt finance—and how exposed the vast majority of us are to the broader economic consequences.