Lawrence Summers on ‘House of Debt’
House of Debt: How They (and You) Caused the Great Recession, and How We Can Prevent It from Happening Again, by Atif Mian and Amir Sufi, University of Chicago Press, RRP£18/$26, 192 pages
Atif Mian and Amir Sufi’s House of Debt, despite some tough competition, looks likely to be the most important economics book of 2014; it could be the most important book to come out of the 2008 financial crisis and subsequent Great Recession. Its arguments deserve careful attention, and its publication provides an opportunity to reconsider policy choices made in 2009 and 2010 regarding mortgage debt.
House of Debt is important because it persuasively demonstrates that the conventional meta-narrative of the crisis and its aftermath, which emphasises the breakdown of financial intermediation, is inadequate. It then goes on to provide a supplementary and in some ways alternative explanation focusing on the deterioration of household balance sheets, an analysis that has profound implications for policy directed both at preventing crises and responding to them when prevention fails.
The book is largely free of equations, jargon, econometrics or data tables. However, no reader should be deceived. It is a summary of a highly serious programme of economic research – one that is in many ways a model for what economists should do. Mian and Sufi, professors at Princeton and the University of Chicago respectively, have examined a profoundly important question: what causes protracted downturns in economic activity? They have marshalled new data – for example, on spending by zip code – to test their hypotheses, assembling such a range of evidence from so many different sources that their conclusions are not susceptible to challenge by those looking to point out statistical errors.
Most in the financial community, the policy community and the commentariat see a breakdown in financial intermediation as the root cause of the financial crisis and Great Recession. The failure to bail out Lehman Brothers is usually viewed as the prelude to crisis intensification. Published accounts by leading actors such as Henry Paulson, Sheila Bair and Timothy Geithner provide a narrative of bailout decisions made with respect to Bear Stearns, Lehman, Fannie Mae and Freddie Mac, Wachovia, Washington Mutual, Bank of America and Citi that brings to mind generals’ war memoirs as they describe the various battles they fought to keep financial institutions, or at least their lending, alive. And the debates about crisis prevention, as illustrated most vividly in the context of the Dodd-Frank regulatory legislation, centre around policies towards banks and the shadow banking system.
Mian and Sufi point out a variety of problems with this approach. First, they note that data on credit spreads suggest that the financial system was fully repaired by late 2009, and that even though the economy at that point was very depressed, growth has been anaemic since. Second, they observe that spending on housing and durable goods such as furniture and cars decreased sharply in 2006 and 2007, well before any financial institution became vulnerable. Likewise, they note that the initial impetus behind recession in the US appears to have been a decline in consumer spending. Additionally, the authors observe that when asked why they were not borrowing more, even small businesses, the sector most dependent on banks, more often than not blamed a lack of customers rather than banks’ unwillingness to lend.
None of this sits easily with what Mian and Sufi call the banking view of the Great Recession. They argue that, rather than failing banks, the key culprits in the financial crisis were overly indebted households. Resurrecting arguments that go back at least to Irving Fisher and that were emphasised by Richard Koo in considering Japan’s stagnation, Mian and Sufi highlight how harsh leverage and debt can be – for example, when the price of a house purchased with a 10 per cent downpayment goes down by 10 per cent, all of the owner’s equity is lost. They demonstrate powerfully that spending fell much more in parts of the country where house prices fell fastest and where the most mortgage debt was attached to homes. So their story of the crisis blames excessive mortgage lending, which first inflated bubbles in the housing market and then left households with unmanageable debt burdens. These burdens in turn led to spending reductions and created an adverse economic and financial spiral that ultimately led financial institutions to the brink.
This interpretation resolves the anomalies that Mian and Sufi highlight. Households do not spend while they are still overly indebted, which precipitates slow growth even after banking is restored to health. Spending slowdowns are caused by household over-indebtedness, so of course they precede problems in the banking system. And, when consumers do not spend, businesses have less need to borrow to finance investment, inventories or receivables.
Their analysis, presented with far more depth and subtlety than I have been able to reflect here, is a major contribution that furthers our understanding of the crisis. It certainly affects what I will examine in trying to predict and forestall future crises. And it should influence policies aimed at crisis prevention by demonstrating the insufficiency of keeping financial institutions healthy and by making a case for macroprudential measures directed at preventing runaway growth in household debt.
When one has a persuasive and novel idea, there is an inevitable temptation to push it a bit too far and to weight it excessively relative to less novel truths. Mian and Sufi succumb to this temptation in the last third of their book, where they discuss the policy responses to the crisis. Like the 2005 homeowners used to capital gains, they are unable to resist doubling down by levering up.
They are sharply critical of what they see as the dominance of the banking view over their preferred levered losses view. At points they are not completely clear but Mian and Sufi mostly, if reluctantly, accept the necessity of action to support the financial system once the fires were burning. It is difficult to see how any prudent policy maker, after witnessing the Armageddon that followed the collapse of Lehman, could have failed to take steps to prevent further collapses. Certainly, plenty of tactical questions can be raised about the steps that were taken. With hindsight some argue that more pain should have been imposed on the financial sector. In the moment, though, the overwhelming imperative was restoring confidence at a time when complete breakdown looked like a real possibility. The government got back substantially more money than it invested. All of the senior executives who created these big messes were out of their jobs within a year. And stockholders lost 90 per cent or more of their investments in all the institutions that required special treatment by the government.
Aside from a few undeveloped questions about whether less could have been done to bail out bondholders while saving the financial system, Mian and Sufi do not really challenge any of this. Instead, they train their fire on policy makers’ failure to take proper account of the levered losses view by providing far more sweeping mortgage relief. They argue that if policy makers had better appreciated their arguments about household balance sheets and been less cognitively captured by banking system concerns, the outcome could have been much better.
Obviously, as the director of President Barack Obama’s National Economic Council in 2009 and 2010, I am an interested party here. It seems to me that Mian and Sufi are naive on policy and contribute much less to the debate on this than they suppose. Their indignation seems to get the better of them as they move from citing the Quarterly Journal of Economics to the National Journal, and from Irving Fisher to Bloomberg News; and, for all their commitment to careful scholarship, they seem not to have spoken to those involved in making policy towards housing during the crisis.
We all believed in 2009 what Mian and Sufi have now conclusively demonstrated – that reducing mortgage debt would spur consumer spending. And there was intense frustration with how few homeowners our programmes were reaching, to the point where I convened all the relevant officials from the Treasury, the Housing and Urban Development department and other agencies every month for two years to challenge them to find ways to accelerate the process and to make sure that they were considering all the various schemes academics and others were suggesting. So Mian and Sufi are not wrong in their dissatisfaction.
Where they do, in my view, go badly wrong is in suggesting that the failure to do more for underwater homeowners reflected a blinkered attachment to an at best incomplete, and at worst harmful, banking-based view of what was happening. Mian and Sufi’s error is a common one among academic economists, many of whom are unwilling to try to understand policy choices that arise from considerations outside simple models.
Start where Mian and Sufi start, with the idea of “cram-down”: the notion that bankruptcy judges should be allowed to write down mortgage debt, and that permitting them to do so would increase the bargaining power of homeowners seeking relief. Although it is more complex than Mian and Sufi suggest, I believe on balance this would have been better public policy, and so argued vigorously in these pages in February 2008.
Why didn’t the administration push for such legislation? Simple. The judgment of the president and his advisers was that there was essentially no chance of it getting the requisite 60 votes in the Senate, where we were not even able to muster a simple majority. Should a more Herculean effort have been made? Perhaps. But the president and his team felt that in a world where many legislative battles lay ahead, a failure on cram-down would be costly in time and political capital. Critics who disagree at this late date are obliged to provide an alternative analysis of the political calculus, not a mere recitation of the arguments for cram-down.
What about the various ideas put forward for bringing about mortgage relief through either coercion or financial inducement? Mian and Sufi’s impressive empirical work demonstrates that households have a propensity to spend perhaps 15 per cent of any debt reduction they receive. So if on very generous assumptions $3tn of mortgage debt was written down by a third, thus reducing $1tn in mortgage debt, this would have added $150bn or 1 per cent of gross domestic product to spending at time when the output gap was close to 10 per cent of GDP. In assessing this benefit, we were very mindful of the Hippocratic Oath’s admonition – “first, do no harm” – and took into account a variety of considerations that Mian and Sufi do not seriously discuss.
First, there was the risk of bringing down the system in an effort to save it. Banks had substantial mortgage holdings and especially large quantities of subordinated second mortgages and home equity lines of credit, which would have been wiped out if mortgage principal had been reduced in a way that respected the seniority of first mortgages. We recognised that large-scale principal reduction would draw in a large number of mortgages that were not delinquent and would otherwise be paid in full. As a consequence, there was the risk of sucking hundreds of billions of dollars out of the banking system. Given that government funds for capital infusions were scarce and that each dollar of bank capital supports $12 of lending, we worried that the spending gains from reducing mortgage debt might well be exceeded by the spending losses from reducing the flow of capital. This fear may have been exaggerated. If they think so, Mian and Sufi owe an explanation as to why.
Second, there was the issue of chilling future lending. Like the Brady plan for Latin America, most debt reduction schemes are voluntary, at least on their face. This is because of a fear that if the government simply starts requiring creditors to write down debts, especially debts that are being serviced, or rearranging creditor priority, the flow of future credit will be inhibited. It did not seem unreasonable to worry at the time that if government forced the write-off of a trillion dollars of mortgage debt, flows of not only mortgage debt but also car loans and credit card debt to consumers would be inhibited as well. This was not a small concern, as the automobile industry was in freefall and consumer confidence was deteriorating very rapidly.
Third, there was the danger of prolonging the housing market’s problems. Even the relatively limited programmes in place have spent as much as a third of their money delaying, rather than avoiding, foreclosures. All that we heard at the time suggested that a significant part of the reason why the housing market was dead was that no one wanted to buy because of a fear that it had further to fall. Delaying inevitable foreclosures with relief risked exacerbating this problem and risked larger foreclosure discounts when houses were ultimately sold.
Fourth, there were regulatory issues. For example, we spent a great deal of time looking at ideas that involved the government buying underwater mortgages from banks on the general model of the Home Owners’ Loan Corporation scheme pursued during the Great Depression. The problem was that in many cases mortgage assets were carried on banks’ books at valuations far above what appeared to be current market value. Buying them at such valuations would have been a massive backdoor subsidy to banks of the kind we would not accept. Forcing writedowns was precluded for regulators who feared what it would do to banks’ capital positions.
Fifth, anyone who has worked in government knows that it is much easier to design policy than to implement it. We were, for example, very aware of the traditional idea that at moments of financial distress it is necessary to convert fixed debt claims into equity and explored all sorts of ideas for writing down mortgages and giving creditors claims on subsequent home price appreciation that would liquefy when the home was sold. Our conclusion was that the administrative capacity to negotiate such arrangements and then to track them through time and collect when homes were sold could not be brought into existence in a reasonable time.
Looking back, there are things I wish had been done differently, such as putting in place more satisfactory leadership at the Federal Housing Finance Agency and more effectively mobilising Fannie Mae and Freddie Mac to help bolster the housing market. The various liberalisations of the Home Affordable Modification Program and the Home Affordable Refinance Program that were put into place over time should, with the benefit of hindsight, have been introduced at the beginning. We may well have misjudged some risks or missed important opportunities to carry out effective policy. Certainly, I have stayed up at night while in government and afterwards worrying over these possibilities. But seriousness requires grappling with the challenges that policy makers perceived, not merely pointing out the importance of debt and citing various plans that were put forward in op-eds. Perhaps this can be the authors’ next project.
The reality that the post-crisis policy challenges were more complex than they recognise detracts only slightly from Mian and Sufi’s accomplishment in House of Debt. All future work on financial crises will have to reckon with the household balance sheet effects they stress. After their work, we can still believe in the necessity of financial rescues; however, we can no longer believe in their sufficiency. And after their work, we have an important new agenda of reforms to consider if future crises are to be prevented.
Lawrence Summers is Charles W Eliot University Professor at Harvard and a former US Treasury secretary
Letters in response to this article: