This is the uncut, extended version of my interview with Atif Mian and Amir Sufi, a shorter version of which appeared on FT.com.
And below is a time guide, followed by some brief thoughts on their excellent new book, “House of Debt”.
Time Guide
[00:00] – [01:30] The systematic patterns showing that localities with high household debt had the biggest subsequent consumption shortfalls
[01:31] – [03:19] Why the credit channel mattered less than people thought.
[03:20] – [04:40] How we know that credit led to home prices rising, not the other way round.
[04:41] – [06:36] Whether their thesis allows for the possibility of complementary narratives.
[06:37] – [08:32] Their preferred policy solution: reducing principal on mortgages, wider access to refinancing
[08:33] – [09:18] On their disagreement with Geithner.
[09:19] – [13:16] On the macro externalities of debt, and sharing risk and losses between debtors vs creditors.
[13:17] – [15:39] Why government subsidies prevent equity-like instruments from becoming more popular.
[15:40] – [17:09] Why the problem of debt is inextricably linked to inequality.
[17:10] – [19:03] We lack a good understanding of why borrowers borrowed so much.
[19:04] – [22:46] On their debate with Brad DeLong and David Beckworth, and whether evidence supports theory.
[22:47] – [23:37] What they’re working on next — inflation expectations, secular stagnation.
House of Debt
The book summarises the work for which the authors are best known, and which is notable as much for its novel methodological approach to the study of debt across local geographies as for its wider conclusions.
Those conclusions are, roughly:
– The above-trend rise in household debt during the years prior to the recession of 2008-2009 included a widespread extension of credit to subprime and low-income borrowers.
– This rise in debt led the home price bubble (not the other way around), though subsequently the price bubble also incentivised borrowers to withdraw funds against their homes. These borrowers — again, many of them in cities with stagnant or even falling nominal incomes — largely used this money for consumption rather than to pay down other debts or to invest.
– The lenders extending this debt had little incentive to worry about underwriting standards, as they expected to package mortgages into MBS and sell them to investors. And because of the global savings glut, there were plenty such investors in search of assets perceived to be safe.
– When mortgage delinquencies started rising and house prices falling, the losses to homeowners’ net worth had two effects: “The first is the concentration of losses on those who have the highest spending sensitivity with respect to housing wealth: debtors. The second is the amplification of the original house-price shock due to foreclosures.”
– These effects produced a demand shock represented by a decline in household consumption, with the other components of growth later following consumption down.
– The responses of monetary policy (limited at the zero lower bound) and fiscal policy (too unfocussed) were helpful, but they were also insufficient and badly targeted, failing to address the root problem of the collapse in households’ net worth — and particularly in those households with the highest marginal propensity to consume.
– Despite some logistical hurdles, a better response would have been to reduce the mortgage principal of these homeowners, and to allow solvent underwater homeowners to refinance at lower fees.
– The credit intermediation channel mattered less than the media and economics profession believed. If anything, policymakers went too far in stabilising the financial system, whose stakeholders should have been punished more harshly (bank shareholders in particular).
– To prevent a similar catastrophe in the future, the financial system should move away from debt contracts that impose all losses on debtors and towards equity-like and hybrid instruments. Future losses must be shared more equitably between creditors and debtors, the expectation of which would convince lenders to act more prudently. A good first step would be for the government to start eliminating subsidies for debt, such as the mortgage interest deduction.
The clever use of micro-level data to discern many of these macro trends makes the authors’ empirical claims — specifically in tracing the causality from increased debt for marginal borrowers to the collapse in homeowner wealth to the sweeping employment effects — highly convincing.
Especially persuasive is the authors’ argument that reducing the mortgage principal for underwater homeowners — through rewriting the bankruptcy code, spending some government money, and taking over for the beleaguered and incompetent servicers — would have been hugely stimulative at relatively low cost to the taxpayer. Their recent response to Tim Geithner’s economic justification for not doing more on housing policy was devastating.
Not all of the work in the book is original. The sections on the savings glut and global pursuit of safe assets, the role of securitisation in encouraging sloppy lending practices, and the fraudulent behaviour of the lenders rely on the studies of others. But they summarise and assimilate the literature neatly into their own narrative.
Although not the immediate focus of the book, inequality is a weighty underlying presence throughout. Given their emphasis on credit-fuelled consumption, their work raises the now-familiar question of whether the only way to keep an economy growing at trend is for savers to funnel money to poorer debtors — regardless of whether the debtors’ income prospects justify it. The authors note at the end of the interview that they want to further study the secular stagnation hypothesis advanced by Larry Summers.
My main quibble with the book is not that its conclusions are wrong but that they are too exclusive.
For instance I suspect that Mian and Sufi go too far in dismissing the importance of the credit intermediation channel. (We discussed this in the interview.) They certainly have a point when they write that stakeholders in the banks deserved more severe punishment than they got. But the fact remains that we never discovered how much worse the recession and subsequent recovery would have been if policymakers hadn’t provided a bulwark to the financial system, and had instead allowed either further bank failures or counterparty defaults.
In addition, Mian and Sufi correctly write that a consumption shortfall led the recession. But the acceleration in the pace of layoffs and the dramatic growth contraction in the fourth quarter of 2008 (an annualised fall in GDP of more than eight per cent) that followed Lehman’s failure were unlikely to have been entirely coincidences. At the very least the possibility that the recession worsened because a financial calamity threatened can’t be ruled out — and therefore we also shouldn’t be too quick to dismiss the credit channel, even if only for its confidence effects on the rest of the economy.
Similarly, Mian and Sufi wave away the argument that monetary policy could have done more if only Ben Bernanke had been more willing. This is difficult to test, especially as the recent history of monetary policy has been a frustrating one of successive starts and premature stops. Yet I still believe that such a protracted recession in concert with a financial crisis generates a widening gap between the demand for safe money-like products and their supply, well beyond the imbalance that already existed before the recession. Overwhelming monetary and fiscal easing, along with a financial intermediation system relatively free of obstruction, is needed to aggressively narrow the gap — an application of Bagehot for the modern era, updated to include the troublesome shadow banking system.
The authors do address this as well in the interview, mainly citing the lack of empirical work needed to substantiate the theory that bolstering inflation and inflation expectations leads households to spend and businesses to invest. They plan to continue studying the effects. As for the current recovery, here too the trouble is with having to speculate about counterfactuals.
Still, this would be the wrong place to conclude. Mian and Sufi have done genuinely original and insightful work in recent years, with an approach to data that will hopefully be tried by others in their profession. This is an important book. I recommend it.
Copyright The Financial Times Limited . All rights reserved. Please don't copy articles from FT.com and redistribute by email or post to the web.