There’s been a lot of interesting commentary on a recent debate surrounding Raghuram Rajan arguments on inequality and the financial crisis at the American Financial Association – see this Economist article for the best writeup. MIT economist Daron Acemoglu’s AFA slides are on his MIT website. Here’s Simon Johnson with commentary. From the Economist:
Governments, he argues, could not simply stand by as the poor and unskilled fell farther behind. Ideally, more should have been spent on education and training. But in the short run, credit was an easy way to prop up the living standards of those at the bottom of the economic pile. This was especially true in America, with its relatively puny welfare state.
Mr Rajan thinks, therefore, that it is no coincidence that America in the early 2000s saw a boom in lending to the poor, including those folks that banks used to sniff at. He points to the pressure the government put on the two state-backed housing giants, Fannie Mae and Freddie Mac, to lend more to poorer people. Affordable-housing targets, slacker underwriting guidelines and the creation of new “low down-payment” mortgages were all used as instruments of public policy.
I think there are two problems with Rajan’s argument.
1. If you look at the history of financial deregulation it’s not clear that it was done first and foremost to help increase access of credit to the poor and middle-class. This could really benefit from some examples. Increased access was often a post-hoc justification, indeed the justification used in the public sphere, but much of what we consider deregulation didn’t impact consumers. Much of the real change in consumer credit law came out of a 1978 Supreme Court ruling, Marquette Nat. Bank of Minneapolis v. First of Omaha Service Corp. Indeed if you look at the bankruptcy reform of 2005, it’s not clear that this was cover to keep a struggling middle-class alive at the expense of capital investors – one could read it as the opposite.
2. The bigger problem is that, in the Economics 101 language, Rajan says that we’ve had a demand-side problem. Why was there so much credit floating? Because people wanted it, their demand increased, to compensate themselves from income stagnation. But for those in the field, MBS looked like a supply-of-credit issue – the orders came in to fill so much paper, not that people were knocking down their doors demanding loans.
How does that fit the data? As always, we can look at loan volume, and prices. I don’t think this excellent paper, Explaining the Housing Bubble (Adam Levitin, Susan Wachter) gets enough attention. (James Kwak covers it here.)
Looking at securitization data, Levitin and Wachter find that the volume of debt increased while the price decreased. Economics 101 quiz: What does that mean? That means that the supply-curve has shifted out. From the paper:
We believe that the cause of the bubble is to be found in the changes in the structure of the housing finance market in 2003-2004, as the market moved from Agency securitization of traditional FRMs to private-label securitization of nontraditional ARMS. It is unquestioned that securitization was the funding mechanism for the housing bubble, but no previous work has examined its pricing in relation to the bubble. We examined the pricing of PLS deals from 2003-2007. Our examination reveals a remarkable trend: even as mortgage risk and PLS issuance volume increased, the spread on PLS over Treasuries that represents their additional risk premium decreased. (See Figures 12 and 13.)
What’s more, spreads on AAA-rated PLS fell during 2004-2007, even as yield spreads on AAA-rated corporate bonds held steady. (See Figure 14.) In other words, the change in spreads was specific to PLS, and did not reflect a general movement in the AAA-rated bond market.
Declining PLS spreads meant that investors were willing to accept more risk for lower returns. In other words, housing finance was becoming relatively cheaper, even as it became riskier. The risk-adjusted price was dropping and quantity was increasing during 2004-2007!…
The movement in PLS spreads and volume—that spreads fell and volume increased even as risk increased, that the spreads fell below corporate bond spreads, and that PLS spread fell while corporate bonds spreads remained static—points to a supply-side explanation of the housing bubble, rather than a demand-side explanation. Simultaneously falling price (spreads) and increasing quantity (volume) means that there had to be an outward (rightward) shift in the housing financing supply curve (from S1 to S2, in Figure 15).
There may also have been an outward (rightward) shift in the housing finance demand curve (from D1 to D2, in Figure 15), as irrationally exuberant consumers sought ever more financing to cope with escalating prices. Such a shift would have resulted in both greater supply (Q2a) and higher prices (P2a), and thus larger PLS spreads. But PLS spreads decreased, even as supply increased. This means that the housing finance supply curve must have shifted outwards (from S1 to S2) enough to offset any outward shift of the demand curve in terms of an effect on price (P2b<P2a). Put differently, even if there was an increase in housing finance demand, there was a greater increase in housing finance supply. Investors’ demand for PLS was outstripping the supply of mortgages.90
I take it for granted that prices are mechanisms for making sure supply finds demand. There’s a lot of morality talk about how American consumers overbinged on credit; but that makes no sense – they were rationally reacting to price mechanisms and increases in the supply of credit. The question is why did so much credit get lent out?
One way to link that back to inequality is through Michael Kumhof and Romain Rancière model in Inequality, Leverage and Crises, where huge gains at the top are lent out to people in the middle of the distribution, increasing their leverage to financial crisis level of instability. We’ll discuss this model and the Levitin paper more in the future.