There is a growing consensus that loose credit and too-clever-by-half financial wizardry sowed the seeds of America’s still-deepening economic malaise. One practice in particular has been singled out for censure — the bundling of loans into assets that could be sold on to investors. The charge is that by breaking the link between those who vet borrowers and those who bear the cost when they default, securitisation led to the lax lending that both fuelled and felled America’s housing market.
Securitisation is not new, and decades of standardisation in areas such as mortgage finance, car loans, and credit cards have removed many of its deficiencies. But new research by Atif Mian and Amir Sufi of the University of Chicago’s business school provides hard evidence that securitisation fostered “moral hazard” amongst mortgage originators, which led them to issue loans to uncreditworthy borrowers.
By examining local variations in housing and mortgage data, the authors were able to identify a causal link running from securitisation to increased credit supply, faster house-price growth and rising default rates. Their findings confirm what many suspected: poor screening by mortgage originators intent on selling on loans was a significant factor in the housing boom and bust.
The researchers’ first task was to establish that the explosion of mortgage growth during the boom was down to easier access to credit. Using detailed data collected by government agencies in 3,000 zip-code districts, they were able to identify areas of pent-up demand by calculating the share of mortgage applications that had been turned down. They found that districts with the highest level of mortgage rejections in 1996, the benchmark year, were those where subsequently there was a disproportionately large increase in the rate of approved mortgages in the boom years between 2001 and 2005.
Could it have been that the economic prospects improved in these neighbourhoods, explaining the increase in the use of credit? That would suggest that more borrowers were better placed to repay a mortgage. No, the Chicago economists contend: they found that the places where it became easier to obtain a mortgage — what they call the “high latent-demand” ZIP codes — also saw declining income and employment growth compared with nearby districts.
Instead, their findings suggested that loans were increasingly made available to willing but marginal borrowers whose poor finances had previously locked them out of mortgage markets. The result is now patently obvious. Lending standards slipped and the borrowers were, on average, far more likely to default on their loans. The biggest increases in debt relative to income, a measure of default risk, were in high latent — demand neighbourhoods, the study says. And the avalanche of mortgage money did not just lower credit quality; it also pushed up the price of houses. Prices rose more quickly in districts where mortgage supply had expanded most.
Behind the mushrooming supply of home loans lay the boom in securitisation. The share of new mortgages sold to investors — excluding loans packaged by government-sponsored mortgage companies, such as Fannie Mae and Freddie Mac — rose from less than a third to more than a half between 2001 and 2005. The increase was particularly prevalent in those districts where initially it had been hardest to obtain a mortgage — the same places where the supply of credit had picked up most and economic prospects had improved least.
The story so far could easily have been a happy one — for a few years, indeed, many Americans savoured the “democratisation” of housing finance that the boom engendered. This was, after all, what financial innovation was supposed to be about& #8212; by spreading risk around the system, it enabled more people who wanted mortgages to get them. Mortgage suppliers were seen as the vital links in a chain that brought lenders and borrowers together.
The weakest link
The ending, however, is more sinister than even many of the sceptics believed. The Chicago economists find strong evidence that the mortgage originators made bad matches between borrowers and lenders. Those areas where they were most willing to increase lending during the boom years have since seen the biggest rise in defaults. Moreover, the increase in default rates was greatest in districts where a larger proportion of mortgage loans were sold on within one year.
Worse, this negative link is most visible where the incentives to check on the creditworthiness of the borrowers were likely to be weakest. Loans securitised by the originators, or sold to non-bank financial firms probably for packaging, were far more likely to end up in default. Intriguingly the study shows that there were fewer defaults than average when loans were sold to an affiliate of the originator. (In sales to banks, the link to the default rate was negligible, though they are believed to have used better screening techniques.) Where the interests of originators and buyers were aligned, subsequent defaults were less likely.
The paper’s authors acknowledge that their analysis might not constitute proof: other unseen factors could have affected the results. But a separate studyÂÂ also shows securitisation has a lot to answer for. It finds that a loan portfolio that fits the criteria for securitisation is far more likely to default than a similar package of loans with a slightly lower credit score that is less likely to be securitised. It is a strong argument for requiring lenders to keep some “skin in the game” — or a share of the securitised bundle of loans on their books.