www.istockphoto.com

When markets crashed in 2008 and 2009, a number of competing hypotheses cropped up as to how it happened and what the major contributors were. 

Some economists concluded that financial leverage in housing markets triggered the downfall, while others just as firmly ruled it out, citing interest rates as the main reason.

The assessments differed so widely, in part, because the market models most economists rely on today have shortcomings: Many use just a few variables that don’t adequately represent the complexity of housing markets, and they typically run on averages and aggregate numbers to find coarse patterns rather than apply the full range of possibilities.

“There are millions of homeowners with widely different tastes and wealth,” says SFI External Professor John Geanakoplos, an economist at Yale. “Aggregating them into a simple average misses the behavior in the tails that causes the big swings up and down like we saw from 1997 to 2009.”

In a recent paper published in American Economic Review, Geanakoplos, SFI External Professors Rob Axtell and Doyne Farmer, and collaborators questioned weaknesses in the current models and asked whether forecasters were using the most telling variables. Their own agent-based model of the U.S. housing market, they wrote, better captures the heterogeneity of agent behaviors.

Their model adds new loan features and incorporates data on household demographics, economic conditions, and housing market behaviors. Starting at the level of the individual homeowner, it follows each through the various costs, likelihood of prepayments and turnovers, and other factors.

Its simulations neatly fit housing prices and other housing market indices from 1997 to 2010. Consequently, they say, manipulating the model (by holding certain variables constant) offers a means to pinpoint the major causes of the 2008-2009 crash. Freezing interest rates, for example, simply reduced the intensity of the boom and bust, but holding leverage (down-payment percentages) constant greatly shrank the boom and wiped out the bust.

“Knowing that mortgage lending with very small down-payments was the primary cause of the 1997-2006 housing boom, and that the sudden spike in required down-payments in 2007 and onward was a primary cause of the housing crash, suggests that the federal reserve ought to rethink its obsession with interest rates as the unique tool for maintaining financial stability,” says Geanakoplos.

Read the abstract from the American Economic Review (May 2012)

Support the Santa Fe Institute here