Beyond leveraged losses: the balance sheet effects of the home price downturn.

ABSTRACT This paper quantifies the impact of declining home prices, increasing mortgage credit losses, and the associated reduction in credit supply on real GDP growth. Using a state-level panel analysis, I first estimate the link between home prices and foreclosures. I estimate that an additional 15 percent home price decline from mid-2008 levels would be consistent with total residential mortgage credit losses over 2007-12 of $750 billion, although the uncertainty is high. I then gauge the impact of such losses on the supply of credit from banks, asset-backed security markets, and the government-sponsored enterprises, and in turn on real GDP growth. In the central scenario, the crisis could lower real GDP growth in 2008 and 2009 by an average of 2.6 percentage points per year. This estimate excludes both adverse multiplier effects (labor market deterioration, global trade repercussions, and credit quality feedback) and policy offsets.

**********

The current housing market downturn weighs on the economy in four main ways. First, the sharp decline in residential construction activity reduces aggregate output directly. From the fourth quarter of 2005 to the third quarter of 2008, declining real residential investment subtracted a cumulative total of 2.5 percentage points from real GDP growth.

Second, declining income in the housing sector in turn has effects on other parts of the economy. Laid-off construction workers and real estate agents cut back on consumer spending, homebuilders (and their subcontractors) invest less in construction equipment, and nonresidential construction firms see less demand for new commercial development. These second-round effects are harder to quantify because they are so spread out through the economy, but they are likely to be significant as well.

Third, declining home prices weigh on aggregate personal consumption through either a negative wealth effect or a mortgage liquidity effect, or both. Households who spent more than they earned during the boom by borrowing against the rising value of their home may be forced to cut back. Even households who did not outspend their income might reduce their consumption in response to a decline in their wealth or their permanent income, or both. Most studies analyzing this issue find evidence for a housing wealth effect, but its size varies widely depending on the time period and the empirical design. (1)

Fourth, losses on mortgage credit deplete the equity capital of leveraged financial institutions and persuade them to reduce their financial leverage. This reduces the supply of credit to households and nonfinancial businesses. David Greenlaw, Hatzius, Anil Kashyap, and Hyun Song Shin (henceforth GHKS) find that an assumed $500 billion in aggregate mortgage credit losses could cut real GDP growth by 1.5 percentage points over a year's time. (2)

This paper focuses on the fourth channel. Building on the study by GHKS, its main contributions are a more detailed empirical analysis of the link between home price declines and mortgage credit losses and a more systematic look at the role of the asset-backed security (ABS) markets. The first section analyzes the links between home prices and foreclosures, and ultimately between home prices and mortgage credit losses, using a state-level panel dataset for the period 1998-2008 to predict foreclosures, The second section discusses the impact of mortgage credit losses on the supply of credit to private nonfinancial borrowers, with a particular focus on on-balance-sheet lending by banks and other leveraged financial institutions, off-balance-sheet lending through the ABS markets, and lending backed by government-sponsored enterprises (GSEs) such as Fannie Mae and Freddie Mac. The third section discusses the potential impact on economic activity, using an instrumental variables approach to estimate the link between credit supply and real GDP growth. The fourth section concludes.