Collateral damage: Foreclosures and new mortgage lending in the 1930s

Price Fishback, Sebastian Fleitas, Jonathan Rose, Kenneth Snowden 18 January 2019

a

A

Severe crises in housing markets contributed to both the Great Recession of 2007 and the Great Depression of the 1930s. Recent research by Gertler and Gilchrist (2018) attributes the impact of housing on the broader economy to financial frictions, whereby decreases in home values generate disruptions in the balance sheets of households and intermediaries. Indeed, during the Great Recession, home foreclosures have been shown to be one specific mechanism that was at work (Mian et al. 2015, Annenberg and Kang 2014). Likewise, home foreclosures also spiked during the 1930s, while housing starts fell to 10% of 1929 levels and nominal housing prices dramatically. But the financial frictions generated by difficulties in housing during the Great Depression have yet to be explored because most literature on the Great Depression financial crisis has followed Friedman and Schwartz (1963) and Bernanke (1983) and focused on commercial banks, which played only a minor role in home mortgage lending at the time. In a recent paper (Fishback et al. 2018), we investigate for the first time the impact that foreclosures had on the supply of new home mortgage loans during the housing crisis of the 1930s.

The leading home mortgage lenders during the 1930s were cooperative building and loan associations (B&Ls). B&Ls accounted for nearly one-half of all institutionally held residential mortgage debt in 1930 (Snowden 2010). As home mortgage specialists, they faced the brunt of mortgage delinquencies and foreclosures, and these problems persisted well past 1933, in contrast to the pattern of suspensions among commercial banks that had largely ceased by mid-1933. 

Using data from individual B&Ls in the 1930s, we examine the immediate and direct impact of foreclosures on new lending. The mechanism that was at work is relatively simple. As foreclosures mounted rapidly in the early 1930s, B&Ls were forced to replace home loans on their balance sheets with real estate. Mortgage loans represented roughly 90% of the assets of a healthy B&L and generated a flow of funds from interest and principal repayments which represented, along with new investments into the association, the pool of funds available for new loans. When a B&L foreclosed on delinquent borrowers, the loans on its balance sheet were replaced with foreclosed real estate that earned lower average returns, that represented unrealized losses, and that was generally illiquid even if a B&L was willing to recognize the loss. As a result, foreclosures by B&Ls directly decreased the pool of loanable funds generated by its existing mortgage portfolio and discouraged new and additional investments into the association.

To examine the link between foreclosures, real estate owned on the balance sheet (REO), and mortgage lending, we collected annual balance sheet data for every state-chartered building and loan association operating between 1928 and 1940 in four states: Iowa, New York, North Carolina, and Wisconsin. These four states capture important regional variation in their housing crises, recoveries, and the evolution of their thrift industries during the 1930s. These specific states were selected because regulators in them reported the annual volume of new mortgage lending for individual B&Ls. Having information about the flow of new lending is unusual and particularly valuable because in its absence we would be forced to rely on annual changes in the stock of mortgage loans reported on the balance sheet to measure the volume of new loans. This proxy is frequently employed in the examination of new lending by banks, but it is a noisy and potentially biased measure for institutions that hold portfolios dominated by mortgages of different vintages and durations when they are dealing with unusually high levels of terminations due to foreclosures. We show that the reduction in measurement error from using the actual value of new loans leads to materially different coefficients and more precisely estimated results. 

Our empirical strategy isolates the impact that yearly shocks to the level of REO at a B&L had on the volume of its new lending in the following year. We estimate the impact using changes from year to year in new loans and REO while controlling for changes in B&L size and the structure of balance sheets, as well as annual shocks to the city or county housing markets where B&L were located. We also examine the relationship while additionally controlling for a variety of new government policies and changes in management of the B&Ls. The relationship between new lending and REO changed very little in all of our robustness checks. For unmeasured factors to be the cause of the relationship we find, they would have to be three times as powerful as the factors for which we already control, which is highly unlikely. 

The results show that an increase in foreclosed real estate on a B&L’s balance sheet had a powerful and negative effect on new mortgage lending during the 1930s. A one-standard-deviation rise in REO as a share of assets was associated with a five-percentage-point drop in new loans as a percentage of assets. Such a drop was nearly one-third of the mean value of the ratio of new loans to assets among B&Ls. 

The impact of REO was felt throughout the crisis but changed in character during the decade. Between 1929 and 1935, the increase in the mean REO share was associated with 30% of the drop in the mean new loan share. In contrast, the REO mean share of B&L assets remained elevated throughout the last half of the decade, with only a small reduction that accounted for 5% of the total increase in new loans. 

Our evidence adds to a growing literature that identifies the 1930s housing and mortgage crisis, and financial frictions within non-bank financial intermediaries, as important, but previously neglected, factors that contributed to the severity and duration of the Great Depression (White et al. 2014). The results also suggest a channel through which the federal government’s Home Owner’s Loan Corporation ameliorated the housing crisis of the 1930s, as has been shown in other recent investigations (Fishback et al. 2013). Between 1933 and 1936, the Home Owner’s Loan Corporation purchased one million distressed mortgages that would have otherwise added substantially to lenders’ REO and, according to our results, depressed local supplies of mortgage credit. On the other hand, our analysis indicates that other New Deal housing policies—the establishment of a mortgage discount lending system within the Federal Home Loan Bank and the introduction of federal savings and loan charters—did not prevent foreclosures from restricting new mortgage lending. 

Beyond the historical contribution, the experience of B&Ls during the 1930s provides an opportunity to examine the impact of foreclosures in a setting where local lenders funded and serviced their own mortgages. The relationship between foreclosures and new mortgage lending is less transparent in the modern era where loan origination, servicing, and funding are separated within complex institutional structures. Nonetheless, the direct costs to lenders of liquidating foreclosed real estate continue to be substantial in the modern market—about 12% of the original loan balance in normal times and rising to 20% and higher during a crisis when foreclosure rates also increase dramatically (Cordell et al. 2015). During a foreclosure crisis, modern lending networks must still absorb the costs and delays associated with the resulting REO, but these are now shared, and often contested, by a variety of participants with different incentives and contractual obligations. And, as in the 1930s, we also saw the supply of mortgage credit tighten considerably during the 2007 housing crisis as foreclosures led to a large excess supply of vacant homes (Federal Reserve 2012). We need to better understand how the impacts of foreclosures are distributed and funded within modern lending channels, whether and by how much they choke off new lending, and how policy interventions can be structured to ameliorate these impacts. 

Author’s note: The views expressed do not necessarily reflect the views of the Federal Reserve Bank of Chicago or the Federal Reserve System.

References

Anenberg, E, and E Kung (2014), “Estimates of the size and source of price declines due to nearby foreclosures”, American Economic Review 104(8): 2527–51.

Bernanke, B (1983), “Non-monetary effects of the financial crisis in the propagation of the Great Depression”, American Economic Review 73(3): 257–276.

Cordell, L, L Geng, LS Goodman and L Yang (2015), “The cost of foreclosure delay”, Real Estate Economics 43(4): 916–956.

Federal Reserve Board (2012), “The US housing market: Current conditions and policy considerations”, white paper.

Fishback, P, S Fleitas, J Rose and K Snowden (2018), “Collateral damage: The impact of foreclosures on new home mortgage lending in the 1930s”, NBER Working Paper 25246.

Fishback, P, J Rose and K Snowden (2013), Well worth saving, Chicago: University of Chicago and NBER.

Friedman, M, and A Schwartz (1963), A monetary history of the United States, Princeton, NJ: Princeton University Press.

Gertler, M, and S Gilchrist (2018), “What happened: Financial factors in the Great Recession”, Journal of Economic Perspectives 32(3): 3–30.

Mian, A, A Sufi and F Trebbi (2015), “Foreclosures, house prices, and the real economy”, Journal of Finance 60(6): 2587–2634.

Snowden, K (2010), “The anatomy of a residential mortgage crisis: A look back to the 1930s”, in L Mitchell and AE Wilmarth (eds.), The panic of 2008: Causes, consequences and proposals for reform, Northhampton MA: Edward Elgar Publishing.

White, E, K Snowden and P Fishback (eds.) (2014), Housing and mortgage markets in historical perspective, Chicago: University of Chicago Press and NBER.

a

A

Thomas R. Brown Professor of Economics, University of Arizona

Assistant Professor of Economics, University of Leuven

Policy Advisor, Federal Reserve Bank of Chicago

Professor of Economics at the University of North Carolina, Greensboro

Events

CEPR Policy Research