Relationship Lending and the Great Depression

Abstract

The Great Depression remains ground zero for studying the non-monetary effects of financial crises. Despite the abundant scholarship on the period, lack of disaggregated data on lending activities has constrained our ability to measure the impact on the real economy of a collapse in long-term lending relationships. We propose here a novel way to extract cross-sectional differences in relationship lending from geographically aggregated financial statements. We find that the banking crises of the early 1930s, by destroying these relationships and the soft yet crucial information garnered from them, explain one-eighth of the economic contraction observed during the Depression. This effect comes specifically from small bank failures which alone explain one-third of the Depression. Large bank failures, on the other hand, were accompanied by a reallocation of deposits towards surviving relationship lenders, leading to economic gains which mitigated the overall negative impact of the banking crises. We show that ignoring cross-sectional differences in continuing relationships on the eve of the Great Depression understates by a factor of 2 the fall in economic activity directly attributable to the banking panics of the early 1930s. We also show that the rebuilding of lending relationships in the mold of those that existed in the 1920s was an important determinant of cross-sectional differences in economic performance during the 1937-38 recession.Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.