After a long review process, my paper with Hans Degryse, Ralph De Haas and Neeltje van Horen “When Arm’s Length Is Too Far: Relationship Banking over the Credit
Cycle” has been conditionally accepted at the Journal of Financial Economics.
The paper assesses the relative effectiveness of different lending techniques during boom and bust times. Theory suggests that while relationship
and transaction-based lenders (where the former rely on soft information and longer-term relationships with their borrowers, while the latter rely more on hard assets and hard information for screening and monitoring) can be substitutes during good times,
relationship lenders have an advantage in crisis times given their deeper information on borrowers. We combine a bank-level survey on lending techniques with firm-level survey information financing constraints across 21 Central and Eastern European countries.
Matching the geographic coordinates of these banks’ branches and firms’ location, we then test whether firms’ financing constraints vary with the relative geographic proximity of relationship lenders, during good times (2005) and crisis
times (2008/9). We find that while relationship lending is not associated with credit constraints during a credit boom, it alleviates such constraints during a downturn. This positive role of relationship lending is stronger for small and opaque firms
and in regions with a more severe economic downturn. Moreover, our evidence suggests that relationship lending mitigates the impact of a downturn on firm growth and does not constitute evergreening of loans.
As always, papers change during
the review process and we have certainly benefitted a lot from this process. One important improvement (and, in my humble opinion) a contribution in itself was to test the consistency of our bank-level survey data on lending techniques with credit registry
data from one specific country. Using credit registry data from Armenia (courtesy Larissa Schӓfer who also has a fascinating
paper on relationship vs. transaction based lenders) we find that banks that classify themselves as relationship lenders engage in significantly longer and broader lending relationships, deal with smaller clients, and are less likely to require collateral,
and thus consistent with what the previous empirical literature on relationship lending has shown. Another improvement was that we show the robustness of our findings to using firms’ balance sheet data rather than survey data, which reassures us that
our findings are not driven by the survey nature of our credit constraint data.