First, we really do not observe genuine lending matchmaking

We pertain our the new approach to 21 nations which have different profile of monetary and you can financial innovation. This just enhances the external authenticity of before show as well as allows us to mine ranging from-nation (together with contained in this-country) variation both in the brand new incidence off relationship credit additionally the strength of your 2008–2009 recession. Furthermore, that with company-level survey research, we are able to separate between financially unconstrained enterprises, businesses that was sometimes refused otherwise disappointed off making an application for a good loan, and providers in the place of borrowing consult. Knowledge having fun with borrowing from the bank registry data face the potential maximum one to low-candidate organizations maybe do not require borrowing, otherwise they need borrowing from the bank but they are annoyed away from implementing. Ultimately, by the studying the character from relationships credit throughout the level and you may trough of the borrowing from the bank years, we are able to demonstrate that, as the predict because of the concept, dating credit is far more extremely important throughout the a good downturn. Regardless of so it different empirical approach, our very own efficiency line up really, and that harden, earlier conclusions to your part regarding dating credit during the downturns.

Around three methodological circumstances have earned comment. However, by using research for borrowing and you will non-borrowing from the bank firms we are able to assess the local standard harmony aftereffect of banks’ financing processes, which we would skip if paying attention merely to the borrowing businesses. 2nd, we believe in survey investigation for agencies and you will financial institutions. When you find yourself business-top survey studies was commonly used on the present books (age.g., Campello mais aussi al., 2010, Popov and you may Udell, 2012, Ongena et al., 2013), questions you will happen from the aspect mistake for the firms’ solutions so you can questions on why they don’t really sign up for financing, or that rejection you can expect to only echo the possible lack of money potential with positive websites present really worth. To help you decrease such concerns, you can expect several robustness testing with different meanings out-of firms’ borrowing restrictions, and tips predicated on harmony sheet investigation.

Third, all of our empirical strategy hinges on the region out of financial institutions and people getting separate of any almost every other. Pursuing the Berger, Miller, Petersen, Ra), i think that the latest banking surroundings close companies imposes a keen exogenous geographic limit towards banking institutions one to providers have access to. You can expect good facts that is consistent with so it expectation. I together with sample to have heterogeneous negative effects of your local presence of banks one to glance at by themselves given that matchmaking loan providers round the different types of enterprises. That it subsequent address the fresh it is possible to endogenous matching off enterprises and you may banks.

Brand new paper proceeds the following. Section dos identifies the information present we blend, and you can Part 3 gifts our very own empirical method. I talk about the standard performance and you can robustness assessment within the Sections 4 and you may 5, respectively. Area 6 up coming considers actual effects, and you may Area seven closes.

Area snippets

The empirical data rests towards the signing up for about three essential bits of suggestions: study toward firms’ credit limits at the different times, brand new geo-coordinates of one’s lender twigs surrounding these firms, and you may data on the lending processes ones banks. I discuss the analysis towards the firms’ genuine results in the Point 6.


To estimate the link between the share of relationship banks near a firm and the probability that the firm is credit-constrained, we use the following model for both the 2005 and 2008–2009 cross section. We hypothesize that relationship banks were particularly helpful once the cycle had turned in 2008. Consider the model Y i j k l = ? 1 X i j k l + ? 2 L j k + ? 3 S h a r e r e l a t i o n s h i p b a n k s j k + ? 4 C k + ? 5 I l + ? i j k l , where Yijkl is a dummy variable equal to one if firm i in locality j of country k in industry l is