Elite Networks Negatively Affect Credit Allocation in Germany

(This interview appeared in SAFE Newsletter Q1 2017)

Rainer HaselmannRainer Haselmann joined Goethe University and the Research Center SAFE as a Professor of Finance, Accounting and Taxation in October 2014. His research focuses on banking regulation, lending and portfolio allocation. Previously, he was a Professor of Finance at the University of Bonn. After having earned a Ph.D. in 2006 from the Leipzig Graduate School of Management, Rainer Haselmann worked as a postdoctoral researcher at the Columbia Business School and the University of Mainz. From 2009 to 2011, he was an Assistant Professor at the Bonn Graduate School of Economics.

In your most recent paper, “Rent-Seeking in Elite Networks”, you examine the role of elite social networks in Germany for the allocation of economic resources. How did you approach this topic?

We focus on the allocation of credit between banks and firms whose officials are members of the same elite service club organization. Based on member data for about 400 service club branches throughout Germany as well as contract-level financial data, we compare the lending relationship between a bank and a firm whose senior managers are members of the same club branch (in-group) to the lending of banks to firms whose officials are not connected via membership in the same club branch (out-group).

What are your results?

We were quite surprised by the enormous magnitude of our findings. First, we observe a drastically higher share of in-group as compared to out-group lending. This effect can be illustrated nicely by focusing on the event when a new entrepreneur joins a given club branch. Subsequent to this event, the firm experiences a 37 percentage points (pp) higher increase in lending from in-group banks than from out-group banks. This increase in in-group lending is not just due to a shift from lending by one bank to another, but due to a boost in total lending, resulting in a higher leverage ratio of the firm whose CEO joins the club.

As banking has to do with asymmetric information, it is, in a way, not surprising that proximity and, with that, possibly more transparent information facilitates credit relations. The interesting question is how these connection-based credits perform.

Yes, the really interesting finding of our paper is that these in-group lending relationships tend to be significantly less profitable for the bank as compared to out-group lending relationships. To look into this question we calculated the return on loans (ROL) that banks generate from in-group vis-a-vis out-group transactions. We find that a given bank generates a 4.4 pp lower ROL on in-group loans compared to out-group loans in the same city. When comparing lending from in-group and out-group banks to the same firm, we observe a 2.7 pp lower ROL for in-group banks. This is a considerable difference, so we were wondering about the reasons for the relatively bad performance of in-group loans. We first looked at the credit conditions, such as interest rates, but we did not find any evidence for the assumption that in-group firms are offered special conditions. Instead, it turned out that the bad performance is a result of too late exits: In-group banks still continue to lend to in-group firms when these firms’ business is turning bad and when out-group banks have already started to withdraw lending (see Figure 1).

On a large scale, this behavior must have a considerable negative impact on the economy as a whole. How is it possible that firms are not able to reap benefits from the generous credit supply?

Overall, we observe that in-group firms do not use the extra financing to make new investments but rather to increase payments to the shareholders. In most cases, this means paying out to the CEO, as our sample consists to a large part of relatively small family-owned firms. This indeed implies that, in a broader context, in-group lending as a consequence of elite networks results in an inefficient allocation of resources and that a developed economy like Germany would likely experience a higher overall economic growth rate if bank officials would not engage in elite networks.

Can you be sure that new or intensified lending relations are a consequence of club membership? One could argue that clubs regularly offer a new membership to the local economy’s shooting stars, so that your findings would display mere correlations rather than causality.

Analyzing the consequences of social networks on economic outcomes is clearly prone to serious selection concerns. However, the highly regulated procedures for membership in this service club organization does not make the timing you suggest very likely. Nevertheless, to strengthen our analysis in this respect, we introduced several tests on this issue. In particular, we focused on a purely exogenous event, namely the election of existing club members as mayors. In Germany, the mayor of a district often becomes chairman of the local state bank's supervisory board. In this capacity, he or she gains great influence on the bank’s loan-granting activity, especially for corporate loans. So, while the elected candidate has been a member of the club branch before, his or her degree of influence changes with the election. What we find is that, after an existing club member is elected mayor and becomes head of the local state bank, this bank increases lending to in-group firms by 49 pp compared to other banks. Additionally, total debt and leverage of in-group firms significantly increase after the election. In contrast, if a club member is elected mayor but is not appointed supervisory board chairman of the local state bank, there is no effect on in-group financing. This quasi-natural experiment with a control group provides a very strong argument for a causal relationship between proximity to a – in this case new – banker and boosting credit relations.

According to your findings, state-owned banks engage more actively in in-group lending than private banks and also face significantly lower returns on these loans. How can you explain this?

One reason for this result might be that the corporate governance of state-owned and private banks differs considerably. The objectives of state banks, as laid down in the respective laws, are more diverse than those of private banks. Also, private banks are controlled by shareholders who benefit from higher profits while state-owned banks tend to be controlled by local politicians. A well designed corporate governance framework seems to be the key to prevent inefficient in-group lending from the perspective of the bank.

Haselmann, R., Schoenherr, D. and V. Vig (2016): "Rent-Seeking in Elite Networks", forthcoming in the Journal of Political Economy and available as SAFE Working Paper No. 132.