The main reason for the existence of social banks is to fund other social enterprises. On that basis, Simon Cornée from the University of Rennes 1, Panu Kalmi from the University of Vaasa and Ariane Szafarz from the Université Libre de Bruxelles propose that social banks can operate profitably and still lend to their borrowers at attractive interest rates when their owners and depositors accept lower returns on their investments.
The three researchers use a large dataset on European banks to investigate this issue and find support for their hypotheses in their article The Business Models of Social Banks that was recently published in Kyklos. Their results help to understand the non-standard objectives of banking organisations.
How can social banks survive?
Social banks are still a relatively small part of the banking sector, but their asset growth during the period studied (1998-2013) was roughly twofold compared to the banking sector on average. Social banks are characterised by their adherence to social goals, transparency of operations, and need to be economically sustainable. They operate in a rather opaque and risky market when they lend to social enterprises. They create strong lending relationships and support to their clientele when they lend below the market prices. How do they manage to be economically self-sufficient?
The authors suggest that this is because their owners and depositors, who support the causes of the social enterprises to which social banks are lending, accept below-market returns on their investments. This leads into three core hypotheses: 1) social banks produce smaller returns to their owners than conventional banks; 2) they remunerate their depositors at below market interest rates; and 3) they grant loans below market interest rates. In their empirical analysis, they find support for all of these hypotheses.
Social banks have received only scant attention in the empirical banking literature, presumably because of their relatively small number. Cornée, Kalmi and Szafarz use a large dataset on European banks from 16 years (1998-2013). After doing extensive searches based on membership lists of social banking institutions, webpages, bank annual reports, and previous literature, they found 29 banks in 11 West European countries of which the primary business is to lend to social enterprises. Allowing for missing data for some banks for some years, they observed around 300 bank-year observations for social banks, which is less than 1% of the total number of observations.
Does ownership explain the results?
An important issue is whether the results related to social banks are due to their status as social banks, or due to the fact that many social banks are also stakeholder banks (cooperative or savings banks).
The authors suggest that even though social banks have many similarities with stakeholder banks and are characterised by lower returns to investors, the interest rate behaviour of social banks may well be rather different. For instance, the depositors of stakeholder banks are usually expecting higher, not lower interest rates.
The authors test this with a subsample consisting only of stakeholder banks and find out that both the interest rates social banks charge on their loans and pay on their depositors are significantly lower than those of stakeholder banks.
Zero interest rates create big challenges for social banks
For banks that operate with lower interest rates on both sides of the market, the movement towards zero nominal interest rates poses special challenges. If the zero interest rate is binding, one would expect that the pricing policies of social banks converge towards those of conventional banks when interest rates are approaching zero. By exploiting the variation of interest rates over time, the authors found support for the argument that the social spread on interest rates is not sustainable when interest rates approach zero. The dataset they use has however only a limited number of observations from the zero interest rate period, so the authors were not able to observe how social banks have adjusted to the prolonged period of zero interest regime.
The results are robust to alternative model specifications
As the main method, the authors employ matching methods. They match social banks with their four nearest neighbours with respect to country, ownership and size. This considerably reduces variation in the sample.
However, are the results specific to this particular way of constructing the sample? The authors further probe their results by using alternative methods on defining key variables, using a full sample instead of a matching sample and examining the selection based on unobservables. All their results remain valid in these robustness checks.