June 11, 2019

Dutch SME bank financing, from a European perspective

This policy brief confirms that Dutch SMEs are applying for relatively few bank loans, and that those applications are relatively often rejected by the banks. This applies to all businesses but more so to SMEs. The 2009–2018 period shows a steady trend, although in recent years the difference with the eurozone average number of applications has grown, while for the number of rejections the difference has decreased.
No title

This policy brief confirms that Dutch SMEs are applying for relatively few bank loans, and that those applications are relatively often rejected by the banks. This applies to all businesses but more so to SMEs. The 2009–2018 period shows a steady trend, although in recent years the difference with the eurozone average number of applications has grown, while for the number of rejections the difference has decreased.

Most businesses that do not apply for a loan do so because they have sufficient financial means at their disposal. In addition, in the Netherlands as opposed to elsewhere in Europe, the expectation of a loan application being rejected appears to be a more important reason for not even applying for one. Dutch businesses also appear to make fewer physical investments. A number of possible other explanations, such as a greater reliance on alternative sources of financing, do not seem to play a role.

Zie hieronder 2 gerelateerde publicaties: 'Bank credit: Dutch versus European firms' en 'Measuring Credit Rationing of Dutch firms'

June 11, 2019
This background document presents and discusses the empirical results included in the CPB Policy Brief ‘Dutch SME-bank financing in a European perspective’. Our goal is to shed light on the differences in access to finance, especially bank loans, for Dutch SMEs in comparison to SMEs in the rest of the Eurozone. Using data on 160,000 SMEs from the Survey on Access to Finance (SAFE), which is conducted semi-annually by the ECB, we see that Dutch firms have a lower chance of being financed by bank loans. This result is a combined effect of both applying less often and being rejected more often. We also explore additional outcomes relevant to this result.

To answer the question of whether there are differences in the access to bank loans between Dutch SMEs and SMEs in the rest of the Eurozone we run a linear probability model with controls using microdata from the Survey on Access to Finance (SAFE). SAFE has been conducted semi-annually starting in the fall of 2009. The entire dataset contains over 200,000 firms of which 160,000 are SMEs. We control for turnover, sector, age of the firm and the number of employees within the firm. On top of that, we add the ‘wave’ (the survey round) as a control variable to correct for time-specific effects.

 

The results found in this analysis are robust to several sensitivity analyses. We use different comparison samples, test whether certain periods could be driving the results and estimate a split-sample regression. In all cases we see that Dutch firms apply less and are rejected more often for bank loans. However, this does not mean that we are able to pin down causal effects in this background document. We only report on correlations. More research is necessary to identify relevant causal relationships. 
 

Authors

Fien van Solinge
June 11, 2019
This background document is associated to the CPB Policy brief on Dutch SME-bank financing in a European perspective. It explains in detail what we do in a microsimulation model that is described in a textframe on credit rationing on page 14 of the Policy Brief. On the basis of firm-level data from Statistics Netherlands and under strong assumptions, we estimate the amount of credit rationing that Dutch firms potentially faced in 2007-2016. Our simulation study is novel, and it highlights that researching credit rationing on the basis of microdata is feasible and that it can yield insights relevant for policy applications.

In this simulation we model credit rationing through bankruptcy costs. To compute the amount of credit rationing we need, firstly, to estimate risks on a firm-by-firm basis, and, secondly, to estimate the bankruptcy costs. We estimate risks using a simple non-parametric approach. We split firms in clusters of similar size and age, and assume that firms within the same cluster face similar risks. We can then use the observed distribution of returns for firms in a given cluster to assess the risks.

Under some scenarios, we obtain that the smallest 20% of Dutch firms might have experienced larger credit rationing than the medium-sized firms. 
It would be interesting to compare these findings internationally, but there are no comparable simulations for other countries.
 

Authors

Andrei Dubovik

Authors

Karen van der Wiel
Andrei Dubovik
Fien van Solinge