Barking up the Wrong Tree? Measuring Gender Gaps in Firm’s Access to Finance

Abstract

The literature on gender-based discrimination in credit markets is recently expanding, but the results are not yet definitive and have not been generally agreed upon. This paper exploits a new dataset on Barbados, Jamaica, and Trinidad and Tobago, which provides detailed information about female ownership and management in firms for investigating the existence of a gender gap in access to finance. The evidence presented herein suggests that more precise measures of the gender composition of the firm show that women-led businesses are more likely to be financially constrained than other comparable firms.

Acknowledgements

This project has benefited from the financial support of the Inter-American Development Bank (IDB). The authors thank the Editor, an anonymous referee, Martin Chrisney, Luana Ozemela, Alberto Zazzaro, and the participants at an IADB seminar for their useful suggestions. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the International Monetary Fund or the IADB, or the policies of these institutions and the countries they represent.

Notes

1. This limitation is specific to the most recent waves of the WBES. In earlier waves, there was a different variable on the gender of the principal owner of the firm (see, for instance, Bardasi et al., 2011; Bruhn, 2009).

2. In this variable, as well as in the following ones, the control group consists of all other firms in the sample, including those fully owned or managed by men and those in which the presence of women in ownership or management is less than or equal to the presence of men.

3. Alternatively, one could merge discouraged and rationed borrowers, building a single dummy variable which is equal to one for rationed and discouraged firms, similar to what is done by Jappelli (1990) and Muravyev et al. (2009). This would avoid the risk of underestimating the share of rationed firms because of self-selection of borrowers who refrain from applying to banks, but would lead to an overestimation of denial rates. The latter issue may generate biased estimates of gender discrimination if men and women differ in their degree of confidence, and thus have different attitudes when choosing to apply for a loan or not. Henceforth, we prefer to include both variables separately in the empirical analysis. A second concern is regarding the group of firms that had not demanded bank credit as they do not need it. While in the baseline regressions those firms are included in the sample, we have tested the robustness of our results excluding this group of (109) non-demanding firms. Results are reported and discussed in the Online Appendix.

4. Unfortunately, we share with the analyses based on the WBES the impossibility to control for other measures that may capture the riskiness of the firm (for example credit rating and financial structure) and the lack of individual-specific information (for example age and education) about the major owner/shareholder or the top manager. Other data are available (for example innovative capacity and labour force education), but their inclusion dramatically reduces the sample size.

5. The limited sample size does not allow for a more granular distinction of the sectors in which the firms operate.

6. The data collected in the three countries are pooled together because the small sample size does not allow for meaningful country-specific analyses. Nevertheless, the pattern in the three countries is indeed very similar, justifying the poolability of the data.

7. The data also highlight that WLBs do not have a significantly different probability of using bank credit than other firms, irrespective of the gender definition. The lack of a gender gap in using bank credit, even without controlling for a firm’s characteristics, is consistent with the evidence discussed by Bruhn (2009) on some Latin American countries.