Bank customers complaints, financial knowledge and gender

ABSTRACT

This work aims to analyse the relationship between financial knowledge and the propensity of consumers to complain in the event of disagreement with a financial institution. A sample of 8,286 individuals from the Survey of Financial Competencies elaborated by the Bank of Spain is analysed. The results reveal that individuals with greater financial knowledge show a greater propensity to complain, presumably due to their heightened ability to detect possible fraud or abusive practices by financial entities. Gender itself is not relevant in the propensity to complain. However, when the sample is divided by gender, financial knowledge is more relevant for women than men. Additionally, the results reveal that in the female sample, propensity to complain is affected by self-confidence, in contrast to the male sample. The results suggest that the authorities should take practical steps to include financial knowledge in the educational curriculum.

KEYWORDS:

• Financial knowledge
• financial literacy
• complaint
• banks
• gender

JEL CLASSIFICATION:

• D14
• D18
• D83
• G21
• G53

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. These authors, as well as co-authored with others (e.g. Lusardi and Mitchell), have previously published several studies on financial literacy, which have not been cited for greater brevity.

2. Specifically, when they analyse the propensity to make a complaint, which they call ‘aggression in handling financial disputes’. See the results in Tables 7 and 8 (pp. 57–58).

3. Data obtained by own elaboration from European Commission, Complaint data 2006–2018.

4. Byrnes et al. (1999) can be consulted for a revision of the theories that explain the gender gap in risk taking.

5. This leads to a possible endogeneity, since a lower propensity to complain may be motivated by an adequate selection of the entity.

6. Bover et al. (2019) offer a detailed explanation of the sample selection process, and Bover et al. (2018) an analysis of the demographic characteristics of the sample by autonomous community.

7. Specifically, we have eliminated the following observation with missing data: 26 in claims, 185 in the possession of a financial product, 51 for self-confidence, 14 for age, 3 for marital status, 2 for educational level and 1 for employment. The sum of these amounts to 282, although the coincidence of the respondent in some observations with missing data reduces it to 268. The income variable has a 10% of missing data, but we use the imputation data provide for the ECF. All variables are explained in the sub-section 3.1 and the ECF codes can be seen in the Table A1 of Appendix.

8. The last item is used by Engels et al. (2020), it is not in OECD methodology.

9. This category includes housewives, students, disabled and those who are not working or looking for work.

10. The Survey of Financial Competences does not specify the amount of household income, but indicates the income bracket in which it falls. This variable has 10% missing data, so we have used the imputed data provided by ECF. An explanation of the imputation procedure can be seen in Bover et al. (2019).

11. For a literature review, see Tronvoll (2007).

12. Liu and McClure (2001) conclude that the incidence of certain demographic characteristics in the behaviour of complainants depends on the type of product. For this reason, it has been decided not to consider studies referring to other products or services.

13. To calculate the descriptive, the sample weights provided by the ECF were used, which have been adjusted to reflect the demographic structure of the Spanish population (Bover et al., 2019).

14. All statistics and estimation have been obtained using the econometric package Stata14. To adjust by weight, the models are estimated using the ‘pweight’ option in Stata´s command.

15. In models 2 and 3 the beta coefficients have been omitted for simplicity.

16. For example, Engels et al. (2020) and Wei et al. (2021) use the educational level of householders’ parents. Unfortunately, this information is not available in our database.