Mobile money adoption and households’ vulnerability to shocks: Evidence from Togo

ABSTRACT

We investigate the determinants of mobile money adoption process and whether its use helps households in Togo to be resilient to predictable and unpredictable life events. Using ordered logit and sequential logit models, our results show that in the adoption process, households benefit from weak ties of social groups such as religious group and informal saving group for the adoption of mobile money. We equally find that being client of banks or microfinance institutions act as powerful channels from one step to another in the process. Besides, our findings reveal that households whoever use mobile money seem to be more resilient to climatic shocks such as drought, irregular rain, soil degradation, erosion and fertility reduction and to shock that affect households’ assets (non-climatic: high prices of agricultural inputs). However, the picture is more contrasted when the individuals are classified by disadvantaged groups such as rural people, women, less educated and people with low incomes.

KEYWORDS:

• Mobile money
• financial inclusion
• economic shocks
• Togo

JEL CLASSIFICATION:

• D14
• O12
• O55

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ In English, FCFA is franc of the African Financial Community (Benin, Burkina Faso, Côte d’Ivoire, Guinea-Bissau, Mali, Niger, Senegal, Togo), 1euro = 656FCFA.

² M-PESA is a mobile transfer service launched in Kenya in 2007 by a mobile operator, Safaricom. The success of this service in Kenya (nearly a quarter of the population adopted it in 2012) led the operator to extend it to other countries such as Tanzania in 2008, Afghanistan in 2008, and Romania in 2014, Chaix and Torre (2015).

³ For income, we always replace the severe outliers (using iqr command) by its median before running any estimation.

⁴ Togo is divided into five major economic regions (Maritime, Plateaux, Central, Kara, Savane). For more details see Figure 2 map and data description.

⁵ Notice that Lomé is considered as a domain because of its density in terms of population.

⁶ Our sample includes 972 individual MFI clients and 545 individual bank clients, see Appendix 1 for more details.

⁷ The coefficient and standard error of the total effects are computed using ‘Delta Method’.