Mobile money, inclusive finance and enterprise innovativeness: an analysis of East African nations

ABSTRACT

This study explores the relationship between firms’ adoption of mobile money services and their innovativeness in the East African countries of Kenya, Uganda and Tanzania. We argue that the use of mobile money by firms in these countries has had a positive indirect impact on their innovativeness by reducing the credit constraints they face. Using data available from the World Bank Enterprise Surveys, we adopt a sequential modelling approach, first estimating the impact of mobile money use on credit constraints and then estimating the impact of credit constraints on binary indicators of product, process and organisational innovation. Innovation is shown to depend on whether or not the firm is credit constrained, and the probability of being credit constrained is shown to depend on the choices made about the use of mobile money. We find that the indirect impact of mobile money on innovation performance is greater for small firms than for medium or large firms.

KEYWORDS:

• Mobile money
• credit constraint
• innovation
• Africa

JEL:

• G2
• O16
• O30

View correction statement:

Correction

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ For the Kenyan National Innovation Agency, see: http://www.innovationagency.go.ke/and for the MOSTI in Uganda, see: https://www.mosti.go.ug/. For the 2010 NSTP in Tanzania, see: http://www.tzonline.org/pdf/National_Research&DevPolicy.pdf.

² When it first introduced M-PESA in 2007 Safaricom in Kenya explicitly targeted the domestic remittances market using the slogan ‘send money home’ (Demirguc-Kunt et al. 2015, p. 42). Also see Munyegera and Matsumoto (2016) who argue for the case of Uganda that by facilitating remittances mobile money resulted in a significant increase in household per capita consumption.

³ In addition to cash-in and cash-out services, agents may also provide customer services as teaching new users how to use the money platform on their mobile phone. Agents, who typically conduct other types of business in addition to mobile money services, take a commission on transactions. See Davidson and Leishman (2010).

⁴ See World Bank, Open DataBank: https://databank.worldbank.org/home.aspx.

⁵ See section 4 below for a discussion of the sampling methodology used in the WBES.

⁶ M-Pesa introduced a deposit and loan service, M-Shwari, in 2013 targeting individual M-Pesa account holders. A similar service, MoKash, was established in Uganda in 2016.

⁷ The literature using the sequential CDM modelling approach focuses on the links between R&D, innovation and productivity. For a recent overview, see Lööf, Mairesse, and Mohnen (2017).

⁸ In order to obtain proper standard errors, we bootstrapped the standard errors (Guan 2003).

⁹ See Greene (2012, p. 746), a noted authority on econometric methodology, who observes with respect to the recursive bivariate probit model, ‘the endogenous nature of one of the variables on the right-hand side of the second equation can be ignored in formulating the log-likelihood.’ Also see Li, Poskitt, and Zhao (2019) who assess the robustness of the maximum likelihood estimator to misspecification of the error distribution. On the basis of Monte Carlo results, they show that when using a recursive bivariate probit model to estimate the average treatment effect (ATE), the quasi maximum likelihood estimator can produce reasonably accurate ATE estimates even in the presence of distributional misspecification.

¹⁰ The samples of firms in each country are constructed following a stratified random selection. For details on the sample frame and survey methodology, see the following link: http://www.enterprisesurveys.org/methodology).

¹¹ The total number of observations used in our regressions is 1686, with 527 for Tanzania, 518 for Uganda and 641 for Kenya.

¹⁹ Agro-industry is defined widely to include food, textiles, garments, leather, wood, paper, publishing and furniture.

¹² Our category of credit constrained firms combines the categories of ‘fully’ and ‘partially’ credit constrained firms in the terminology of Kuntchev et al. (2012, p. 10). They define partially credit constrained firms as firms that while meeting the conditions in the definition above did make use of external finance during the previous fiscal year and/or had an outstanding loan at the time of the survey. Our results presented in Section 5 below are robust to defining credit constrained firms as being those that are ‘fully’ credit constrained. For the sake of brevity, the regressions using the alternative more restrictive definition have not been included but are available upon request.

¹³ Other studies using the number agents in close proximity to the enterprise as an instrument include Islam, Muzi, and Rodriguez Meza (2018) and Jack and Suri (2014).

¹⁴ We would like to express our thanks to Asif Islam of the World Bank Global Indicators Group for providing us with the geocoded data.

¹⁵ We would like to express our thanks to Michael dos Santos and Petronella Tizora of Insight2Impact (i2i) for kindly providing us with the geographical data on the number of mobile money agents. Insight2Impact is a global resource centre dedicated to providing data services in support of financial inclusion. See: https://i2ifacility.org.

¹⁶ As observed above, while micro credit services were introduced by MNOs operating mobile money platforms in all three countries more recently, this occurred after the period for which our data measuring the use of mobile money was collected by the World Bank. Demand for credit by individuals and firms was unrelated to the investments the MNOs had made up to 2013 in establishing their agent networks.

¹⁷ The literature addressing this issue is vast and is addressed in the research on clusters and in that on regional and national innovation systems. For a recent overview, see Ozman (2009).

¹⁸ For purposes of comparison the Table 5 in Appendix shows the results for a simple probit regression without instrumenting the intensity of mobile money use.