Informality and credit constraints: evidence from Sub-Saharan African MSEs

ABSTRACT

The attributes of micro and small enterprises (MSEs) influencing access to credit, in particular the level and role of firm informality, are analysed in the article. The puzzle is the push for MSEs to join the formal sector and the tug to avoid the extra burden it places on the firm. It is important to know more clearly what forces are at work and the sources of the causal effects. This study uses data from the World Bank Enterprise Surveys for five low-income countries (LICs) in Sub-Saharan Africa. The method is empirical and as we find informality to be endogenous to credit constraints, an instrumental variable approach is estimated. Further, to address the possibility of reverse causality, an instrument for the informality variable is required; not registered with Inland Revenue (tax office) is the chosen instrument variable. The findings reveal that as the probability of a firm operating in the formal sector increases, there is greater access to external credit. The causality relationships are tested providing a strong platform for the formalization of polices to reduce the informality of the MSE sector. These are discussed in the context of the research findings.

KEYWORDS:

• Credit constraints
• informality
• micro and small enterprises
• low-income countries
• endogeneity

JEL CLASSIFICATION:

• G32
• O17

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

¹ According to the Enterprise Surveys, Sub-Saharan MSEs have 0–19 employees. Firms with 0–4 employees are categorized as micro, and firms with 5–19 employees are categorized as small (Enterprise Surveys 2013).

² Sub-Saharan Africa is a very diverse region, whether measured by population income levels, or composition of output. Forty-five countries, excluding Sudan geographically lie south of the Sahara Desert and belong to the Sub-Saharan Africa group (Regional Economic Outlook 2012). International Monetary Fund puts the 45 countries into 4 sub-groups: oil exporters, middle-income countries, low-income countries and fragile countries. In this study, we have considered Burkina Faso, Democratic Republic of the Congo, Madagascar, Mali and Rwanda as low-income countries in Sub-Saharan Africa (IMF 2015).

³ In a statistical model, a parameter or variable is said to be endogenous when there is a correlation between the parameter or variable and the error term (Wooldridge 2002).

⁴ ivprobit – Probit model with continuous endogenous regressors. It is typically for this reason that generalized linear models, like probit or logit, are used to model binary dependent variables in applied research, and an approach that extends the probit model to account for endogeneity. See the probit techniques proposed by Rivers and Vuong (1988).

⁵ The World Bank’s Informal Enterprise Surveys (IFS) collect data on nonregistered business activities in every region of the world. The IFS are implemented in parallel to the World Bank’s Enterprise Surveys (ES), which interview formal, private, non-agricultural firms in countries around the world (Enterprise Surveys 2013).

⁶ The World Bank’s Enterprise Survey website provides details as to how the survey was conducted. (Enterprise Surveys, http://www.enterprisesurveys.org, The World Bank.) An Enterprise survey is a firm-level survey of a representative sample of an economy’s private sector. The surveys cover a broad range of business environment topics including access to finance, corruption, infrastructure, crime, competition and performance measures.

⁷ The following group of variables is controlled for at the individual firm level: security of property rights, firm-related general characteristics, firm owner’s characteristics and business environment.

⁸ Following Maddala (1983), it is widely believed in the literature that in the joint estimation of Equations 1 and 2, parameter vectors are not identified in the absence of exclusionary restrictions. However, Wilde (2000) argues that the joint model is identified as soon as both equations have a varying exogenous regressor. Monfardini and Radice (2008) also state that identification of this model does not require any additional instruments. But note that in the absence of exclusionary restrictions, identification relies heavily on the functional form. Therefore, estimation with additional instruments yields parameters’ estimates that are more robust to distributional misspecifications. Hence, I rely on identifying instruments in the empirical analysis.

⁹ Wald’s (1940) method of fitting straight lines was specifically developed to overcome errors-in-variables problems. Durbin (1954) showed that Wald’s method is a special case of instrumental variables (see also Geary 1949). In this symposium, Hausman provides an overview of measurement error problems.