A historical perspective on informal institutional and international entrepreneurship

The study investigates why international entrepreneurship varies across African countries, focusing on how informal, ethnicity-related institutions and their historical roots shape new venture internationalization. While prior work emphasizes formal institutions and culture in largely Global North contexts, Africa remains underexamined. The authors introduce ethnic fractionalization as an informal institution that can diminish trust and information-sharing, thereby constraining new ventures’ ability to discover, evaluate, and exploit foreign market opportunities. They adopt a history-based approach, positing that the precolonial prevalence of the tsetse fly—an epidemic disease vector harmful to humans and fatal to livestock—created persistent institutional disruptions that increased ethnic fractionalization. The paper asks: How do the historical roots of ethnicity-related informal institutions affect international entrepreneurship in Africa? The authors predict that higher ethnic fractionalization reduces both the propensity and intensity of new venture internationalization, that historical tsetse prevalence directly depresses internationalization today, and that the tsetse effect operates in part through ethnic fractionalization.

International entrepreneurship research links cross-national variation in entrepreneurial internationalization to political, economic, cultural, and social institutions, typically modeling institutions as exogenous predictors while overlooking their historical origins. A historical perspective suggests traumatic shocks (wars, disasters, epidemics) can permanently alter institutional paths, influencing contemporary entrepreneurial behavior. The tsetse fly’s historical distribution in Africa provides exogenous variation in epidemic disease exposure with documented long-run effects on development. Prior studies show disease environments weakened institutional quality (e.g., property rights and contract enforcement) and trust, which are crucial for entrepreneurship. Ethnic fractionalization, reflecting the probability that two randomly selected individuals belong to different ethnic groups, is theorized to reduce social cohesion, trust, and information-sharing, lowering entrepreneurial capacity to internationalize. The authors develop three hypotheses: H1, ethnic fractionalization is negatively associated with internationalization propensity and intensity; H2, precolonial tsetse prevalence is negatively associated with internationalization propensity and intensity; H3, ethnic fractionalization mediates the negative relationship between tsetse prevalence and internationalization outcomes. Mechanistically, tsetse-induced disease shifted livelihoods, intensified competition over scarce resources, increased outgroup exclusion, and fostered social distance, raising ethnic fractionalization that persists via cultural transmission and path dependence.

Data: Firm-level data come from the World Bank Enterprise Survey (WBES), 2009–2019. New ventures are firms ≤10 years old at the time of interview. After excluding older firms and matching to country-level sources (Harvard Dataverse, Global Competitiveness Index, World Governance Indicators), the final sample includes 33,621 firm-year observations across 40 African countries.

Measures: (1) New venture internationalization: (a) Propensity—binary indicator equal to 1 if the firm exported in the previous fiscal year, 0 otherwise; (b) Intensity—share of foreign sales in total sales. (2) Ethnic fractionalization: country-level Historical Index of Ethnic Fractionalization (Harvard Dataverse; Alesina & Zhuravskaya), defined as EF_ct = 1 − Σ_i S_ict^2. Historical EF uses the earliest available year (1945–1970 per country); modern EF uses the most recent year aligned to the WBES firm-year. (3) Precolonial tsetse prevalence: Tsetse Suitability Index (TSI) from Alsan (2015), a standardized measure of steady-state tsetse population based on climate-driven insect growth models; aggregated to country-level using population-weighted regional averages. (4) Controls: firm-level—venture size (ln employees), managerial industry experience (ln years), product quality (international certification dummy), venture innovation (percent sales from main new or significantly improved product/service); country-level—domestic market size (Global Competitiveness Index), economic development (ln GDP per capita), government effectiveness (World Governance Indicators). Industry, year, and country fixed effects included; standard errors clustered by country.

Model specification: For internationalization propensity (binary), probit regressions are used. For internationalization intensity (continuous), OLS regressions are used, with an inverse Mills ratio from the propensity model to address potential selection. To test mediation by ethnic fractionalization, a two-stage approach is employed: First stage regresses EF on precolonial TSI (and uses settlement duration as an instrument for EF following Ahlerup & Olsson, 2012). Second stage regresses internationalization outcomes on predicted EF (and controls). Robustness includes structural equation modeling to decompose TSI’s total effect into direct and indirect (via EF).

• Ethnic fractionalization negatively predicts internationalization of new ventures (supports H1): Historical EF → lower export propensity (β ≈ -1.984, p<0.001) and lower export intensity (β ≈ -0.955, p<0.001). Modern EF → lower export propensity (β ≈ -1.334, p<0.001) and lower export intensity (β ≈ -0.856, p<0.001).
• Precolonial tsetse prevalence negatively predicts modern internationalization (supports H2): TSI → lower export propensity (β ≈ -0.103, p<0.01) and lower export intensity (β ≈ -0.179, p<0.001).
• Mediation by ethnic fractionalization (supports H3): First stage shows TSI positively predicts both historical and modern EF (β ≈ 0.101 and 0.084, p<0.001). Indirect (mediated) effects are statistically significant: TSI → EF (historical) → export propensity (indirect ≈ -0.041, p<0.01) and → export intensity (indirect ≈ -0.008, p<0.05); TSI → EF (modern) → export propensity (indirect ≈ -0.043, p<0.01) and → export intensity (indirect ≈ -0.008, p<0.05).
• Controls: Venture innovation positively associated with both outcomes; product quality and managerial experience positively associated with intensity; larger domestic markets and higher GDP per capita often relate to lower export intensity; selection correction (inverse Mills ratio) is positive and significant in intensity models.
• Scope: Findings are based on 33,621 firm-year observations from 40 African countries (WBES 2009–2019).

The findings address the central question by demonstrating that a historically rooted informal institution—ethnic fractionalization—constrains new venture internationalization in Africa through reduced information-sharing and lower trust, thereby lowering both the likelihood of exporting and the share of foreign sales. Moreover, the precolonial prevalence of the tsetse fly exerted a traumatic shock with persistent institutional consequences, increasing ethnic fractionalization and directly and indirectly depressing modern internationalization outcomes. This history-based identification helps mitigate endogeneity concerns common to cross-country institutional analyses and clarifies causal pathways consistent with path dependence. The results expand institutional theory in international entrepreneurship by highlighting the role of ethnicity-related informal institutions, integrate historical context into explanations of modern firm behavior, and contribute new evidence from an underresearched African context. Policy implications suggest that complementing formal institutional reforms with initiatives to enhance trust and information-sharing across ethnic lines (e.g., cross-community business networks) may support early-stage internationalization.

This study shows that ethnic fractionalization, as an informal institution with deep historical roots, negatively affects both the propensity and intensity of new venture internationalization in Africa. It also demonstrates that the precolonial prevalence of the tsetse fly has a persistent negative relationship with modern internationalization, operating in part through increased ethnic fractionalization. The work advances international entrepreneurship theory by contextualizing informal institutions, integrating a historical perspective that links traumatic shocks to contemporary entrepreneurial outcomes, and enriching evidence from Africa. Future research could broaden internationalization measures (speed, scope), examine other forms of informal institutions (e.g., religion, tradition), and incorporate host-country institutional configurations and cross-country configurational or mixed-method approaches to further unpack how history and institutions shape entrepreneurial internationalization.

• Internationalization is measured via export propensity and intensity; other dimensions such as speed and scope are not captured.
• Focus on ethnic fractionalization may omit other informal institutions (e.g., religion, tradition) that also shape internationalization.
• Analysis centers on home-country institutions and their historical roots; host-country institutional arrangements and bilateral distances are not modeled.
• Although historical shocks help mitigate endogeneity, causal links cannot be established definitively; unobserved factors may remain.