When Economic Freedom and Bank Credit Deepen Africa’s Income Divide

Africa's markets may be opening faster than opportunity is spreading. Bank credit can rise, regulations can ease and investment can expand while informal firms, workers and low-income households remain excluded from the channels that create wealth. When assets, collateral, skills and legal protection are concentrated, liberalisation can generate more economic activity without closing the income gap.

In the study Economic Freedom, Financial Development and Inequality Dynamics, published in Economies, University of Johannesburg researcher Margaret Rutendo Magwedere analyses 26 African economies to trace this divide. The preferred estimates associate greater economic freedom and financial development with higher inequality, while education moves in the opposite direction, suggesting that reforms can amplify existing advantages when broader access is missing.

Liberalisation Does Not Erase Structural Advantage

The study covers the period reported as 2002 to 2024 and combines inequality data from the Standardized World Income Inequality Database with the Heritage Foundation's economic freedom index. It also uses World Development Indicators for private-sector bank credit, economic growth, inflation and education. The country sample reflects data availability rather than complete continental coverage.

Economic freedom in the analysis is a composite of 12 areas, including property rights, judicial effectiveness, business and labour freedom, tax burden, public spending, trade, investment and financial freedom. Such conditions can lower barriers to enterprise and improve resource allocation. They do not, however, determine whether households begin with comparable assets, skills, information or political influence.

The preferred system generalised method of moments estimate finds a positive and statistically significant relationship between economic freedom and inequality. A one-point rise in the index is associated with a 0.0002103 increase in the Gini measure. The coefficient is small, but its direction remains positive in the alternative estimation reported by the paper.

Magwedere interprets the result through the structural character of inequality in many African economies. Better-connected firms and wealthier households can act quickly when regulations ease or markets open because they already possess capital, collateral and networks. People working informally or living far from commercial centres may gain little from the same reform, at least initially.

The research does not establish that economic freedom inevitably causes inequality, nor does it show that restricting freedom would improve distribution. It indicates that formal freedom can coexist with unequal practical capacity. Rights and market access may exist on paper while the cost of using them remains prohibitive for large parts of the population.

More Credit Can Still Mean Less Access

Financial development produces a similarly uncomfortable result. The study measures it through bank credit to the private sector as a share of gross domestic product. In the preferred model, a one-unit increase in this measure is associated with a 0.000226 rise in inequality, statistically significant at the one percent level.

More lending, in other words, does not necessarily mean more inclusive lending. Banks may expand their balance sheets by serving governments, major companies, salaried customers and borrowers with conventional collateral. Informal enterprises, small farmers, women-owned businesses and low-income households can remain outside the system even as national indicators of financial depth improve.

Digital finance does not automatically resolve the imbalance. Mobile accounts and payment platforms can lower transaction costs and extend services beyond bank branches, but early adoption often favours urban, connected and digitally literate users. Automated credit scoring can widen access, yet poor data, opaque models or biased proxies can reproduce exclusion at greater speed.

The financial-development result also deserves caution. The paper's comparison model, feasible generalised least squares, produces a negative coefficient rather than a positive one. Magwedere gives priority to the system GMM model because it better addresses reverse causality and unobserved country differences, but the sign reversal shows that the relationship is sensitive to estimation choices.

Such sensitivity weakens any sweeping claim that financial expansion inherently widens inequality. A more defensible reading is that financial depth and financial inclusion are different outcomes. Policymakers need to ask who receives credit, at what price, for what purpose and on what terms instead of treating the size of the financial sector as evidence of democratised opportunity.

Education Sends the Clearest Equalising Signal

Income inequality also displays persistence in the study. The lagged inequality measure has a positive and significant coefficient of 0.129 in the preferred model, suggesting that present disparities are strongly shaped by earlier ones. Unequal economies do not reset when growth accelerates or regulations improve; accumulated differences in assets, schooling and market position carry forward.

Education provides the analysis's clearest equalising signal. Higher primary and secondary enrolment, measured through a gender parity indicator, is negatively and significantly associated with inequality in the preferred estimate. Better access to learning can strengthen mobility by improving skills and opening routes into more productive, better-paid work.

Enrolment alone cannot capture educational quality, completion, digital access or whether labour markets can absorb graduates. Even so, the result reinforces a practical point: people need capabilities before they can convert economic freedom into income. Liberalisation without human-capital investment may expand the returns to scarce skills while leaving the underlying scarcity untouched.

Other findings are less secure. Inflation is negatively associated with inequality in the preferred model, contrary to the usual expectation that rising prices hurt poorer households most, but it becomes statistically insignificant in the comparison estimate. Economic growth is not significant in the preferred specification, underscoring how weakly headline growth can describe distributional change.

The inflation finding should not be read as a policy argument for tolerating price instability. Low-income households generally have fewer savings and weaker protection against food, fuel and housing shocks. The paper treats the result as context-specific, and its lack of stability across estimators calls for further investigation rather than a change in inflation policy.

Reform Must Measure Who Gets In

The policy response is not to retreat from economic freedom or financial development.

• Governments need to build inclusion into the mechanics of reform. Credit guarantees for viable smaller firms, lower identification and collateral barriers, reliable credit information, stronger consumer protection and effective competition rules can widen participation without suppressing private initiative.
• Financial regulators should publish lending data by firm size, income, gender and location wherever privacy safeguards allow. Account ownership, total credit and transaction volumes reveal scale but not distribution.
• Development agencies and investors should judge programmes by affordability, sustained use and productive outcomes, including whether finance helps households build resilience or firms create decent work.
• Education, social protection and capable institutions remain part of the same economic architecture. Reliable courts and property systems must work for small entrepreneurs as well as established investors. Tax systems must finance public goods without placing disproportionate burdens on low-income households, while safety nets must prevent temporary shocks from becoming permanent losses of assets and opportunity.

The research has credible methodological features. System GMM is appropriate for a dynamic relationship in which inequality, finance and policy can influence one another. Reported tests support the instrument set and find no second-order serial correlation; the model uses 22 instruments across 26 country groups, reducing concern about excessive instruments.

Several limitations narrow the conclusions. The 26-country sample cannot represent every African institutional setting, and selection based on complete data may introduce bias. Bank credit measures financial depth rather than affordability or reach, while the composite freedom index cannot identify which of its 12 components shapes the observed association.

An internal reporting issue also needs clarification. A fully balanced annual panel covering 2002 through 2024 would ordinarily contain 598 country-year observations across 26 economies, yet the descriptive analysis reports 546 and the preferred model uses 494. Missing years or lag structures may explain the difference, but the paper does not reconcile the stated span with those totals.

Future research should separate property rights, trade openness, taxation, regulatory efficiency and other dimensions of economic freedom rather than treating them as a single package. It should also distinguish financial access, depth, efficiency and stability, then test whether institutional quality or inclusion thresholds change their relationship with inequality.