ARCHIVES

This study tests whether the association between private credit and subsequent economic growth is nonlinear and whether the shape differs between household and non-financial corporate borrowing. A transparent availability rule yields an unbalanced annual panel of 28 economies. The data window is 1995–2019; the common one-year-lagged estimation sample contains 665 country-years over 1996–2019. Fourth-quarter total, household, and corporate credit-to-GDP ratios are BIS series preserved in the ALFRED vintage dated 30 April 2022. Growth and controls come from the IMF April 2022 World Economic Outlook archive. Two-way fixed-effects regressions use country-clustered standard errors. Quadratic optima are screened for statistical curvature, observed support, observations on both sides, and delta-method uncertainty; supplementary panel threshold regressions use a trimmed grid and a country-cluster wild bootstrap. The evidence does not identify an inverted-U relation. Total credit is negatively associated with next-year real GDP-per-capita growth in the linear model, while the quadratic term is positive and insignificant. Household credit has a significantly more negative linear association than corporate credit. The household quadratic is positive and its implied stationary point lies above support; corporate curvature is imprecise. Bootstrap threshold tests do not reject linearity at 5%. Thus credit composition is an important correlate of growth, but the data do not identify universal growth-maximizing credit thresholds. The results favor cautious macroprudential attention to household leverage and credit allocation, subject to endogeneity and cross-country heterogeneity.