What explains Latin America's poor investment performance during the 1980–2001 period?: a panel unit root analysis

Abstract

This article analyses the theoretical and empirical links between key economic variables and private investment spending in Latin America during the 1980–2001 period. The Seemingly Unrelated Regression (SUR) estimates for the pooled investment model suggest that real (lagged) public investment, the output gap, lagged domestic credit to the private sector, and the national savings rate have a positive and significant effect on private capital formation, while the standard deviation of the real exchange rate index has a negative effect on private capital formation. A major contribution of the study is the application of recently developed panel unit roots test on the stacked residuals of the pooled regressions. The tests indicate that the included variables have a stable, non‐spurious (cointegrated) relationship. All in all, the findings in this article make an important contribution to the ongoing debate about which policies need to be promoted to raise and sustain the rate of private capital formation – Latin America's future source of employment and income creation.