Escaping the Poverty Trap: Expectations, Adaptive Learning and Human Capital Accumulation

This paper studies the transition from poverty to prosperity using a model of multiple equilibria. Existing theoretical work in this area has been dominated by perfect foresight models. In a stochastic growth model with human capital externalities, I study the role of expectations in transition dynamics. In particular, I assume that agents are boundedly rational and examine the effect of econometric learning on stability and transitional dynamics. I find two equilibria, of which only the poverty trap with no education is locally stable under learning. However, even with low existing levels of human capital, shocks to expectations permit the economy to escape the poverty trap under learning. The escape path does not lead to the upper steady state. Instead, adaptive learning generates endogenous oscillations in investment in physical and human capital. ∗Department of Economics, Northeastern University, 360 Huntington Avenue, Boston, MA 02115, USA. Email: [email protected]. I thank George Evans, Shankha Chakraborty, Jeremy Piger, Yue Fang, Bill Dickens, Maria Luengo-Prado and seminar participants at the University of Oregon, University of Oklahoma, Northeastern University, Middlebury College, Rutgers University Newark, University of Georgia, Lawrence University, and California State University Sacramento for insightful comments, suggestions and stimulating conversations out of which this product emerged. All remaining errors are mine.