This dissertation consists of three independent essays that explore different aspects of the relation between aggregate demand and growth. The first chapter revisits the debate on unit roots in macroeconomic time series and discusses the endogeneity of the natural rate of growth. It analyzes how different theories of fluctuations interpret unit roots and offers an alternative interpretation, relating unit roots to the persistence of aggregate demand shocks. Finally, it presents two empirical exercises: one tests for unit roots in output for twelve Latin American countries using panel data. The results suggest that GDP series are non-stationary and therefore shocks may have persistent effects in the economy. The second exercise tests the hypothesis of an endogenous natural rate of growth, and the estimations suggest that the potential output has been influenced by the actual level of economic activity in Latin American economies. The second chapter focuses on the effects of aggregate demand on productivity and growth, as presented in the literature on cumulative causation models of growth in the Kaldorian tradition. It discusses the controversies related to the estimation and interpretation of Verdoorn's Law and provides estimations of the law using panel data for the seven largest economies in Latin America (1985-2001). The results confirm the existence of increasing returns in manufacturing, and the possibility of cumulative growth cycles in the region. The third chapter focuses on some economic policy implications of demand-led growth, and discusses the potential long-term growth effects of inflation targeting regimes. The hypotheses of procyclical and asymmetrical monetary policy were tested by using a VAR model for Brazil, Chile and Mexico (1999-2005). The results suggest that inflation targeting is likely to produce a downward bias in aggregate demand, and generate negative effects on long-run growth rates. The analysis and results provided here support different aspects of demand-led growth, and convey economic policy implications. In particular, it supports the idea that recessions may have long-lasting and irreversible effects on output and employment, and reinforces arguments against the promotion of sharp contractions in response to financial or currency crises in the developing world.