This paper analyzes the relationship between financial development and economic growth in Latin America with a Granger causality test and impulse response functions in a panel vector autoregression model. Using annual observations from a sample of 18 countries from 1962 to 2005, it is shown that while economic growth causes financial development, financial development does not cause economic growth. This finding is robust to different model specifications and different financial indicators. Interestingly, when the sample is divided according to different income levels and institutional quality, there is two way causality between financial development and economic growth only for the middle income group and for countries with stronger rule of law and creditor rights. The impulse response functions show that a shock to financial development has a positive impact on economic growth only for these subsamples, but the net effect of financial development on growth is relatively small. * Acknowledgements: This paper benefited greatly from comments by Robin and Kevin Grier; all their help is greatly appreciated. I am thankful to Inessa Love for making her code for panel VAR available to me and Haichun Ye for her help in the empirical part. I also thank James Prieger for helpful comments, Michelle Isenhouer for research assistance, and Luis Rodriguez for graphics editing. Two anonymous referees have made valuable comments, which helped to improve the paper significantly. Any remaining errors are my own.