This paper examines the effect of banking on economic growth in modern Russia. To overcome simultaneity and selection, we exploit regional banking variation induced by the creation of “specialized banks” (spetsbanks) in the last years of the Soviet Union (1988-1991). Consistent with the qualitative work of Joel Hellman  and Juliet Johnson , we show that these reforms generated an ideal natural experiment in that the concentration of spetsbanks is jointly uncorrelated with 15 predictors of future growth, including pre-banking income, education, anti-market sentiment, institutional quality, and government interference in the economy. Results indicate that while the presence of one additional spetsbank per million inhabitants increased total within-state lending to private firms and individuals by 14 to 26 percent in the early 2000s, it had no effect on investment or per capita income. In contrast, we find that spetsbanks increased employment. Additional results indicate that spetsbanks increased growth in regions in which they were less connected to government and were generally more similar to non-spetsbanks, as well as in regions that were better at protecting property rights. Our results thus strongly suggest that bank origins, political connections, and property rights are important determinants of effective finance.