OBJECTIVES: High drug prices present substantial challenges to providing good, affordable healthcare, especially in resource poor settings. To combat rising drug prices, in 2002, Argentina implemented “The Generic Law” mandating that prescribers write prescriptions with the International Nonproprietary Name instead of brand name to promote generic drug use and lower overall costs. We examined whether this policy was effective in lowering drug prices. METHODS: We used IMS price data of 192 drug products from the WHO essential medicines list in Argentina from 1995-2010 along with limited utilization data. The study employed a longitudinal interrupted-time-series analysis of price and drug use before and after The Generic Law was implemented. We estimated a fixed effects regression model both with real price and log of real price as the dependent variable and time, policy, number of drug products on the market, and brand or generic status and brand policy interaction as independent variables, with residual tests for robustness. RESULTS: The robust pooled log price trends of 192 drug products show that prices decreased over all years by 1.1% (CI=0.0067-0.0151,p<0.00). We showed a robust interaction effect between policy and brand with the policy causing a 7.9% decrease (CI=-0.1359--0.0221,p=0.006) in generic drugs prices over brand prices (1.5 pesos).  The policy by itself did not have a significant effect over all drug prices, nor did density of market share.  Brand drugs across all time periods had a 27% increase (CI=0.0895-0.4525,p=0.003) in price (5.13 pesos).  Visual inspection of utilization trends indicated that brand utilization decreased while generic use stayed stable or increased. CONCLUSIONS: Argentina’s generic policy had its intended differential effect between brand and generic prices with generic prices decreasing significantly with respect to brand prices, although this difference was small. These findings provide some evidence supporting implementation of generic drug policies in other countries, but should be validated.