Since the mid 90s 20 US states and DC have legalized medical marijuana, and similar reforms are being contemplated in several other states. To evaluate the pros and cons of medical marijuana reform it is important to know its impact on the well being of society as a whole. In the present thesis I hypothesize that medical marijuana legalization has lead to lower violence rates, based on a review of prior research suggesting that stricter illicit drug law enforcement may increase violence rates, and evaluate this hypothesis empirically. The impact of legalization on various city level violence rates as well as heroin/cocaine distribution and possession is estimated using a robust fixed effects framework. Additionally the synthetic control group approach is used to estimate the impact on the state homicide rate. The data employed is a panel of 540 US cities divided over 12 legalizing and 34 non-legalizing states (1980-2010). The results suggest that medical marijuana legalization is on average associated with a drastic decrease in drug-related and alcohol induced homicides and a large contraction of the heroin/cocaine market. The estimated impacts on these variables are larger in states with lenient Medical marijuana legislation and high user rates, and significantly negative and very large estimated impacts are found in some of these states also for the overall homicide rate. The impact of legalization is null or in some cases even positive in states with stricter medical marijuana laws. These results indicate that medical marijuana reform can produce substantial positive externalities to violence rates and hard drug use, but suggest that these will not be realized if the boundaries of medical marijuana legalization are too clearly defined and/or legalization is coupled with stricter enforcement of hard drug laws.