Drug patents are distorted. Unlike most other inventors, drug inventors must complete years of testing to the government’s specifications and seek government approval to commercialize their inventions. All the while, the patent term runs. When a drug inventor finally launches a medicine that embodies the invention, only a fraction of the patent life remains. And yet, conventional wisdom holds — and empirical studies show — that patent life is essential to innovation in the pharmaceutical industry, perhaps more so than any other inventive industry. Congress tried to do something about this in 1984, authorizing the Patent and Trademark Office to “restore” a portion of the patent lost to premarket testing. PTO doesn’t restore all of the lost time, though, which raises the question whether the U.S. legal system may steer researchers away from drugs that take a long time to develop. This Article focuses on that question. It examines every grant of patent term restoration for a new drug or biologic from the scheme’s 1984 enactment to April 1, 2018. And it fills a conspicuous gap in the literature: few scholars have considered patent term restoration from an empirical perspective, none has used a dataset of this size and scope, and none has addressed the questions this Article addresses. Two significant conclusions stand out. First, longer clinical programs lead to shorter effective patent life, even after PTO has granted patent term restoration. The results are strongly statistically significant and contribute to a growing body of literature raising the alarm that the U.S. legal system may be systematically skewing drug research incentives away from the harder problems — such as a cure for Alzheimer’s Disease and interventions at the early stages of cancers. Second, Congress decided to allow drug companies to apply patent term restoration to continuation patents, specifically because this would increase the chances of reaching 14 years of effective patent life. Ten years later Congress changed the way patent terms are calculated without considering the effect on patent term restoration. Selecting a continuation patent no longer has the same effect. Today a drug company is most likely to achieve the 14 years of effective patent life by securing a new, original patent that issues late in clinical trials. Policymakers and scholars complain when companies secure these later-expiring patents, but the findings in this Article suggest those patents may be necessary to accomplish what Congress intended in 1984.