We estimate the causal effect of partisan media on voters’ choice by exploiting a natural experiment in which Sky News Australia – a conservative 24-hour news channel – became freely available to 8 million people. Using comprehensive national polling-station-level electoral data and variation in broadcast signal strength, we implement a difference-indifferences design comparing voting patterns before and after its 2018 free-to-air launch. We find that exposure to Sky News increased the conservative party's lead by 1.5 percentage points in the 2019 election. Compulsory voting allows us to identify the effect of partisan news in a context unaffected by voter turnout.