In the 2016 general election, 21 counties in Utah administered voting entirely by mail, while eight counties administered traditional polling place-based voting. Using vote propensity scores to control for voters’ pre-existing differences in likelihood to vote, we show that the advent of vote-by-mail increased turnout by 5-7 points. Low-propensity voters, including young voters, showed the greatest increase in turnout in vote-by-mail counties relative to their counterparts in non-vote-by-mail counties. We find similar results by zooming in on specific geographic areas within Utah where vote-by-mail counties are bordered by non-vote-by-mail counties, with magnitudes of 4-9 points of increased turnout. In one mountaintop community that happened to be bisected by a county line, the increase in turnout due to vote-by-mail may have been as high as 12.5 points.