Paying for Integers

Abstract

Leveraging the unique context of New York City taxi rides, we estimate that passengers are 25 to 30 percentage points more likely to tip a suggested amount when presented an integer tip suggestion. By combining our empirical results with a behavioral model, we find that our results are consistent with integers acting as focal points for giving and not menu opt-out costs or left-digit bias in warm-glow. Focusing on a 2012 rate fare change that increased the probability of integer tip suggestions by 700%, our estimates imply a transfer from riders to drivers of 2.9 million dollars in 2013.

Keywords: Integers, Defaults, Tips, Suggestions, Taxi