With a bit of analytical effort, you can crunch the numbers from 100 million New York City taxi rides to pinpoint how passengers tip their drivers. You can even figure out how many miserly yellow cab riders leave no tip at all. But sometimes the data doesn’t quite make sense no matter how you scrutinize it.
Here’s the chart showing how much people tip based on the fare:
(In case you’re wondering if time of day matters here: No, it does not. This same pattern occurs at all times of the day.)
A few distinct trends emerge, and all of them are baffling:
• Tip percentages tend to go down as the fare increases …
• … and yet tip percentages increase as fares move from $20 to $30.
• But for larger fares of $40 or more, there is a very clear trend that fares ending in the digits 0 or 5 have a much smaller average tip than fares with similar values.
What are the reasons for this? We surveyed a range economists and behavioral experts for their thoughts, including: Dan Hamermesh from University of Texas, Richard Thaler from University of Chicago, Andrew Lo from Massachusetts Institute of Technology, Dan Ariely from Duke, Cass Sunstein from Harvard, Bryan Caplan from George Mason, and pricing consultant Rafi Mohammed. Here are some theories to explain this taxi tipping mystery:
1. As far as the overall trend for smaller tip percentages, we can all agree this makes sense. For very cheap fares under $15, people just give a couple of bucks without adjusting much for the increase in fare. The suggestive power of the 20 percent default tip on the payment screens in New York City taxis doesn’t hold much sway. “As the fare goes up,” Thaler said, “the impetus to deviate from the 20 percent default goes up. Twenty dollars is a big tip. People may think a $10 tip is fine even for a $75 fare.”
Ariely offered a two-model theory to explain rider behavior—one based on percentages and another based on a maximum amount—with passengers switching between the two depending on which costs less money. If that were true, he says, at some level the tip amount will become independent of the fare. Rafi Mohammed agrees, saying that above $50 people may feel that a $10 tip is good enough regardless of the final fare.
Another theory—not from an economist, but from my boss—is that fares above $50 tend to be airport trips with a large proportion of foreign travelers who might come from countries where tipping isn’t customary. These relatively pricey fares from the hotels of Manhattan might therefore be matched with stingy tips.
2. The rising trend for tip percentages in the $20-$30 range has no easy explanation. In fact, the theories from No. 1 above would now fail to explain this phenomenon. Some hypothesized that fares near $30 might involve trips between boroughs, and so passengers might feel they’re getting more service from the driver. Another theory is simply that people are bad at math and can’t figure out what the right tip is at $20 vs. $30. This doesn’t explain why bad math is particularly evident between $20 and $30.
Another thought is that business travel and corporate card use might skew higher tips in that range, but again that doesn’t explain why it stops exactly at $30. The intriguing part of the $20-$30 curve is how smooth it is. The tip percentages go up for every single fare along the way, without any sudden changes. Higher prices and higher tip percentages, but only for this one 10-dollar range.
None of our economists had a good reason for this. “This is really weird,” one panelist concluded. “I really have no idea.” Another said he would need to devote “a lot of hours to the task of studying the data in great detail” before he could speculate. Another suggested that people dislike fractional dollar amounts and have a general predilection toward whole numbers. None of these theories, however, can explain why tips on fares above $30 go back up for a big chunk of the curve—until we get to the next mystery of huge drops at specific numbers.
3. The most bizarre behavior is around the fares ending in the numbers 0 and 5. The only explanations people can come up with for this are vague “round number” discussion points—people just like to leave a tip or produce a final bill that ends in a round number. But this is not what actually happens.
When a fare hits a new round number, passengers are twice as likely to give no tip at all. One idea is people are so upset that the first number jumped up (moving from the $59 to the seemingly more painful threshold of $60), so they feel compelled to skip the tip altogether. That behavior disappears at $61 again and people resume tipping normally. But we also see the same behavior at $65 and $75, so that still doesn’t explain it.
The trick for taxi drivers is to not let the fare hit that round number. The average tip at $60 is $8.82, but the average tip at $59 is $10.33. So in fact, going from $59 to $60 resulted in a loss of $1.50 in tip—more than the difference in the fare. The driver would have been better off just stopping the meter a minute earlier, making up the difference with a bigger tip. You see this pattern every time. The tip on a $70 ride is more than two dollars less than the tips at $69 or $71.
The pattern gets so skewed that the average tip on a $64 ride ($11.41) is higher than the average tip on a $100 ride ($11.21).