The Pokemon GO Death Count

Have you ever wondered what economists do fun in Indiana? You might think that they could benefit from the global supply of digital goods like everyone else without having to worry about local effects. But you would be wrong. They are always thinking local.

For Purdue University economists, Mara Faccio and John McConnell, they observed the many playing Pokemon Go obsessively and thought, “I wonder if there could be a good instrument in that game?” In particular, how many people has this game likely killed in the first 148 days of its release?

It turns out they have an answer in this paper. Well, at least on the value of lives lost, injuries and damage done, of between $5.2 million and $25.5 million in Tippecanoe County, Indiana, alone. If that translates up US wide that is between $2 billion and $7.3 billion making it one of the largest negative externalities ever seen. My guess is that no Kaldor-Hicks compensation is going to correct that one.

Let’s step back a moment in case you are perplexed by this digital disaster, and explain how this game works. Basically, you go around picking up various weapons and stuff from things called Pokestops which are places that you can only see on the Pokemon Go game if you are right next to them. For kids, that meant you had to grab your iPhone and wander around the neighbourhood, inadvertently getting a ton of exercise, in order to play this game. For adults, you could use a motor vehicle. So that meant, the authors hypothesised, that those adults were gaming while driving and that this was likely to be a tad unsafe.

Armed with that hypothesis, an ordinary person might delve into the Tippecanoe County police reports of accidences and see if you could see a blip after July 6, 2016, when Pokemon Go was released to the public. And you would find that there was an increase in crashes near PokeStops after that date beyond what was occuring elsewhere. The data showed that of the 286 additional crashes that occured in the county after July 6 compared with the same period of time before, 134 were near PokeStops. These PokeStops were common but in a spatial sense not that common, so that is a big number. In terms of lives lost, this translated to 2 people.

Of course, these weren’t ordinary people. They knew that it wasn’t necessary people playing the game while driving that caused the accidents but instead people parking hoping out of their cars and running around in increased numbers that meant they might just be in the wrong place at the wrong time more often. Yes, Pokemon Go was to blame but it was as much to blame as a football match around Purdue stadium before and after game time. So how to work out if people were really idiots or not?

It turns out that the designers of Pokemon Go had themselves anticipated the potential idiocy of their customers and put in controls that prevented the game being used by driving. Thankfully — for the economist’s identification strategy that is — they didn’t do this for all PokeStops but for super-stops called Gyms where all the real action took place. If people were doing to park and get out of their cars and play the game and then get run over, this was where they were going to do it. For the economists, this meant they could see if this made a difference.

And it did. And what is more, despite the intensity of game play near Gyms, the number of crashes was significantly higher near PokeStops rather than Gyms! So it was morons playing while driving that was responsible for all of this.

So what should we do about about this? The authors offer the following:

Using these numbers as a basis for policy recommendations is tempting. The immediate impulse is to recommend further bans on the use of mobile phones while driving. The cautions associated with such a recommendation are three-fold. First, in response to concerns about potential crashes due to users playing the game while driving, in an update of the game, Niantic added a pop up message saying “You’re going too fast! Pokémon Go should not be played while driving.” This message pops up when the game detects the player to be in a rapidly moving vehicle. A message further asks the player to confirm that they are a passenger. Thus, in an effort at self-regulation, the game cautions users against playing the game while driving. Second, policy recommendations require a consideration of both costs and benefits and we have made no attempt to calculate the economic benefits of using mobile devices while driving. We acknowledge, though, that identifying any economic benefits of playing Pokémon GO while driving stretches our imaginations. Third, as concluded by Abouk and Adams (2013), the Highway Loss Data Institute (2010), and Lim and Chi (2013), the effect of, if any, bans on the usage of mobile phones (including texting) while driving appears to be short-lived or limited to specific subsets of drivers. [emphasis added]

In other words, their recommendation is: “just stop it! Stop it! You idiots! It is just a f**king game!” I believe this is a sensible balanced approach in this instance.