A couple of weeks ago I asked why people might avoid finding out how worthy a charity is. People know that they’ll give to worthy charities and yet they shy away from finding out that a charity is worthy so that they avoid giving. Why might this be?
There aren’t many situations in which rational agents choose to avoid obtaining information costlessly. The times when they might are when it gains them a strategic advantage. However, you’re not playing a game against anyone else when you pass a collector on the street. The only person that you could be said to be playing against is another temporal self. Carillo and Mariotti’s paper on strategic ignorance explains how a person might choose to avoid information in order to gain an advantage over their future self. If you’re keen to get quasi-technical about it, read on.
The idea behind strategic ignorance is that an agent who discounts in a hyperbolic fashion will sometimes not want her future self to be aware of some information. The reason is that the future self would, if she had the information, act in a manner contrary to the current self’s best interests.
In the donation example there are warm fuzzies to be gained from stopping to learn what the collection is for and from anticipating that you might do a good deed. If you do learn that the collection is for a good cause then you have to incur the cost of donating. Imagine an agent, who discounts the next period by , walking along the street before seeing the collector. She discounts both the benefit and the cost by and would choose not to give to a collector. However, as she slows for the collector, she starts to experience the undiscounted benefit and only discounts the cost by . Now the cost is smaller relative to the benefit and she might decide to donate if the charity is good enough. For some probability that the charity is worthy, the past self wouldn’t have donated, but the current self would donate.
Now comes the key part. Suppose that people tend to believe that charities aren’t usually particularly worthy but they suspect that, if they were to check, the charity would have a higher probability of being worthy than their prior estimate. Now if they slow for the collector they might find out that charities are fairly worthy after all and choose to stop and donate, which would be bad from their current point of view. Then they’ll be better off if they don’t find out the true likelihood of the charity being worthy, because that would induce their future self to donate, which their current self doesn’t want. Hence, a person might donate when faced with a worthy charity, yet choose not to disabuse themselves of the notion that most charities aren’t worthy.