Proximity Bias

Location, Location, Location

Here’s the start of a new academic paper, which I hope to post soon. I am finding the topic unusually interesting.

The policy upshot is that proximity really, really matters for so many things: eg, medical care, public assistance, voting, vaccination, and court attendance. If, for example, a hospital is not proximate, people might die. If a court is not proximate, people might not show up. A lack of proximate X, where X matters, can be devastating.

Ready? Here goes (warning: academic prose, and footnotes, ahead):

“And if you can’t be with the one you love, honey/Love the one you’re with”

- Stephen Stills

Proximity matters. For grade school students, random seating assignments have a significant effect on friendships: If one student is seated next to another, the likelihood of a friendship between them two increases significantly.[1] Physical proximity increases the likelihood that people will become coauthors.[2]

More broadly, proximity is an important and insufficiently studied aspect of choice architecture,[3] and hence has relevance for law and policy. Greater proximity, or less proximity, can be a nudge. For many purposes, closer is thought to be better.[4] Might that be a heuristic? Might it lead to a bias? Might any such bias reduce welfare?

Suppose that you need to see a doctor for a serious medical problem, and that you have two options. A superb doctor, Dr. Jones, is forty-five minutes away by car. A good doctor, Dr. Smith, is ten minutes away by car. You might choose Dr. Smith because of his proximity. You might be travel-averse. If the difference between the two doctors does not much matter to your health, your choice seems fully rational; the proximate choice might be the welfare-promoting choice. If the difference between the two doctors greatly matters to your health, and if you are overweighting proximity in terms of your own welfare, we can describe your choice as a product of proximity bias.

We might also speculate that in hard cases, or even not-so-easy ones, people use the proximity heuristic: The more proximate choice is deemed to be the better choice. Proximity might operate through a process of attribute substitution[5]: Instead of asking what is best (a not-so-easy question), people ask what is proximate (a much easier question). Of course the proximity heuristic might have ecological rationality. It might be entirely sensible, in many contexts, to think: Choose the closer.

For restaurants, shopping, and gas stations, that might be the best way to decide; it might reduce the costs of decision and the costs of error. If people are uncertain as between Option A and Option B, selection of the closer might be optimizing. Use of the proximity heuristic might increase experienced well-being, if people genuinely dislike traveling longer distances. Use of the proximity heuristic might also increase anticipatory well-being, if people would dread traveling longer distances, even if they would not actually dislike travel, in the event, as much as they expect.[6]

It would be easy to imagine a series of tests for whether people care mostly or entirely about proximity as such, or whether they are using proximity as a heuristic in difficult cases. A proximity heuristic will sometimes produce proximity bias. Even if no heuristic is at work, people might show proximity bias by overweighting proximity relative to other factors. It is also true that by contrast, people might show proximity neglect: They might choose the less proximate alternative and give too little attention to the negative welfare consequences of having to get there.

To know whether proximity bias (my focus here) or proximity neglect is at work, it is necessary to know more about the situation and about people’s preferences. Suppose, for example, that in the case given above, the seventy additional minutes in the car would be bothersome and boring, but that they would not produce massive welfare losses. Alternatively, suppose that those seventy additional minutes would be extremely costly, perhaps because they would interfere with work obligations or perhaps because they would interfere with family responsibilities.

Preferences are also highly variable here. Some people much dislike travel, even for relatively short distances; some people do not much dislike travel, especially over short distances, and might even like it. There might or might not be nonlinearities. People might show proximate bias as between a local option and a close-but-not-so-local option, but not between a quite distant option and an immensely distant option, or vice-versa.

People might be indifferent between traveling fifteen minutes and traveling forty-five minutes, but care greatly as between travelling an hour and travelling three hours. They might not much care as between traveling eight hours and eleven hours, or they might care greatly about that. We can imagine a wide range of possibilities on this count.

To know whether there is proximity bias (or proximity neglect), we need to know more about what would be lost and what would be gained by choosing one or another doctor. Outsiders will usually know far less than choosers do. What outsiders see as a bias might be nothing of the kind. Experiments would be necessary to test whether choosers are rationally taking account of proximity, or whether we are genuinely dealing with proximity bias or proximity neglect. When proximity bias (or proximity neglect) is at work, people are likely to be displaying a hedonic forecasting error[7]; they expect to suffer higher (or lower) welfare losses than they actually do.

When it exists, proximity bias should be seen as the spatial equivalent of present bias.[8] Indeed, present bias is a kind of proximity bias, where proximity is defined by reference to time rather space. In some cases, proximity bias is literally present bias: People might care about proximity only because they care about short-term welfare costs. Caring about those costs is not, of course, a bias. We are dealing with a bias if people give a great deal of weight to the short-term welfare costs of getting from one place to another, and do not give much weight to the long-term welfare benefits of choosing the more distant option.

---

[1] See Julia Rohrer et al., Proximity Can Induce Diverse Friendships, PloS One (2021), available at https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0255097

[2] See Arianna Salazar Miranda and Matthew Claudel, Spatial Proximity Matters, PloS One (2021), available at https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0259965

[3] See Richard H. Thaler and Cass R. Sunstein, Nudge: The Final Edition (2021).

[4] There is an evident relationship between proximity bias and the endowment effect; the endowment effect might even be seen as a special case, though it is of course distinct. See Richard H. Thaler, Quasi-Rational Economics 7-10 (1991).

[5] See Daniel Kahneman and Shane Frederick, Representativeness Revisited: Attribute Substitution in Intuitive Judgement, in Heuristics and Biases (Thomas Gilovich et al. 2012).

[6] There is a plausible evolutionary explanation for proximity bias: In hunter-gatherer societies, closer is probably less dangerous. If you venture far, you might be killed.

[7] See generally Timothy D. Wilson & Daniel T. Gilbert, Affective Forecasting, in 35 Advances in Experimental Social Psychology 345 (Mark P. Zanna ed., 2003).

[8] Ted O’Donoghue and Matthew Rabin, Present Bias: Lessons Learned and to be Learned, 105 Am Econ Rev 273 (2015).