When the rules of a game are changed the results can be hard to predict. People may not react to in the way perfectly rational agents would be predicted to, and people’s reaction can evolve over time as they figure out strategies and respond to the behavior of others. Predicting perfectly may be difficult, but often policies are implemented without thinking about incentives at all, beyond the obvious intended effect. If you get in the habit of thinking just a bit about how strategic players may try to exploit a possible rule change, you can be way ahead of the average person.

The Wikipedia article on **perverse incentives** gives a list of interesting historical examples where new laws or policies ended up having the opposite of the desired effect, because they incentivized people to take counterproductive actions. Some of these plans backfired for surprising reasons, but most of their failures (at least in hindsight) seem like could have been predicted in advance by just thinking one step ahead about incentives.

Here are some examples I think are particularly glaring from the article:

• In the ‘Great Hanoi Rat Massacre’ of 1902, in order to curb the city’s rat population the colonial government offered a bounty for each rat killed. However, people only needed to bring in rat tails to collect the bounty, and so rat catchers began cutting the tails off of captured rats and then releasing them back into the sewers to produce more rats.
• In 2021 the US Congress enacted strict regulations to prevent sesame from cross-contaminating foods that did not contain it as a listed ingredient. Many companies then found it simpler and cheaper to just add sesame to the list of ingredients, instead of comply with the regulation.
• Under construction

How could thinking one step ahead have avoided these mistakes? I’ll cover each example in turn. Of course hindsight bias is strong, but hopefully these examples are clear enough that they are still convincing.

• If your true goal is to for people to perform task A (killing rats) but you reward them for task B (bringing rat tails), you must be very certain that task B can only be done by also accomplishing task A. Perhaps people in the government did not realize that rats could survive with their tails removed, if so they should have double checked this.
• The goal here, I assume, was to have more safe food options available for people allergic to sesame. They may have succeeded in a sense, because the regulation essentially made it so if a food does not have sesame listed as an ingredient then it must have been carefully protected from cross-contamination. Market forces will then cause some food companies to save money and list sesame as an ingredient, and some will spend the money to capture the market for sesame-allergic consumers. But I imagine this was not exactly the outcome to lawmakers were intending. Why not require companies to actually have sesame as an ingredient to be able to list it? Let’s think one step ahead again. What will frugal companies probably do in response? My guess is just add a tiny amount of sesame to the food to get around the rule. In my opinion, the regulation that would make sense is to have different labels based on the amount of sesame contamination: to qualify as ‘sesame free’, it must have been made with the strict rules, and another label is ‘possible sesame contamination’.

Back to Pages