incentivesI am interested in creating a complex system of incentives. I do not want to get into the details of this system, nor my context, as I want this question to be interesting for as many experts as possible.
My question is, what methods I could use to estimate the effects of these incentives. My most reliable method is to ask myself how I would react to the various incentives. This has gotten me quite far, but I want to get closer before I start testing this in real life, as testing this would cost a large sum of money. I could maybe do a scaled down test, but I am hoping that there could be some cheaper methods I could use first.
I have tried asking people how they would react, and I very often find that they cannot really put themselves in the situation and predict their reactions, they often give me answers, though, but I do not trust that they would act according to their answers. Somehow, their answers seem too simple. I am certainly also interested in methods on how to ask these questions if there are any.
Without more detail, it is a bit difficult to answer. But, the literature does provide some guidance.
For example, economists often try to elicit people’s willingness to pay for some good (often an environmental amenity). Interestingly, we often find that their WTP for the amenity is less than what they are willing to accept to lose it (WTA). We would expect the two measures to be about the same (see Willig’s famous 1974 AER paper).
Indeed, WTA for a loss is almost always higher than WTP for the corresponding gain (see Table 1 in Brown and Gregory (1999) and, more famously, Hanemann (1991)). In general, there is an income effect with WTP questions that is not well-accounted for, while WTA answers often involve some risk aversion (that is, upsetting the status-quo). There is a large literature here–in addition to what I’ve listed, see Howowicz and McConnell.
But, the point is that how one words the question can be very important. If the incentives are posed as a loss (e.g. payments NOT to do something your subjects already do) instead of as a gain (payments to do something that offsets a current behavior), you may get radically different results for people’s estimates of what they might do given your proposed policy. You might try finding a WTP (or, as applicable, WTA) equivalent to your incentives and see if you get different answers.
There is very little innovation. Most of what we have is improvements or changes on existing products and processes. You can get a decent projection by looking at the effects of similar incentives in other markets. One thing to note here is that you must take into account the marketing of the incentive. Some incentives are more successful because of their marketing than the actual draw of the incentive.
All content is licensed under CC BY-SA 3.0.