utility-functionIn the usual models, every person has a utility function and tries to maximize it.
Do any models account for the existence of different long- and short-term utility functions? For example, if I get drunk today, I’ll increase my short-term utility function (feel good), but reduce my longer-term utility function (hangover next morning).
Or is the single utility function a weighted average of functions for different “terms”?
It might help to look for “time inconsistent preferences”. Covers topics like addiction, procrastination. On the latter, the classic reference is “Procrastination and Obedience” by Akerlof in the American Economic Review.
Discounted utility is the usual way of modeling utility that is summed over time. For example, assume tomorrow’s utility should only be considered 90% as important as today’s utility (discount factor β=0.9), and the next day’s utility should be considered 90% x 90% = 81% as important as today’s utility, so that the sum over time is:
U = ∑ β^t u(t)
For a more complex model, you could replace β^t by a more precise function reflecting people’s actual preferences today for particular outcomes in the future. On the other hand, many economic models take a simpler approach by using only two time periods, ie U = u(t=0)+βu(t=1).
It is addiction centered (the utility of a specific good at time t, depends on the comsumption of this good at previous times) but here is a reference : http://www.drugtext.org/Economics/a-theory-of-rational-addiction.html
Even if this is not exactly the kind of model you are asking for, something similar should to it to model long/short term utility.
All content is licensed under CC BY-SA 3.0.