Diminishing marginal utility of income and “big ticket” items

One of the main justifications for redistribution policies is “diminishing marginal utility”. We have already discussed that this doesn’t really make sense as we can’t compare peoples “utility”. For example, people that receive higher utility from consumption will work more and thereby will earn a higher income – from here we cannot tell whether the change in welfare from taking a dollar off them and giving it to someone who earns less will be positive or negative.

Furthermore we have the fact that two people with the same lifetime income level the person with a more variable annual income will be taxed more than people who do not have a variable income – implying that DMU does not work as a defense here!

However, there is a further complication to the DMU story. Even if everyone has the same “utility function” we cannot necessarily assume that marginal utility will be diminishing in income.

Why? Well because of the cost of “big ticket” items. If we are in a situation where a market to “hire” big ticket items does not exist (or suffers from some type of market failure) it may be the case that a “high income person” would receive a greater amount of utility per dollar than a low income person if they can afford to buy the big ticket item. This would occur if the purchase of the big ticket item created a huge increase in the persons happiness.

If redistribution implies that neither agent can buy the big ticket item, then the total welfare in society will be lower, following this redistribution. The marginal utility curve in this case would look something like this (apologises for the terrible graph – my photoshop is messing up 🙂 ):

  • CPW

    You’re just arguing that the individual marginal utility curve might be kinked in various ranges, aren’t you? I think this is plausible enough, but I don’t see any reason to assume that the kinks are located in any income range in particular – so I don’t see this as being a point for or against DMU based transfers. But keep those technical objections coming…

  • “I think this is plausible enough, but I don’t see any reason to assume that the kinks are located in any income range in particular – so I don’t see this as being a point for or against DMU based transfers”

    Having the appropriate distribution for it to matter is as realistic as assuming that everyone has the same Utility function. As people constantly assume that to promote transfers why can’t I assume that the cost of big ticket items fall in the appropriate range 🙂