Note that allowing a murderer to, well, murder, improves his economic welfare - it increases his economic utility. Yet murdering is a net negative in the ethicist's utility function.
Economics makes normative claims because economists typically have some relatively uncontroversial normative assumptions - like maximizing economic welfare is a good thing. This is by and large true, but see my counter example above. Also, economists aren't trying to prove that the values they assume are the correct ones. They are assuming certain ethical values and proposing policies that maximize these values.
The two types of utility functions look very similar - the math is the same, both describe goal seeking behavior, etc., but the difference is the preference sets that each describe. Murder can increase utility in the economist's utility function, but not in the ethicist's (under normal circumstances).
Murder can increase utility in the economist's utility function
That is really immaterial though and computationally moot. Ok so his "utility function" is negative. Is that it, is that the difference? Besides, I would argue that reevaluating it on those terms does a poor job of actually describing motivation in a coherent set.
Yet murdering is a net negative in the ethicist's utility function.
It isn't in the economists? These things aren't neutral.
The broader aspect that economists seek is normative. You said it yourself in the economists a...
Recently I argued that the economist's utility function and the ethicist's utility function are not the same. The nutshell argument is that they are created for different purposes - one is an attempt to describe the actions we actually take and the other is an attempt to summarize our true values (i.e., what we should do). I just ran across a somewhat older post over at Black Belt Bayesian arguing this very point. Excerpt: