If $u(x)=\frac{x^{1-\mu}}{1-\mu}$, then the elasticity defined as $\frac{\partial u}{\partial x}\frac{x}{u}=1-\mu$.

However, when we define $u(x)=\frac{x^{1-\mu}-1}{1-\mu}$ (usual def. for isoelastic utility), we have $\frac{\partial u}{\partial x}\frac{x}{u}=(1-\mu)\frac{x^{1-\mu}}{x^{1-\mu}-1}$. And this doesn't seem to give a constant elasticity. So, why do we call this utility isoelastic?

Any help would be appreciated.


Iso-elastic utility is defined as a function $U(w)$ where for all $k \gt 0$ you have $$U(kw)=f(k)U(w)+g(k)$$ for some functions $f(k)$ and $g(k)$ independent of $w$, i.e. an affine transformation of the original utility function

This is iso- because this kind of utility function leads to identical (or at least proportionate) decisions: it means that the optimal risk/reward behaviour at given level of wealth $w$ is also optimal for another level of wealth $kw$ if you multiply all the original investments by $k$. For example, if at some wealth level you would put a third of it into a particular risky investment and two-thirds into a particular safe investment, then with an iso-elastic utility function you would do the same with at any wealth level

Both your formulations are iso-elastic in this sense, since they only differ by a constant $\frac{1}{1-\mu}$ and so lead to identical distribution decisions. The difference between them is that the first has $u(0)=0$ and $u(1)=\frac{1}{1-\mu}$ while the second has $u(1)=0$, and since that difference has no real-life implications it might be unreasonable to claim the two utility functions had different properties

Those who use the second formulation often associate it with a logarithmic utility function as the limit function as $\mu \to 1$. This limit cannot be done without subtracting the $\frac{1}{1-\mu}$ (no longer a constant, since $\mu$ is changing), so they take the second formulation for arithmetical convenience

  • 1
    $\begingroup$ "This is iso- because..." yes but why is it elastic? By the same reasoning I could also call it isocost utility. $\endgroup$ – Giskard Dec 21 '16 at 7:13
  • $\begingroup$ You regard constant elasticity as something which, for example, leads to a constant proportionate change in demand for a given proportionate income change no matter what income you start from, or a constant proportionate change in supply for a given proportionate price change no matter what price you start from. Both are observable behaviours if true. These isoelastic utility functions lead to a constant proportionate distribution of wealth as wealth changes no matter what level of wealth you start from, again an observable behaviour if true. But I did not invent the name $\endgroup$ – Henry Dec 21 '16 at 9:34
  • 1
    $\begingroup$ @denesp look at Henry's comment. He forgot to tag you, I think. $\endgroup$ – An old man in the sea. Dec 21 '16 at 20:54
  • $\begingroup$ @Anoldmaninthesea. Thank you. I saw it, I think the system sends a notification to the previous commenter. $\endgroup$ – Giskard Dec 21 '16 at 23:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.