I'm working on a consumer choice model involving a consumer with one good and a numeraire. In this model, the price of the numeraire is assumed to be one. My objective is to identify the utility function that results in the following optimal allocation:

$$m^* = A \left(\frac{p}{\omega}\right)^\alpha$$

Here, $\omega$ represents the consumer's income, and $A$ and $\alpha$ are constants and $-1 < \alpha < 0$. The consumer's goal is to maximize their utility subject to the budget constraint:

$$\max u(m,x) \quad \text{s.t.} \quad \omega = x + p \cdot m$$

The first-order conditions (FOCs) for this optimization problem are:

$$u_x = \lambda$$

$$u_m = \lambda \cdot p$$

$$x^* = \omega - m^* \cdot p$$

Given this setup, I'm seeking assistance in determining the utility function $u(m,x)$ that leads to the optimal allocation $m^{*}$ as specified above.

Here's what I've tried so far:

  • I've attempted to reverse-engineer the utility function starting from the FOCs but got stuck at integrating the conditions to find a cohesive function.
  • I also looked into some standard utility functions (like Cobb-Douglas), but I'm not sure how to adapt them to fit this particular problem.

Thank you in advance for your help!

  • $\begingroup$ Hi! Is $\alpha$ positive or negative? Seems like unless $-1<\alpha<0$, you are going to run into trouble with extreme values of $p$. $\endgroup$
    – Giskard
    Commented Mar 23 at 18:46
  • $\begingroup$ Also, do you have reason to believe that such a utility function exists, or do you just assume it? $\endgroup$
    – Giskard
    Commented Mar 23 at 18:49
  • $\begingroup$ Hi, thank you for your comment. I have no proof that such a utility function exists. I have also added that $-1 < \alpha < 0$. $\endgroup$ Commented Mar 24 at 20:23

1 Answer 1


I think your formula is still too general, so what you want will not be possible. Given $-1 < \alpha < 0$ and $$ m^* = A \left(\frac{p}{\omega}\right)^\alpha, $$ we have $$ \frac{m^*p}{w} = A \left(\frac{p}{\omega}\right)^{1+\alpha}. $$ The left hand side is the ratio of the consumer's money spent on this good $m$. But $$ \lim_{\omega \to 0} A \left(\frac{p}{\omega}\right)^{1+\alpha} = \infty, $$ so if you make the income of this consumer small enough, their optimal good allocation implies that they spend more than 100% of their income on this good $m$, which should not be possible. If $1+\alpha$ were negative, you would run into a similar problem by increasing income.

In the special case of $1+\alpha = 0$, a solution actually exists; Cobb-Douglas: $$ U(m,x) = m^Ax^{1-A} $$ yields this exact demand function for $m$ with $\alpha = -1$.

A general lesson is that if demand is proportional to a power of income, problems arise if it is not specified that the formula only works in a certain range/interior cases, unless demand is linear in income (homothetic utility functions).

  • $\begingroup$ Thank you for your response. Your explanation is clear, especially since the empirical relationship has been identified within a specific income range. I plan to apply this utility function for numerical applications within the same income bracket. In this context, could you advise if it's possible to identify a utility function that results in a constant elasticity of the price-to-income ratio? $\endgroup$ Commented Mar 27 at 11:03
  • $\begingroup$ What you describe is not entirely clear to me, but I never did any empirical work with utility functions. I recommend posting a new question with some additional details on parameters such as $\alpha$ and the desired income range. $\endgroup$
    – Giskard
    Commented Mar 27 at 13:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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