I'm trying to derive the utility possibility frontier of a economy whose consumption contract curve is $$y_A = \frac {y} {x} x_A$$ and $$y_B = \frac {y} {x} x_B$$where $x_A + x_B = x$ and $y_A + y_B= y$.

$$U_i = \sqrt {x_iy_i}$$

While deriving the UPF. they have added up $U_A$ and $U_B$, and I don't understand why. Why do we add up utilities while deriving UPF?

(I gave all the details of the question because I'm not sure if they added the utilities only because of the peculiarities of the question or if its an algorithm. )

  • $\begingroup$ I don't understand either. Who are they, and why do they provide the contract curve when you can easily derive it from the utility functions? $\endgroup$
    – Giskard
    Jun 18, 2015 at 18:38
  • $\begingroup$ This was a really long part by part question in my exam. How do you calculate UPF from the utility functions? $\endgroup$
    – dexter
    Jun 18, 2015 at 18:41
  • 1
    $\begingroup$ That is one of the possible definitions of social welfare function. It is called "Benthamite social welfare function". $\endgroup$
    – AnilB
    Jul 19, 2015 at 19:03

2 Answers 2


Given that $u_A(x_A, y_A) = \sqrt{x_Ay_A}$, $u_B(x_B, y_B) = \sqrt{x_By_B}$ are the utility functions of A and B, and total endowment of X and Y in this pure exchange economy is $\omega_X$ and $\omega_Y$, utility possibility frontier (UPF) is the set of all utility pairs $(\mu_A, \mu_B)$ such that the there exist a Pareto optimal allocation $((x_A, y_A), (x_B, y_B))$ with the property: $\mu_A = u_A(x_A, y_A)$ and $\mu_B = u_B(x_B, y_B)$. So to find UPF, we will use the following conditions:

  • Optimality:

$\displaystyle \frac{y_A}{x_A} =\frac{y_B}{x_B}$

  • Feasibility:

$x_A + x_B = \omega_X$ and $y_A + y_B = \omega_Y$

  • Utility:

$\mu_A = \sqrt{x_Ay_A}$ and $\mu_B = \sqrt{x_By_B}$

Solving the above system of equations we get the UPF as: $\mu_A + \mu_B = \sqrt{\omega_X\omega_Y}$.


In general, one could use the following procedure to find UPF (in a two-person case):

  1. fix one person's utility at some arbitrary level, say $U_A=\bar u$
  2. maximize the other person's (i.e. person $B$) utility subject to the resource constraints, as well as $U_A=\bar u$
  3. $B$'s maximum utility, call it $U_B^*$, found in step 2 should be a function of $\bar u$, and other parameters of the problem: $U_B^*(\bar u,\dots)$
  4. lastly, you can vary the value of $\bar u$ (in the range of acceptable values for $\bar u$) to trace out the UPF.

There's no a priori reason that points on the UPF should equal to the sum of individual utilities, since utility is ordinal (unless assumed otherwise). It may be that the square root utility chosen for this question is what's causing the result. But I haven't done the math to verify it.

  • 2
    $\begingroup$ I think it is worthwhile to mention that any point on the UPF will also be Pareto-optimal (otherwise you could increase someone's utility), so the contract curves are still useful. And it seems that since both utility functions are homogeneous functions of degree 1, indeed all Pareto-optimal points will be on the UPF. $\endgroup$
    – Giskard
    Jun 18, 2015 at 21:16

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.