The first theorem of welfare economics states that "Every Walrasian Equilibrium (WE) is Pareto Efficient (PE)" - Microeconomic Theory Nicholson, Synder 11th ed. p477. But PE is quite a weak condition. Is WE a stronger condition or is does WE just mean any PE allocation?

I was specifically wondering; suppose we have an exchange economy with agents $N$, and there is a unique competitive equilibrium (WE) solution, $q^*$. Does this also maximize aggregate social welfare amongst possibly many PE outcomes? Where aggregate social welfare is defined as:

$$ SW(q^*) = \sum_i^N U_i(q_i^*) $$

To ask this question only makes sense if WE $\subset$ PE allocations, because if there is a single PE allocation then it must by definition maximize aggregate social welfare, because there is no other feasible solutions.

Any solution maximizing social welfare has to be PE, but not all PE solutions are SW maximizing. For example, at a given PE allocation agent $i$ could lose 1 util in exchange for agent $j$ gaining 2.

  • $\begingroup$ The two halves of the question are unconnected...please post separate questions as separate questions. 1) Are WE aggregate utility maximizing? No, not necessarily. 2) Is the set of WE the same as the set of PE? It depends on what you mean exactly by the set of WE. See the second theorem of welfare economics for details. $\endgroup$
    – Giskard
    Feb 3, 2018 at 13:07
  • $\begingroup$ They are not separate questions. One follows from the next but question 2 should come first, because question one depends on the answer to question 1. I have updated the question, to switch the questions around. $\endgroup$
    – spinkus
    Feb 3, 2018 at 23:42

1 Answer 1


Using the 'utility monster' critique of aggregate utility maximization it is pretty easy to show that gen. comp. eq. does not always maximize aggregate utility.

Imagine a pure exchange economy with two agents, $A$ and $B$ and two types of goods, $x_1$ and $x_2$. Agent $A$ has nothing whereas agent $B$ has one unit of each good. The agents respective utility functions are $$ U_A(x_1^A,x_2^A) = 2x_1^A + 2x_2^A \hskip 20pt U_B(x_1^B,x_2^B) = x_1^B + x_2^B. $$ Trivially the initial distribution is a competitive equilibrium and also trivially it does not maximize aggregate utility.

I am not sure I understand @spinkus's comment, as obviously not every solution maximizes the aggregate utility, but I will address the other point: Not trading. That is not at all important for me, I just wanted to make the example trivial.

Suppose given some strictly monotonic utility functions $U_A,U_B$ and some initial endowments $w_A,w_B$ some allocation $a^*$ is a competitive equilibrium. If this allocation does not maximize aggregate utility we are done. If it does, consider the following monotonic transformation of $U_B$: $$ \hat{U}_B \triangleq c \cdot U_B $$ where $c$ is a positive number. In a pure exchange economy with $U_A,\hat{U}_B$, initial endowments $w_A,w_B$ the competitive equilibrium will again be $a^*$ as the preferences are unchanged. Yet even if $a^*$ maximized $U_A+U_B$, it clearly will not maximize $U_A + \hat{U}_B$ if $c$ is small enough.

  • $\begingroup$ OK, sure. But given any initial endowment in this economy, it's never rational for any agent to trade. Every allocation is Pareto efficient, and also a competitive equilibrium. Every solution does in fact maximize social welfare w.r.t. the feasible outcomes. $\endgroup$
    – spinkus
    Feb 2, 2018 at 12:01
  • 1
    $\begingroup$ @spinkus What do you mean by "feasible outcomes"? If you're a social planner distributing endowments, then you're not bound by the constraints in the answer, you can change the initial endowments. $\endgroup$
    – Ege Erdil
    Feb 2, 2018 at 12:40
  • $\begingroup$ stated in the question, I can't change endowments. I'm not a social planner. By feasible outcomes, I mean that the "social planner" allocation would not arise from the economically rational behavior of agents in a free market. Therefore is not feasible. $\endgroup$
    – spinkus
    Feb 2, 2018 at 23:35
  • $\begingroup$ @spinkus Okay, I give up, I no longer understand what you mean. Even after reading your edit. $\endgroup$
    – Giskard
    Feb 3, 2018 at 6:54
  • $\begingroup$ @denesp What if said "Does the equilibrium solution also maximize surplus" instead of "... maximize aggregate welfare"? $\endgroup$
    – spinkus
    Feb 3, 2018 at 7:39

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.