Why Compare Allocations Person by Person?
An allocation describes how available goods or productive resources are assigned and what outcomes they generate. Comparing two allocations raises a question that total output or total surplus alone cannot answer: does the change make anyone worse off? The Pareto criterion answers this by comparing the welfare of each person separately. It is useful for identifying changes that benefit at least one person without imposing a loss on anyone else.
The criterion is deliberately limited. It does not rank every pair of allocations, and it does not decide whether the distribution of resources is fair. It provides a precise test for an improvement that is unambiguous from the standpoint of individual welfare.
If person \(i\)'s preferences can be represented by a utility function \(U_i\), the test is expressed as follows. Here \(x\) and \(y\) denote complete allocations, including every person's assigned bundle and, where relevant, the resulting production.
This utility test assumes that the utility functions represent the individuals' preferences: a higher value means the person strictly prefers the allocation, and equal values mean indifference. Utility levels are used to compare the same person's welfare across allocations. They are not, in general, comparable across different people: a utility increase of 2 for one person cannot automatically be treated as equal in value to an increase of 2 for another.
Feasibility Comes Before Welfare Comparison
A proposed change counts only if it can actually be achieved with the available goods, resources, and production opportunities. A plan that would make everyone better off but requires more resources than exist is not a Pareto improvement. Conversely, an allocation that leaves usable resources unassigned may permit a feasible improvement without taking anything away from anyone.
The test is about individuals, not a sum of utilities. If one person gains and another loses, the change is not a Pareto improvement, even if the gain is numerically larger than the loss. That change might be judged favorably using some other criterion, but that would require additional assumptions about how gains and losses should be compared.
Worked Example: Unassigned Goods in an Exchange
Two people, Ada and Ben, share an endowment of 10 identical units. Each unit assigned to a person gives that person one utility unit, so \(U_A=q_A\) and \(U_B=q_B\). The current allocation gives Ada 4 units and Ben 4 units, leaving the other units unassigned. Determine whether assigning the unassigned goods can improve the allocation.
Method: check the resource constraint, then compare each person's utility. The available total is 10. The current allocation uses \(4+4=8\) units, so \(10-8=2\) units are unassigned.
Consider assigning both unassigned units to Ben. The proposed allocation is \(q_A=4\), \(q_B=6\), with no units left unassigned. It is feasible because \(4+6=10\).
Ada's utility remains \(U_A=4\). Ben's utility rises from \(U_B=4\) to \(U_B=6\). Thus Ada is no worse off and Ben is strictly better off. This is a Pareto improvement, so the original allocation is not Pareto efficient.
Check: The proposal uses exactly the 10 available units, and neither person's utility falls. The key is that the gain comes from previously unassigned units, not from taking goods away from Ada.
Exchange Can Create Mutual Gains
Even when all goods are assigned, the initial allocation need not be Pareto efficient. People may value the goods differently, so a feasible exchange can improve both people's welfare. To test such a proposal, calculate each person's utility before and after the exchange and verify that the total quantity of each good is preserved.
Worked Example: A Mutually Beneficial Exchange
Two people, Cora and Dev, share 5 units of good \(X\) and 5 units of good \(Y\). Their utility functions are \(U_C=2x_C+y_C\) and \(U_D=x_D+2y_D\). Initially, Cora has \((x_C,y_C)=(2,3)\) and Dev has \((x_D,y_D)=(3,2)\). Test a proposed exchange that leaves Cora with \((3,2)\) and Dev with \((2,3)\).
Step 1: Verify feasibility. Initially, the total of good \(X\) is \(2+3=5\), and the total of good \(Y\) is \(3+2=5\). After the exchange, the totals are \(3+2=5\) units of \(X\) and \(2+3=5\) units of \(Y\). The exchange is feasible.
Step 2: Calculate initial utilities. Cora's initial utility is \(2(2)+3=7\). Dev's initial utility is \(3+2(2)=7\).
Step 3: Calculate utilities after the exchange. Cora's new utility is \(2(3)+2=8\). Dev's new utility is \(2+2(3)=8\).
Both people gain one utility unit: \(8-7=1\) for Cora and \(8-7=1\) for Dev. The exchange is therefore a Pareto improvement. The initial allocation is not Pareto efficient.
Check: The proposed allocation preserves both goods' totals, and each person's utility rises from 7 to 8. A common mistake is to check only one person's gain or to overlook whether the exchange uses goods that are actually available.
Production, Unused Capacity, and Trade-Offs
The same reasoning applies when production determines what can be allocated. A production constraint describes which output combinations are feasible. If a feasible output combination leaves productive capacity unused, it may be possible to raise someone's consumption without reducing anyone else's. But at a fully used constraint, increasing one person's output may require reducing another's.
Worked Example: From Slack Capacity to a Trade-Off
A workshop can produce at most 12 identical service units in a day. It assigns output to two customers, Mina and Noor. Each person values one assigned unit as one utility unit: \(U_M=q_M\) and \(U_N=q_N\). The current allocation is \((q_M,q_N)=(5,5)\). Compare it with \((5,7)\), and then assess whether the latter allocation is Pareto efficient under the stated production constraint.
Step 1: Check the current use of capacity. The current output is \(5+5=10\), which is below the capacity of 12. The unused capacity is \(12-10=2\) units.
Step 2: Check the proposed change. At \((5,7)\), total production is \(5+7=12\), so the proposed allocation is feasible. Mina's utility remains 5; Noor's rises from 5 to 7. This is a Pareto improvement on the current allocation.
Step 3: Test whether further improvements are possible at \((5,7)\). Under the assumed technology, feasible outputs satisfy \(q_M+q_N\le12\). The proposed allocation uses all 12 units. Since each person's utility equals that person's assigned output, a strict gain for Mina requires \(q_M>5\). Feasibility then requires \(q_N\le12-q_M<7\), which makes Noor worse off. A strict gain for Noor similarly requires \(q_N>7\), which would require total output above 12 if Mina stayed at 5. Thus no feasible change can make at least one person better off while keeping the other no worse off. This allocation is Pareto efficient under these assumptions.
Check: The initial allocation is inefficient because capacity is unused; the later allocation is efficient because it exhausts capacity and, with these preferences, extra output for one customer must come at the other's expense. Full capacity alone would not establish efficiency if some produced output were wasted or if preferences differed from those specified.
Why the Pareto Test Works—and What It Does Not Say
The utility formulation follows directly from the definition. A utility representation records how an individual ranks outcomes. Therefore, \(U_i(y)\ge U_i(x)\) means person \(i\) is at least as well off at \(y\), while \(U_i(y)>U_i(x)\) means that person is strictly better off. Requiring the weak inequality for every person prevents anyone from losing; requiring a strict inequality for at least one person ensures that the change is a genuine improvement rather than simply an alternative allocation that leaves everyone indifferent.
Pareto efficiency is the absence of such a feasible change. To test an allocation, one must consider the alternatives allowed by the resource and production constraints—not just the alternatives listed in one example. If no feasible alternative weakly improves everyone and strictly improves someone, the allocation is Pareto efficient. If even one such alternative exists, the allocation is not.
This conclusion is not a statement that the allocation is equitable, desirable in every respect, or the best available by a social standard. An allocation can be Pareto efficient even if one person receives nearly all the goods, provided that giving some goods to someone else would make the first person worse off. The Pareto criterion also cannot decide between two allocations when one benefits one person and harms another. It avoids making interpersonal comparisons of utility, but that restraint means it leaves many policy choices unresolved.
The earlier material on allocative and productive efficiency considered whether resources or output are used efficiently. Pareto efficiency gives a general welfare test across feasible allocations: it asks whether any change can help someone without harming anyone. The concepts are related, but applying this test requires specifying who is affected, what counts as feasible, and how each person's welfare changes. It does not require calculating a total utility or treating a transfer between people as automatically beneficial.
Common Mistakes
- Ignoring feasibility. A proposed allocation that exceeds available goods or production capacity cannot be a Pareto improvement.
- Calling any gain an improvement. If at least one person gains but another loses, the change fails the Pareto test, regardless of the size of either person's utility change.
- Adding utility changes across people. The definition requires individual comparisons, not a sum of utility numbers. Utility scales need not be comparable across people.
- Equating efficiency with fairness. Pareto efficiency means no further Pareto improvement is feasible; it does not mean that resources are distributed equally or justly.
- Assuming all efficient allocations are identical. There may be many Pareto-efficient allocations, with different distributions of goods and welfare.
Check Your Understanding
For each problem, check feasibility first and then apply the person-by-person test.
- Two people share 9 identical units. Their utilities equal their assigned quantities. An allocation assigns 3 units to Lee and 4 to Pat, leaving the rest unassigned. Is assigning both remaining units to Pat a Pareto improvement? (Answer: Yes. The feasible allocation changes utilities from \((3,4)\) to \((3,6)\), so Lee is unchanged and Pat gains.)
- Two people have utility levels \((10,6)\) at allocation A and \((10,8)\) at feasible allocation B. Is B a Pareto improvement on A? (Answer: Yes. The first person's utility is unchanged and the second person's utility rises.)
- At allocation C, two people's utilities are \((7,9)\). A feasible alternative gives utilities \((8,8)\). Is the alternative a Pareto improvement? (Answer: No. The first person gains, but the second person's utility falls from 9 to 8.)
- A workshop can produce at most 15 units. It assigns 6 units to one customer and 7 to another, with each person's utility equal to assigned units. Is the allocation Pareto efficient under this production constraint? (Answer: No. It uses \(6+7=13\) units, leaving 2 units of capacity; assigning both to either customer improves that person's utility without reducing the other's.)
- Two customers have utilities equal to their assigned quantities, and a workshop can produce at most 15 units. It assigns 8 units to one and 7 to the other. Can any feasible change be a Pareto improvement? (Answer: No. All capacity is used, and any increase in one customer's assigned output requires a decrease in the other's; the allocation is Pareto efficient under the stated assumptions.)