Why Redistribution and Efficiency Are Separate Questions
A competitive market can allocate resources efficiently while distributing income in a way society regards as undesirable. The Second Welfare Theorem addresses a different question from the First Welfare Theorem: rather than asking whether a competitive equilibrium is efficient, it asks whether a desired efficient allocation can be reached through competitive markets after an appropriate redistribution of initial wealth.
The distinction matters for policy. If a government can make a suitable lump-sum transfer—one that changes wealth but does not change the reward from an additional unit of work, saving or production—it may change consumers' budget sets without distorting their choices at market prices. Under the theorem's assumptions, competition can then support the chosen efficient allocation.
A Pareto-efficient allocation is feasible and has no feasible alternative that makes at least one person better off without making anyone worse off. Efficiency does not specify who receives which goods, nor does it imply that the allocation is equitable by any particular social standard. The theorem says that a suitable distribution of purchasing power can support an efficient allocation; it does not say which efficient allocation society ought to choose.
What the Theorem Assumes—and What It Does Not Promise
The precise assumptions depend on the economic model. In the standard competitive setting used to interpret the theorem, the economy has well-defined consumers and resources, consumers and firms take prices as given, and markets are complete enough for relevant goods to be traded. Preferences are convex, and production possibilities are convex. There are no unpriced external effects that make private choices diverge from social costs or benefits. Regularity conditions are also needed so that the relevant prices and choices exist.
Convexity is important because it helps ensure that an efficient allocation can be supported by a common price system. A suitable redistribution then gives each consumer enough wealth to afford the bundle assigned to that consumer at those prices. The theorem is not a claim that every feasible allocation can be supported: the target must be Pareto efficient and the assumptions must permit supporting prices.
The redistribution is not the same as changing market prices to force consumers to take particular quantities. It changes initial purchasing power. Once that change is made, consumers remain free to choose their preferred affordable bundles, and firms continue to make profit-maximizing decisions. In the ideal model, the transfer is lump-sum rather than tied to the quantity a person buys, sells or produces.
How a Redistribution Supports a Target Allocation
The economic logic has two parts. First, for a Pareto-efficient target, the model's assumptions must allow a common price vector that makes each consumer's assigned bundle a utility-maximizing choice and each firm's assigned production plan a profit-maximizing choice. Second, each consumer must have income equal to the value of the assigned bundle at those prices. A suitable redistribution aligns incomes with those values.
The budget condition for consumer \(j\), whose assigned bundle is \(x_j^*\) and whose price vector is \(p\), is:
Here \(m_j\) is the consumer's income after redistribution. The assigned bundle is affordable because its value equals income. If it is utility-maximizing in the consumer's budget set, the consumer chooses it. If all assigned bundles and production plans are feasible together, markets clear. This is the core decentralization logic: prices support choices, and redistribution supplies the required purchasing power.
This is a conditional explanation, not a claim that any arbitrary prices will work. The demanding part is finding prices that support the target choices in the first place. Convexity and the other assumptions are what make the theorem's existence result possible in the standard model.
Each consumer's bundle and each firm's production plan must fit within the economy's available resources and technology.
At those common prices, the target choices must be optimal for consumers and firms under the model's assumptions.
Redistribute initial endowments or wealth so each consumer can afford the assigned bundle at the supporting prices.
Because the target is feasible, the chosen bundles and production plans use the available resources consistently.
Worked Applications
Worked Example: Supporting an Efficient Allocation with a Transfer
Two consumers, A and B, trade two goods in a pure exchange economy. Each has utility \(u(x_1,x_2)=\sqrt{x_1x_2}\). The feasible target gives A the bundle \((6,3)\) and B the bundle \((2,1)\). Initial endowments are \(e_A=(8,0)\) and \(e_B=(0,4)\). Determine whether a transfer can support the target as a competitive allocation, and calculate it.
Method. First check that the target uses the total endowment. The total of good 1 is \(6+2=8\), and the total of good 2 is \(3+1=4\). The target is feasible.
For this utility function, the marginal utilities are \(MU_1=\tfrac{1}{2}\sqrt{x_2/x_1}\) and \(MU_2=\tfrac{1}{2}\sqrt{x_1/x_2}\). Thus \(MU_1/MU_2=x_2/x_1\). At both target bundles, the marginal rate of substitution is \(3/6=1/2\) for A and \(1/2\) for B. Choose prices \(p_1=1\) and \(p_2=2\), so \(p_1/p_2=1/2\), matching both consumers' marginal rates of substitution.
At these prices, A's target bundle is worth \(1(6)+2(3)=12\), while B's is worth \(1(2)+2(1)=4\). Their initial endowment values are \(1(8)+2(0)=8\) for A and \(1(0)+2(4)=8\) for B. A transfer of 4 units of money from B to A therefore gives them the required incomes of 12 and 4.
To check consumer choice, each consumer with this utility spends half of income on each good at an interior optimum. A with income 12 demands \(x_1=12/(2\cdot1)=6\) and \(x_2=12/(2\cdot2)=3\). B with income 4 demands \(x_1=4/(2\cdot1)=2\) and \(x_2=4/(2\cdot2)=1\). Aggregate demand is \((6+2,3+1)=(8,4)\), exactly the available total.
Conclusion and trap avoided. The transfer supports the target at the stated prices, and markets clear. A common mistake is to compare only the target bundles with the original endowments and conclude that no adjustment is possible. The theorem allows initial wealth to be redistributed; it does not require the target to be supported by the original distribution.
Worked Example: A Feasible Allocation That Cannot Be Supported
Two consumers have identical utility \(u(x_1,x_2)=\sqrt{x_1x_2}\). Total resources are \((10,10)\), and a proposed allocation assigns \((8,2)\) to A and \((2,8)\) to B. Can any redistribution of wealth make this allocation a competitive equilibrium with both consumers consuming positive amounts of both goods?
Method. Check the necessary common-price condition at the proposed interior bundles. For each consumer, utility maximization at positive prices requires \(MU_1/MU_2=p_1/p_2\). For this utility function, \(MU_1/MU_2=x_2/x_1\).
At A's bundle, the ratio is \(2/8=1/4\). At B's bundle, it is \(8/2=4\). Supporting A's bundle would require \(p_1/p_2=1/4\), while supporting B's would require \(p_1/p_2=4\). No single positive price ratio satisfies both conditions.
This is not just a budget problem. The marginal rates of substitution differ, so the allocation is not Pareto efficient. For a small trade, let A give up \(\varepsilon\) units of good 1 and receive \(\varepsilon\) units of good 2. The first-order change in A's utility is proportional to \(-MU_{1,A}+MU_{2,A}\), which is positive because \(MU_{1,A}/MU_{2,A}=1/4<1\). B's first-order change is proportional to \(MU_{1,B}-MU_{2,B}\), which is positive because \(MU_{1,B}/MU_{2,B}=4>1\). For sufficiently small positive \(\varepsilon\), both consumers gain.
Conclusion and trap avoided. No redistribution can make this target a competitive equilibrium at common positive prices. A common error is to apply the theorem to any feasible allocation. Its target must be Pareto efficient; changing wealth cannot repair conflicting marginal conditions at the target bundles.
Worked Example: Different Preferences and Different Target Incomes
In a two-good exchange economy, A has utility \(u_A=x_1^{3/4}x_2^{1/4}\), and B has utility \(u_B=x_1^{1/4}x_2^{3/4}\). The target bundles are \((9,3)\) for A and \((1,3)\) for B. Initial endowments are \(e_A=(10,0)\) and \(e_B=(0,6)\). Show how a transfer supports the target at prices \(p_1=p_2=1\).
Method. Check feasibility, price values, and each consumer's optimizing demand. The target uses \(9+1=10\) units of good 1 and \(3+3=6\) units of good 2, so it exhausts the available resources.
For Cobb-Douglas utility \(x_1^a x_2^{1-a}\), an income \(m\) at positive prices \(p_1,p_2\) gives demands \(x_1=am/p_1\) and \(x_2=(1-a)m/p_2\). These demands follow from the budget condition and the first-order condition \(MU_1/MU_2=p_1/p_2\). At prices \((1,1)\), A's target is worth \(9+3=12\), and B's is worth \(1+3=4\).
A's initial endowment is worth 10, and B's is worth 6. A transfer of 2 units of money from B to A changes their incomes to 12 and 4. A then demands \(x_1=(3/4)(12)=9\) and \(x_2=(1/4)(12)=3\). B demands \(x_1=(1/4)(4)=1\) and \(x_2=(3/4)(4)=3\). These are exactly the target bundles, and aggregate demand remains \((10,6)\).
The target also has equal marginal rates of substitution: for A, \((3)(x_2/x_1)=3(3/9)=1\); for B, \((1/3)(x_2/x_1)=(1/3)(3/1)=1\). Thus both consumers' marginal tradeoffs match the price ratio \(p_1/p_2=1\). Under the stated interior, convex-preference setting, the calculation is consistent with supporting an efficient target.
Conclusion and trap avoided. The target is supported after a transfer of 2, despite unequal preferences and unequal target incomes. Do not assume that equal prices imply equal incomes or equal bundles; prices reflect relative scarcity, while redistribution determines purchasing power.
Limits and Interpretation
The Second Welfare Theorem is a separation result: in the ideal model, society can select a distribution of initial wealth to support an efficient outcome, while competitive prices coordinate individual choices. It does not itself provide a rule for choosing among efficient allocations. An allocation that gives one consumer much more than another may still be Pareto efficient, and the theorem alone does not label it fair.
The assumptions limit its practical interpretation. If preferences or production possibilities are nonconvex, supporting prices may fail to exist for an efficient allocation. If there are externalities, missing markets or market power, private choices may not respond to prices in the way the result requires. In addition, real-world taxes and transfers may affect behavior. A tax that changes the marginal reward for work or production is not the undistorting lump-sum redistribution assumed in the clean theorem.
For welfare analysis, the theorem therefore provides a benchmark rather than a blanket endorsement of redistribution or of every market outcome. The correct sequence is to identify an efficient feasible target, check whether the model permits common supporting prices, and then ask whether wealth can be redistributed to give consumers the necessary budgets. Efficiency and distribution are related, but they are not interchangeable.
Check Your Understanding
For each problem, identify the supporting-price or income condition that matters before stating your conclusion.
- Two consumers have utility \(\sqrt{x_1x_2}\). At prices \(p_1=1\), \(p_2=2\), a target assigns \((4,2)\) to A and \((2,1)\) to B. Initial endowments are \((6,0)\) and \((0,3)\). Find the transfer needed to support the target and verify aggregate demand. (Answer: A's target costs 8 and B's costs 4; initial incomes are 6 and 6, so transfer 2 from B to A. Demands are \((4,2)\) and \((2,1)\), totaling \((6,3)\).)
- A consumer with utility \(x_1^{2/3}x_2^{1/3}\) is assigned \((8,4)\) at prices \((2,1)\). Find the bundle's value and determine whether it is the consumer's utility-maximizing demand at that income. (Answer: Value \(=2(8)+4=20\); demands are \(((2/3)(20)/2,(1/3)(20)/1)=(20/3,20/3)\), so the assigned bundle is not utility-maximizing and these prices do not support it.)
- Two consumers with identical utility \(\sqrt{x_1x_2}\) receive \((6,2)\) and \((2,6)\). Can positive common prices support both interior bundles, regardless of redistribution? (Answer: No. Their marginal rates of substitution are \(2/6=1/3\) and \(6/2=3\), so no common price ratio supports both.)
- A has utility \(x_1^{1/2}x_2^{1/2}\), B has utility \(x_1^{1/4}x_2^{3/4}\), and prices are \((1,1)\). The target bundles are \((4,4)\) for A and \((2,6)\) for B. Calculate the incomes needed to support the bundles and state whether a transfer from initial incomes 10 for A and 10 for B can achieve them. (Answer: Required incomes are 8 for A and 8 for B; their total is 16, equal to initial total wealth 20 only if 4 is not assigned elsewhere. A transfer alone cannot change total wealth, so these specified initial incomes cannot both become 8 through a zero-sum transfer.)