Deriving the Second Welfare Theorem’s Redistribution Result

From an Efficient Target to a Competitive Outcome

The Second Welfare Theorem addresses a different question from the First Welfare Theorem. Instead of asking whether a competitive equilibrium is efficient, it asks whether an efficient allocation can be made a competitive outcome. The key is to separate prices, which guide choices, from purchasing power, which determines which choices consumers can afford.

In an exchange economy, each consumer starts with an endowment of goods. A proposed efficient allocation may not be what consumers choose at market prices if they keep those initial endowments as their incomes. The theorem’s redistribution result is to give each consumer income equal to the value of that consumer’s bundle in the target allocation. If suitable prices support the target, consumers then choose the target bundles, and feasibility ensures that markets clear.

Definition: A price vector \(p=(p_1,\ldots,p_L)\) supports an allocation \(x^*=(x_1^*,\ldots,x_I^*)\) if, for every consumer \(i\), every feasible bundle \(y_i\) weakly preferred to \(x_i^*\) costs at least as much as \(x_i^*\): \(p\cdot y_i\ge p\cdot x_i^*\). For implementation as a consumer choice, we also need that every bundle strictly preferred to \(x_i^*\) costs strictly more: \(p\cdot y_i>p\cdot x_i^*\) whenever \(y_i\succ_i x_i^*\). The second condition rules out a strictly better bundle that is affordable at the target’s income.

Assumptions and the Redistribution Result

Consider a finite exchange economy with \(I\) consumers and \(L\) goods. Consumer \(i\) has endowment \(e_i\) and receives a feasible consumption bundle \(x_i\) from the set of nonnegative bundles. A target allocation \(x^*\) is feasible when the total quantity allocated of each good equals the total endowment of that good:

\sum_{i=1}^{I}x_i^*=\sum_{i=1}^{I}e_i.

Preferences are assumed to be convex and locally nonsatiated, and we use supporting prices \(p\ne 0\) with nonnegative components. Convexity is the condition that permits supporting-price arguments for an efficient target under suitable regularity assumptions. For the implementation argument below, the relevant support condition must exclude strictly preferred bundles from the target’s budget set. One sufficient setting is continuous preferences with a strictly positive supporting price vector \(p\gg 0\): a strictly preferred bundle on the same price hyperplane would have a nearby, slightly cheaper bundle that remains strictly preferred, contradicting support.

Theorem: Suppose \(x^*\) is feasible and there is a common price vector \(p\) such that, for every consumer \(i\), no bundle strictly preferred to \(x_i^*\) costs \(p\cdot x_i^*\) or less. Give consumer \(i\) income \(m_i^*=p\cdot x_i^*\). Then \(x_i^*\) is a utility-maximizing choice for that consumer at prices \(p\) and income \(m_i^*\). Because the target allocation is feasible, these choices clear all markets. The required transfer from consumer \(i\)’s initial wealth is \(T_i=m_i^*-p\cdot e_i\), and the transfers sum to zero.

The target income is not an arbitrary payment. It is exactly the cost of the target bundle at the supporting prices. The transfer adjusts each consumer’s initial wealth to that target income. A positive \(T_i\) is a transfer received; a negative \(T_i\) is a contribution made. The identity that makes this redistribution budget-neutral follows from feasibility:

\sum_{i=1}^{I}T_i =\sum_{i=1}^{I}p\cdot x_i^*-\sum_{i=1}^{I}p\cdot e_i =p\cdot\left(\sum_{i=1}^{I}x_i^*-\sum_{i=1}^{I}e_i\right) =0.

Why the Target Becomes a Consumer Choice

Here is the central justification. Fix a consumer \(i\), and give that consumer income \(m_i^*=p\cdot x_i^*\). The target bundle is affordable because its cost equals income. Suppose, for contradiction, that some affordable bundle \(y_i\) were strictly preferred to the target. Affordability would mean \(p\cdot y_i\le m_i^*=p\cdot x_i^*\), but the strict support condition requires \(p\cdot y_i>p\cdot x_i^*\). These inequalities cannot both hold. Thus no strictly preferred bundle is affordable, which is precisely what is needed for the target to be a utility-maximizing choice. Other bundles may tie with the target in utility; that does not prevent the target from being a maximizer.

The budget of each consumer is balanced at the target income, and the sum of those incomes equals the value of the economy’s total resources. Since the target bundles exhaust those resources, choosing the target for every consumer clears each market. This proves implementation; it does not claim that the target is the only equilibrium allocation or that consumers’ initial endowments would have produced the same allocation without transfers.

Start with the target.
Check that the target is feasible and Pareto efficient, and identify supporting prices for its consumers’ preferences.

Check strict-preference exclusion.
Verify that no bundle strictly preferred to a consumer’s target bundle costs that target bundle’s value or less.

Set target incomes.
For each consumer, calculate \(m_i^*=p\cdot x_i^*\), then compare it with initial wealth \(p\cdot e_i\).

Verify implementation.
Show that each target is a utility-maximizing affordable choice, transfers sum to zero, and aggregate target quantities equal aggregate endowments.

Worked Examples

Worked Example: Implementing a Two-Consumer Target

Two consumers exchange two goods. Their utilities are \(u_A(x_1,x_2)=x_1^{1/4}x_2^{3/4}\) and \(u_B(x_1,x_2)=x_1^{3/4}x_2^{1/4}\). The proposed target bundles are \(x_A^*=(4,6)\) and \(x_B^*=(6,1)\); their endowments are \(e_A=(2,3)\) and \(e_B=(8,4)\). Find supporting prices, target incomes, and transfers, and check implementation.

Method. For Cobb–Douglas utility \(u=x_1^\alpha x_2^{1-\alpha}\), the marginal rate of substitution is \(\mathrm{MRS}=\frac{\alpha}{1-\alpha}\frac{x_2}{x_1}\). At consumer A’s target, \(\mathrm{MRS}_A=\frac{1/4}{3/4}\frac{6}{4}=\frac{1}{3}\cdot\frac{3}{2}=\frac12\). At consumer B’s target, \(\mathrm{MRS}_B=\frac{3/4}{1/4}\frac{1}{6}=3\cdot\frac16=\frac12\). Thus both target bundles have the same marginal rate of substitution. Choose the normalized prices \(p=(1,2)\), for which \(p_1/p_2=1/2\).

Check feasibility. Total target quantities are \((4+6,6+1)=(10,7)\). Total endowments are \((2+8,3+4)=(10,7)\), so the allocation is feasible.

Calculate target incomes. Consumer A’s target costs \(1(4)+2(6)=4+12=16\), so \(m_A^*=16\). Consumer B’s target costs \(1(6)+2(1)=6+2=8\), so \(m_B^*=8\).

Calculate initial wealth and transfers. A’s endowment is worth \(1(2)+2(3)=2+6=8\), so \(T_A=16-8=+8\). B’s endowment is worth \(1(8)+2(4)=8+8=16\), so \(T_B=8-16=-8\). The transfers balance: \(+8-8=0\).

Verify choices. At these prices, A spends one quarter of income on good 1 and three quarters on good 2. With income 16, the demands are \(x_{A1}=4\) and \(x_{A2}=12/2=6\). B spends three quarters of income on good 1 and one quarter on good 2. With income 8, the demands are \(x_{B1}=6\) and \(x_{B2}=2/2=1\). Both demands equal their targets. Strictly convex Cobb–Douglas preferences make these interior demands unique. The market-clearing check above completes the verification. The common mistake this method avoids is setting \(p_1/p_2\) equal to 1 merely because a target bundle looks balanced; the price ratio must match the marginal rate of substitution.

Worked Example: Why Weak Support Alone Is Not Enough

A consumer has lexicographic preferences over two nonnegative goods: bundles are ranked first by \(x_1+x_2\), and ties are ranked by \(x_2\). The target is \((1,0)\), and prices are \(p=(1,1)\). Check whether the weak supporting inequality alone makes the target a utility-maximizing choice at income 1.

Check weak support. A bundle weakly preferred to \((1,0)\) must either have \(x_1+x_2>1\), or have \(x_1+x_2=1\) and \(x_2\ge0\). In either case its cost \(x_1+x_2\) is at least 1, the target’s cost. Thus the weak supporting inequality holds.

Check the budget set. The bundle \((0,1)\) costs \(0+1=1\), so it is affordable at income 1. Its total quantity is also 1, but its second coordinate is 1 rather than 0. It is therefore strictly preferred to the target. The target is not utility maximizing at that income.

Conclusion and trap avoided. Weak support does not by itself rule out a strictly preferred bundle with exactly the target’s cost. The implementation proof needs strict-preference exclusion, or assumptions that ensure it. This example’s lexicographic preferences are not continuous, so it does not satisfy the sufficient regularity conditions described above. It shows why the equality case must be checked rather than silently treated as a contradiction.

Worked Example: Finding Transfers from a Different Target

Two consumers have utilities \(u_C(x_1,x_2)=x_1^{1/3}x_2^{2/3}\) and \(u_D(x_1,x_2)=x_1^{1/2}x_2^{1/2}\). Their proposed target bundles are \(x_C^*=(3,6)\) and \(x_D^*=(5,5)\), and their endowments are \(e_C=(5,2)\) and \(e_D=(3,9)\). Find a supporting price vector normalized to \(p_2=1\), and calculate the transfers.

Find the price ratio. For C, \(\mathrm{MRS}_C=\frac{1/3}{2/3}\frac{6}{3}=\frac12\cdot2=1\). For D, \(\mathrm{MRS}_D=\frac{1/2}{1/2}\frac{5}{5}=1\). Both target bundles are supported by \(p_1/p_2=1\); with \(p_2=1\), take \(p=(1,1)\).

Check feasibility. The target totals are \((3+5,6+5)=(8,11)\), and the endowment totals are \((5+3,2+9)=(8,11)\).

Calculate incomes and transfers. C’s target income is \(3+6=9\), while initial wealth is \(5+2=7\), so \(T_C=9-7=+2\). D’s target income is \(5+5=10\), while initial wealth is \(3+9=12\), so \(T_D=10-12=-2\). Transfers sum to zero.

Verify choices. C’s Cobb–Douglas expenditure shares are one third on good 1 and two thirds on good 2. At income 9 and unit prices, this gives \((9/3,18/3)=(3,6)\). D spends half on each good, giving \((10/2,10/2)=(5,5)\). The target is implemented, and its quantities clear both markets. The calculation avoids confusing an allocation’s physical feasibility with the income distribution needed to support it.

What the Theorem Does—and Does Not—Say

The result is a possibility statement: with suitable supporting prices and a suitable redistribution of purchasing power, an efficient target can be decentralized as a competitive outcome. It does not say that every redistribution will work, that arbitrary prices support a target, or that an efficient allocation is automatically the outcome under the original endowments. The prices must align with the target’s supporting conditions, and the transfers must give each consumer the target bundle’s value.

A useful exam check is to keep three questions separate. First, is the allocation feasible? Second, do common prices support the target choices, including the strict-preference condition needed for consumer maximization? Third, do the proposed incomes and transfers equal the target values and balance in total? Passing only one or two of these checks is not enough to establish implementation.

Key takeaway: A Pareto-efficient target can be implemented when common prices support it and consumers receive income equal to the value of their target bundles. Transfers reallocate purchasing power, not goods: feasibility makes the transfers sum to zero and ensures that target choices clear markets. Weak support alone can fail if a strictly preferred bundle is affordable at exactly the target’s cost.

This treatment is aligned with the SOA VEE Economics (Validation by Educational Experience) syllabus.

Check Your Understanding

For each problem, distinguish the price-support calculation from the income and feasibility checks.

A consumer has utility \(u=x_1^{2/5}x_2^{3/5}\) and target bundle \((5,3)\). What price ratio \(p_1/p_2\) makes the target’s marginal rate of substitution equal the price ratio? (Answer: \(p_1/p_2=2/5\).)
A consumer has utility \(u=x_1^{1/3}x_2^{2/3}\), target \((6,3)\), and prices \(p=(1,2)\). Is the target supported by these prices? Compare the marginal rate of substitution with the price ratio. (Answer: No; the MRS is \(1/4\), while \(p_1/p_2=1/2\).)
At prices \(p=(2,1)\), consumer A’s target is \((2,4)\) and endowment is \((1,5)\); consumer B’s target is \((3,2)\) and endowment is \((4,1)\). Calculate both transfers and check that they balance. (Answer: \(T_A=+1\), \(T_B=-1\), and the sum is zero.)
A consumer with utility \(u=x_1^{1/4}x_2^{3/4}\) has target \((4,8)\), endowment \((1,4)\), and supporting prices normalized to \(p_2=1\). Find \(p_1\) and the required transfer. (Answer: \(p_1=2/3\), target income \(32/3\), initial wealth \(14/3\), and transfer \(+6\).)
At prices \(p=(2,1)\), two consumers’ target bundles are \((1,4)\) and \((5,2)\), and their endowments are \((3,1)\) and \((3,5)\). Check feasibility and find the transfers. (Answer: Both aggregate vectors are \((6,6)\); the transfers are \(-1\) and \(+1\), which sum to zero.)

What You'll Learn