Pareto Efficiency Explained: Pareto Improvements, Fairness and Allocative Efficiency

An allocation is Pareto efficient (also called Pareto optimal) when no one can be made better off without making at least one other person worse off. A Pareto improvement is any change that makes at least one person better off and nobody worse off, so an allocation is Pareto efficient exactly when no Pareto improvement is left.

The idea is named after the Italian economist Vilfredo Pareto, and it is the most widely used test of efficiency in economics because it needs no judgment about whose well-being counts more. That same feature is its biggest limit: Pareto efficiency says nothing about fairness. This guide works through a small numeric example, shows how the concept connects to allocative and productive efficiency and to perfect competition, and covers the mistake students make most often, which is treating "efficient" as a synonym for "good". For the short definition, see the glossary entry for Pareto efficiency.

The vocabulary in one table

Two features of this definition matter for everything that follows. First, a Pareto improvement needs nobody to lose, not just a net gain. Second, the test compares each person with themselves. It never asks whether one person's gain is bigger than another person's loss.

A worked example: Maya, Leo and a shared fridge

Two roommates share 4 slices of pizza and 4 cans of soda. Each has a simple scoring rule for their own satisfaction:

• Maya values each slice of pizza at 3 points and each soda at 1 point.
• Leo values each slice of pizza at 1 point and each soda at 2 points.

Each person's points rank only their own options. Maya's 3 points and Leo's 2 points are not on a common scale, so adding them together would mean nothing.

### Step 1: score the equal split

Allocation A gives each roommate 2 slices and 2 sodas.

• Maya: (2 × 3) + (2 × 1) = 6 + 2 = 8 points
• Leo: (2 × 1) + (2 × 2) = 2 + 4 = 6 points

### Step 2: look for a Pareto improvement

Maya likes pizza relatively more, and Leo likes soda relatively more. Suppose Maya trades 1 soda for 1 of Leo's slices.

• Maya now has 3 slices and 1 soda: (3 × 3) + (1 × 1) = 9 + 1 = 10 points, up from 8.
• Leo now has 1 slice and 3 sodas: (1 × 1) + (3 × 2) = 1 + 6 = 7 points, up from 6.

Both are better off and nobody is worse off, so this trade is a Pareto improvement. That proves allocation A was not Pareto efficient, even though it is a perfectly equal split.

### Step 3: keep trading until no gains remain

Repeat the swap once more. Allocation B gives Maya all 4 slices and Leo all 4 sodas.

• Maya: (4 × 3) + (0 × 1) = 12 points
• Leo: (0 × 1) + (4 × 2) = 8 points

### Step 4: check that B is Pareto efficient

Any further change has to take pizza from Maya or soda from Leo, because those are the only things each holds.

• Take 1 slice from Maya and give it to Leo: Maya loses 3 points. To make her whole, Leo would have to hand over 3 sodas, which costs him 3 × 2 = 6 points while the slice gives him only 1 point. Leo ends up 5 points worse off.
• Take 1 soda from Leo and give it to Maya: Leo loses 2 points. To make him whole he would need 2 slices back from Maya, which costs her 2 × 3 = 6 points while the soda gives her only 1. Maya ends up 5 points worse off.

Every possible change makes someone worse off, so allocation B is Pareto efficient. Checking it a second way: Maya would give up to 3 sodas for a slice, while Leo would give up a slice for anything above half a soda. Once Maya holds all the pizza and Leo all the soda, neither has anything left that the other values more than its owner does.

### Step 5: compare several allocations

Check C and D: Leo in D scores (4 × 1) + (4 × 2) = 4 + 8 = 12, and any change to C or D has to take something from the person who holds everything, so both are efficient. Allocation E scores (4 × 1) = 4 for Maya and (4 × 1) = 4 for Leo; swapping all the goods turns E into B and makes both better off, so E is inefficient.

The table carries four lessons:

1. There are many Pareto efficient allocations, not one. B, C and D all pass.

2. An efficient allocation can be extremely unequal. In C, Leo has nothing, yet C is Pareto efficient because helping Leo requires taking from Maya.

3. Efficient is not the same as better. Moving from A to C is not a Pareto improvement, because Leo falls from 6 points to 0. An inefficient allocation (A) and an efficient one (C) can simply be unranked by the Pareto test.

4. The test cannot choose among B, C and D. Picking one requires a value judgment about how to weigh Maya against Leo. That is the job of a social welfare function, compared in Pareto efficiency vs social welfare function.

With two people and two goods, economists draw every possible split in an Edgeworth box, where the set of Pareto efficient allocations forms the contract curve.

Pareto efficiency vs fairness

Efficiency is about waste: are there gains nobody is collecting? Fairness is about distribution: who gets what? These are different questions, and the Maya and Leo table separates them cleanly:

• Equal but inefficient: A splits everything evenly, yet both roommates could be better off.
• Efficient but one-sided: C and D pass the Pareto test while one roommate gets nothing.
• Efficient and better for both than the equal split: B, reached by trading from A.
• Neither: E is inefficient, and both roommates do worse than under A.

You do not have to choose between the two. Starting from the equal split A and trading gets the roommates to B, which is efficient and better for both than A. The second fundamental theorem of welfare economics generalizes this: under suitable conditions, any efficient allocation can be reached by choosing how to redistribute starting endowments and then letting voluntary exchange do the rest, so the choice of starting point decides which efficient outcome you reach. Economists usually use the Pareto test to find waste, and then argue about distribution separately.

How Pareto efficiency connects to allocative and productive efficiency

AP Microeconomics tests two specific kinds of efficiency. Both are pieces of the broader Pareto idea, applied to production and markets.

### Productive efficiency

Productive efficiency means producing at the lowest possible cost: a firm at the minimum of average total cost, or an economy on its production possibilities curve (PPC). A point inside the curve is Pareto inefficient for a simple reason. The economy could make more of at least one good without making less of anything, and that extra output could go to someone without taking from anyone else.

### Allocative efficiency

Allocative efficiency means producing the mix of goods people value most, which in a single market happens where price equals marginal cost (P = MC). A quick numeric check shows why that is a Pareto condition:

• A buyer values one more unit at $8 and it costs $5 to make, but the unit is not produced.
• Trade it at $6. The buyer gains 8 - 6 = $2 and the seller gains 6 - 5 = $1.
• Someone gains and nobody loses, so leaving that unit unproduced was a missed Pareto improvement.

The reverse works too. If a unit costs $9 to make and the buyer values it at only $6, stopping production saves $9 of resources. Handing the buyer $6 of that leaves her no worse off, with $3 left over for someone else. That missed gain, summed over every unit, is deadweight loss. For how the two AP efficiency tests differ, see allocative vs productive efficiency.

### The full set of conditions

An economy is Pareto efficient only when all three hold at once. Producing at minimum average total cost is the long-run perfectly competitive outcome that AP calls productive efficiency at the firm level; it is not needed for an allocation to be Pareto efficient.

Perfect competition and the first welfare theorem

The first fundamental theorem of welfare economics says that a perfectly competitive equilibrium is Pareto efficient, provided there are no externalities, no public goods, full information and no market power. In long-run equilibrium, perfectly competitive firms produce where P = MC, which delivers allocative efficiency, and at the minimum of average total cost, which delivers productive efficiency. Every buyer faces the same price, so every buyer values their last unit at that same price and no trades between buyers are left over.

When one of those conditions fails, unused Pareto improvements appear. A monopoly sets price above marginal cost and stops short of the efficient output. In the oligopoly example in Cournot, Bertrand and Stackelberg Models, a Cournot duopoly produces 80 units where 120 would be efficient, and the 40 missing units each had a buyer willing to pay more than they cost. Externalities and public goods create the same kind of gap, which is why market failure is often described as a market outcome that is not Pareto efficient.

Common mistakes

1. Calling an efficient outcome good or fair. Pareto efficiency only says no free gains remain. Allocation C, where one roommate gets nothing, passes the test.

2. Assuming any move to an efficient allocation is a Pareto improvement. Moving from A to C makes Leo worse off, so it is not one, even though C is efficient and A is not.

3. Thinking there is only one Pareto efficient outcome. There are usually many, and the test cannot rank them against each other.

4. Adding up points across people. Saying B (12 + 8 = 20) beats C (16 + 0 = 16) is a social welfare judgment, not a Pareto comparison.

5. Treating Pareto efficiency and allocative efficiency as identical. P = MC is the single-market condition. Pareto efficiency is the broader test, and it also needs productive efficiency (the economy on its PPC) and no unused trades between consumers.