Why Green Consumers’ Understanding of Tax Rebates Matters

A moral question inside an economic model

Environmental taxes conventionally aim to make buyers face the damage their purchases impose on others. A Pigovian tax sets the charge per unit equal to its environmental cost. But what if consumers already restrain themselves for moral reasons?

In Moral Consumers and the Environment, a theoretical preprint submitted to arXiv on November 6, 2026, Mogens Fosgerau and Jörgen Weibull examine how that restraint interacts with taxes, rebates and emissions caps. Their findings are mathematical implications of a model, not estimates of how actual shoppers respond.

What “moral” means here

Consumers decide whether to buy a unit of an environmentally harmful good. They differ in how much they value it, what they know about its environmental cost and their degree of moral concern. Environmental harm is initially linear in consumption and shared equally across the population. Sellers compete, and the government returns all environmental-tax revenue through equal lump-sum transfers.

The authors use “homo moralis” preferences. A consumer evaluates a purchasing rule by asking what her own material payoff would be if others hypothetically followed that rule too, applying it to their own circumstances. Her degree of morality determines the weight of this thought experiment.

This is neither a prediction that others will imitate her nor simply pleasure from performing a good deed. It is a particular representation of Kantian reasoning, not a claim to capture every ethical motivation.

For the baseline finite-population model, the authors establish an equilibrium that is unique apart from decisions at events of probability zero. Subsequent policy analysis mainly uses a large-population approximation.

Why understanding rebates changes the answer

The central distinction is between consumers who understand the tax-and-transfer system and those who see the purchase price but do not recognize the returned revenue. Both perceive prices correctly.

When consumers understand the system, their moral calculation includes both environmental harm and the revenue their hypothetical collective purchases would generate for the public. At a Pigovian tax, those two amounts cancel within that calculation. Buyers then purchase when their valuation exceeds the price, which already includes environmental damage.

If everyone has the same degree of morality, below the purely Kantian extreme, the Pigovian tax is uniquely optimal regardless of how accurately consumers know individual goods’ environmental costs. Moral preferences have not disappeared; the policy has removed the net external cost to which they respond.

With differing degrees of morality, the conclusion is narrower. Pigovian taxation always satisfies the model’s first-order optimality conditions for aware consumers. The paper proves unique global optimality under additional restrictions, including perfectly informed consumers, exponentially distributed valuations and an upper bound on morality. A stationary point alone does not establish a general optimum.

When lower taxes emerge

Consumers unaware of rebates count environmental harm in their moral reasoning without counting the offsetting public revenue. Under common morality and perfect environmental information, the optimal tax is the environmental cost multiplied by one minus the degree of morality. Voluntary restraint therefore reduces the corrective charge needed.

If these consumers know only average environmental harm, the optimal tax instead subtracts morality times that average from each good’s actual damage. For sufficiently low-damage goods, this produces a subsidy: consumers otherwise overestimate their harm and buy too little. This remains a model result, not a recommendation to subsidize any particular product.

With intermediate information, the corresponding benchmark can require taxes conditioned on private signals the government cannot observe, making it unimplementable. With heterogeneous morality, one uniform charge cannot generally align everyone’s choices with the efficient allocation.

Mixed awareness creates another mismatch. For perfectly informed consumers sharing a positive, non-extreme degree of morality, the optimal tax lies between the wholly aware and wholly unaware cases. Aware consumers then buy too much relative to the efficient benchmark, while unaware consumers buy too little.

Caps and information have qualified effects

For consumers who understand a binding allowance system, a sufficiently stringent cap admits an equilibrium where everyone buys according to valuation versus price, regardless of morality. Collective restraint can affect allowance prices without changing capped emissions.

This result requires a cap at least as tight as the emissions level under Pigovian taxation, morality below the purely Kantian extreme, and uniformly returned allowance revenue. Among those caps, the best reproduces the Pigovian allocation. The analysis does not establish that actual emissions-trading systems always neutralize voluntary action.

Nor do environmental labels necessarily reduce harm. In a specified no-tax setting, better information increases consumption, while its environmental consequences can go either way.

What remains uncertain

The baseline assumes equally shared damage, independent valuations and environmental costs, frictionless rebates and a planner who knows actual damage. Welfare includes moral preferences evaluated using realized outcomes; that choice matters under heterogeneous morality. Extensions address continuous demand and convex damage, but not a full household budget with multiple goods and income effects. No new consumer experiment tests the mechanisms. The contribution is a conditional policy argument: moral motivation and policy understanding must be analyzed together.