Dynamic pricing, efficiency, and the second-best problem
Dynamic pricing improves allocation efficiency but can worsen total welfare when production externalities remain unpriced.
Instacart is under FTC investigation for dynamic pricing. That's the headline. It won't be the last firm to experiment with individualized pricing, and it's not the most interesting part of the story.
The more interesting question is what happens when dynamic pricing pushes markets toward allocation efficiency in sectors with unpriced negative externalities.
Economics 101 tells us that dynamic pricing, particularly pricing based on an individual's willingness to pay, reduces deadweight loss. Transactions that fail under uniform pricing now clear. Output rises. In a narrow, private-market sense, this is efficient.
But allocation efficiency is not the same thing as socially optimal outcomes.
That distinction matters because many real-world markets are not missing efficiency; they are missing prices. Specifically, they are missing prices for external costs imposed on third parties: pollution, congestion, emissions, environmental damage. When those costs are unpriced, private markets systematically overproduce relative to the social optimum.
In those settings, deadweight loss can play an unexpected role.
The second-best problem
This is a classic second-best problem. In the presence of one distortion (unpriced externalities), the introduction of a second distortion (deadweight loss from uniform pricing) can partially offset the welfare loss from the first. Removing the second distortion without correcting the first can worsen total welfare. The "efficient" outcome becomes inferior to the "inefficient" one.
Under uniform pricing, some marginal transactions fail to occur not because their private benefit is zero, but because price does not reflect true social cost. Fewer units are produced, fewer transactions clear, and output remains below the private optimum. From a textbook perspective, this is inefficient. From a welfare perspective, it can function as an accidental constraint on socially harmful production.
Some welfare loss is avoided by accident rather than by design.
This is a failure of policy creating an accidental second-best outcome, not a feature to celebrate. Like a broken speedometer that reads low and prevents speeding tickets. Technically helpful, but you'd still rather fix the speedometer than rely on the defect. (Or, just drive slower ...)
Perfect or near-perfect price discrimination removes that constraint. When firms can extract each buyer's full willingness to pay, marginal transactions that previously failed now clear. Production expands to serve buyers whose private benefit is positive, even when the social cost of that production exceeds that benefit.
Output moves closer to the private optimum, and further from the social optimum.
Capacity vs. new production
Whether this matters in practice depends on what type of production dynamic pricing enables. The welfare impact is not uniform across contexts.
When dynamic pricing fills existing capacity (e.g., an airline seat that would fly empty, a rideshare vehicle already on the road) it spreads fixed externalities across more transactions without creating new harm. Emissions per passenger decline. This is likely welfare-improving.
When dynamic pricing enables new marginal production (e.g., additional flights, more delivery routes, expanded manufacturing runs) it creates transactions where private benefit exceeds private cost but falls short of social cost. This is likely welfare-reducing.
The net effect is an empirical question, not a theoretical certainty. But in sectors where marginal production is elastic and externalities are high, the pattern appears repeatedly.
Airlines fill marginal seats, making additional routes and flights viable, increasing emissions. Delivery platforms serve marginal demand, increasing miles driven, congestion, and local pollution. Pollution-intensive industries expand output to price-sensitive buyers, increasing discharge and environmental damage.
In each case, private surplus rises. Measured efficiency improves. Total welfare may not.
The claim that dynamic pricing materially increases externality-generating production is theoretically sound but empirically unverified. This is conjecture worth testing, not established fact. The magnitude of the effect (whether we're discussing a 2% increase in emissions or 20%) remains unknown. But the directional logic holds: when externalities are unpriced, pricing mechanisms that push output toward the private optimum intensify the gap between private and social outcomes.
Scale and automation change the stakes
What makes this relevant now is scale and automation. Algorithmic pricing powered by machine learning makes individualized WTP extraction feasible across entire sectors in real time.
Not as an occasional experiment, but as the default operating mode.
What was once a theoretical pricing strategy is now deployable infrastructure. The welfare implications aren't new. The velocity and reach are.
Dynamic pricing removes the accidental brake.
We get allocation efficiency in the private market while intensifying the very externality problem that pricing theory assumes away.
This outcome might follow logically. That doesn't make it desirable.
Importantly, this is not a story about irresponsible firms or bad corporate behavior. Firms that price dynamically based on revealed willingness to pay are doing exactly what competitive pressure and fiduciary logic demand. There is no realistic expectation that firms will voluntarily leave surplus on the table to protect social welfare, nor should public policy be designed on that assumption.
Dynamic pricing doesn't create the welfare problem. It exposes it.
By accelerating output toward the private optimum, individualized pricing makes the cost of missing externality prices visible. It removes the camouflage provided by market frictions and forces the underlying misalignment into the open.
First-best vs. second-best responses
In a first-best world, the response would be straightforward: price the externalities directly. Carbon pricing, congestion pricing, pollution taxes. Align private incentives with social costs and let markets do the rest.
But in the real world, where externality pricing remains politically constrained, adjustment occurs through second-best mechanisms. When governments cannot price the harm, they increasingly regulate the behavior that amplifies it.
Absent credible externality pricing, pressure will shift toward constraining pricing mechanisms themselves. Not necessarily because they are inefficient, but because they reveal inefficiencies elsewhere.
What does that constraint look like? Price caps that limit discrimination? Algorithmic transparency requirements? Sector-specific bans on individualized pricing in high-externality industries? Each creates second-order distortions.
Caps reduce firm incentive to serve marginal customers. Transparency invites coordination. Bans push discrimination underground or into adjacent unregulated behaviors. None are first-best. All are predictable responses when first-best policy remains off the table.
The distributional question
There is a separate distributional problem worth naming explicitly.
Price discrimination extracts surplus from consumers with high willingness to pay—often those who are desperate, time-constrained, or informationally disadvantaged. Willingness to pay is not the same as ability to pay without harm; it can reflect constrained choice rather than surplus wealth.
A parent ordering groceries for a sick child pays more than a leisure shopper. A business traveler pays more than a vacationer. Those with fewer options pay more than those with flexibility.
This is wealth transfer, not welfare loss in the strict economic sense. Total surplus rises. But the distribution shifts systematically toward firms and away from consumers least able to avoid extraction. Allocation efficiency improves. Equity does not. Whether that trade-off is acceptable is a political question, not an economic one. But it is a question dynamic pricing forces into the open.
Instacart just happens to be today's case study.
Sarah Holder