Skepticism toward algorithmic pricing is widespread, but both the economic theory and empirical evidence strongly support the expectation that algorithmic pricing is welfare-enhancing. Algorithmic pricing–whether the real‑time dynamic fares you see in a ride‑share app or the personalized discounts your grocery app quietly applies at checkout– builds on a century of research showing that matching supply and demand in real time improves efficiency and total welfare, while personalized pricing can expand output and raise welfare. As early as the 1980s, Hal Varian demonstrated that as long as a firm sells more units under personalized prices than it would under a single posted price, consumer welfare can rise alongside profits. He also offered a simple observational test: if output quantity goes up, total surplus is likely higher.
The underlying logic is straightforward: if a firm can charge only a single price to all consumers, and consumers’ willingness to pay varies, then it’s likely that some mostly poor consumers will be priced out of a market entirely by the single price, while other mostly affluent consumers will achieve windfalls by paying far less than they are willing to pay. On the other hand if customers can be charged different prices based on their willingness to pay, then it is possible to serve all consumers with a willingness to pay covering the marginal cost, meaning a larger share of poor consumers can consume in the market, while still increasing profits by charging higher prices to consumers with a higher willingness to pay. This basic logic goes back to coupon clipping and perhaps further, though algorithmic pricing tools in the connected economy now make it significantly cheaper to pass that textbook logic through to the real-world economy.
What exactly is “algorithmic pricing”?
We group the practice into two broad categories:
Dynamic pricing adjusts a price in response to changes in supply‑and‑demand conditions. Dynamic pricing is unambiguously more efficient and welfare enhancing than having prices that no longer match supply and demand, as it reduces the risk of shortages or wasted excess inventory. Short-term surge pricing and discounts on ride-sharing apps are recognizable examples of dynamic pricing, without which rider wait times would be extremely long during busy periods and many drivers would waste time with no fares in less-busy periods.
Personalized pricing offers different consumers different prices based on predictive signals about their willingness to pay. As a general rule, so long as personalized pricing is associated with an increase in output, it is increasing total welfare by allowing more consumers to make purchases in a market than they would be able to with a single price.
Forms of both dynamic pricing and personalized pricing preceded algorithmic pricing tools based on computer code. What is new with algorithmic pricing is not the theory, but the combination of cheap computational power, rich data, and intense digital‑era competition that allows prices to change at granular time intervals and more closely match individual budgets.
Economically, the key change is that more firms than ever before can now profitably serve buyers whose peak willingness-to-pay sits only pennies above marginal cost. Bringing those buyers into the market expands output quantity, trims deadweight loss, and—crucially—targets the benefits toward households who would otherwise be priced out.
Evidence from dynamic markets
Airlines: A 2022 National Bureau of Economic Research working paper on U.S. airlines finds that algorithmic pricing heuristics raise revenues 4‑5 percent and increase consumer surplus about 3 percent compared with a uniform optimal fare. Those gains are driven by cheaper tickets sold well in advance to leisure travelers, a group that skews lower‑income and more price‑sensitive than last‑minute business passengers.
Ride‑hailing: Juan Camilo Castillo’s 2024 paper on Uber surge pricing estimates that moving from a flat tariff to real‑time pricing raises total welfare by 2.15 percent of gross revenue. Rider surplus alone climbs 3.57 percent, while the platform’s profits and driver earnings fall slightly. Importantly, riders at all income levels benefit because periods of artificially long wait times are eliminated.
Electricity: Dynamic pricing based on the balance between power grid demand and generation supply shows similar patterns in a regulated electric utility setting. Ito, Ida, and Tanaka’s field experiment paper links opt‑in dynamic pricing to measurable welfare gains: price‑sensitive households—often those with tighter budgets—are more likely to enroll and reap larger savings, so aggregate welfare rises even after accounting for selection. A survey of 15 additional pilots confirms that residential customers cut usage during peak‑price periods without appreciable rebound, improving efficiency and lowering system costs.
Across these three very different markets—perishable seats on a plane, point‑to‑point urban transport, and kilowatt‑hours flowing over a grid—the empirical result is remarkably stable: algorithmic pricing reallocates demand toward slack capacity and away from scarcity, lowering effective prices for the most budget‑constrained consumers.
Personalized pricing: empirical nuance and redistributive upside
Because personalized prices are visible only to the buyer and the seller, they can feel mysterious—sometimes even unfair—to outsiders. Rigorous field experiments complicate that intuition. A randomized controlled trial at a large digital platform shows that more than 60 percent of individual consumers actually pay less under personalization. Who benefits? Primarily shoppers with lower valuations—those on tighter budgets or with less urgent needs—who would have walked away at a single posted price.
Armstrong and Vickers’ theoretical work on personalized pricing in competitive markets helps point to how competitive dynamics in the presence of personalized pricing can benefit consumer welfare. When rivals can identify your customers, they can target them with poaching discounts, forcing incumbents to match or lose the sale—an effect that pushes average prices down and service quality up. Dynamic competition therefore reallocates surplus toward the price‑sensitive segment.
Why low‑income households gain the most
Consider a single mother ordering groceries. Her shopping app knows she lives in a ZIP code with lower median income, has not purchased protein in two weeks, and historically responds to 25‑percent discounts but not 10‑percent coupons. An algorithm can offer just enough of a markdown—financed in part by a premium charged to a high‑earning professional—to make chicken breast affordable this week. Scale that logic across streaming subscriptions, prescription drugs, cloud storage, and weekend travel, and the aggregate inclusion effect is large.
Multiple empirical settings corroborate the intuition. The Uber study finds similar percentage welfare gains for riders regardless of neighborhood income, meaning the absolute dollar benefit is larger for poorer households that take fewer trips. The airline paper shows the biggest price drops on leisure itineraries disproportionately purchased by lower‑income families. And electricity pilots consistently report that lower‑usage, lower‑income customers save the most in percentage terms under peak‑pricing.
Competitive pressure dominates in practice
Popular narratives sometimes assume that algorithms enable silent collusion. Yet empirical studies of actual deployments regularly report either no price increase or modest price decreases. The NBER airline paper explicitly models competitive interaction and still finds positive surplus for consumers. A 2025 Mercatus Center review catalogues dozens of markets where personalized or surge pricing lowered effective prices for at least one segment without raising prices for all others.
Conclusion
The heart of the consumer‑welfare standard is whether a policy leaves people better off. Across airlines, ride‑sharing, retail, and electricity, algorithmic pricing has repeatedly expanded output, trimmed deadweight loss, and—most important—brought lower‑income households into markets they once could not afford. The scholarly record shows both theory and evidence pointing in the same direction: when firms can vary prices with precision and face vigorous rivalry, consumers—especially the price‑sensitive—win. Freezing every buyer into a single queue at a single price would not make the economy fairer; it would simply ration access by income and time, with the least‑resourced losing out first.
