Consumers May Emerge as Collateral Damage when Lawmakers Target Algorithmic Pricing
As persistent price inflation of basic household goods continues to stress U.S. consumers’ budgets, one idea politicians have seized upon is that the algorithms various companies use to set and update prices are somehow to blame. In particular, lawmakers and regulators are looking to curtail the use of consumer data to offer personalized algorithmic pricing, frequently branded “surveillance pricing” by opponents. Unfortunately, these concerns have led to legislative and regulatory proposals that, in many cases, have the potential to harm consumers—all in order to prevent worst-case pricing practices that are (so far) more imaginary than real.
State of Play
California passed a ban on “common pricing algorithms” last year. Now, in 2026, several dozen states have considered legislation banning or regulating surveillance pricing. Four states have passed bills of varying scope. These concerns have reached federal lawmakers as well, with a Senate hearing titled “The Consumer Cost of AI Surveillance Pricing” and new enforcement guidelines for personalized pricing proposed by the Federal Trade Commission (FTC).
California’s nation-leading bill, AB 325 (the Cartwright Act) was written so vaguely that one policy analyst commented, “[I]t is entirely unclear whether AB 325 accidentally regulate[s] effectively all market transactions.” Most of the bills passed this year are narrower in scope and at least attempt to refrain from banning algorithmic pricing practices that benefit consumers. New York opted to start by passing mandatory disclosures for how personal data is used in algorithmic pricing, while a bill to ban surveillance pricing outright has passed the legislature but remains stalled. Connecticut enacted a similar ban, while Colorado passed a personalized pricing ban ultimately vetoed by Gov. Jared Polis. Meanwhile, New Jersey and Maryland banned personalized pricing for grocery retailers and food delivery only.
These bills vary; however, Polis identified a common flaw amongst them in his veto statement, expressing concern about “discouraging perfectly acceptable uses of technology to set an appropriate price or wage, or the use of technology to save consumers money through discounts that may not exactly fit within the bill’s definition of an acceptable discount.”
Poorly Defined Terms Lead to Bad Policy
Public discourse surrounding the use of data and algorithms in setting prices has been confused by a series of terms used loosely and interchangeably. “Algorithmic pricing” can encompass any use of automated calculations that adjust a company’s prices. “Dynamic pricing” refers to adjusting prices in response to constantly changing economic inputs, often done in real time via price adjustment algorithms. Meanwhile, personalized pricing (or “surveillance pricing”) refers to using consumer data to offer customized prices to certain individuals or groups.
The term “surveillance” adds a connotation of menace to what is otherwise the banal practice of monitoring consumers’ shopping behavior. Tracking consumers’ purchase habits, whether online or in a physical retail setting, helps retailers figure out what to stock, how customers shop, and how sales and offers can entice customers to purchase products. Using this data to provide personalized offers to certain customers—whether by group (e.g., membership discounts) or by individual (e.g., personalized coupon offerings, rewards programs)—is hardly a new practice.
Opponents seem to envision a particular worst-case use for personalized pricing in which a retailer uses granular data about individual consumers to raise the price of certain goods to the highest amount each individual is willing or able to pay. Consumers would certainly find that level of individual price gouging annoying, and it would create an easy market for competitors to advertise their consistently low prices. Nevertheless, many of the policy proposals advancing across the country appear to start from the presumption that personalized pricing amounts to a form of exploitation that consumers cannot escape without government intervention.
Unclear Economics Surrounding Personalized Pricing
Not all algorithmic pricing is personalized, and many personalized pricing tactics actually benefit consumer welfare. For example, depending on how it is implemented, personalized pricing can particularly benefit lower-income consumers. Even the economic studies cited by the FTC show that whether individualized pricing favors consumers or companies depends a great deal on the details, which ought to recommend against sweeping regulations of personalized pricing writ large.
In fairness, it is also possible to use algorithmic pricing for genuinely anticompetitive collusion or illegally discriminatory shifting of costs. But it is important to consider that, as a Mercatus Center study observes, “[T]he pricing algorithms themselves do not cause collusion—they are merely neutral tools that can facilitate it.” Consumer protection and antitrust laws already exist to prosecute retailers who use algorithmic pricing to collude to raise prices or to freeze out competition if those activities prove harmful to consumers. Similarly, no new law is needed to target businesses that claim to offer one standard price while secretly employing individualized pricing.
As mentioned earlier, common business tactics to attract customer loyalty (e.g., rewards points, membership discounts, birthday rewards, customized coupons) are examples of personalized or individualized pricing that few would regard as predatory or unfair. Yet several bills that ban surveillance pricing struggle to find a definition that does not also encompass these practices—to the point where many bills simply resort to exempting these programs one by one.
Questionable Benefits and Unintended Consequences
Online retailers have used algorithms to adjust their prices in real time for many years, frequently giving them an advantage over physical retail stores that still rely on manual adjustments to physical price tags. On large online retail platforms like Amazon or Walmart, price adjustments can only be done viably by algorithm, with fluctuations delivering either savings or increases in the same way the local grocery store’s prices adjust to reflect supplier costs. The recent introduction of electronic shelf labels (ESLs), which can be updated in real time, offers physical retail stores a way to keep up with their online competitors. But anxieties about how ESLs could theoretically be used to adjust prices on an individual basis have placed them in the crosshairs of these algorithmic pricing bans. For example, New Jersey’s surveillance pricing ban places a temporary moratorium on ESLs until a state study determines whether to recommend a longer-term ban.
This is just one example of how worst-case thinking can lead to delays or bans on pricing practices that could prove beneficial. Even when personalized pricing bans include carve-outs for obviously beneficial practices like loyalty rewards programs, senior discounts, or personalized coupons, any new ideas not specifically enumerated would likely require companies to obtain permission before trying them out.
This mother-may-I approach to personalized pricing is made worse by the increasing number of individual states and municipalities seeking to regulate algorithmic pricing. This results in yet another nationwide regulatory patchwork that will increase compliance costs for regional and national retailers. The difficulty navigating such an environment favors large incumbents while discouraging innovation.
Disclosure mandates are not necessarily helpful or effective as an alternative to outright bans on personalized pricing policy either. For example, New York’s new disclosure mandate requires any retailers using algorithmic price inputs to display a disclosure warning consumers “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA.” At the federal level, the FTC’s new proposed guidelines contemplate labeling any use of personalized pricing as an “unfair or deceptive practice” unless companies disclose an incredibly granular level of detail regarding how personal data was used to inform the price offered. New York’s law will likely produce an obnoxious number of disclosure signs and pop-ups, and it is potentially impossible for any business to comply with the FTC proposal.
Conclusion
As R Street recently commented to the FTC, policies aiming to curtail personalized pricing—even those intended to increase transparency for consumers—are imperfect solutions to a largely hypothetical problem. Existing consumer protection laws should largely suffice in the event that companies use algorithmic pricing to deceive consumers outright. Concerns about data collection and use by companies ought to be addressed as a discrete issue via a federal data privacy framework. And to the extent that state and federal officials feel the need to act, regulations should target specific, defined consumer harms rather than algorithmic or personalized pricing in general.