Regulatory Limits on Personalized Pricing

The FTC has proposed a new rule governing personalized pricing, the increasingly feasible practice of using information about an individual customer to determine the price they are offered. Importantly, the FTC does not have authority to ban personalized pricing. Instead, sellers would have to disclose when personal information was used to determine a customer's price. The economics is familiar: by estimating each customer's willingness to pay, firms can practice direct price discrimination, charging higher prices to customers willing to pay more and lower prices to those who otherwise might not buy. This allows firms to capture more consumer surplus, but it can also increase sales by bringing low-willingness-to-pay customers into the market.

In some ways, personalized pricing takes us back to the nineteenth century. Before price tags became common, once a customer had selected some items at the general store, the clerk and the patron would start haggling. Different customers could therefore pay different prices for the same item, depending partly on the merchant's assessment of their willingness to pay. The spread of department stores and mass retailing replaced this individualized bargaining with posted, take-it-or-leave-it prices. Digital technology may now be reversing that evolution. Instead of a nineteenth-century shopkeeper sizing up the customer across the counter, an algorithm can use browsing histories, locations, previous purchases, and other data to more precisely size up willingness-to-pay. The FTC's proposal would still allow this twenty-first-century version of an old pricing strategy but would require firms to tell customers when it is happening.

How Low Should Seasonal Discounts Be?

Retailers often mark down seasonal merchandise to clear inventory, but how big should the discount be? An Impact Analytics article describes a clothing retailer that initially discounts winter coats by 20%, observes the resulting sales, and then increases the discount to 35% when inventory remains too high. This is a simple pricing decision in which lowering price sacrifices margin on units that would have sold anyway but generates additional sales. Keep cutting price as long as the marginal benefit from the additional sales exceeds the marginal cost from the lower price.

The pricing calculation can must compare benefits to costs. Selling another coat frees shelf and warehouse space, releases capital tied up in inventory, and avoids being stuck with merchandise that will be worth even less when the season ends. On the other hand, deeper markdowns may damage the brand or teach customers to wait for discounts. Include all the consequences of changing price, including possibly hidden changes in costs. The profit-maximizing markdown isn't the one that clears the shelves; it is the one where the benefit of cutting the price a little further no longer exceeds the cost.

Compensating differentials outlawed in Birmingham

 From MarginalRevolution:

In 2010 an employment tribunal ruled that Birmingham City Council had discriminated against thousands (6,000) of female workers — cooks, cleaners, care assistants, caretakers — who were denied bonuses paid to the mostly male binmen, gardeners, and gravediggers (400). Why were the binmen given bonuses? Well, refuse collection is filthy, heavy, outdoor work and not many people want to be gravediggers.

Birmingham went bankrupt when they had to pay the bonuses to the 6,000 women and they went bankrupt.  Then they eliminated the bonuses paid to the men, who then went on strike.  Then:

The council hired an outside contractor to take over its rubbish collection and it now pays roughly triple its pre-strike outsourcing bill..

A Pac-Man Defense Against Predatory Pricing?

After the story of how Southwest Airlines broke an attempt at predatory pricing with booze, I learned of the Dow vs. Die Deutsche Bromkonvention episode. See here for an interesting narrative of events. To summarize, at the turn of the twentieth century, the German chemical industry was dominant and the various producers there had a comfortable cartel selling bromine at 49 cents a pound. Herbert Dow had developed a way to produce bromine and sold it in the US for 36 cents. The Germans might have been OK if he stayed in the US, but the cartel swung into action when he began exporting in 1904. 

It poured bromine into America at 15 cents a pound, well below its fixed price of 49 cents, and also below Dow’s 36 cent price.  

Dow simply pulled out of the US market and sold abroad. Moreover, he had his agents secretly buy up hundreds of thousands of pounds at that 15 cents, repackage it, and resell it in Europe at 27 cents. Instead of being driven out of business, he profited from the cartel's actions! At first the Germans were confused by the insatiable demand in the US and the lost orders from Europe. Was one of their cartel members offering secret discounts? Accusations were made. Even after they discovered Dow's ploy, they were not sure how to respond. In the end, the bromine monopoly was broken.

Perhaps Dow got lucky, but perhaps predatory pricing may not be so easy.

The Cost of Going to Space

Terzi & Nicoli have a fascinating new paper analyzing over 4,400 orbital launches since 1960 to estimate "Wright's Law." This law states that a new technology's unit costs fall in proportion to cumulative experience. Essentially, it is an expression of learning-by-doing. They show:

... that the average cost of sending a kilogram to orbit has dropped from 87,023 USD in 1960 to 3,868 USD in 2025. Using a Wright's Law framework, we estimate that for each doubling of cumulative payload to orbit, the average cost of sending a kilogram to orbit decreases by 21.2% ...

This is faster than many past technologies e.g., 19th century Steam engines and recent solar photovoltaics. There is a positive feedback loop because, as launch costs fall, more stuff is launched. Their forecast is that it will cost reach $300/kg by 2040. For reference, sending a kg from LA to London by overnight express costs $5-$50/kg depending on package size. I get giddy trying to imagine all the potential new opportunities.

Incentive Misalignment in Communism: the case of Vietnam

Rainer Zitelmann writes in How Capitalism Beat Communism in Vietnam in Reason Magazine,

The state collectives did not reward members for the amount of rice they produced, but instead counted how many days they had worked. If you worked 30 days, you got 30 points, which gave you the right to a defined share of the harvest. If you worked 20 days, you got 20 points and correspondingly less.

Days worked is cheap to verify, hard to dispute, immune to weather. But it was not what the country needed, which was rice. Output was the right target, but no individual's contribution was measurable, so the Collectives graded attendance instead.

Farmers responded rationally — they showed up. Real effort migrated to private plots where they could keep what they made. While cooperatives absorbed 40 percent of state agricultural funding, over 60 percent of cooperative members' income came from the 5 percent of land left in private hands. In 1980 the country grew 14 million tons of rice against the 16 million it needed.

Private Property (under the Doi Moi reforms) fixed the problem: households leased land and kept what they produced. Vietnam is now among the world's largest rice exporters.

BOTTOM LINE: People will optimize whatever you count; and Private property aligns incentives and makes the proverbial pie bigger.  

Using Demand Signals to Re-Price

A CNBC story indicates that Starbucks repeatedly raised menu prices over the past several years. Initially, this strategy worked well. Higher prices more than offset modest declines in customer traffic, increasing average revenue per transaction. More recently, however, the company's quarterly earnings have shown declining customer visits, and new CEO Brian Niccol has shifted the company's emphasis away from broad price increases and toward improving service, simplifying the menu, and restoring the in-store experience, i.e. product differentiation. While many factors undoubtedly contributed to Starbucks' slowdown, the company's strategic pivot suggests that management believes customers have become increasingly willing to substitute away from Starbucks when prices rise. The available evidence is consistent with demand becoming more price elastic than it was only a few years ago.

Demand elasticity is not a fixed characteristic of a product or brand. Firms may consider premium brands' demand characteristics as permanent, but elasticity depends on the availability of substitutes and customers' willingness to switch. As time passes after a price increase, more consumers become aware that the price has changed, become more willing to experiment with competing brands, and discover acceptable alternatives. With better informed customers, future price increases can trigger much larger reductions in sales than before. The Starbucks experience is less a story about charging "too much" for coffee than about how repeated price increases may inadvertently encourage customers to search for substitutes, making demand progressively more elastic. Every price increase is also an experiment that provides information for future pricing decisions.

Specrtum Auctions Address Economies of Scope

The Federal Communications Commission (FCC) auction can demonstrate economies of scope. The FCC periodically auctions off additional spectrum licenses for advanced services. It recently raised $3.5 billion from recent AWS-3 (Advanced Wireless Services) auctions and expects to raise $30-$6 billion from upcoming Upper C-Band auctions. In telecom, these frequencies are not independent assets; their value is multiplicative. This additional spectrum will allow for greater capacity to handle the exponentially growing demand for wireless data transmission.

Spectrum auctions are usually run as Simultaneous Multiple-Round Auctions (SMRAs). US coverage is divided into many distinct geographic areas that are auctioned simultaneously across many rounds. An aggregation problem arises when a bidder risks winning fragmented licenses that are less valuable without their complementary counterparts. Since transmission has origin and a destination, it is more valuable to win a license in, say, Chicago if the operator can also secure one for Dallas. SMRAs allow bidders to cobble together desired regional capacity by aggregating multiple licenses.

The cell towers, base stations, and cables that link them are huge fixed costs. The ability to defray these costs over more volume reduces average costs. Economies of scale in these fixed costs generate economies of scope across these fixed costs in different regions.

Curating Complements: Apple & USB-C

One of the strongest arguments for industry standards is that they increase competition. A common charging connector means consumers can choose from hundreds of competing cable manufacturers instead of being locked into a single supplier. The resulting competition lowers prices and increases consumer welfare. When the EU required Apple to transition from its proprietary Lightning connector to USB-C in 2023, it illustrated that this logic, while generally sound, is incomplete. The WSJ reports that one independent repair specialist saw Apple iPhone failures associated with poorly designed or non-compliant USB-C cables and chargers. While these failures appear to be relatively uncommon, they highlight an economic tradeoff that is easy to overlook: greater competition among complementors can also increase quality variation. Under the Lightning ecosystem, Apple exercised tight control over accessory manufacturers through its Made for iPhone (MFi) certification program. Consumers paid more for certified cables, but they also purchased into a curated ecosystem with stronger quality assurance. By requiring Apple to adopt the industry-standard USB-C connector, regulators and policymakers may have increased competition, but they may also have exposed consumers to a much broader range of accessory quality.

Firms sometimes vertically integrate to solve coordination and quality-control problems, and not to exploit market power. Apple had strong incentives to ensure that its phones, chargers, cables, and software functioned reliably as an integrated system because failures reflected directly on the Apple brand, regardless of who manufactured the accessory. A proprietary standard gave Apple greater ability to police quality and exclude unreliable suppliers. This is likely the case for many of the Apple iPhone’s complements. An open standard like USB-C creates incentives for lower prices, interoperability, and reduced electronic waste. It also may have shifted more responsibility to consumers, who must distinguish between high-quality and low-quality accessories in a crowded marketplace. In this case, it is unclear if the pre-purchase consumer search costs and post-purchase repair costs were comparable to the competitive benefits from allowing open entry. But the episode demonstrates that a tradeoff could exist.

Can you Go to Jail if your AI Engages in Collusion?

The stereotype of price fixing is fat, cigar-chomping executives meeting secretly in smoke-filled rooms to agree to raise prices. But the Justice Department's recent consent decree with RealPage anticipates that AI collusion may look different. RealPage sold revenue-management software that used confidential pricing information from competing landlords to recommend rents. Although the software generated the recommendations, the DOJ alleged that the system facilitated unlawful coordination among competitors. I had my doubts. Nevertheless, the settlement requires RealPage to stop using certain competitively sensitive data and to change features of its pricing software.

AI can still recommend prices. Firms have used sophisticated pricing software for decades. A broader implication is that firms cannot avoid antitrust liability by outsourcing pricing decisions to an algorithm. If competing firms provide confidential information to a common AI system that helps coordinate pricing decisions, regulators may view the arrangement much like traditional collusion.

This distinction will become increasingly important as companies deploy AI agents to make autonomous business decisions. An AI pricing system that independently analyzes a firm's own costs, demand, and inventory is generally very different from one that relies on competitors' confidential information or otherwise facilitates coordination among rivals. The RealPage consent decree is consistent with antitrust law focusing on economic outcomes rather than how those outcomes were produced.