In the past few years, artificial intelligence (AI) has sparked a range of emotions ranging from euphoria to dread. AI is often the scapegoat for various societal and economic problems. A recent lightning rod is algorithmic pricing. San Diego is the latest city blaming AI for high rents. San Diego City Council President Sean Elo-Rivera recently proposed a ban on “algorithmic price fixing.” That followed San Francisco, which in September became the first city to outlaw “algorithmic devices” in setting rents. They seem to think a berserk computer from “The Terminator” or “The Matrix” movies has sparked recent increases.
In the US, housing prices, like prices for other products, derive largely from market factors. Price levels are affected by supply and demand, and particularly by inequities or skewed distribution in wealth and property ownership. For rental prices, in particular, a strong influence is the market consolidation in rental properties. Over 225,000 homes are owned by 3 large companies: Invitation Homes, Progress Residential, and Blackstone. Similarly, 3 companies – Equity Residential, AvalonBay, and Camden Property Trust – own about 225,000 units across 770 properties. Crucially, these are concentrated in prime areas.
Price increases (or reductions) are usually a response to market dynamics. For instance, the covid19 pandemic led to rental price reduction in certain areas such as Silicon Valley and New York City, due to an outflux of residents from previously hot property markets in proximity of strong employment locations. Similarly, a large influx of new companies or new jobs can often lead to price increases, for instance in areas that have become popular for natural gas extraction via hydraulic fracture (“fracking”). Reversal of these short-term shocks can, in return, reverse those price changes.
In areas where rental prices are high, or where they have increased rapidly, the primary contributing factors therefore are general economic conditions (e.g., the very high inflation and interest rates in the early 2020s), shortfall in housing supply, and the market power of property owners, either because of high demand (e.g., in strong employment areas like Silicon Valley) or due to concentrated ownership in specific areas. How they set prices or the types of algorithms they use does not have a strong contributing effect.
Policy makers have recognized the role of market factors, and in some cases, it is appropriate to introduce economic regulations to protect renters. In California, the Tenant Protection Act restricts annual price increases, using an inflation-linked formula. For San Diego in particular, there is an 8.6% rent increase cap in 2025, and thanks to new tenant protection laws that went into effect on January 1, San Diego’s renters have more federal protection than ever before. However, some of the city’s lawmakers are still not satisfied. But, pointing the dagger towards the method of price calculation (i.e., algorithms) is inappropriate and unwise.
For any product, price levels must reflect market realities. There is always some method, some formula (or “algorithm”) that factors in these realities to compute prices. As various types of market data have become digitally available in recent years, price-setters can often employ machine-learning (or AI) algorithms that are more automated and require less human input. The concern with algorithmic pricing is that algorithms that learn from market data can tacitly collude and set supra-high prices through some sort of reward-punishment strategies.
However the evidence for algorithmic price collusion is thin, often based on experiments rather than real-world factors, and lacks proof. The White House Council on Economic Advisers’ research paper on the effect of pricing algorithms estimates that price algorithms imposed a cost of $3.8 billion on renters (which amounts to an average of $85 a year across 44.5 million rental units), and even this estimate has several caveats including several simplifying assumptions and limited data.
In reality, algorithmic price collusion is hard to sustain because it does not capture market competition factors well: too-high prices are not in the seller’s interest (they reduce sales) and will not persist without coordination (other laws exist to deal with price coordination).
Certainly, algorithmically-set prices will occasionally be super-high (e.g., when two competing sellers set their prices as proportional to the competitor’s price), just like they can sometimes be super low (airlines seats that have sold for pennies on the dollar). But these are mere aberrations. Moreover, such collusion is more relevant for items with fast moving prices such as airline seats, hotel rooms, or some products on e-commerce marketplaces, vs. apartment prices that generally change once per year.
Besides setting prices, technology platforms that use AI can help property owners better estimate the value of their homes and units. But this software also benefits the renter, who gets a clear offer on a unit based on market rates. And it helps the landlord, who is more likely to make a fair profit, which then is used to invest in improving the property and constructing new properties (increasing supply), which in time reduces prices. In November, rents dropped 0.7% nationally from the year prior, are at 6.2% below the August 2022 levels. Across the country, changes ranged from a 10.6% increase in Cleveland to a 12.4% decrease in Austin (see Rent.com’s November 2024 Rent Report “Rents Fall as Affordability Improves.”).
Blaming artificial intelligence algorithms for apartment rent levels is flogging the wrong horse. It can also dilute attention from other underlying factors such as market concentration in ownership, inflation, and the need to build – and fairly allocate – new housing. Moreover, unnecessarily blaming AI, or banning AI algorithms for rents, can lead to similar bans in many other fields and cause harm to California and US economy. The US is in a vicious AI competition with China, and artificial restrictions on AI applications here will merely shift power away from US (and its AI epicenter, California) towards China.
Copyright © 2025 California Business Journal. All Rights Reserved.
For California Business Journal Disclaimers, go to https://calbizjournal.com/terms-conditions/.
