Amazon is quietly engineering a profound shift in how cities and school districts buy routine supplies. The company has persuaded thousands of local governments to abandon competitive bidding and fixed-price contracts — the core safeguards of public-sector purchasing. In their place, agencies have signed agreements with Amazon Business, the company’s corporate- and government-facing platform, that specify dynamic pricing. As one agreement puts it, “This contract will not need to be amended when prices fluctuate.”As a result, cities and school districts that rely on Amazon pay wildly varying prices for the same items, even on the same day. On January 10, 2023, the Pittsburgh Schools bought two cases of Kleenex for $57.99 each; that same day the Denver Schools paid just $36.91 for a single case. On September 15 that year, two Iowa City, Iowa, school employees each ordered identical cartons of FritoLay snacks. One paid $26.22, the other $34.31.

My colleagues and I uncovered thousands of similar examples as part of an investigation into Amazon’s growing role in public purchasing. We analyzed transaction records from 128 cities, counties, and school districts, reviewed contract documents, and interviewed public officials and vendors. The findings are detailed in a new report, Turning Public Money into Amazon’s Profits: The Hidden Cost of Ceding Government Procurement to a Monopoly Gatekeeper.

We estimate that local governments spent $2.2 billion with Amazon Business in 2023 — nearly quadruple what they spent (after adjusting for inflation) in 2016.

For the vast majority of this spending, what governments pay on Amazon Business is determined by whatever Amazon’s algorithms decide to charge at the moment a public employee clicks “buy.” Because these prices are constantly shifting and Amazon provides no transparency into how they are set, public agencies cannot readily discern when they’re being overcharged.

While it’s impossible to know how much Amazon’s pricing systems may be inflating costs, the evidence suggests the impact could be substantial. We found that the Denver Public Schools, which spent $5.7 million with Amazon in 2023, could have saved 17 percent — roughly $1 million — had the district consistently received the lowest prices observed in our dataset.

How has Amazon persuaded local officials to sign up for opaque and fluctuating pricing? Its pitch is that its marketplace provides all the competition needed. “Amazon Business makes it easy for public sector customers to find the price competition they need — in a marketplace that shows offers from multiple sellers,” the company tells governments. Many officials have accepted this idea.

But while Amazon’s platform might seem to have the hallmarks of genuine competition, it is, in fact, only a semblance of it.  Its seller fees — which average 45 percent of each sale in the U.S. — impose high costs that sellers must pass along. And its own pricing algorithms, along with the automated repricing tools used by many third-party sellers, are built to maximize profits.

Recent academic work shows that multiple pricing algorithms in an online market can interact in ways that drive prices up rather than down, especially when one firm dominates the market. With the ability to adjust at lightning speed, a firm like Amazon can instantly undercut any competitor that lowers its price, teaching rivals that discounting will not win them additional market share. Sellers learn to keep prices elevated.

The contrast with established procurement norms could not be more stark. Traditionally, local governments solicit competitive bids. To win a contract, suppliers compete to offer the lowest prices they can in exchange for a multi-year commitment. This process not only locks in low prices and appropriate volume discounts but also provides transparency: Bid documents are public records, allowing anyone to see how a contract was awarded and at what prices.

As Amazon’s reach grows, the small and mid-sized businesses that have long supplied schools, cities, and other public institutions are disappearing, along with the local jobs and tax revenue they create. Over the last decade, the number of independent suppliers of office products and janitorial goods has fallen from 1,300 to 900. Contrary to common assumptions, we found that these suppliers often beat Amazon on price and delivery speed. As their numbers shrink, Amazon’s leverage over schools and cities will only intensify.

Two policy steps are essential. First, cities must restore competition to their procurement systems and deliberately seek out local suppliers, who are a crucial source of market diversity in an increasingly consolidated landscape.

Second, states must ban dynamic pricing in public procurement and require all contracts signed by local governments and state agencies to specify fixed prices. Without such safeguards, public institutions will become increasingly beholden on an opaque, algorithmic pricing regime designed to benefit a single dominant company at the expense of taxpayers.