New Math Proof: Two Extra Sellers Can Make Any Market Run Perfectly
Just two more traders makes a marketplace as good as one that knows everything.
WHAT HAPPENED: A team of eight computer scientists and economists proved something surprisingly tidy about two-sided markets — places like eBay, Uber, or a stock exchange where buyers and sellers are matched up. If you add just two more participants to whichever side is smaller, a very simple pricing rule suddenly performs as well as a hypothetical all-knowing auctioneer who could see inside everyone's head. Add only one extra trader, they show, and it falls apart.
WHY IT'S HARD: In real markets, nobody knows what a buyer is truly willing to pay or what a seller truly needs. Designing a rule that gets good outcomes anyway is the central puzzle of this field. The famous trick here — the "Bulow-Klemperer" result — says that in one-sided auctions, extra competition can substitute for knowing everyone's private values. This paper extends that idea to two-sided markets, where buyers and sellers both have secrets. The winning rule, "Trade Reduction," roughly means: let the obvious deals happen, then shave off the most expensive one so the platform doesn't lose money.
WHY YOU MIGHT CARE: Platforms spend fortunes on clever algorithms to set prices and match trades. This proof says something cheaper works: recruit a couple more people on the thin side. For a farmer's market with three sellers and twenty buyers, or a local gig app with too few drivers, that's a concrete, low-cost lever. It also sets a hard floor — you genuinely cannot do it with one, no matter how clever the rule. So the answer to "how much extra competition do we need?" is a clean, memorable two.
THE CATCH: This is pure theory. It assumes buyers and sellers are drawn from stable, well-behaved statistical populations, and that buyers generally value the goods more than sellers do — a big assumption. Real marketplaces are messy, people behave strategically, and platforms often have their own fees and rules that break the math. No app will change next month. But regulators and marketplace designers now have a sharp new benchmark to argue about.
- Adding just two extra participants to the smaller side of a market recovers the ideal outcome — one extra is provably not enough.
- The winning rule needs no knowledge of anyone's private valuations, so it's cheap and untuned to set up.
- It's a theory result with strict statistical assumptions, so treat it as a design guide, not a product launch.
Why It Matters
Thin marketplaces — few sellers, few drivers — may be fixed cheaply by recruiting just two more participants.