Seoul stopped paying buses per passenger and the race for fares vanished

Seoul's buses before 2004 were run by private companies that kept the fares they collected — which made every bus a competitor of every other bus at the same stop. The incentive structure produced exactly what it rewarded: drivers racing to reach passengers first, dangerous driving and high accident rates, oversupply on profitable corridors, no service on unprofitable ones, and firms sliding toward insolvency as the subway took riders. Service quality was the residual of a fare-grabbing contest.

The conventional fixes — stricter enforcement, route licensing, subsidising losses firm by firm — all left the core incentive intact: revenue arrived through the farebox, so behaviour optimised for the farebox.

On 1 July 2004 Seoul rewired who gets paid for what, in one stroke: the city pooled all fare revenue and paid operators for service delivered — kilometres run to standard on their contracted routes — not for passengers carried, keeping ownership private but operation quasi-public. With revenue decoupled from ridership, the reform bundle became enforceable: routes reorganised into trunk-and-feeder with median bus lanes, fares fully integrated with the subway through the T-money smart card (transfers free), and service allocated by network need rather than profitability.

Pooled all fare revenue and paid operators per service-kilometre run to standard, not per passenger, so the race for fares had no prize and compliance became the income.

Payment per passenger made racing the direct path to income, so safety rules fought the operator's wallet.

Pooling fares and paying per service-km made compliance the income, aligning operators with the network.

With revenue decoupled from ridership, free transfers and feeder routes became affordable, improving system design.

The reform required auditing service standards, which Seoul did with GPS and smart-card data, keeping the standard real.

When agents chase revenue, change what revenue pays for: pay for delivered service to standard, not per sale, to kill pathological races at the root.

Bus accidents fell dramatically, speeds on median-lane corridors jumped, ridership and satisfaction rose — at the price of a standing operating subsidy (~$270M/yr initially). Became the reference case for gross-cost bus reform — studied via UN-Habitat and World Bank, and copied in outline by cities across Asia and Latin America.