The Downs–Thomson paradox (named after Anthony Downs and John Michael Thomson), also known as the Pigou–Knight–Downs paradox (after Arthur Cecil Pigou and Frank Knight), states that the equilibrium speed of car traffic on a road network is determined by the average door-to-door speed of equivalent journeys taken by public transport.
It is a paradox in that improvements in the road network will not reduce traffic congestion. In fact, improvements in the road network can make congestion worse if the improvements make public transport more inconvenient or if it shifts investment, causing disinvestment in the public transport system.
The general conclusion, if the paradox applies, is that expanding a road system as a remedy to congestion is ineffective and often even counterproductive. That is known as Lewis–Mogridge Position and was extensively documented by Martin Mogridge in the case study of London on his book Travel in towns: jam yesterday, jam today and jam tomorrow?
A 1968 article by Dietrich Braess pointed out the existence of the counterintuitive occurrence on networks: the Braess' paradox states that adding extra capacity to a network, when the moving entities selfishly choose their route, can in some cases, reduce overall performance.
There is a recent interest in the study of this phenomenon since the same may happen in computer networks as well as traffic networks. Increasing the size of the network is characterized by behaviors of users similar to that of travelers on transportation networks, who act independently and in a decentralized manner in choosing optimal routes between origin and destination is an extension of the induced demand theory and consistent with Downs' 1992 theory of "triple convergence", formulated to explain the difficulty of removing peak congestion from highways. In response to a capacity addition three immediate effects occur: drivers using alternative routes begin to use the expanded highway; those previously traveling at offpeak times (either immediately before or after the peak) shift to the peak (rescheduling behavior as defined previously) and public transport users shift to driving.