Public Transit Capacity and Users' Choice: An
Experiment on Downs-Thomson Paradox
Résumé
We study the Downs-Thomson paradox, a situation where an additional road capacity
can cause an overall increase in transport generalized cost and therefore a decrease in welfare
for transport users. To this end, we build an experiment based on a double market-entry
game (DMEG) where users have to choose between road and public transit after that the op-
erator has choosen public transit capacity. The optimal strategy for operator is to minimize
capacity, and the equilibrium for users depend on the endogeneous public transit capacity
compared to exogeneous road capacity. The most important result is that we observe the
Downs-Thomson paradox empirically in the laboratory: An increase in road capacity causes
shift from road to rail and, at the end, increases total travel costs. But the contrary is
not true: A decrease in road capacity does not cause lower total travel costs, which is in
contradiction with our theoretical model. Results also show that the capacity chosen by
operator di¤ers from Nash prediction, levels being signi cantly higher than those predicted
by our model. Moreover, users coordinate remarkably well on Nash equilibrium entry rate
while capacity has been chosen by operator.
can cause an overall increase in transport generalized cost and therefore a decrease in welfare
for transport users. To this end, we build an experiment based on a double market-entry
game (DMEG) where users have to choose between road and public transit after that the op-
erator has choosen public transit capacity. The optimal strategy for operator is to minimize
capacity, and the equilibrium for users depend on the endogeneous public transit capacity
compared to exogeneous road capacity. The most important result is that we observe the
Downs-Thomson paradox empirically in the laboratory: An increase in road capacity causes
shift from road to rail and, at the end, increases total travel costs. But the contrary is
not true: A decrease in road capacity does not cause lower total travel costs, which is in
contradiction with our theoretical model. Results also show that the capacity chosen by
operator di¤ers from Nash prediction, levels being signi cantly higher than those predicted
by our model. Moreover, users coordinate remarkably well on Nash equilibrium entry rate
while capacity has been chosen by operator.
Origine :
Fichiers produits par l'(les) auteur(s)
Loading...