Modelling new pricing strategies for the Santiago Metro

After twenty years of operation the Santiago Metro decided to start a new strategy of price differences by period in order to decrease congestion in the morning peak and to postpone the acquisition of new equipment. In this paper we report the results of a model specifically designed to test the impact of different price levels on patronage by period. Stated preference (SP) data was collected, focusing on three aspects: change in time of travel, price differences and comfort improvements. A goods-leisure microeconomics framework was adopted to decide on users segmentation and model specification. Two methodological aspects were explored, the non-linear marginal valuation of time displacements and the different known ways to analyse SP rating experiments. In this quest, models were calibrated both with a standard and an optimal probabilistic interpretation of the semantic scale and with the ordinal probit and binary logit models. Comparisons were made and conclusions reached about the more appropriate approach.Finally, the current time-of-day pricing strategy of the Santiago Metro, which commenced operation in February 1994, was simulated using the best models reported in the paper. This was done in order to test their validity against an observed reality.