Congestion tolling — Dollars versus tokens: Within-day dynamics. (October 2022)

• Record Type: Journal Article
• Authors: Seshadri, Ravi; de Palma, André; Ben-Akiva, Moshe
• Is Part Of: Transportation research. Volume 143(2022)
• Publication Date: 2022-10
• Subjects: D61 -- R41 -- R48; Tolls -- Tradable credit schemes -- Mobility permits -- Congestion -- Dynamic models -- Efficiency -- Equity; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2022.103836
• Languages: English
• ISSNs: 0968-090X

Abstract

Abstract: Tradable credit schemes (or tolling in tokens) are a form of quantity control, which promise to be an appealing alternative to congestion pricing (or tolling in dollars) owing to considerations of revenue neutrality, equity, reduced infrastructure costs, and political acceptability. The comparative performance of the two instruments under uncertainty in demand and supply has only recently received attention in the transportation setting, despite being widely studied for emission markets. In this paper, we add to this literature by considering a tradable credit scheme in a departure time context wherein users are provided an initial endowment of tokens by the regulator and incur a token charge (determined prior to all departures) to travel in a specific time-period. Tokens can be bought and sold within a marketplace at a price determined endogenously by token demand and supply. Two key features of the market model are: (1) the time-of-day dynamics of price is explicitly modeled through a smooth market clearing mechanism in each period, and (2) the selling decisions of users, which determine the distribution of token supply in the market over the day are explicitly modeled. This enables us to study the impact of selling behavior on performance of the credit system. Travel demand is modeled using a logit mixture model and supply consists of static congestion. Extensive experiments under stochastic demand show that when the tolls (in dollars and tokens) are not day-to-day adaptive, tolling in tokens outperforms tolling in dollars when congestion effects are more severe (e.g., realistic BPR models and steep congestion functions, high demand levels and high day-to-day variability). Importantly, we find that this result is robust with respect to selling behavior in the market, although there can be welfare losses in the quantity control system when selling behavior in the market is excessively irrational. These findings underscore the importance of examining disaggregate market behavior when designing tradable credit schemes. Moreover, when the supply of tokens can be adapted from day to day, the credit system was found to be superior in all tested scenarios, provided the selling behavior of individuals is rational. Finally, even in the case when toll revenues in the price instrument are equally redistributed (often difficult in practice), tolling in tokens (when tokens are equally distributed) is marginally more equitable in scenarios where congestion effects are more severe. These findings make a case for tolling in tokens. Highlights: We compare tolling in dollars versus tokens in a within-day dynamic setting. Selling behavior of individuals in the token market is explicitly modeled. Tolling in tokens is more efficient when congestion effects are severe. The findings make a case for tolling in tokens.