Thread by @robmun12

Check out my new working paper on using supply shocks to quantify traffic congestion externalities! You can find it at papers.ssrn.com/sol3/papers.cf… or on my website. Here follows a thread: (1/13)

Motivation: Traffic congestion is one of the most challenging issues of urban agglomeration and bears substantial costs. Congestion costs are often higher than their socially optimal levels because of a missing market problem: roads are generally not priced (2/13)

Cities implement policies that limit the use of on-demand vehicles to address this problem. Their effectiveness depends on how much they approximate the optimal size of a Pigovian tax on road usage (3/13)

In this paper, I quantify two relatively unknown parameters needed to design an optimal congestion policy. I estimate the effect of a vehicle on congestion and document its substitution patterns to other transportation modes in NYC during 2009-2017 (4/13)

The # of vehicles in a city is endogenous! Identification requires varying the # of vehicles holding demand constant → we would randomly take cars out of the streets in a way that is not anticipated by demand → difficult to implement, so I exploit a natural experiment (5/13)

Large share of for-hire drivers is potentially Muslim while only 3% of NYC’s population is Muslim → Islam has two major holidays and their dates change every year following the lunar calendar → Difficult to anticipate unless you’re Muslim (6/13)

Identification assumption: we can expect that the # of vehicles in the city decreases during Muslim holidays because many drivers do not work on those days, while the rest of the city is oblivious to them (7/13)

Results: during Muslim holidays the number of active taxis decreases by about 1,000 (9.1% of the total), which decreases the time per mile traveled (the inverse of speed) by 0.46 minutes (7.1% of a baseline of 6.5 minutes per mile) (8/13)

Substitution patterns during Muslim holidays:
-taxi and other for-hire trips ↓ between 1.6% and 16.4%
-the # of passengers per trip ↓by 1.2% suggesting increased waiting time
-travel distance in taxis ↑ between 2.1% and 3.4% suggesting switching to walking (9/13)

Welfare: change in consumer surplus is ambiguous. Those who still travel via vehicle, welfare increases because travel time is reduced. Those residents who face increased wait times or who switch to a less-preferred transportation mode will suffer from reduced welfare (10/13)

Welfare calibrations that consider this welfare trade-off suggest that consumer surplus in the city increases between $8 and $13 million per day depending on assumptions about wait time (11/13)

Takeaways:
-Estimates of speed capture congestion spillovers and can be directly used to design policies that reduce congestion externalities
-Estimates relevant in other contexts, as traffic characteristics in NYC by borough and day time extend to many cities worldwide (12/13)

-Quantifying the welfare trade-offs of the reduction of vehicles during Muslim holidays provides a relevant bound to guide the design of congestion policies and evaluate their political viability

Check the paper for more details! (13/13)