20. May 2024
On August 10, 2023 an accident caused the closure of the Gotthard base tunnel, a key connection for travel between the North of Switzerland and the canton Ticino. The closure lasted until December 10, 2023. Thereafter, a reduced schedule of 31 trains per weekend was implemented. Since March 2024 further 7 trains per weekend pass the base tunnel. See the information of the SBB.
The closure implied that travel to the canton Ticino from the North of Switzerland lasted approximately one hour longer, as trains were rerouted via the panorama route. In this blog we investigate whether the longer travel time has led to lower expenditure growth for hotels and restaurants in Ticino. To estimate the effect we use the expenditure data from Monitoring Consumption Switzerland (MCS) to compare the change of expenditures in the categories accommodation and food & beverage services in the time window around accident on August 10, 2023 (corresponding to March 2023 to end of February 2024, i.e., five months before and seven months after the closure).
To account for common changes of expenditures in Switzerland in these categories during the time window, we compare the changes of expenditures in Ticino to those of Bern (BE), St.Gallen (SG), Thurgau (TG) and Solothurn (SO).[1] To account for the seasonality of expenditures within a calendar year, which differs across regions in Switzerland (e.g., because of a different importance of the tourism versus manufacturing sector), we compute the year-on-year growth rate of expenditure for each month during the considered period. We then estimate how the base tunnel closure affected these growth rates.
The analysis follows a classic differences-in-differences design. We find that the control region exhibited similar trends in expenditure growth rates as the canton Ticino (the treated region) in the period before the accident occurred, i.e., in the period March 2023 to August 2023. A possible concern may be that expenditure shifted from Ticino to the control region after the tunnel closure, which may have inflated the difference in expenditure growth rates between Ticino and the control region after the tunnel closure. It is worth keeping this in mind when inspecting the results.
Figure: Year-on-year growth rates per month in Ticino and the control region (BE, SG, TG, SO)
The figure illustrates the effect on expenditure growth rates in accommodation and food & beverage services in Ticino relative to the cantons of Bern, St. Gallen, Thurgau and Solothurn. The figure shows that expenditure growth rates fluctuate somewhat both before and after the tunnel closure. After the tunnel closure an interesting pattern emerges, however. In the months of September and October 2023, expenditure growth rates for hotels and restaurants were substantially lower in Ticino than in the control region. From November 2023 onwards, the expenditure growth of Ticino caught up, exceeding the growth observed in the control region. This effect was particularly pronounced for food & beverage services. Indeed, the point estimates for the changes in expenditure growth after the tunnel closure in Ticino relative to the control region are negative but not statistically significant at conventional levels if we consider the period until the end of February 2024.
Overall, these findings suggest that the additional hour of travel time (two hours for a return trip) only had a temporary negative effect on expenditure growth for hotels and restaurants in Ticino. Indeed, the negative impact of the tunnel closure in September and October 2023 seems to have been compensated by stronger expenditure growth thereafter.
[1]The assumption of a parallel trend before the tunnel closure cannot be rejected for this control region. We have experimented with alternative control regions, containing the largest cantons in terms of population size in the German speaking part of Switzerland, i.e., the cantons ZH, BE, AG, SG, LU, BL, TG, SO, or subsets of these cantons. Our preferred specification was to use BE, SG, TG and SO as control region because including other cantons implied different trends from those in Ticino or more volatile expenditure growth in the considered expenditure categories before the tunnel closure.