Switching Costs in the Brazilian Airline Sector

Abstract

This paper calculates the switching costs in the context of the Brazilian airline industry by considering an indirect approach, and econometrically evaluates its determinants during the 2005-2017 period. Evidence suggests non-negligible switching costs that display a decreasing trend over time, which is consistent with an increase in the competition in the sector, as also exhibited in other countries. An instrumental variables estimation suggests that variables referring to the number of flights, a frequent flyer program and the entry/presence of competitors play a role in determining the magnitude of switching costs.

Keywords:

• Switching costs
• airline industry
• frequent flyer programs
• air transport
• industrial organization

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 The websites of the airlines that operate in Brazil commonly report that onboard services (except snacks and non-alcoholic drinks) are only available for flights of more than one hour or slightly longer.

2 Refer to OAG (2018) for a presentation of the rankings on the number of flights on the routes and the punctuality of airlines.

3 For more details on UPP, refer to Shy (2002).

4 Events introduced by law 11.182. For more details, refer to Brasil (2005).

5 Based on ANAC (2018b).

6 The companies PANTANAL and TEAM operated with only occasional flights.

7 Based on ANAC (2018b).

8 It is reasonable to assume that the creation of this last program did not impact the competition in the Rio de Janeiro/São Paulo route considering that AZUL only had a few flights on weekends.

9 See European Economic Authorities (2005).

10 The focus of the regulatory agency in Brazil (ANAC) is mostly on safety and punctuality issues, and the deregulated market has prevailed for some time except for restrictions on the entry of foreign firms. As for antitrust policies, implemented by the competition agency CADE, a light-handed approach has prevailed [see CADE (2017)].

11 While the price data refer to the date of the sale of the tickets, the passenger data paid refers to the date of the flight. As a result, when there were bankruptcies and acquisitions, the remaining flights accounted for paid passenger data without a counterpoint in terms of ticketing price data. In these cases, the average price charged for tickets in the last three months was used as a proxy. It seems reasonable to suppose that the remaining flight tickets were sold in that period.

12 Through equations 9 and 10, it is possible to see that the entry of a firm that practices even lower prices in the $N_k$ position should reduce the switching cost $S_k,$ but should increase the switching cost of all other $i$ competitors, $S_i,$ holding all else constant.

13 The conceptual framework of this test was conceived by Breusch (1978) and Godfrey (1978) and further commented on in Breusch and Pagan (1980). Since then, different applications can be found in the context of panel two-stage least squares.