Are on-time performance statistics worthless? An empirical study of the flight scheduling strategies of Brazilian airlines

https://doi.org/10.1016/j.tre.2020.102186Get rights and content

Highlights

  • We develop a method to decompose the extra schedule block times into operational and strategic factors.

  • We estimate the impact of extra times on flight delays, allowing for the moderation effects.

  • We test the hypothesis of the existence and effectiveness of schedule padding.

Abstract

Airlines may manage their on-time performance by lengthening schedules with engineered increases in planned flight times. We develop an econometric model of high dimensional sparse (HDS) regression to decompose the extra schedule block times into operational and strategic factors. We estimate the impact of extra times on flight delays, allowing for the moderation effects of runway congestion, slots, and propagated delay. We test the hypothesis of the existence and effectiveness of schedule padding practices. We find that a 2012 on-time disclosure rule may have induced carriers’ padding behavior. In contrast, slot regulation may prevent the formation of extra block times.

Introduction

Airlines are taking a little more conservative approach to ensure they’re going to arrive on time.” (…) “It’s part of their marketing campaigns, part of their affinity programs to develop consumer loyalty.” - Sean Cassidy (vice president of the US Air Line Pilots Association labor union and Alaska Airlines pilot).1

Flight scheduling is one of the most important tools of airline network management, as it is a key driver of operating costs. However, carriers may have strong incentives to plan scheduled flight times not only based on cost conditions but also on the status of service quality competition in the market. Setting longer flights confers airlines with more flexibility to deal with unexpected delays and still accomplish the scheduled arrival times, a strategy labeled “schedule padding” in the industry. By padding their schedules, airlines have often been accused of improving their on-time performance (OTP) in a spurious way.2 This gaming behavior may damage the passengers’ perception of service quality and eventually mask the entire air transportation system’s inefficiencies.

However, scheduling longer travel times may be inevitable for an airline from a flight operations standpoint. Actually, the flight management systems on modern airliners optimally determine the cruise speed of a flight in line with the cost index parameter (CI), a ratio between time-dependent costs and fuel costs. Each flight cruise is typically assigned a speed that is between the “maximum range cruise speed”—a low speed consistent with a null CI level—and the “maximum permissible cruise speed”—a high speed consistent with the maximum CI level.3 In the first case, the time-dependent cost is low relative to the unit fuel cost, allowing for longer flight duration and less fuel consumption; in the second case, the time-dependent cost strongly dominates the fuel cost, leading to flights with shorter duration and higher fuel consumption. In this sense, the task of scheduling flight times is strictly dictated by the relative operating costs of the airline, and as a result, not all extra times added to scheduled flight times constitute real strategic buffer times.

The objective of this paper is to empirically decompose the extra times incorporated by carriers into their schedules into strategic and operational determinants. We also aim to assess the efficiency of extra times in enhancing OTP by estimating their impacts on the odds of flight delays. We analyze the Brazilian airline industry from 2001 to 2018. In this period, the country witnessed relevant variations in its OTP records. In 2008, the São Paulo/Guarulhos (GRU) airport, a key international gateway in the country, was considered one of the most delayed airports in the world.4 Ten years later, however, the situation had completely changed, with GRU now ranking number 10 among the top 20 major airports with respect to OTP.5 While the occurrence of delays in the Brazilian market dropped from 27.5% in 2008 to 15.8% in 2018, the episodes of early arrivals increased from 0.4% to 44.3% in the same period.6

We inspect the market incentives of airlines to engage in strategic flight scheduling by means of an econometric of high dimensional sparse (HDS) regression that estimates the drivers of the extra scheduled block times in Brazil. We also investigate the impact of three regulatory reforms: the 2012 on-time disclosure resolution—which made it mandatory for airlines to publish their delay and cancellation statistics for each flight on their websites and other sales channels; the 2014 slot reform at major airports in the country; and finally, the introduction of a major Air Traffic Management (ATM) innovation—the implementation of performance-based navigation (PBN) procedures in Brazilian airports since the late 2000 s.

This paper aims to contribute to the recent econometric literature on airline strategic scheduling, including works by Skaltsas, 2011, Forbes et al., 2018, Fan, 2019, Yimga and Gorjidooz, 2019, Brueckner et al., 2019. In particular, we study the decoupling of scheduled block times from the “unimpeded” block times, i.e., the gate-to-gate travel times accomplished under ideal flight circumstances.7 We then test whether a lengthening in flight duration by carriers is motivated by a set of market-related drivers that are ceteris paribus to the cost index-related factors. To the best of our knowledge, this is the first study that empirically distinguishes the strategic buffer time from the operational extra time of airlines. We therefore examine whether carriers set longer scheduled block times purely as a schedule padding practice, by adding a strategic buffer to intentionally improve the reported OTP, or if such extra time is actually an unavoidable consequence of changes in operating conditions.

The literature has already emphasized the problems associated with the commonly utilized measures of flight delays - Forbes et al., 2018, Yimga and Gorjidooz, 2019 are recent examples. El Alj (2003) discusses the limitations of using scheduled times as a reference when assessing OTP and airport and airspace system congestion, since airlines may anticipatedly plug expected delays into their schedules. Our study contributes to the literature by being the first to decompose airlines’ extra block times into operational and strategic –i.e., competition-driven–factors. In this sense, we are the first to investigate the effects of market concentration and low-cost carriers’ participation on the incentives that airlines have to engage in schedule padding to avoid damages in their on-time performance. Additionally, we assess the extent of the impact of schedule padding on OTP indicators that are relevant to authorities, operators, and the consumer in general. Finally, we also contribute by estimating the effect of an on-time disclosure rule on schedule padding. This issue has been raised by the previous literature but has not been estimated in a ceteris paribus way, i.e., isolated from the operational causes of block time prolongation.

This paper is organized as follows. Section 2 provides a discussion of the literature on airline scheduling decisions and the determinants of flight on-time performance. Section 3 presents the empirical model. Section 4 presents the estimation results, which are followed by the conclusions.

Section snippets

Airline scheduling and the duration of flights

The schedule planning of an airline is a complex process, as it seeks to simultaneously optimize the exploitation of the resources to meet the demand across the carrier’s multiple routes (Holloway, 2008) and to maximize revenues (Abdelghany et al., 2017). Scheduling becomes even more challenging due to the uncertainties of the system that affect demand, airlines’ operations, and prices (Belobaba et al., 2009). Most existing scheduling models ignore the uncertainties in actual operations by

Application

We consider the domestic Brazilian airline industry from 2001 to 2018 to develop empirical models of extra scheduled block time and flight delays. Deregulation and low-cost carrier entry produced a notable intensification of price competition in this industry in the studied period. As a result, the market has expanded significantly, from 29.9 million domestic passengers transported in 2001 to 93.7 million in 2018.12

Estimation results

Table 3, Table 4 present the results of our empirical models of extra scheduled block time (EXTBT) and arrival flight delays (ODDSDEL) in Brazil. To simplify the exposition, we omit indexes k and t. In both tables, the respective Column (4) contains our preferred model; Columns (1) to (3) present a set of subspecified versions of the main model, in which we drop some key variables; and finally, Columns (5) to (8) display the results of some robustness checks. In both tables, it is possible to

Conclusion

This paper investigates the main drivers of the extra times incorporated into the flight schedules of Brazilian airlines. We utilize an econometric method of high dimensional sparse (HDS) regression that employs the least absolute shrinkage and selection operator (LASSO) of Tibshirani (1996) to select adequate controls from among a vast set of available variables. Our estimation has the methodological contribution of decomposing the extra schedule block times into operational and strategic

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The first author wishes to thank the Coordination for the Improvement of Higher Education Personnel (CAPES) – Finance Code 001; the third author wishes to thank the São Paulo Research Foundation (FAPESP) - grants n. 2013/14914-4 and 2015/19444-1; National Council for Scientific and Technological Development (CNPq) - grants n. 301654/2013-1, n. 301344/2017-5.

References (35)

  • N. Kafle et al.

    Modeling flight delay propagation: a new analytical-econometric approach

    Transp. Res. Part B: Methodol.

    (2016)
  • L. Kang et al.

    Behavioral analysis of airline scheduled block time adjustment

    Transp. Res. Part E: Logistics Trans. Rev.

    (2017)
  • V.A.P. Miranda et al.

    Airport slots and the internalization of congestion by airlines: an empirical model of integrated flight disruption management in Brazil

    Transp. Res. Part A: Policy Practice

    (2018)
  • G. Santos et al.

    Determinants of delays at European airports

    Transp. Res. Part B: Methodol.

    (2010)
  • Ö. Şafak et al.

    Accommodating new flights into an existing airline flight schedule

    Transp. Res. Part C: Emerg. Technol.

    (2019)
  • Y. Wang et al.

    Scheduled block time setting and on-time performance of US and Chinese airlines—A comparative analysis

    Transp. Res. Part A: Policy Practice

    (2019)
  • J. Yimga et al.

    Airline schedule padding and consumer choice behavior

    J. Air Transp. Manage.

    (2019)
  • Cited by (0)

    View full text