Price discrimination in Australasian air travel markets

Abstract

This paper explores oligopolistic pricing behaviour in the context of widespread price discrimination, as observed in domestic NZ and trans-Tasman air travel markets. These markets are of particular interest as being the domain of three substantial competition policy cases between 2002 and 2010, each involving proposals by Air New Zealand to form ‘alliances’ (cartels) with one of its competitors – first, Qantas, and then Virgin Blue. Theoretical predictions are tested with internet-sourced pricing data covering the 2004–2006 period. It is found that market structure matters more for pricing in the ‘time-sensitive’ (business) sector of the market than for ‘price-sensitive’ leisure travellers. The implications drawn are that the regulatory authorities were correct to refuse permission for Air New Zealand's proposed alliances with Qantas, and may also be justified in their qualified approval of the 2010 proposal with Virgin Blue as a partner.

Keywords:

• airline markets
• price discrimination
• competition policy

Acknowledgements

The author thanks two referees for excellent comments. Research support from the University of Auckland Faculty Research Development Fund, and from the NZ Institute for the Study of Competition and Regulation is gratefully acknowledged.

Notes

1. Note that the ‘no-arbitrage’ necessary condition for successful price discrimination is met by the universally enforced condition that an airline ticket is not transferable to another person.

2. Fewer than nine observations were made if the flight sold out before the flight date. In this case, for the purposes of aggregation the last observed fare is recorded as a notional fare in subsequent weeks.

3. Fifth freedom regulatory rights allow long-haul carriers who access Australia or NZ to also carry passengers travelling between Australia and New Zealand (ACCC, 2010, para 4.7, p. 15).

4. Emirates flies daily on four Tasman routes – three from Auckland; one from Christchurch. Three other long-haul airlines used 5th Freedom rights – Royal Brunei, Garuda, LanChile – but each on only one route, and with frequencies of one flight/day or fewer, such that their combined share of the trans-Tasman market is a little above 2%. The pricing of these airlines appears to be similar to that of Emirates, and results including them are not shown in this paper, for simplicity.

5. Unfortunately we don't know how many tickets were purchased at each offered price. More detail on the data and variables is given in Hazledine (2010a), which uses a subset of this paper's database to estimate average price models. Note that using a simple unweighted average of the airfare observations gives quite similar results.

6. Airline product differentiation can be due to national carrier loyalty; to lock-in to a particular FFP or airline alliance, and to the extent of the airline's regional network. Route-level differentiation could be generated by on-time record and by frequency of flights on a route.

7. In Hazledine (2010a) I tell ‘stories' for each of the 19 routes that in my opinion plausibly explain the (exogenous) market structure of all but one or two of them

8. A referee asked if the airline dummy coefficients are robust to a change to a non-linear (quadratic) specification of the HHI. It turns out that they are.

9. de Roos et al. (2010) are quite reticent about the source of their data, but the very large number of observations suggests that their data were obtained by downloading from the database of a cooperative travel agency, rather than by manual interrogation of websites, as we had to rely on.

10. See Hazledine (2004) for more detail on this case and the decisions.

11. Emirates is not itself a LCC, but it is believed to have low variable (opportunity) costs of flying daily across the Tasman, because its wide-body jets would otherwise spend the day sitting on the tarmac in Australia, on their turn-around between long-haul international flights.

12. In November 2007, Pacific Blue commenced flying on some domestic NZ routes, pulling out of the domestic market nearly three years later, in September 2010, having made substantial losses (http://www.stuff.co.nz/the-press/news/4030081/Pacific-Blue-flies-coop).

13. The last-minute fares of Air New Zealand and Qantas will usually be in fare classes offering the flexibility to change flights and even cancel travel, with refund. Flexibility is not in general such an attraction for leisure travellers, and not to any travellers using Emirates or Pacific Blue, who each only offered one daily flight on the routes they served.

14. See ACCC (2010, Table 4.5, p. 21). In making use of this table I don't count the market shares of the three fringe 5th Freedom carriers, which amount to just over 2%, in total.

15. The ACCC was concerned about competition effects on four smaller routes (Wellington–Brisbane, Wellington–Sydney, Wellington–Melbourne, Dunedin–Brisbane), two of which are not currently served by Qantas or Jetstar. The Commission imposed conditions that Air NZ/Virgin maintain and grow their current capacity levels on these routes. I note that not reducing capacity is a necessary but not sufficient condition for the suppression of price increases (the airlines could always choose to fly with empty seats).

16. The application of the labels ‘time-‘ and ‘price'-sensitive to business and leisure travellers has been standard form throughout these cases. See, for example, ACCC (2010, Para 5.9, p. 24).

17. Just as Air New Zealand appears to have a particularly strong position in the ex-New Zealand business market, as evidenced by the price premium it sustains in day-before fares, as revealed by the airline dummy variable coefficients in the P08k regression model of Table 1.