![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
A supramonopoly is the market structure of a homogeneous good that is priced higher than the monopoly level. We present evidence of supramonopoly in the US air passenger service. For four airline mergers during the period of 1993 to 2009, we identify routes that dropped fares for no other reason than the increase in market power due to merger. Therefore each of those routes had fare higher than the monopoly level before the merger to monopoly. We explain supramonopoly as the result of a cartel with strongly punitive matching rules. We also discuss the striking implications for antitrust and regulatory policies.
Notes
1. The merger was approved by antitrust authorities on 29 October 2008 and consummated a few hours after the approval.
2. During an interview with the Wall Street Journal, Delta CEO Richard Andersen said that the merger could ‘eliminate at least $1 billion in cost’ in 2012, which is four years from the date of the announcement of the merger (Wall Street Journal, 22 April 2008, B1).
3. This is the reason we removed mergers involving Mountain Air Express, CCAIR, Paradise Island Airlines, Business Express Airlines, etc.
4. For example, in July 2006, Northwest Airlines bought operating certificates and related assets from FLYi Inc (Independence Air).
5. For example, mergers between Delta Airlines and Atlantic Southeast Airlines in March 1999, between Delta Airlines and Comair Holdings in October 1999, between Northwest Airlines and Mesaba Airlines in December 2006.
6. In the Q3 2009 raw data, the average ‘itinerary yield’ is $0.22 per mile and 90% of records have an itinerary yield less than $0.39 per mile. Tickets exceeding $3/mile represent 0.051% of the data.
7. The assumption here is that a flight between points A and B with two stops at points C and D is qualitatively different from a direct flight, from a flight with one stop only, and from a flight between A and B with stops at E and F. Flights with more connections often have lower fares. See Singal (1996, p. 244).
8. Every record contains information on ‘the reporting carrier’ that submits data to DOT, ‘the ticket carrier’ that sold the ticket, and ‘the operating carrier’ that flew the airplane. In the Q3 2009 data, 293,921 records have multiple ticket carriers, representing 5.62% of the total raw data. In addition, 1,678,439 records have multiple operating carriers, representing 32.11% of the total raw data. The reporting carrier is the same as the ticket carrier or the operating carrier or both in 74.08% of records. All three carriers are different in 15.97% of records. In the construction of subsamples, the reporting carrier has priority. If that does not work, then either the operating or the ticket carrier will be used instead.
9. Kim and Singal (Citation1993, p. 553, fn 16) drop routes with few than 30 observations in their data (probably DB1A, rather than DB1B) and Peters (Citation2006, p. 635) drops routes with fewer than five observations in his data. These data sets are all 10% samples.
10. The routes were deleted if served by other carriers than merging airlines in the quarter of merger announcement.
11. In the Q3 2009 raw data, we found that 18.03% of routes were duopoly. Among them, 70.23% had observations fewer than 30, and were removed from the route sample. Moreover, we only examine duopoly routes served by merging airlines on which each had at least 10% market share. This restriction further rules out a large portion of sample routes.
12. The formula allows time for market power to change but no time for integration. The formula is more stringent than the one that Kim and Singal (Citation1993, p. 555) use to calculate fare change due to greater market power. They compute changes from the quarter before the (announcement) period to the quarter after the (announcement) period. Thus they allow a lag of one quarter.
13. Duopoly routes in the merger between American Airline and Trans World Airlines had higher relative fare change under HHI.
14. Thirty‐eight incidents of cutting fares in Appendix 2 and 33 in Appendix 3 entail supramonopoly fares.
15. In Appendix 2, there are seven incidents of supramonopoly involving raising prices: two in the American/Trans World merger, one in the United/Aloha merger, and four in the Delta/Northwest merger. The supramonopoly fares are 0.15% to 4.91% lower than those of the control groups in these incidents; however, the fare reductions range from 0.80% to 27.21% on those routes where the fares were cut while the control groups raised fares. In Appendix 3, there are five incidents of supramonopoly involving raising prices: one in the American/Trans World merger, one in the United/Aloha merger, and three in the Delta/Northwest merger. The supramonopoly fares are 0.01% to 4.91% lower than those of the control groups in these incidents; however, the fare reductions range from 0.80% to 27.21% on those routes where the fares were cut while the control groups raised fares.
16. This suggests that matching behaviors fall into three categories: ones that make the firm absolutely better off, ones that make the firm worse off but better off than if no matching, and ones that makes the firm absolutely worse off. Osborne (Citation1976, p. 839) observes that retaliation in a perfect cartel can never make the firm absolutely better off.
17. The matching rule takes the mathematical form of a derivative, in this case dq 2/dq 1 = 4.
18. The perfect cartel is calculated as follows. The marginal revenue for a firm sharing the market with another firm is MR = 6 − 4q. Its marginal cost is MC = 5.5 − 6q. Thus the two have intercepts 6 and 5.5 on the price axis, and intercepts 1.5 and 0.917 on the quantity axis, respectively. The MR is everywhere higher than the MC. The MC becomes zero at output 0.917 and remains zero beyond. The profit maximizing output then is 1.5, where MR becomes zero also. It follows that price equals 3. To calculate profit, subtract total cost for 0.917 units of output (MC is zero beyond) from revenue for 1.5 units of output. That is 1.48 a firm.
19. DB1A is not available from DOT’s website and we are not sure how it differs from DB1B.
20. Addyston Pipe and Steel Company v. United States, 175 U.S. 211 (1899).
21. See American Tobacco Company v. United States (1946) and Theatre Enterprises, Inc. v. Paramount Film Distributing Corp. (1954).