Universities Research Conference

May 15, 2009
Severin Borenstein and Dennis Carlton, Organizers

Silke J. Forbes, UC, San Diego, and Mara Lederman, University of Toronto Does Vertical Integration Affect Firm Performance? Evidence from the Airline Industry

Forbes and Lederman investigate the effects of vertical integration on operational performance in the U.S. airline industry. All of the large U.S. network carriers use regional partners to operate some of their short- and medium-haul routes. However, some regional partners are owned while others are independent and managed through contracts. Using detailed flight-level data and accounting for the potential endogeneity of integration decisions, the researchers estimate whether an airline’s use of an owned - rather than independent – regional partner at an airport affects its delays and cancellations on the flights that it itself operates out of that airport. Their results indicate that integrated airlines perform systematically better than non-integrated airlines at the same airport on the same day. Furthermore, they find that the performance advantage of integrated airlines more than doubles on days with highly adverse weather conditions. Because adverse weather requires airlines to make real-time adjustments to their planned schedules, this finding suggests that the benefits of ownership are particularly high when firms need to make a greater number of non-contracted adaptation decisions. The authors believe that this work is one of the first to both document the performance implications of vertical integration decisions and to shed light on the underlying source of these differences.


Steven Puller, Texas A&M University and NBER, ANIRBAN SENGUPTA and STEVEN WIGGINS, Texas A&M University
Testing Theories of Scarcity Pricing and Price Dispersion in the Airline Industry

Puller, Sengupta, and Wiggins use a unique new dataset – ticket transaction data – to test between two broad classes of theories regarding airline pricing. The first group of theories, as advanced by Dana (1999b) and Gale and Holmes (1993), postulates that airlines practice scarcity based pricing and predict that variation in ticket prices is driven by differences between high demand and low demand states. Dana's theory predicts that airlines sell tickets with higher and more dispersed prices in unexpectedly high demand states; Gale and Holmes predict that more discounted "advance purchase" seats are sold in off-peak demand periods. Both of these groups of theories predict substantially higher shares of low price tickets in off-peak versus peak flights. The second group of theories, as advanced in the yield management literature, indicate that fare variation is driven by differences in ticket characteristics as associated with price discrimination. The researchers use a census of ticket transactions from one of the major computer reservation systems to study relationships between fares, ticket characteristics, and flight load factors. The central advantage of their dataset is that it contains additional variables not previously available. These variables measure both the ticket characteristics central to the price discrimination theory and information on load factor and peak/off-peak travel times needed to test the scarcity pricing theory. They find only modest support for the scarcity pricing theories – the fraction of discounted advance purchase seats is only slightly higher on off-peak flights and fare dispersion is nearly the same. However, ticket characteristics that are associated with second-degree price discrimination drive much of the variation in ticket pricing.


James Dana, Northeastern University and Eugene Orlov, Compass Lexecon
Internet Penetration and Capacity Utilization in the US Airline Industry

Airline capacity utilization, or load factors, increased dramatically between 1993 and 2007, after staying fairly level for the first 15 years following deregulation.Dana and Orlov argue that consumers’ adoption of the Internet, and their use of the Internet to investigate and purchase airline tickets, explains this increase. They find that differences in the rate of change of metropolitan area Internet penetration explain differences in the rate of change of airline-airport-pair load factors. Consistent with that explanation, they also find that, all else equal, changes in Internet penetration have a bigger impact on load factors on flights in more competitive markets and on flights with fewer total passengers. They argue that a significant part of the associated $3 billion reduction in airlines’ annual capacity costs represents a previously unmeasured social welfare benefit of the Internet.

Alessandro Gavazza, New York University
The Role of Trading Frictions in Real Asset Markets

Almost all real assets trade in decentralized markets, where trading frictions could inhibit the efficiency of asset allocations and depress asset prices. Gavazza uses data on commercial aircraft markets to empirically investigate whether trading frictions vary with the size of the asset market. Intuitively, it is more difficult to sell assets that have a thin market. As a result, firms find it optimal to hold on longer to assets with a thinner market in case their profitability rises in the future. Thus, when markets for firms’ assets are thin, firms’ average productivity and capacity utilization are lower, and the dispersions of productivity and of capacity utilization are higher. In turn, prices of assets with a thin market are, on average, lower and have a higher dispersion, since prices depend on firms’ productivity and capacity utilization. The empirical analysis confirms that trading frictions vary with the size of the market, as aircraft with a thinner market have: 1) lower turnover; 2) lower capacity utilization; 3) higher dispersion of utilization levels; 4) lower mean prices; and 5) higher dispersion of transaction prices.


Jan K. Brueckner, UC, Irvine
Price vs. Quantity-Based Approaches to Airport Congestion Management

Brueckner analyzes price and quantity-based approaches to management of airport congestion, using a model in which airlines are asymmetric and internalize congestion. Under these circumstances, optimal congestion tolls are differentiated across carriers, and a uniformity requirement on airport charges (as occurs when slots are sold or tolls are uniform) distorts carrier flight choices. Flight volumes tend to be too low for large carriers and too high for small carriers. But quantity-based regimes, where the airport authority allocates a fixed number of slots via free distribution or an auction, lead carriers to treat total flight volume (and thus congestion) as fixed, and this difference generates an efficient outcome as long as the number of slots is optimally chosen.


Itai Ater, Tel Aviv University
Internalization of Congestion at U.S. Hub Airports

Congestion externalities arise when airlines do not consider that scheduling another flight may result in flight delays for other airlines. Open questions are whether and how carriers take into account the delays a flight inflicts on other flights operated by the same carrier. Ater addresses these questions by studying congestion during high-volume time periods, used by hub-carriers to reduce the layover time of connecting passengers (also known as flight banks). To facilitate the empirical application, he uses individual flight time data, as well as airport, aircraft, and weather information, and explicitly identifies the characteristics of banks, such as their time length and the number of flights offered by each carrier. The empirical analysis proceeds in two main steps: he shows first that banks dominated by one airline are longer and are characterized by lower flight density. Second, he finds that longer banks are associated with shorter flight delays. These findings imply that hub-carriers internalize congestion by scheduling longer banks. Furthermore, these findings may suggest that congestion management solutions implemented at hub airports dominated by one airline could have a limited impact on congestion itself.