Does it make sense for an online company to post two different
prices for the same product on the same website, hoping that some
buyers will buy at the higher price? According to the traditional view,
it should not because the shoppers of a company setting a low and a
high price for the same product (e.g., in two different parts of its
website) would very quickly learn to buy only at the low price, given
the price transparency of the Internet. In this paper we present
evidence of differing prices being posted by the same e-seller on the
same website at the same time for exactly the same product.
Our data are taken from the websites of six European Low Cost
Carriers (hereafter, LCCs) and pertain to both United Kingdom's (UK)
domestic and European international flights. We focus on the possible
difference in the fares denominated in two different currencies (those
of the countries at a route's endpoints) available to two individuals
who, each in their own country, are simultaneously trying to book a
seat on the same flight.
Our analysis reveals the use of three basic online pricing scenarios.
In the first, the two individuals are offered uniform fares: the ratio of
the fares in the two currencies is very close to the prevailing exchange
rate and thus the Law of One Price holds (Goldberg and Knetter,
1997). The second online scenario entails perfect segmentation
between the two markets, with one individual being offered a price
significantly higher than the other: due to the way the booking system
works, the former individual cannot buy at the other, lower price. In
the third scenario the two (groups of) buyers are also segmented, but
the airlines' online booking system contains a hidden discount. That is,
by engaging in a simple search it is possible for the high-fare
individual to access and possibly (depending on the net value of the
discount) buy at the lower fare.
About 34% of fares in our sample of more than two million observations
pertain to either the second or the third scenario; for the
largest LCC in our sample, Ryanair, that proportion increases to
slightly above 50%. This paper thus highlights a source of price dispersion in airline markets that is exclusively due to the disconnection
between the two fares' ratio and the prevailing exchange
rate.1 To our knowledge, no previous study indicates the use of
different currencies as a source of online fare dispersion.2 Furthermore,
this study also documents a noticeable source of online price
dispersion within each single airline's website and thus fills a gap in
the literature, which has produced scant empirical evidence of online
intra-firm price dispersion and has mainly focussed on the sources of
online inter-firm price dispersion (see, inter alia, Baylis and Perloff,
2002; Baye et al., 2004; Brynjolfsson and Smith, 2000; Clay et al.,
2001; Clemons et al., 2002; Smith and Brynjolfsson, 2001; and the
surveys by Baye et al., 2006; Ellison and Ellison, 2005; Stole, 2007). A
notable exception is the work by Cabolis et al. (2007) showing
evidence of international price discrimination for the economics
textbook sold by Amazon on its US and British websites, with
American prices being substantially more expensive.3
To shed further light onto the properties and the motivations
underlying the previously described pricing scenarios, the present
paper draws from and aims to contribute to two generally distinct
strands of literature, one studying price dispersion in e-commerce setups,
the other in airline markets.
With regard to the former, Baye et al. (2004) investigate whether
online inter-firm price dispersion is a temporary disequilibrium
phenomenon that can be observed for short time periods, i.e., until
prices are driven towards the perfectly competitive level because
consumers are able to search costlessly for the lowest price available
online. In their analysis of the prices posted on a comparison site, a
product's difference between the two lowest prices tends to persist
over time and can therefore be characterised as an equilibrium
phenomenon.4 In terms of our research, establishing that temporal
persistence characterises the second and third pricing scenarios
entails that they represent an intrinsic feature of the airlines' pricing
approach; i.e., the airlines do not apply them sporadically and in
isolation but, rather, they intentionally embed such pricing schemes
into their revenue management systems. In other words, temporal
persistence would suggest an ex-ante commitment to a specific
pricing rule that, following Dana (1999), appears to be justified by the
need to manage anticipated idiosyncratic demand conditions at the
single flight's level. We assess the presence of temporal persistence by
investigating whether, for the same specific flight, an airline changes
its pricing approach during a fixed period of time preceding a flight's
departure. The econometric evidence from a dynamic panel data
setup suggests that the second and third online pricing scenarios tend
to persist over a relevant time horizon before a flight's departure. That
is, the same pricing tactic is maintained persistently over the period
preceding a flight's departure, thereby lending support to the notion
that such tactics are deliberately chosen by the airlines and reflect
their expectation on a given flight's demand conditions.
Many empirical studies document the impact of market structure
on price dispersion in airline markets (see Borenstein and Rose, 1994;
Gerardi and Shapiro, 2009; Hayes and Ross, 1998; Stavins, 2001 for
the case of the US market; Gaggero and Piga, 2011; and Giaume and
Guillou, 2004 for the European one). The discussion hinges around
two opposing competitive mechanisms. In the “monopoly effect”, an
increase in competition makes it harder for an airline to price discriminate,
so that the mark-up on the price-inelastic (e.g., business)
passengers reduces to a level more in line with the level of fares
charged to price-elastic (e.g., leisure) passengers. The result is that
competitive pressure reduces dispersion; i.e., under the “monopoly
effect” we should observe a positive relationship between dispersion
and market concentration. This prediction is reversed under the
“brand effect”, where price dispersion increases with the competitive
pressure. This is because competition reduces the prices offered to
non-loyal (leisure) travellers more than the fares to loyal (business)
customers (Borenstein, 1985; Holmes, 1989). From an empirical
viewpoint, the results in Borenstein and Rose (1994) and in Stavins
(2001) are consistent with the brand effect. However, Gerardi and
Shapiro (2009), using the same measure of market concentration as in
Borenstein and Rose (1994) but a different econometric approach,
find support for the monopoly effect. Furthermore, they show that a
positive relationship between market concentration and fare dispersion
is particularly strong on those routes characterised by the copresence
of a large number of both business and leisure travellers,
since the difference in demand elasticities of these groups constitutes
a textbook pre-requisite for profitable price discrimination.
As far as the pricing approaches illustrated in this study are
concerned, the second and the third scenarios share a feature, which is
typical of third-degree price discrimination, i.e., one group is charged
a higher price than the other. It should then be expected that the
intensity with which these pricing scenarios manifest themselves
should be larger in markets with a strong presence of business travellers.
Furthermore, the third scenario also embodies elements of
second-degree price discrimination (menu pricing) since it exploits
customers' heterogeneity in the propensity to search for the best
price. Business travellers are generally assumed to have a high
opportunity cost of time and therefore should find it very costly to
search, whilst leisure travellers, given their relatively higher price
sensitivity and lower opportunity cost of time, are more likely to make
the extra effort to look for a discount opportunity. Therefore, the
approach followed by Gerardi and Shapiro (2009) to distinguish
predominantly leisure routes from more mixed ones (i.e., leisure and
business) provides a useful framework for our investigation of
whether the two dispersive pricing scenarios are positively or negatively
related to market concentration. Such a relationship is
expected to be stronger in routes with a more heterogeneous demand.
In this paper, to test the link between market structure and online
fare dispersion, we construct a specific and direct measure of
travellers' heterogeneity on a route in terms of purpose of travel. An
advantage of our approach is that it uses publicly available data, which
can be consulted by the airlines in their design of pricing strategies.5
We find support to the hypothesis of market concentration being a
crucial factor in increasing online dispersion. As in Gerardi and
Shapiro (2009), the intensity in the use of price segmentation strategies
of both second and third-degrees appears to be positively
related to market concentration mostly in routes with a highly heterogeneous customer basis. Furthermore, the empirical evidence
reveals a different propensity across companies to use segmentation
strategies, in line with the view outlined in Borenstein and Rose
(2007) according to which the airline industry is, and has been,
characterised by a very high-level of business model experimentation
leading to a high degree of heterogeneity in pricing and competitive
strategies across carriers.
To our knowledge, no previous study has jointly investigated the
economic rationales underlying an airline's pricing strategy and its
actual implementation in an online setting. Because we find that the
use of discriminatory pricing intensifies in routes characterised by a
more heterogeneous customer basis, our analysis offers a compelling
case of how the Europ