Papers

Decision making under risk in Deal or no Deal ( 446.7 KB )
Latest Version: September 2006 with Yianis Sarafidis
Abstract
We analyse the choices of 399 contestants in the Australian version of the television game show Deal or no Deal. We calculate risk-aversion bounds for each contestant, revealing considerable heterogeneity. We then estimate a structural stochastic choice model that captures the dynamic decision problem faced by contestants. To address individual heterogeneity, we nest the dynamic problem within the settings of both a random effects and a random coefficients probit model. Our structural model produces plausible estimates of risk aversion, confirms the role of individual heterogeneity and suggests that a model of stochastic choice is indeed appropriate. We also examine generalisations to expected utility theory, finding that the rank dependent utility model provides substantially im- proved explanatory power. Finally, we do not find strong evidence in favour of an endowment effect for lotteries.


Political campaigning with momentum effects ( 300.0 KB )
Latest Version: March 2011 with Yianis Sarafidis
Abstract
This paper considers a game between two players (e.g., politicians) engaged in an advertising race which is subject to momentum effects. Momentum is modeled as a complementarity between current and past advertising spending in a way that is reminiscent of models of addiction and habit formation: the more effective a player's past spending has been, the more effective her current and future spending will be. We analyze a two-agent zero-sum differential game. We show that for symmetric games with a Cobb-Douglas payoff function the closed-loop equilibrium leads to an outcome that is equivalent to that of an open-loop equilibrium of the same game, but in which the momentum effect is more pronounced. In the polar case where the advertising stock does not decay over time, the advertising path is a decreasing power function. We also look at nonsymmetric games, for which we compute the open-loop equilibria numerically by solving the discrete-time analog of the continuous-time game. Our numerical results suggest that in games where one politician has more resources than her opponent, the gap between the politicians' advertising intensity increases during the race.


The determination of voter turnout ( 61.1 KB )
Latest Version: November 2004
Abstract
A panel of off-year general elections is employed to examine the institutional determinants of voter turnout in the United States. We employ instrumental variables and correlated random effects to contend with unobservable fixed determinants of voter turnout. We find only mixed support for the role of institutional factors in influencing turnout. Some support is found for the notion that constraining the powers of policy makers and the voting public tends to reduce turnout. Finally, the variation in cross-section regression results suggests that some care is required in examining individual cross-sections.


Pricing dynamics in the Australian airline market ( 345.7 KB )
Latest Version: March 2007 with Gordon Mills and Stephen Whelan
Abstract
We harness a unique dataset of airfares to examine price discrimination in the Australian air travel market. Price discrimination is achieved by restricting the menu of available fares, and through variation in fares with the day of booking. Our fixed effects estimator allows us to characterise both of these mechanisms. Price discrimination by booking day is particularly dramatic, with fares rising rapidly in the week before travel. We also find an ambiguous relationship between market structure and price discrimination. We find the greatest discrimination on routes involving competition between the two main airlines, Qantas and Virgin. However, we find greater discrimination on monopoly routes than routes pitting Virgin with Qantas? subsidiary, Jetstar.


Do firms play Markov strategies? ( 86.1 KB )
Latest Version: April 23, 2010
Abstract
The observation of Edgeworth-like petrol price cycles is routinely taken as evidence for the Maskin and Tirole (1988) price commitment model. We exploit an unusual policy innovation in the Western Australian retail petrol market that permits us to observe the precise timing of market actions. We find that, not only is the Maskin and Tirole model rejected in this setting, but its rejection hinges centrally on the restriction to Markov strategies.


Retail petrol price cycles in Western Australia ( 959.7 KB )
Latest Version: March 29, 2010 Hajime Katayama
Abstract
We employ a Markov-switching regression model to characterise pricing dynamics in the Western Australian retail petroleum market. We find the existence of Edgeworth price cycles of a similar nature to those recently observed in some other retail petrol markets. Cycles are frequent, asymmetric, and of substantial amplitude. Importantly, we observe cycles in an environment inconsistent with the timing specification of the leading theory of Edgeworth cycles. A recent change in policy imposed a discrete time environment on petroleum retailers in Western Australia, allowing us to rule out the Maskin and Tirole (1988) model. We also discuss episodes of disruption and evolution of the price cycle.


Cheap Tuesdays and the demand for cinema ( 532.8 KB )
Latest Version: Jordi McKenzie
Abstract
Many movie markets are characterised by extensive uniform pricing practices, hampering the ability to estimate price elasticities of demand. Australia presents a rare exception, with most cinemas offering cheap Tuesday ticket prices. We exploit this feature to estimate a random coefficients discrete choice model of demand for the Sydney region in 2007. We harness an extensive set of film, cinema, and time-dependent characteristics to build a rich demand system. Our results are consistent with a market expansion effect from the practice of discounted Tuesday tickets, and suggest that cinemas could profit from price dispersion by discounts based on observable characteristics.


Examining models of collusion: the market for lysine ( 275.0 KB )
Latest Version: January 2009
Abstract
We compare the experience with collusion in the market for lysine with the predictions of theory. The lysine market provides an ideal setting following the confessions of cartel participants in anti-trust investigations. Data availability allows demand and cost functions to be estimated and observed mark-ups compared with predictions. We find that several integral aspects of collusion in the lysine market are not adequately addressed in the literature: the dynamics associated with entry and investment; persistent asymmetries between firms; the cartel?s bargaining problem; and the existence of cheating in equilibrium. These issues are likely to have much wider applicability beyond the lysine market.


Collusion with a Competitive Fringe: An Application to Vitamin C ( 175.7 KB )
Latest Version: January 2004
Abstract
In a recent high profile case of collusion in the market for vitamin C, the cartel initially accommodated a fringe competitor before ultimately collapsing under the competitive burden. In this paper, the cartel's decision of when to dissolve is endogenised within a dynamic model of collusion. Demand estimates and cost information from the vitamin C market are used to calibrate the model. Results are intimately linked to the dynamic nature of the model. A cartel is found to persist only while fringe competitors remain small; fringe competitors invest heavily while a cartel is operating; entry deterrence is mitigated if cartel members are able to accommodate entry; and firms of an intermediate size are the most likely to accommodate entry.


The causal effect of substance use on adolescent sexual activity ( 205.8 KB )
Latest Version: March 2007 with Alex Acworth and Hajime Katayama
Abstract
Using the National Longitudinal Survey of Youth 1997, we estimate the causal effects of alcohol and marijuana use on youth sexual behaviour. Three primary aspects distinguish this study from those in the literature. First, we exploit panel data to control for unobserved heterogeneity through a difference in difference estimator. Second, we control for observed heterogeneity by adopting propensity score matching methods, and adopting a more complete set of control variables than has been considered to date. Third, we separately examine the effect of initiating substance use and ceasing substance use, and thus differentiate the likely effect of policy on current substance users and non-users. The results indicate striking differences across gender. For males, initiating alcohol or marijuana use increases the likelihood of engaging in sexual intercourse and uncontracepted sexual intercourse, while ceasing use decreases the likelihood only for alcohol. For females, there is no causal link in most cases. The soundness of our estimation method is confirmed by Rashad and Kaestner?s (2004) test; our method identifies no causal relationship between smoking cigarettes and sexual activities.