Speaker: John Hey, University of York
"This work starts from the well-documented observation that many consumers stick with an oldsupplier for longer than is necessary, thus leading to a significant waste of money. This is truein a large number of fields: gas supplier; electricity supplier; bank; telephone supplier; insuranceprovider (for example, Giulietti et al, 2005; Honka, 2014). People seem to find it difficult toswitch from the old and familiar to the new. There are two market frictions that are well knownin the literature for preventing people from switching to a new and more convenient provider: (1)the cost of searching information about new suppliers, that is the search cost, and (2) the cost ofmoving to a new supplier, that is the switch cost. Previous experimental and theoretical literaturehas studied these two frictions in isolation. However, field evidence show that these two frictionsfrequently occur together, and a recent theoretical article by Wilson (2012) suggests that switchand search cost should be studied in unison in order to fully understand the consequence of thesefrictions on individual and market behaviour, and suggests that the search costs have a greaterimpact on individual purchase choices and then on market outcomes. The experiment tests ifthe individual behaviour under search and switch costs is in line with the optimal choice rule de-scribed in the Wilson paper. The results of the experiment will be analysed in two parts and thefollowing questions asked: (1) do subjects appear to be using a reservation rule to determine theirsearch, and are they computing this optimally? (2) when subjects stop searching, do they switchoptimally? These two question regarding the optimality of the individual decision making areanswered through a comparative statics analysis based on Wilson's optimality rule predictions,and fitting the data with Wilson Model. In addition, we also fit the data with an extension ofWilson Model that incorporates risk-aversion, which generally provides provide a better fit to the data."