Social learning and the theory of search

Lead Research Organisation: University of Kent
Department Name: Sch of Economics

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

Publications

10 25 50
publication icon
Cartwright E (2014) Strategic delay and information cascades in Journal of Economics

publication icon
Cartwright, E. (2009) Two agents and the theory of search

publication icon
Cartwright, E. (2007) Search, social learning and hterogeneity in payoffs

 
Description Search is a fundamental part of economic behaviour. For example, consumers search for products that they like or shops that offer good value for money while firms search for cost saving technologies or the most productive employees. In many contexts, individuals and firms learn not only from their own experience but also by observing the actions of others.

While there is plenty of literature available on search and social learning, the literature modelling social learning and search together is surprisingly sparse. So this project was motivated by asking how social learning impacts on search behaviour.

We sought a simple way to get some understanding of how social learning influences search. They considered the 'standard model of search' with two rather than the standard one agent, which they likened to considering two consumers who could sample shops in order to see what the different shops had to offer. A large part of the project was devoted to observing and comparing consumer search behaviour while alone with similar behaviour while observing others.

The project has confirmed some intuitive results that might be expected and also led to some more surprising insights, such as the potential benefits of preference heterogeneity for social learning.

There was an incentive to free-ride: in other words, it could be optimal for a consumer to search less, avoid the costs of sampling and eventually imitate the other consumer.

There was potentially an incentive to over-search; in other words it could be optimal for a consumer to search more than he would do on his own because this could encourage the other consumer to search more.

Free-riding was expected but the possibility of over-search was not expected and is a particularly interesting insight from the research.

We found that one consumer's optimal search depended on the other consumer's search resulting in multiple equilibria. They also found some equilibria to be 'unfair' in the sense that one consumer was doing most of the search and received a lower payoff while the other had a 'free ride' and received a higher pay-off.

At the aggregate level they found that social learning resulted in higher pay-offs and typically in finding a 'better shop, faster'.

One objective was to look at how heterogeneity influenced search. Sampling heterogeneity means that a consumer who is 'better at finding good shops' should do more search and get a relatively lower pay-off. Preference heterogeneity can lead to less free-riding but more over-search. Combinations of preference heterogeneity and sampling heterogeneity can lead to wildly different search behaviour.

Another objective was to explore how sequential search differs from simultaneous search. It was found that preference heterogeneity could increase the expected gains from social learning in sequential search because it led to more sequential sampling. This illustrates the importance of considering search, social learning and preference heterogeneity in combination.
Exploitation Route The possibility of free-riding in search is particularly important in terms of innovation. Our findings are, therefore, of interest in understanding how to encourage R&D and other innovative activities.
Sectors Digital/Communication/Information Technologies (including Software),Manufacturing, including Industrial Biotechology
 
Description The results have been or are being developed for publication in top economics journals. The results have also been widely disseminated at conferences and workshops.