One of two recently awarded British Academy grants, awarded to Dr Eugene Jeong, aims to develop theoretical models to understand when governments face the correct incentives to disclose the statistics.
Whilst COVID-19 has shone a spotlight on statistics to do with infection rates, concerns have been raised about the accuracy of statistics, both in terms of measurement, and in terms of exactly what governments choose to disclose.
The theoretical analysis will be complemented with experiments to see how people respond to over or under-reporting versus truthful reporting. The project will develop valuable evidence on incentives to truthfully disclose statistics and the implications of not doing so. It is shown that uncertainty led by underreporting can incur a higher expenditure for the government than when there is no uncertainty. The excess expenditure increases as either there is more uncertainty, or the number of cases grows.
The second grant is for Dr Matthew Polisson’s research into experimental tests of rational decision making. Rational decision making under risk/uncertainty and over time is central to economic analysis, and it is natural that experimentalists should want to test both (older) mainstream and (newer) behavioural models of choice within these contexts.
Generally speaking, many experiments to-date have collected a relatively small number of decisions from each subject, with these problems somewhat narrowly tailored to refute particular theories (or components of those theories). The emphasis in this research is to present subjects with problems that are representative (in a statistical and economic sense) in order to provide a better positive account of preferences.
In order to do this, the researchers develop tests which are (a) comprehensive (in the sense that they apply to each model taken as a whole) and (b) nonparametric (in the sense that they avoid making auxiliary functional form assumptions). They then apply our tests to rich individual-level data which are amenable to statistical analysis.