Data science in political economy
This thesis provides three applications of Data Science methodologies in Political Economy, combining them with established techniques in the literature. Its aim is to show how important is to keep an open dialogue with different disciplines to broaden the standard toolkit of an empirical economist and be able to tackle research questions in a novel way.
In chapter 2, I study the effects of terrorist attacks on British politicians' immigration rhetoric on social media. I scrape the Members of the Parliament's Twitter accounts and identify the immigration-related Tweets, which are then leveraged to frame a natural experiment. Looking at the 2017 Manchester bombing as my main event study, I find a substantial decrease in the expected number of immigration-related Tweets after the incident. I hypothesise that this "muting effect" results from the risk-averse attitude of politicians during the election campaign.
Chapter 3 explores strategic voting in the United Nations General Assembly (UNGA). We first predict the expected behaviour of country representatives in the UNGA. Next, we construct a network that describes the structure of the deviations underlying the observed votes. The graph is used to compute a Reciprocity Index. Through this statistic, we find that deviations from the expected votes are systematically not reciprocated. The conclusions are consistent with a narrative of vote buying and question the unweighted voting system of the Assembly.
In chapter 4, we investigate government's information processing and its implications for policy responses. To study how the Mexican government processes a specific signal (opinion pieces from newspapers), we devise a News Index that creates a link between informational inputs and policy outputs. We find that changes in the index are associated with policy overreactions. The findings are further assessed through a natural experiment. Overall, the results are consistent with the dominant theory of disproportionate information processing in government's decisions.
History
File Version
- Published version
Pages
155.0Department affiliated with
- Economics Theses
Qualification level
- doctoral
Qualification name
- phd
Language
- eng
Institution
University of SussexFull text available
- Yes