Text analysis to identify partisan influence in achieving sustainable development goals (SDGs)

Abstract

Partisan influence on public policy has been studied cross-nationally in macro level by examining variables such as GDP and democracy level. Political scientist like Tufte (1992) have argued parties do matter in OECD countries as leftist parties are inclined to spend more on social policy and equality, while Huber et al. (2008) refuted partisanship does not matter in social policy spending in Latin American countries. With regards to Latin American OECD countries, partiesdo-matter and parties-do-not-matter hypotheses were continuously debated as both seemed to correspond. As Mexico went through a huge political partisanship movement from non-left party to left party in the 100 years after the Mexican revolution, this paper chooses micro level countrybased approach research on partisan influence on public policies with the case of Mexico. Under this sudden shift, this paper questions if partisan influence in shaping and implementing public policy is attainable through examining both incumbents’ speeches. Unlike the traditional methodology identifying partisan influence, the purpose of the research is to introduce a new methodological approach analyzing partisan influence by using stenographic records of presidential speeches and discourses as a data set. Also, for the policy categorization, policies were aligned based on Sustainable Development Goals (SDGs)’ social, economic, and environmental pillars. By sorting each left and non-left presidential speeches and discourses as per 17 different categories of SDGs through computer-based unsupervised learning text analysis, the study has figured out that left incumbent in Mexico tends to put more emphasis on social policies than nonleft party in rhetoric level. This is the first such analysis by using a new methodological tool scrutinizing political party’s public policy direction aligned with SDGs, based on presidential speeches and discourses as a data set.