Q&A: Machine-learning Model Tracks Trends in Public Finance Research
In celebration of its 40th anniversary, the journal Public Budgeting & Finance recently published an article that explores the leading topics in public finance and budgeting over the past four decades. Can Chen, an associate professor of public management and policy at Georgia State University, along with his former doctoral students Shiyang Xiao and Boyuan Zhao, employed a machine-learning technique called structural topic modeling (STM) to analyze 1,028 articles published in the journal between 1981 and 2020. The researchers identified 15 latent topics within the fields of public budgeting, public finance, and public financial management.
By comparing these topics to the content covered in standard exams for Certified Public Finance Officers (CPFO), Chen and his team discovered substantial overlap. However, they also noticed that some topics were mentioned less frequently, indicating potential areas of underexplored research and suggesting future research agendas in the field of public budgeting and finance.
Chen’s motivation for conducting this study arose from the inquiries of his doctoral students, who often sought guidance on the major trends within their field. He recognized the need to comprehend the overall landscape of public budgeting and finance, including recent topics. Consequently, Chen decided to employ machine learning to conduct a thorough analysis of the journal’s archives, utilizing the technological capabilities available through Georgia State University’s Digital Landscape Initiative.
While many other fields have already embraced machine learning for data analysis, Chen’s research is one of the first efforts to introduce this methodology into the study of public budgeting and finance. By leveraging ideas from various disciplines and applying them to their own field, Chen and his colleagues aim to expand the boundaries of knowledge and enhance the understanding of public finance topics.
The findings of this study hold significant implications for both scholars and practitioners in the government budgeting and finance domain. The authors noted a decline in practitioner engagement in the journal over time, emphasizing the importance of re-establishing connections with practitioners and fostering knowledge exchange between academics and professionals in the field.
Moreover, this research aids scholars, practitioners, and students in gaining a deeper understanding of the existing literature and identifying potential areas for collaboration. The study also provides valuable insights for doctoral students, assisting them in identifying new research topics and shedding light on the field’s historical evolution.
As the journal looks toward the future, Chen believes it is crucial to concentrate on pressing challenges such as healthcare, technology, and climate change. Conducting further research in these areas within the context of public finance and budgeting will contribute to effectively addressing these global challenges.
In conclusion, Chen’s innovative application of machine learning and text-mining techniques has facilitated a comprehensive analysis of the trends in public finance and budgeting research over the past 40 years. By employing this methodology, the study has illuminated the historical development of the field and identified potential areas for future exploration. This research not only benefits scholars but also encourages collaboration with practitioners and promotes the exchange of knowledge and expertise within the field of public budgeting and finance.
