Abstract

This presentation delves into the use of artificial intelligence (AI) and machine learning in municipal budget analysis, offering a new perspective on public finance. The presentation includes two key case studies. In the first, focusing on the United States, AI tools analyze local government budgets to predict various local characteristics such as age distribution, rural population, education levels, employment outcomes, racial composition, and political preferences. Notably, the results indicate a marked increase in the predictive power of local budgets for national political trends, suggesting a closer link between local fiscal decisions and broader political shifts. The second case study examines public policies decisions of 8,000 Italian municipalities transforming budget data into vectors and calculating the cosine similarity between pairs of municipalities. This approach reveals patterns in how resources are allocated, with the cosine similarity highlighting the degree of similarity or divergence in fiscal policies among different localities. The study also introduces the Municipal Fiscal Divergence Index (MFDI), quantifying the deviation of local budgets from expected norms based on local characteristics. Together, these case studies demonstrate the profound impact AI and machine learning can have on understanding and interpreting municipal budgets