Research & Projects
Research & Projects
Publications
A Python Package to Assist Macroframework Forecasting: Concepts and Examples with Sakai Ando and Sultan Orazbayev. IMF Working Papers 2025, 172 (2025)
Abstract: In forecasting economic time series, statistical models often need to be complemented with a process to impose various constraints in a smooth manner. Systematically imposing constraints and retaining smoothness are important but challenging. Ando (2024) proposes a systematic approach, but a user-friendly package to implement it has not been developed. This paper addresses this gap by introducing a Python package, macroframe-forecast, that allows users to generate forecasts that are both smooth over time and consistent with user-specified constraints. We demonstrate the package’s functionality with two examples about forecasting US GDP and fiscal variables.
Work In Progress
AI Meets Fiscal Policy: Mapping Government Spending Actions Across 64 Countries with Davide Furceri, Adrian Peralta Alva and Nikhil Patel.
Abstract: We develop an AI‑assisted narrative method to construct a database of fiscal actions and use it to compute fiscal multipliers for 64 countries. A fixed GPT‑4.1 prompt applied to Economist Intelligence Unit Country Reports classifies, for each country–quarter since 1952, the stance and motivation of discretionary spending. We validate this series using Romer and Romer (2019)’s episodes, lead–lag dynamics of spending and macro fundamentals, and existing action‑based consolidation data. We use it as an internal instrument in VARs to estimate government‑spending multipliers. The average cumulative multiplier is about 0.64 at one year and 0.84 at two years: advanced economies display modest multipliers below 1, and larger for spending contractions relative to expansions. Emerging and developing economies, on the other hand. have larger one and two‑year multipliers (above 1 at two years). Multipliers vary systematically with trade openness, public debt, and political uncertainty, while a detailed narrative measure of political support has little explanatory power.