The Composition of Fiscal Consolidation

Abstract: This paper examines the differential effects of tax-based versus spending-based fiscal consolidations on debt sustainability and economic growth using narrative consolidation data for 17 advanced economies from 1978 to 2016. The Adler et al. (2024) dataset, published as International Monetary Fund (IMF) Working Paper 24/210, provides a comprehensive revision of the Alesina et al. (2015, 2019) narrative data, extending coverage to 17 countries including the Netherlands and updating shock magnitudes throughout the sample period. Macroeconomic outcome variables are drawn from World Bank, Organisation of Economic Co-operation and Development (OECD), and IMF primary sources. The narrative shock sample is bounded in 2016 to ensure the three-year debt outcome window remains pre-pandemic; local projections use shocks through 2014 to guarantee all five horizons avoid COVID contamination. 

Employing country and year fixed-effects panel regressions with clustered standard errors, local projections methods, and linear probability models, this paper reveals that spending-based consolidations are associated with higher probabilities of debt reduction and more favorable growth outcomes than tax-based consolidations. A one-percentage-point tax-based consolidation reduces GDP growth contemporaneously by 0.59 percentage points, versus 0.29 percentage points for spending-based consolidations. Local projections reveal that tax-based consolidations impose increasingly negative effects over time (reaching −2.94 percentage points cumulatively five years after the event), while spending-based consolidations generate a trajectory that turns positive after two years and reaches +1.81 percentage points after five years. Tax-based consolidations reduce the probability of substantial debt reduction by 9.5 percentage points and of debt stabilization (debt-to-GDP ratio does not increase over three years) by 12.4 percentage points. Spending-based consolidations significantly raise the probability of substantial debt reduction by 11.1 percentage points. These results are robust to heterogeneity by initial debt level.

Find the paper here.