Fiscal Adjustments in Latin America: The Role of Composition and Public Debt

Abstract

Fiscal consolidations are widely regarded as contractionary, yet whether the composition of adjustment matters remains an open question, particularly in emerging economies. We study the macroeconomic effects of fiscal consolidation in Latin America using a narrative dataset of 61 deficit-driven adjustment episodes identified from IMF Article IV reports for nine economies over 1989–2024, constructed with the assistance of large language models. A 1% of GDP consolidation reduces real GDP by approximately 0.9% on impact and by 0.5% after two years.

In the unconditional sample, tax-based and spending-based plans have indistinguishable output costs. Once we condition on the initial level of public debt, however, a sharp asymmetry emerges: at elevated debt levels, the output cost of a tax-based consolidation is approximately 4.7 times that of a comparable spending-based plan. Spending-based consolidations also raise the probability of a durable improvement in the primary balance by 9 percentage points at
the two-year horizon and 16 percentage points at the three-year horizon, while tax-based plans have no detectable effect. Our findings suggest that fiscal adjustment is fundamentally state dependent: the composition of consolidation becomes a first-order determinant of both its macroeconomic costs and its fiscal effectiveness precisely when public debt is high.