撒哈拉以南非洲的财政乘数叙述（英）

The Fiscal MultipliersNarrative of Sub-Saharan Hany Abdel-Latif, Khalil Bechchani, Antonio David, Thibault Lemaire WP/26/12 2026JAN IMF Working PaperAfricanDepartment Prepared byHany Abdel-Latif,Khalil Bechchani,Antonio David,Thibault Lemaire* Authorized for distribution byAmadou Nicolas Racine Sy IMF Working Papersdescribe research in progress by the author(s) and are published to elicitcomments and to encourage debate.The views expressed in IMF Working Papers are those of the ABSTRACT:This paper estimates the macroeconomic effects of fiscal consolidations in sub-Saharan Africausing a newly constructed narrative dataset of discretionary fiscal policy actions for 14 countries over 1990–2024.The dataset, documented in Abdel-Latif et. al. (2025), identifies fiscal measures undertaken for reasons unrelatedto current or prospective economic conditions, providing a credible basis for estimating fiscal multipliers. Theresults show that a fiscal consolidation of 1 percent of GDP reduces output by about 0.54 percent after two years,a larger effect than what is found using alternative identification methods. Fiscal consolidations also reduce RECOMMENDED CITATION:Abdel-Latif, H. & Bechchani, K.& David, A., & Lemaire, T. (2026).“The FiscalMultipliers Narrative of Sub-Saharan Africa.”IMF Working Paper No. WP/26/12. WORKING PAPERS The Fiscal Multipliers Narrative of Hany Abdel-Latif,Khalil Bechchani,Antonio David,Thibault Lemaire 1. Introduction Overlapping shocks and policy slippages have heightened debt vulnerabilities in sub-Saharan Africa (SSA),compelling many countries in the region to undertake substantial fiscal consolidation efforts. The successof these adjustment plans requires a carefully calibrated, evidence-based approach to guide decisions on This paper addresses this gap by exploiting a newly developed narrative dataset of discretionary fiscalpolicy actions for 14 SSA countries over 1990–2024, documented in detail in Abdel-Latif et al. (2025). Thedataset is constructed using a combination of human examination of contemporaneous IMF staffreportsand artificial-intelligence–assisted text analysis to identify fiscal measures explicitly motivated by long-termobjectives rather than by current or prospective economic conditions, following the methodology outlined in and Kano, 2023a; Badru et al., 2025), typically lower than 0.5 (compared to estimates close to 1 in AdvancedEconomies, Fig. 1). This could possibly be explained by economic factors, such as the prevalence of alarge informal sector (Lemaire, 2020; Colombo et al., 2024; Bechchani, 2025) and lower tax levels relative to However, these low estimates may also be linked to biases associated with conventional identificationstrategies used to disentangle the exogenous effects of fiscal policies. Identification problems are particularly This paper contributes to the literature in several ways. First, it provides the first application of the narrativedataset on fiscal consolidation developed by Abdel-Latif et al. (2025) to estimate fiscal multipliers in SSA andoffers new evidence on the dynamic effects of fiscal consolidations on output. Second, while prior research onthe consequences of fiscal tightening has focused mainly on output effects, this paper also explicitly examinesexternal adjustment. In particular, building on the implication of open-economy macroeconomic models We find that a fiscal consolidation of 1 percent of GDP reduces output by about 0.54 percent after two years.Consolidations also lead to a reduction in imports, an improvement in the current account balance, and adepreciation of the real effective exchange rate. These findings are in line with standard Keynesian effects Furthermore, the evidence suggests that the composition and timing of fiscal adjustments further influencetheir macroeconomic effects. Over a three-year horizon, spending cuts have larger multipliers than tax The remainder of the paper is structured as follows. Section 2 reviews the related literature and discussesidentification challenges in estimating fiscal multipliers in SSA and other emerging market economies.Section 3 summarizes key features of the narrative dataset introduced in Abdel-Latif et al. (2025). Section 4 2. Literature Review Conventional approaches to identify exogenous fiscal policy actions—including changes in the cyclicallyadjusted primary balance (Alesina and Ardagna, 2010; Badru et al., 2025), structural VAR models (Ilzetzki Gorodnichenko, 2013a; Arizala et al., 2021)—often fail to isolate discretionary fiscal actions from endogenousmovements in revenues and expenditures and are subject to significant measurement error (David et al.,2022), since they may capture "statistical" or "mechanical" variation in fiscal balances rather than discretionary The narrative approach à la Romer and Romer (2010), tackles these identification challenges by directlyexamining the stated motivations behind policy actions.By focusing on fiscal meas