世界银行-欧盟结构和投资基金支出的地区财政乘数是否很大？：证据的重新评估（英）-2024.1-32页_659kb

Summary: Regional Fiscal Multipliers on EU Structural and Investment Fund Spending

Introduction

This paper reassesses the short-term fiscal multipliers of EU Structural and Investment Funds, finding little evidence of large multipliers at either the national or subnational (NUTS2) level.
While EU funds boost investment almost euro-for-euro, their stimulatory effect on GDP in the short term is modest, often below one or statistically insignificant.

Methodology

• Features NUTS2/ country-level data (2000-2018).
• Uses "modeled" EU payment data to better align with actual spending timing.
• Employs instrumental variables (IV) with a "leave-one-out predicted disbursement schedule" approach to address potential endogeneity by leveraging variations in neighboring regions/countries.
• Controls for time and region fixed effects, and includes lagged GDP growth.
• Estimates contemporaneous (impact), cumulative (2-3 years), and spillover effects.

Key Findings / Results

1. Multipliers are generally small or insignificant: The primary conclusion is a lack of evidence for large short-term fiscal multipliers on EU fund spending.

   • Results hold across different estimation methods (OLS vs IV) and time horizons (contemporaneous vs cumulative).
   • At the NUTS2 level, a preferred IV specification yields a precisely estimated zero multiplier. Other OLS/IV estimates are typically less than 1 or insignificant.
   • At the country level, multipliers are usually small and insignificant (estimated less precisely).

2. Investment Effects are Robust: Despite small GDP effects, EU funds significantly boost investment contemporaneously.

   • Investment multipliers reach close to or significantly exceed 1, particularly in Central and Eastern Europe (CEE).
   • The difference between modeled and raw payment data often biases IV investment multipliers downwards, but modeled payments are preferred.

3. Spillovers and Anticipation Effects:

   • Little evidence of significant spillover effects on neighboring regions' GDP.
   • Some tentative negative "anticipation effects" were found at the country level (pre-spending GDP growth), but less so at NUTS2. Anticipation effects on investment were more pronounced in non-CEE subnational regions, suggesting potential "dynamic crowding out."

4. Heterogeneity:

   • The contemporaneous OLS multiplier was slightly larger for CEE countries.
   • Specific funds (like Cohesion Fund - CF) had significant positive investment multipliers but not always for GDP.
   • Multipliers were generally insignificant across different programming periods, time splits, and regional (North/South, richer/poorer) disaggregations.

Policy Implications

• Policy makers should have realistic expectations about the short-term stimulatory effects (fiscal multipliers) of EU funds on regional GDP.
• The focus should be on the longer-term benefits of EU funds in achieving sustainable green growth and digitalization, consistent with their original purpose.
• Multiplier estimates are relevant for relative effects (growth difference between regions) rather than predicting absolute growth boosts from aggregate fiscal spending under all monetary conditions.

Comparison with Literature

This paper contrasts with recent studies (Coelho, Durand & Espinoza, Canova & Pappa) that often report higher multipliers. Key differences in methodology (data, controls, treatment of endogeneity) and specifications may explain these divergent findings, particularly regarding time period selection and macroeconomic conditions.

Conclusion

Contemporary EU fund spending does not generate substantial short-term fiscal multipliers for regional GDP growth, despite strong effects on investment spending. Policy focus should be on long-term productive capacity building.