2013年-IMF国际货币组织全球_Fiscal_Policy_and_Lending_Relationships_48页_1mb

报告摘要

Summary of "Fiscal Policy and Lending Relationships"

Core Content

This working paper investigates the impact of fiscal policy on loan market conditions in the United States, with a focus on how government spending shocks influence the bank spread and lending behavior. It combines empirical analysis using a Structural Vector Autoregression (SVAR) model with a theoretical Dynamic Stochastic General Equilibrium (DSGE) model that incorporates lending relationships as a key friction in the financial system.

Main Points

Empirical Analysis (SVAR):
The paper estimates the response of the US economy to a 1% government consumption expenditure shock using quarterly data from 1954q1 to 2007q4. The results show that:
- Real output increases significantly, peaking at around 1.75 after two years and a half.
- Lending increases, with a peak of almost 5% after three years.
- The bank spread declines by about 50 basis points immediately and reaches a negative peak of 70 basis points after two years.
- Private consumption is crowded in, while private investment initially falls but eventually experiences a crowding-in effect.
- Hours worked and real wage increase, and the price mark-up decreases over time.
Theoretical Model (DSGE):
A Real Business Cycle (RBC) model is developed that integrates lending relationships as a mechanism through which fiscal shocks are transmitted. The model:
- Incorporates deep habits in consumption and lending, reflecting how past consumption and lending patterns affect current behavior.
- Demonstrates that stronger lending relationships amplify the financial accelerator effect of fiscal policy.
- Matches the empirical findings of the SVAR model, showing that government spending shocks lead to better credit market conditions, which in turn boost economic activity.
Key Concept: Financial Accelerator Effect
The paper argues that lending relationships create a friction in the loan market that enhances the transmission of fiscal shocks. This is because:
- Banks have an informational advantage over borrowers, leading to a hold-up problem.
- Borrowers find it costly to switch lenders, which results in a financial accelerator effect where fiscal policy influences lending and credit spreads more effectively.

Key Information

Methodology:
- The paper uses a SVAR model to estimate the effects of government spending shocks on the economy.
- A DSGE model is developed with a focus on lending relationships and deep habits in consumption and borrowing.
Data and Variables:
- Quarterly US data from 1954q1 to 2007q4 is used.
- Variables include real per-capita government consumption, real per-capita output, average marginal tax rate, private lending, bank spread, private consumption, private investment, hours worked, real wage, and price mark-up.
Identification of Shocks:
- Government spending shocks are identified using the Blanchard-Perotti (BP) approach, which accounts for implementation and decision lags.
- Robustness checks are performed using SPF/Greenbook forecasts to purify shocks of any predictable component.
Model Features:
- Deep habits in consumption and lending are modeled as external habits, which are more tractable than internal ones.
- The model is extended to include sticky prices and a central bank, forming a New Keynesian (NK) version.
- The equilibrium conditions and steady-state are derived to ensure consistency with empirical results.

Structure and Components

I. Introduction

Highlights the shift in focus from monetary to fiscal policy during the Great Moderation and the recent crisis.
Notes the importance of credit market conditions in business cycle dynamics.
Outlines the paper's contributions: empirical and theoretical analysis of fiscal policy transmission through lending relationships.

II. Empirics

Presents the SVAR model and its results.
Discusses the robustness of the model to various specifications and anticipation effects.
Provides impulse response plots for all variables in response to a government spending shock.

III. Model

Describes the DSGE model with households, government, entrepreneurs, final good firms, and banks.
Details the deep habit formation in consumption and lending.
Explains the role of the banking sector and the financial accelerator mechanism.
Outlines the functional forms and parameter choices used in the model.

IV. Results

Discusses the financial accelerator effect and how it is amplified by stronger lending relationships.
Shows the impact of fiscal policy on economic activity via improved credit conditions.

V. Effects of Model Features

Analyzes the sensitivity of results to deep habit persistence and the degree of lending relationships.
Examines the implications of including government consumption in the utility function.

VI. Conclusion

Summarizes the main findings: fiscal policy has significant effects on the loan market, and lending relationships enhance the transmission of these effects.
Suggests future research directions, including the role of monetary policy and the microeconomic foundations of lending relationships.

Key Findings

Empirical Evidence:
- A government spending shock leads to a decline in the bank spread and an increase in lending.
- These effects are consistent with the financial accelerator mechanism.
- The SVAR results are robust to various specifications and anticipation effects.
Theoretical Implications:
- Lending relationships are modeled as a friction that affects the informational asymmetry between banks and borrowers.
- Deep habits in consumption and lending are used to capture this friction without explicitly formalizing asymmetric information.
- The model shows that stronger lending relationships increase the effectiveness of fiscal policy.
Policy Relevance:
- The paper suggests that fiscal policy can be more effective in economies with strong lending relationships.
- It provides a theoretical framework for understanding the transmission mechanism of fiscal shocks through the financial sector.

References and Appendices

References:
- Includes key studies on fiscal policy, credit spreads, and lending relationships.
- References to the SVAR and DSGE literature are provided for context.
Appendices:
- Appendix A: Robustness checks of SVAR results, including substitution of tax measures, inclusion of inflation and interest rate variables, and purification of government spending shocks.
- Appendix B: Sensitivity exercises for the DSGE model, testing the impact of deep habit persistence and lending relationships.
- Appendix C: NK extension of the model with sticky prices and monetary policy.
- Appendix D: Symmetric equilibrium and steady-state conditions for the DSGE model.

Figures

Figure 1: Impulse responses of the SVAR model to a 1% government spending shock.
Figure 2: Additional impulse responses for the same shock.
Figure 3: Illustration of the financial accelerator effect.
Figure 4: Effects of model features on the response of the economy.
Figure A.1: Robustness checks for the SVAR results.
Figure B.1–B.4: Sensitivity analyses of the DSGE model.
Figure C.1: Comparison of flexible and sticky price versions of the model.