国际清算银行-外汇市场主权风险与银行贷款：1999年土耳其地震的证据（英）-2023.4-59页_3mb

Sovereign Risk and Bank Lending: Evidence from the 1999 Turkish Earthquake

Key Points:

1. Research Objective: To identify the transmission of government risk to bank lending via banks' balance sheets.
2. Natural Experiment: Utilized the 1999 Marmara Earthquake as an exogenous fiscal shock, boosting Turkey's sovereign risk.
3. Empirical Method: Employed a difference-in-differences approach, comparing banks with varying exposures to government bonds (holdings as % of assets) before and after the earthquake.
4. Core Findings:
   • Banks with higher government bond exposure experienced greater declines in net worth and lending post-earthquake.
   • This effect is attributed to the balance sheet channel, where reduced bond values (due to sovereign risk) constrained bank lending capacity.
   • The channel is estimated to explain ~50% of the total lending decline observed between August and November 1999.
5. Robustness & Identification:
   • Accounted for alternative channels (e.g., demand shock, financial repression).
   • Pre-trends showed no significant performance differences between high/low bond exposure banks, validating the DID design.
6. Policy Implications: Highlights the importance of the sovereignty-bank nexus during fiscal distress, particularly in the context of rising public debt post-COVID.

Methodology & Significance:

• Heterogeneous treatment effects identified based on government bond holdings.
• Addressed endogeneity and reverse causality concerns through the natural experiment framework.
• Provides evidence for the bank-sovereign doom loop as a key transmission mechanism in sovereign debt crises.