|
|
|
| How Do Implicit Guarantees by Local Governments Affect the Real-Time Credit Risk of Small and Medium-Sized Banks? |
| ZHU Xiaoquan, CHEN Zhuo, XU Tong, HE Zhiguo
|
China School of Banking and Finance, University of International Business and Economics;
PBC School of Finance, Tsinghua University;
HKU Business School, The University of Hong Kong;
Graduate School of Business, Stanford University |
|
|
|
Abstract With the risk event of Baoshang Bank breaking the expectation of rigid payment, regulators have realized that the key premise for mitigating the risks of small and medium-sized banks (thereafter SMBs) lies in identifying those risks. Due to the delayed and possibly manipulated disclosure, it is difficult to monitor the SMBs' credit risks in real time through traditional financial indicators such as non-performing loan ratios. In this context, our paper first investigates whether the spread of Interbank Negotiable Certificates of Deposits (IBNCD) can serve as a real-time measure of credit risk for SMBs. Furthermore, the development of SMBs has led to the possibility that IBNCD pricing could be influenced by factors beyond the issuer's fundamentals, with the implicit guarantees from local governments being the most typical example. The key research question of the paper is how the capabilities and willingness of implicit guarantee from local governments affect the IBNCD pricing of regional SMBs.
From a theoretical perspective, the paper includes a preliminary hypothesis, which posits that the IBNCD spread can serve as a forward-looking indicator of SMBs' credit risks. On the one hand, an increase in the IBNCD spread reflects a rise in the bank's interbank market financing costs, which ultimately transmits to the credit risks. Therefore, the IBNCD spread should be consistent with low-frequency credit risk indicators (e.g., non-performing loan ratios). On the other hand, after the global financial crisis, the interaction between shadow banking risks and on-balance-sheet risks deepened, and the rapid growth of off-balance-sheet wealth management products increases banks' rollover risk. If the IBNCD spread reflects SMBs' credit risks in real time, it should also predict risk indicators with higher frequency, such as the yield shortfall of wealth management products.
Building on this preliminary hypothesis, the paper outlines the theoretical foundation of how local government implicit guarantees influence the credit risk of regional SMBs. From an investor's perspective, the financial strength of local governments reflects their capacity to provide guarantees. The weaker a local government's ability to provide guarantees, the higher the risk premium demanded by investors for local SMBs, which is reflected in higher IBNCD spreads. On the other hand, we argue that the expansion of off-balance-sheet liabilities of local governments is a stronger signal of their willingness to provide guarantees. According to the too-big-to-fail theory, local governments in regions with higher off-balance-sheet debt growth, i.e., a larger expansion of Local Government Financing Vehicle (LGFV) debt, are more likely to provide implicit guarantees. More importantly, LGFV uses financial resource from regional SMBs to help local government to boost economic growth; thus, a larger expansion of LGFV debt indicates stronger willingness of local governments to provide implicit guarantees, which is reflected in lower IBNCD spreads.
Our sample includes publicly issued IBNCD by urban and rural commercial banks in prefecture-level cities between 2014 and 2019. We examine the predictive power of IBNCD spreads for bank credit risk indicators. The results demonstrate that widening IBNCD spreads among SMBs are associated with elevated non-performing loan ratios, heightened customer concentration, and increased loan risk in the subsequent quarter, alongside more pronounced shortfalls in wealth management product returns. This indicates that the IBNCD spread has an advantage in capturing the real-time credit risk of SMBs.
Furthermore, we explore how the capabilities and willingness of local governments to provide implicit guarantees affect the credit risk premia of SMBs. We use the fiscal deficit rate as a measure of the local government's guarantee capacity and the additional growth of LGFV debt (i.e., the residual from regressing the increment of LGFV debt on fiscal pressure) to measure the local government' s willingness to provide implicit guarantees. The results show that the fiscal burden of the cities where the issuing banks locate significantly increases the SMBs' credit risk. In contrast, the additional expansion of LGFV debt is seen as a strong signal of the willingness to provide implicit guarantees, which lowers the IBNCD spread.
Finally, we try to address the endogeneity issues through two methods. First, we use the guarantee capacity and willingness of other cities in the same province (but excluding the provincial capital cities) as instrumental variables, and the results are consistent with OLS. Second, we examine the Baoshang Bank event as a shock, finding that after the breakdown of the implicit guarantee expectation, the spread on IBNCD issued by SMBs in regions with weaker guarantee capacities increased more significantly, and the issuance success rate decreased more than regions with stronger guarantee capacities.
Our findings have key policy implications. First, an early identification and risk warning system for problematic banks is crucial, with the market-oriented characteristics of IBNCD serving as valuable complement to regulatory tools. Second, future policies should avoid distorting incentives for SMBs through the evaluation of local economic development. Third, regulators must address non-performing loans in SMBs. While this may increase the burden of supporting the financing of small enterprise in the short term, it helps authorities to manage the monetary creation in the long term by curbing the arbitrage in the interbank market. Future research can track the impact of implicit guarantees on IBNCD pricing during the debt reduction process.
|
|
Received: 08 August 2023
Published: 22 May 2025
|
|
|
|
|
|
|
|
