Evaluating financial fragility: a case study of Chinese banking and finance systems

The paper addresses how to evaluate and monitor financial fragility, with a focus on China’s banking system and broader financial system. It motivates the work by noting that bank fragility threatens economic stability and can precipitate crises. Key bank metrics—non-performing loan (NPL) ratios, average return on total assets (ROAA), liquidity ratios, and capital adequacy ratios (CAR)—are highlighted as central to stability. The authors argue that when accumulated vulnerabilities reach critical thresholds, crises can become unavoidable, underscoring the need for reliable fragility indicators. The study’s purpose is twofold: (1) to develop a financial fragility evaluation index system and apply factor analysis to cross-sectional data for 15 Chinese banks (2018) to score bank fragility; and (2) to construct and track a composite Financial Fragility Index (FFII) for China’s financial system using subsystem indicators spanning 2007–2022, thereby capturing temporal dynamics and informing policy. The paper emphasizes the importance of a strong economic environment, robust market early-warning mechanisms, effective financial monitoring, and managing external exposure in sustaining financial stability.

The study situates its contribution within a broad literature on bank risk and financial fragility. Prior research links loan pricing and borrower default risk to bank fragility (Yu et al., 2015), liquidity risk to higher NPLs and lower profitability (Djebali & Zaghdoudi, 2020), and underscores the roles of CAR and asset quality in risk assessment (Hajar & Habib, 2020). Some studies suggest banks with lower liquidity risk may take aggressive actions that heighten vulnerability (Smaoui et al., 2020), while bank balance sheet structures—low current assets, high foreign liabilities, leverage—predict crisis risk in developing countries (Haan et al., 2020). Increased credit derivatives holdings may raise systemic risk (Halili et al., 2021). NPL stocks adversely affect bank lending and stability (Serrano, 2021), whereas stronger capital and profitability reduce fragility (Kanga et al., 2020). Broader determinants include macroeconomic conditions, policy, market concentration, monetary policy, wealth distribution, regulation, and foreign exchange exposure (Laura et al., 2015; Kim et al., 2020b; Sushanta & Ricardo, 2013; Fabio & Claudio, 2014; Mitkov, 2019; Danilo, 2020; Georgiadis & Zhu, 2021). Methodologically, the authors draw on factor analysis foundations (Spearman; Qin & Lin, 2021; Kim et al., 2020a) and fragility/early-warning frameworks (Kaminsky, 2006; Gobert et al., 2002; Karadima & Louri, 2020).

The study develops a two-part methodology. 1) Bank fragility (cross-sectional, 2018): Using data for 15 major Chinese banks (ROAA, liquidity ratio, CAR, NPL ratio), the authors apply factor analysis to derive fragility scores. To ensure comparability and align the direction of indicators so that a lower transformed value reflects lower fragility, performance-type indicators (ROAA, liquidity, CAR) are transformed accordingly. Factor extraction with Varimax rotation yields three principal components explaining 92.88% of variance. Each bank receives scores on the three components and a total fragility score (F). 2) Financial system fragility (time series, 2007–2022): The authors construct a Financial Fragility Index (FFII) from four subsystems: A) Economic and environmental fragility (e.g., GDP growth, fixed asset investment growth, M2 growth, CPI); B) Financial market early-warning fragility (e.g., P/E ratios for Shanghai and Shenzhen, total stock market value/GDP, securities index volatility for both exchanges, debt dependence); C) Financial surveillance fragility (e.g., M2/M1, one-year actual deposit interest rate, growth of loans by financial institutions, fiscal deficit/GDP); D) Financial outward-oriented fragility (e.g., external debt ratio, foreign trade dependence, current account balance/GDP, debt service ratio, import cover). Each indicator is mapped to degree-of-fragility (DOF) bands (Safe/Normal/High/Dangerous) with specified boundary ranges (Table 1). Subsystem composite scores are computed by averaging indicators within each subsystem (Fi = ΣFij/j). The overall FFII is a weighted average with greater emphasis on early warning and monitoring: FFII = (FA + 2FB + 2FC + FD)/6. The FFII is then categorized into DOF bands: Safe (0–17), Normal (17–25), High (25–35), Dangerous (>35). Data sources include the China Statistical Yearbook, China Financial Yearbook, China Macro Statistics Database, and the National Bureau of Statistics.

• Bank fragility rankings (2018): Using factor analysis, Bank of Ningbo has the lowest fragility score (F = -0.89), attributed to stronger ROAA and CAR and a lower NPL ratio. Banks with low fragility also include Bank of Communications, China Construction Bank, and Industrial Bank. China Minsheng Bank has the highest fragility score (F = 0.83), driven by lower CAR and higher NPLs. The three principal components together explain 92.88% of the variance in the banking indicators. - Drivers within banks: Liquidity ratio and NPL ratio are key determinants of banks’ fragility scores. Banks with better liquidity and lower NPLs exhibit lower fragility. - System-wide fragility dynamics (2007–2022): The FFII displays notable fluctuations tied to major events: (1) 2007–2009 subprime/global financial crisis—FFII reached dangerous levels; (2) 2015 market turbulence—elevated early-warning fragility; (3) 2019 COVID-19 shock—another fluctuation in fragility. - Subsystem patterns: • Economic/environmental fragility trended mostly Normal before 2011, moving below the Safe line thereafter. • Financial market early-warning fragility showed episodes of High/Dangerous levels, notably around 2015, driven by stock market volatility, total market cap/GDP, and debt dependence. • Financial surveillance fragility remained largely within Safe ranges from 2007–2022, indicating effective regulatory monitoring and supportive fiscal/monetary policy. • Financial outward-oriented fragility trended downward over time and stayed broadly at Normal levels; external trade dependence and debt metrics generally improved, though import cover was at a Dangerous level due to high foreign exchange reserves. - Composite index interpretation: The FFII classification bands (Safe 0–17; Normal 17–25; High 25–35; Dangerous >35) contextualize yearly scores; examples from reported years show transitions from Dangerous to Safe/Normal as policies stabilized post-crisis.

Findings indicate that a focused, indicator-based framework can effectively capture both cross-sectional bank fragility and time-varying system fragility. At the bank level, liquidity management and credit risk (proxied by NPLs) are pivotal levers of fragility, aligning with literature that links solvency and asset quality to stability. At the system level, the FFII tracks macro-financial stress episodes and policy responses: after the 2008 crisis, policy actions stabilized fragility; the 2015 equity market episode raised early-warning fragility; and the COVID-19 shock introduced new fluctuations. Persistent strength in financial surveillance suggests regulatory upgrades and prudent fiscal/monetary policies have reduced monitoring-related vulnerabilities. Meanwhile, a supportive economic environment is central to system resilience, and external exposure has been managed to keep outward-oriented fragility broadly at Normal levels. The framework offers policymakers a composite, interpretable gauge to inform timing and focus of macroprudential, fiscal, and monetary measures.

The study introduces a practical evaluation index system for financial fragility and applies it to Chinese banks and the broader financial system. Contributions include: (1) a bank-level fragility scoring model using factor analysis on ROAA, liquidity, CAR, and NPL data for 15 banks; and (2) a composite FFII that aggregates four subsystems—economic environment, financial market early warning, financial monitoring, and financial outward orientation—over 2007–2022. Results show that lower fragility at the bank level is associated with stronger liquidity and lower NPLs, while system-level stability is anchored by a healthy economic environment, improved early-warning mechanisms, and effective financial monitoring. Policy recommendations: government agencies should rigorously analyze market information before policy changes, ensure policy stability and transparency, strengthen supervision to avoid under- or over-implementation, and communicate policies clearly to reduce information asymmetries. Banks should prudently develop non-interest income businesses within regulatory limits, carefully identifying and managing associated risks to avoid elevating fragility. The proposed framework has practical significance for comprehensive monitoring of financial vulnerability in China.

The authors note several constraints: (1) the stability framework for China’s banking system requires deeper development and may have design shortcomings; (2) indicator selection is limited by data availability, potentially omitting relevant variables; (3) the bank-level analysis uses 2018 pre-pandemic data as representative, which may not capture post-2020 structural changes. Future research should deepen the banking stability framework, expand and refine indicators as data improve, incorporate post-pandemic dynamics, and consider regional (local government) financial risks and broader financial market stability analyses.