Liquidity indicators,early warning signals in banks,and financial crises

This research compares the performance of three liquidity indicators, namely liquidity ratio (LiqR), liquidity creation (LiqC) and net stable funding difference (NSFD), for sending early warning signals for distressed banks. Recent evidence has shown that LiqR appears incapable of measuring the liquidity condition of banks. However, LiqC and NSFD have not yet been fully examined. Thus, which indicator is more useful in an early warning model becomes an interesting issue. We classify distressed banks as banks that have experienced a bank run, bailout, or failure. Sample data are collected from the United States and the European Union from before and after the financial crisis. We then estimate model predictive value using the sample before the crisis to predict liquidity shortages. Evidence shows that the academic (LiqC) and officially recommended indicators (NSFD) outperform LiqR as early warning signals. Furthermore, LiqC performs best when banks actively engage in income diversification but not fund diversification. Therefore, a well income-diversified bank with high LiqC tends to have high distress probability in the next period.     

Currency crisis early warning systems: Why they should be dynamic

Traditionally, financial crisis Early Warning Systems (EWSs) have relied on macroeconomic leading indicators when forecasting the occurrence of such events. This paper extends such discrete-choice EWSs by taking the persistence of the crisis phenomenon into account. The dynamic logit EWS is estimated using an exact maximum likelihood estimation method in both a country-by-country and a panel framework. The forecasting abilities of this model are then scrutinized using an evaluation methodology which was designed recently, specifically for EWSs. When used for predicting currency crises for 16 countries, this new EWS turns out to exhibit significantly better predictive abilities than the existing static one, both in- and out-of-sample, thus supporting the use of dynamic specifications for EWSs for financial crises.     

Systemic financial risk early warning of financial market in China using Attention-LSTM model

We propose an Attention-LSTM neural network model to study the systemic risk early warning of China. Based on text mining, the network public opinion index is constructed and used as a training set to be incorporated into the early warning model to test the early warning effect. The results show that: (i) the network public opinion is the non-linear Granger causality of systemic risk. (ii) The Attention-LSTM neural network has strong generalization ability. Early warning effects have been significantly improved. (iii) Compared with the BP neural network model, the SVR model and the ARIMA model, the LSTM neural network early warning model has a higher accuracy rate, and its average prediction accuracy for systemic risk indicators has been improved over short, medium and long terms. When the attention mechanism is included in the LSTM, the Attention-LSTM neural network model is even more accurate in all the cases.     

Contagion risk in the Czech financial system: A network analysis and simulation approach

This paper examines the potential for contagion within the Czech banking system via the channel of interbank exposures of domestic banks, enriched by a liquidity channel and an asset price channel, over the period March 2007 to June 2012. A computational model is used to assess the resilience of the Czech banking system to interbank contagion, taking into account the size and structure of interbank exposures as well as balance sheet and regulatory characteristics of individual banks in the network. The simulation results suggest that the potential for contagion due to credit losses on interbank exposures was rather limited. Even after the introduction of a liquidity condition into the simulations, the average contagion was below 3.8% of the remaining banking sector assets, with the exception of the period from December 2007 to September 2008. Activation of the asset price channel further increases the losses due to interbank contagion, showing that the liquidity of government bonds would be essential for the stability of Czech banks in stress situations. Finally, the simulation results for both idiosyncratic and multiple bank failure shocks suggest that the potential for contagion in the Czech banking system has decreased since the onset of the global financial crisis.     

Measuring financial systemic stress for Turkey: A search for the best composite indicator

In this study, we aim to construct a single financial stress indicator (FSI) for Turkey adopting weekly data from between April 2005 and December 2016. To do so, we compose 15 different FSIs using 14 variables that will represent five different markets, i.e. the money market, the bond market, the foreign exchange market, the equity market and the banking sector. We aggregate these five different markets using a variety of techniques, including principal component analysis (PCA), basic portfolio theory, variance equal weights and the Bayesian dynamic factor model. We compare 15 different FSIs on the basis of their relation to, and the forecasting power of, different variables such as the growth rate of industrial production, the OECD business condition index and the OECD composite leading indicator for Turkey. Our results suggest that there is no simple best indicator for Turkey to measure financial systemic stress. Some indicators offer good forecasting power for economic growth while others have a stronger correlation with systemic risk. Therefore, we offer a final FSI for Turkey conducting a model averaging method via a rolling correlation based weighting scheme to benefit from the information content of all the FSIs and observe that the final FSI successfully indicates the tension periods.     

A non-parametric index of corporate governance in the banking industry: An application to Indian data

This paper presents a methodological framework for constructing a non-parametric index of corporate governance for banks. The index is constructed by aggregating six distinct dimensional indices capturing different dimensions of corporate governance, namely board effectiveness, audit function, risk management, remuneration, shareholder rights and information, and disclosure and transparency. For aggregation, a tailored version of data envelopment analysis (DEA) approach which is popularly known as constrained ‘Benefit-of-the-Doubt (BoD)’ model is employed. This approach is unique and distinctive in the sense that it requires no a priori knowledge of weights, and assigns endogenous weights obtained from actual data to individual dimensions of bank governance in order to construct a composite index of corporate governance. This methodological framework has illustrated by applying it for a data set of 40 Indian banks operating in the year 2017. The data set has been compiled using 58 governance regulations as defined by relevant jurisdictions.