Aurora Borealis
1 September 2014 CFA Institute Journal Review

A Multiperiod Bank Run Model for Liquidity Risk (Digest Summary)

  1. Vassilis Efthymiou, CFA

With the use of a rigorous bank-run model, the authors deconstruct the possibility of default into its insolvency risk and illiquidity risk components. The computation of default probability because of the bank’s failure to endure a bank run appears to have important implications from a regulatory and risk management perspective.

What’s Inside?

The authors evaluate the contribution of illiquidity risk to a bank’s total default probability as opposed to considering insolvency as the only source of default risk. They investigate a bank’s exposure to illiquidity risk when it is faced with bank runs by short-term debtors who decide to refrain from rolling over their funding to the bank. They conclude that this illiquidity risk is actually higher than the risk of the bank not having sufficient capital in the future. Moreover, through the illiquidity risk corridor, significant volatility of risky assets, high frequency of rollover opportunities, and excessive reliance on short-term debt all seem to increase the probability of default for the bank.

How Is This Research Useful to Practitioners?

Illiquidity risk is defined as the probability of a bank not being capable of paying short-term debt in the event of a bank run at any future rollover date. By quantifying illiquidity risk as a separate component of total default risk, the authors showcase the importance of expectations in the current and future decisions of short-term debtors with respect to the probability of surviving a bank run at a specific rollover date. They also empirically find that as the underlying illiquidity risk increases, the value of risky assets becomes more volatile. Additionally, rollover opportunity is more frequent, and short-term debt weights are higher among tapped funding sources. Hence, these results will guide risk managers in more effectively hedging corresponding risks.

The authors contend that the implications of their model suggest that regulators’ focus should be diverted from strengthening banks’ solvency to establishing liquidity monitors for the banks. In this respect, they indicate that the recent introduction of the liquidity coverage and stable funding ratios by Basel II contributes to this end. Finally, they argue that in times of tight liquidity, along with interest rate cuts, lower haircut rates for risky assets posted as collateral to the European Central Bank lending facilities should also be used to increase leverage and, therefore, liquidity in the market. The intuition underlying this approach is that the bank will be able to raise more funds with the same amount of risky assets. These funds can then be used to pay short-term debt that is not rolled over, thereby reducing the bank’s probability of default as a consequence of illiquidity.

How Did the Authors Conduct This Research?

Under the framework of structural credit risk models, the authors create a dynamic model that is based on multiple dates on which short-term debtors decide whether to roll over their funding or “run on the bank.” At each rollover date, the authors set the expected return from rolling over equal to the expected return from withdrawing short-term funds and investing them outside the bank, with the goal of computing three important thresholds: the expected asset value below which a bank run is triggered, the value barrier below which a bank run leads to default as a result of illiquidity, and the minimum value of assets needed by the bank to remain solvent.

Moreover, the authors use balance sheet and credit default swap data from Merrill Lynch to calibrate the model’s parameters, assuming that asset values follow a geometric Brownian motion. More specifically, by running two simulations on the Merrill Lynch data of a bank run by short-term creditors, they derive the direction and strength of the relationship between the asset value volatility, rollover frequency and amount of short-term debt, and probability of default following a bank run.

Abstractor’s Viewpoint

Using backward optimization of a binomial process, the authors effectively manage to split total default probability for a bank into its insolvency and illiquidity components. The separation significantly contributes to the understanding of the severe repercussions bank runs may have on the viability of a bank. Moreover, it facilitates the identification of those factors capable of making default more or less probable. Nevertheless, to the extent that such a theoretical approach fails to incorporate marketwide risk aversion and bank stress contagion, it might be regarded as imperfect in accurately predicting a default as a result of a bank run.

We’re using cookies, but you can turn them off in Privacy Settings.  Otherwise, you are agreeing to our use of cookies.  Accepting cookies does not mean that we are collecting personal data. Learn more in our Privacy Policy.