Measuring each bank’s contribution to systemic risk on a daily basis

Summary:

– Volatility poses a threat to financial stability.

– Measuring systemic banking risk and volatility has been crucial since thesubprime crisis, and many new tools have been developed as a result.

– Risk measures based in particular on value at risk have made it possible to track the contribution of banks to systemic risk on a daily basis using NYU Stern’s VLAB.

Does my bank contribute significantly to systemic risk? What impact would the failure of my bank have on the financial system as a whole? This article introduces the value of analyzing the volatility of financial asset returns and various risk measures such as value at risk. It then details the method and operation of NYU Stern’s VLAB (Volatility Lab) for tracking the daily contribution of banks and countries to systemic risk.

What is volatility?

The volatility of a financial asset is a measure of the variance in the distribution of a financial security’s returns. Volatility is therefore a measure of risk because it represents the variability of a financial asset’s returns over a given period. The greater the volatility, the greater the fluctuation in its price and therefore the greater the risk associated with that asset. It should be noted that this measure takes into account price variation without prejudging whether this variation is positive or negative.

There are many different names for volatility (historical volatility, future volatility, realized volatility, implied volatility, etc.) which correspond to different calculation methods.

The first family of volatility measures is the historical volatility family. This measure of variability corresponds to a measure of the variability of a security over a given past period. It can be measured simply using the standard deviation of a security’s past returns or using other types of models that employ more complex calculation methods (GARCH, E-GARCH, etc.) that take into account the observed characteristics of volatility and thus improve its predictive capabilities. For example, a GARCH (generalized autoregressive conditional heteroskedasticity) model will be used when volatility clusters are observed, i.e., periods during which volatility remains low and periods of high volatility. Or an E-Garch model if we consider that these clusters are not distributed symmetrically in both directions (i.e., when markets fall, variability is greater than when they rise). There are therefore a multitude of models associated with the different characteristics observed in volatility (which we will not detail here) that improve the measurement of a financial asset’s variability and the predictability of future changes in that security.

The other major family of volatility is the family of implied volatilities. These volatilities are described as implied because they are not derived from measurements of price changes but are derived from option pricing models. The price of an option or financial derivative depends on the variability of the underlying financial security. For example, a derivative that allows oil to be purchased at the end of the month is more expensive if the price of oil varies greatly than if it does not vary, because the seller of this product takes more or less risk depending on whether the price of oil is highly or lowly variable. This provides a measure of the implied volatility of the price of oil, based on the prices of oil derivatives. Within the implied volatility family, there are different types of volatility depending on the period considered. We refer to future implied volatility if the implied volatility is calculated for a future period, past implied volatility if it is calculated using derivatives that have matured in the past, historical volatility if the volatility is measured by the variance of a security’s past returns, etc.

Other risk measures

Volatility is a risk measure that provides an approximate assessment of the financial risk of an asset. However, in recent years, other risk measures have been developed to take into account other components of financial risk. In particular, we will detail the correlation between value-at-risk and expected shortfall (expected loss).

Correlation measures the strength of the relationship between two assets, with the linear correlation coefficient ranging from -1 to 1. If it is 1, then the relationship between the two assets is very strong and the financial securities vary in the same direction. Correlation therefore makes it possible to measure risk because it assesses the contagion effects between assets (i.e., if one asset falls sharply, how will the other assets evolve?).

Value-at-risk corresponds to the extreme values of the distribution of returns on a financial security. For example, value-at-risk at a threshold of 5% over a 30-day period corresponds to the minimum loss incurred in the most unfavorable 5% of cases. This model is a classic measure of risk, but it tends to underestimate extreme risks. Expected shortfall, which corresponds to the average loss for the worst 5% of cases [1], allows for better consideration of extreme risks [2].

A practical tool for measuring systemic risk in banks: NYU’s VLAB!

The aim of this section is to introduce you to a risk measurement tool developed by New York University (NYU Stern), which you can find via the following link: (http://vlab.stern.nyu.edu/). This VLAB, or volatility laboratory, led by Nobel Prize-winning economist Robert Engle (2003), analyzes daily variations in numerous markets using the latest techniques developed by university research teams. Created in 2008, the laboratory provides real-time data on the correlation and variability of financial assets, as well as risk analysis.

There are many risk measurement models based on volatility and return distribution, as we have seen previously, and this laboratory allows these risk measures to be studied for many assets. We will not go into detail here about risk measures relating to volatility, correlation, and value at risk, but we will focus in particular on systemic risk measures and the importance of systemic risk and its characteristics . The objective of this analysis is to show whether financial institutions are likely to have capital needs when the market faces difficult conditions and, therefore, their contribution to systemic risk. This analysis is similar to stress tests and consists of three steps.

The first step is to estimate the sensitivity of the firm’s share price when market conditions are difficult, i.e., a 2% decline in this case. This measure, called Marginal Expected Shortfall, takes into account many dimensions of risk because it considers investors’ perception of the bank’s situation, the bank’s exposure to the market, and the volatility of the stock. Second, this measure is extrapolated by considering a financial crisis scenario. Third, the bank’s capital losses during a crisis are combined with the current value of the firm and the amount of debt to determine the amount of capital needed to weather the crisis.

The systemic risk measure therefore corresponds to the bank’s capital requirements when it is exposed to adverse market conditions. These amounts are presented in the SRISK column of the following table, which ranks the systemic risk of each bank. As can be seen by clicking on the link, the major French banks are at the top of this ranking, which is confirmed by France’s significant contribution to overall systemic risk, largely due to the size of its banking sector. It is hardly surprising to find large banks in this ranking and on the list of systemic banks under special supervision by the Basel Committee. Nevertheless, the systemic risk of banking institutions has been significantly reduced since August 2011. This notable decline can be explained by the gradual implementation of regulatory reforms associated with Basel III and recapitalization programs. I therefore invite you to follow the evolution of banking institutions’ contribution to systemic risk step by step using vlab.