Measuring Financial System Risk
Expected Shortfall (ES) seek to measure the potential loss in an extreme event” (Acharya et al. , 2008). 1. Value-at-Risk (VaR) The most important factor in any system is the estimation of any value and the risk of that value being lost. In any financial system value is defined in terms of assets possessed by the bank and its interconnections. “Even as risk management assumes an increasingly central role in financial institutions, the actual of financial risk continues to be problematic” (Jorion, 2000).
In financial systems, the main question the system risk regulators would ask is “How much would the firm lose with a certain asset investment or asset movement? ” To answer this question, system risk regulators would use the Value-at-Risk (VaR) model to analyze the level of risk that would occur over a given period of time if they traded the given portfolio. Financial and commercial firms often use the VaR model to measure the potential loss of value exposure a given firm would have in the event of trading its portfolio.
Adverse market flexibility in the economic framework necessitates can create a liquidity crisis in one firm and cause a spillover to other firms in the financial system. This spillover of liquidity crisis would the lead to an economic failure or financial crisis such as the current crisis. There are three main risk measuring fundamentals of the Value-at-Risk (VaR) model namely; precise intensity of loss in value, a predetermined time in which risk is evaluated and a confidence period. 2. Expected-Shortfall (ES)
The expected VaR may not be adequate to measure systems risk as it only “looks at the probability of the shortfall of claims over capital being positive. But the size of the shortfall certainly matters: someone will have to pay for the remainder. ” (Kaas et al. , 2008). In order to measure and cover the risk exposure due to the derivable shortfall, the expected-shortfall model is used to give evidence of “how bad is bad” (Kaas et al. , 2008). To measure the total expected ES, historical data on a firm’s losses is collected from each sector in the firm and clustered into quarters.
Data on the quarters when the firm made the most losses is then evaluated to show the effect of those quarters’ losses on the aggregate losses – this is the financial system risk caused by the firm’s losses and is deemed as the marginal expected shortfall. In all sectors, its important to measure this ES as it clearly shows a firm’s risk exposure in the past and data from this evaluation can be used to estimate future probability of having an expected shortfall.
This estimation can then assist regulators to employ policies to counter this from happening. With this model daily financial system risk evaluation is done to define a firm’s position financially. This helps identify areas that require allocation of resources in the firm’s divisions, ensure overall management of available firm resources and track performance across the firm division by division. Value-at-Risk vs. Expected-Shortfall
VaR’s main shortfall is in it being nonsubadditive – this means that the model, though it helps weigh the probable maximum loss, it may not give an exact figure or size of the loss in given incidents. Because of this shortfall of the VaR model, the ES model is becoming more and more popular among system risk analysts as a more in-depth and accurate measurement model to use in measuring accurate systemic risk. As Diebold & Santomero (1999) reveal;
In the face of the new reality, firms in the global financial marketplace have been scrambling, once again searching for appropriate tools and managerial approaches to guide their organizations. This is occurring against a backdrop of risk managers spending the last decade increasing their focus on firm-level risk management systems, spending tens of millions of dollars on trading systems, real time position reporting, and VaR risk management system (p. 6-7)
This means that whether regulators calculate the aggregate risk of two variables separately and adds up the risk estimate, it would be the same as if the two variables were joined and their risk aggregate calculated i. e. with VaR its not always the case that VaR (X+Y) ? VaR(X) + VaR(Y) as Saita (2007) elaborates. This is the quality of subadditivity possessed by the ES model and that VaR lacks as mostly of the time if used it would give variant risk estimates. VaR is, however widely known and used that ES should then be considered as complimenting of Value-at-Risk.
The Expected-shortfall model can be used to calculate the exact risk size after employing VaR estimation model to ensure that regulators can have clear, accurate and additional information on the potential and actual risk exposure posed by a particular firm in order to impose regulatory policies to curb loss spillovers that would risk the collapse of a financial system and the world economy. Acharya et al. (2008) reveal that like any model the risk assessment models are hampered by limitations and that: Stress tests can be used to assess risk concentration and interconnected counterparty risks.
The regulator could estimate the consequences of the failure of a large institution. In addition, scenario analysis can limit excessive risk taking in good times. After prolonged periods of low volatility, statistical measures of risk go down. As a result, risk taking becomes pro-cyclical, and this increases the likelihood and severity of financial crisis (p. 6). Risk scenarios can also be dependent on time and regulators must then ensure they conduct adequate analysis of how risk posed by firms evolves time after time. This plays a key role in cushioning the economy against future failure.