AIMs: Explain the various approaches for estimating VaR. Describe the implications of regime switching on quantifying volatility. (Linda Allen Chapter 2)
Questions:
324.1. Which of the following approaches to value at risk (VaR) estimation is the LEAST dependent (if at all) on the historical return series?
a. Parametric
b. GARCH(1,1)
c. Hybrid of parametric and nonparametric
d. Implied volatility
324.2. Which of the following best summarizes the key difference(s) between these two approaches to VaR estimation: exponentially weighted moving average (EWMA; aka, RiskMetrics) versus the HYBRID approach?
a. Hybrid approach does not utilize exponentially declining weights
b. Hybrid is parametric and EWMA is non-parametric
c. Hybrid estimates the VaR as a quantile (percentile) of ordered (but weighted) historical returns, but EWMA does not sort returns
d. There is no difference: EWMA is the hybrid approach
324.3. According to Linda Allen, what is the most important difference between a parametric approach to VaR and historical simulation (a non-parametric method)?
a. Expected shortfall (ES; aka, conditional VaR) is non-trivial under a parametric approach; but ES is "easy and useful" in historical simulation
b. Parametric methods use all data efficiently; historical simulation makes very inefficient use of data
c. Historical simulation is better suited to instances of regime-switching
d. Parametric methods impose normality, but "normality cannot be salvaged."
Answers:
Questions:
324.1. Which of the following approaches to value at risk (VaR) estimation is the LEAST dependent (if at all) on the historical return series?
a. Parametric
b. GARCH(1,1)
c. Hybrid of parametric and nonparametric
d. Implied volatility
324.2. Which of the following best summarizes the key difference(s) between these two approaches to VaR estimation: exponentially weighted moving average (EWMA; aka, RiskMetrics) versus the HYBRID approach?
a. Hybrid approach does not utilize exponentially declining weights
b. Hybrid is parametric and EWMA is non-parametric
c. Hybrid estimates the VaR as a quantile (percentile) of ordered (but weighted) historical returns, but EWMA does not sort returns
d. There is no difference: EWMA is the hybrid approach
324.3. According to Linda Allen, what is the most important difference between a parametric approach to VaR and historical simulation (a non-parametric method)?
a. Expected shortfall (ES; aka, conditional VaR) is non-trivial under a parametric approach; but ES is "easy and useful" in historical simulation
b. Parametric methods use all data efficiently; historical simulation makes very inefficient use of data
c. Historical simulation is better suited to instances of regime-switching
d. Parametric methods impose normality, but "normality cannot be salvaged."
Answers:
