Market Risk

Question 1

Alice, an assistant in a risk management department, has received the following information to estimate the market risk of a portfolio with both the arithmetic returns with normal distribution and the geometric returns with lognormal distribution assumptions:

Assumption	Return
Annualized average of arithmetic returns	$15\%$
Annualized standard deviation of arithmetic returns	$25\%$
Annualized average of geometric returns	$12\%$
Annualized standard deviation of geometric returns	$27\%$

She has been told that both daily arithmetic returns and daily geometric returns are serially independent of data analysts in the company.

Assuming there are $252$ trading days in a year, which of the following statements is true?

A. The 1-day normal $95\%$ $\text{VaR}$ is equal to $1.96\%$ and the 1-day lognormal $95\%$ $\text{VaR}$ is equal to $1.76\%$.
B. The 1-day normal 95% $\text{VaR}$ is equal to $2.25\%$ and the 1-day lognormal $95\%$ $\text{VaR}$ is equal to $2.30\%$.
C. The 1-day normal 95% $\text{VaR}$ is equal to $2.53\%$ and the 1-day lognormal $95\%$ $\text{VaR}$ is equal to $2.71\%$.
D. The 1-day normal 95% $\text{VaR}$ is equal to $3.02\%$ and the 1-day lognormal $95\%$ $\text{VaR}$ is equal to $3.31\%$.

Answer: C
Assume arithmetic return is normal distributed
$\text{VaR}(\alpha )=-({\mu }_r-z_{\alpha }\times {\sigma }_r)$

Assume geometric return is lognormal distributed
$\text{VaR}(\alpha )=(1-e^{{\mu }_r-z_{\alpha }\times {\sigma }_r})$

• $\alpha$ is the significance level
• $(1-\alpha )$ is the confidence level
• $z_{\alpha }$ is the standard normal variate corresponding to $\alpha$
• $5\%$ significance level, $z$ is $1.645$, $1\%$ significance level, $z$ is $2.326$

1-day normal $\text{VaR}(5\%)=-(\frac{0.15}{252}-1.645\times \frac{0.25}{\sqrt{252}})=2.53\%$

1-day lognormal $\text{VaR}(5\%)=1-e^{(\frac{0.15}{252}-1.645\times \frac{0.25}{\sqrt{252}})}=2.71\%$

---

Question 2

An analyst working in an investment bank uses the $\text{Ho-Lee Model}$ to estimate short-term interest rates. The following data has been collected to calculate forecasted short-term rates.

Term	Data
Annualized drift in month 1	$0.06$
Annualized drift in month 2	$0.066$
Annualized volatility of the interest rate	$0.01375$

The current level of short-term interest rate is $3\%$. What is the interest rate in the upper node in month $2$?

A. $4.83\%$
B. $5.05\%$
C. $5.56\%$
D. $6.12\%$

Answer: A
The drift in $\text{Ho-Lee Model}$ is time-dependent, changes from date to date. The time-dependent drift over each time period represents some combination of the risk premium and of expected changes in the short-term rate.

$$dr={\lambda }_t{d}_t+\sigma d_w={\lambda }_t{d}_t+\sigma ϵ\sqrt{d_t}$$

Node 1: $r_0$
Node 2: $r_0+{\lambda }_1{d}_t+\sigma \sqrt{d_t}$
Node 3: $r_0+{\lambda }_1{d}_t-\sigma \sqrt{d_t}$
Node 4: $r_0+({\lambda }_1+{\lambda }_2)d_t+2\sigma \sqrt{d_t}$
Node 5: $r_0+({\lambda }_1+{\lambda }_2)d_t$
Node 6: $r_0+({\lambda }_1+{\lambda }_2)d_t-2\sigma \sqrt{d_t}$

The interest rate in the upper node in month 2 is
$r=r_0+({\lambda }_1+{\lambda }_2)d_t+2\sigma \sqrt{d_t}=3\%+(0.06+0.066)\times \frac{1}{12}+2\times 0.01375\times \frac{1}{\sqrt{12}}=4.83\%$

---

Question 3

A regulatory analyst at a worldwide commercial bank has been reviewing regulatory requirements from the Basel Committee’s $\text{FRTB}$ guidelines. The main purpose is to explore the difference between $\text{FRTB}$ and previous market risk regulatory requirement from the $\text{Basel I}$ and $\text{Basel II.5}$ frameworks. Which of the following is correct regarding the FRTB’s requirements?

A. Under $\text{FRTB}$, standardized approach and internal models approach are both accepted for calculating market capital.
B. Expected shortfall ($\text{ES}$) with a $99\%$ confidence level is proposed, rather than $\text{VaR}$ with a $97.5\%$ confidence level.
C. The FRTB introduces 5 different liquidity horizons to replace the earlier 10-day horizons used in Basel I and 20-day horizons used in Basel II.5.
D. In common with previous market risk regulatory requirements, $\text{FRTB}$ requires that market risk be calculated at the trading desk level.

Answer: A
Market risk capital in $\text{Basel I}$ is based on $\text{VaR}$ for a 10-day horizon with a $99\%$ confidence level.
Market risk capital in $\text{FRTB}$ is is based on stressed expected shortfall with $97.5\%$ confidence level.

Under $\text{Basel I}$ and $\text{Basel II.5}$, the 10-day horizon used i to reflect the liquidity of the market variable.
Under $\text{FRTB}$, the changes to market variables would take place in stressed market conditions over periods of time. Five different liquidity horizons are used: 10 days, 20 days, 40 days, 60 days and 120 days.

---

Question 4

Mike is evaluating several interest-rate models to develop an appropriate term-structure model, including both time-dependent drift and time-dependent volatility characteristics, in the fund’s options pricing practice. Which of the following is a correct description of the specified model?

A. In the $\text{Vasicek Model}$, when the rate is above its long-run equilibrium value, the drift would be negative or positive, depending on whether the short-term rate is mean-reverting.
B. In the $\text{Cox-Ingersoll-Ross Model}$, the basis-point volatility of the short-term rate is proportional to the interest rate.
C. In the $\text{Model 3}$, the volatility of the short-term rate is assumed to decline exponentially to zero.
D. Same as the $\text{Model 2}$, in the $\text{Ho-Lee Model}$, the drift of the interest rate process is assumed to be time-varying.

Answer: C
$\text{Model 2}$ adds a fixed drift to $\text{Model 1}$, in order to obtain aa richer model in an economically coherent way.

$$dr=\lambda dt+\sigma dw=\lambda dt+\sigma ϵ\sqrt{dt}$$

• $\lambda$ is annual drift in interest rates.
• The drift in the risk-neutral process is a combination of the true expected change in the interest rate and of a risk premium.

In contrast to $\text{Model 2}$, the drift in $\text{Ho-Lee Model}$ is time-dependent, changes from date to date.
The time-dependent drift over each time period represents some combination of the risk premium and of expected changes in the short-term rate.

$$dr={\lambda }_{t}dt+\sigma dw={\lambda }_{t}dt+\sigma ϵ\sqrt{dt}$$

In the $\text{Vasicek Model}$, short-term rates will be characterized by mean reversion, assuming that the economy tends toward som equilibrium.
When the rate is above/below its long-run equilibrium value, the drift is negative/positive, driving it down/up toward it long-run value.
The greater the difference between $r$ and $\theta$, the greater the expected change in the short-term rate toward $\theta$.
The drift combines both interest rate expectations and risk premium.

$$dr=k(\theta -r)dt+\sigma dw$$

• $k$ is the speed of mean reversion.
• $\theta$ is the long-run mean reversion level.
• $r$ is current interest rate level.

$\text{Model 3}$ is a special case of time-dependent volatility models. The volatility of the short rate starts at the constant $\sigma$ and then exponentially declines to zero.

$$dr=\lambda (t)dt+\sigma e^{-at}dw$$

• $a$ is decay rate

In the $\text{Cox-Ingersoll-Ross Model}$, the interest rates are bounded from below by zero.

$$dr=k(\theta -r)dt+\sigma \sqrt{r}dw$$

• $\sigma$ is yield volatility, which is constant.
• $\sigma \sqrt{r}$ is an annualized basis-point volatility, which is proportional to the square root of the rate, implying basis point volatility increases at a decreasing rate.

---

Question 5

Golden Finance’s current CRO plans to improve the efficiency of the current model used to measure the bank’s risk. Due to the complexity of the previous model, the CRO suggests simplifying the $\text{VaR}$ estimation process by narrowing down the range of risk factors in order to find key drivers in the bank’s trading positions. Therefore, the risk management team decides to use $\text{VaR}$ mapping to accomplish the task. Which of the following is most likely appropriate to map the given position?

A. Mapping each position in a corporate bond portfolio to a set of government bonds with the closest maturity.
B. Mapping USD/CNY forward contracts to the USD/CNY spot exchange rate.
C. Mapping zero-coupon government bonds to government bonds with regular coupons.
D. Mapping corporate bonds with regular coupons to credit default swap with the same corporate bonds as underlying assets.

Answer: B
对于选项A,任何债券都必须映射到最能代表其当前状况的收益率上,而公司债券和政府债券由于风险不同,收益率有所差异(一般情况下,期限相同,政府债券收益率低于公司债券)。因此其无法成为最佳映射。该选项描述错误,不符合题意,为错误选项。

对于选项B,将多个美元/人民币远期合约映射到美元/人民币即期汇率是合适的,因为所有外汇远期头寸都受即期汇率这个主要风险因素的影响。该选项描述正确,符合题意,为正确选项。

对于选项C,将一个简单的单一不确定性来源(到期时的收益)映射到多个不确定性来源(息票支付和到期时的收益) ,违反了VaR映射的原则。该选项描述错误,不符合题意,为错误选项。

对于选项D,支付票息的公司债与信用违约互换的性质并不相同。公司债是一项可以稳定获得收益的资产;而信用违约互换则需要购买者按期支付现金流,如果标的债券发生违约则从信用违约互换的卖出方收回本金。该选项描述错误,不符合题意,为错误选项。

---

Question 6

Value-at-risk ($\text{VaR}$) models are only useful insofar as they predict risk reasonably well. This is why the application of these models always should be accompanied by validation which is a general process of checking whether a model is adequate. Which of the following is correct?

A. When the model is perfectly calibrated, the number of observations falling outside the $\text{VaR}$ should be consistent with the confidence level.
B. $\text{VaR}$ backtesing is typically based on the returns over a longer period of time, as the longer the period, the more information it covers, including fees, commissions, spreads and net interest income.
C. Whenever the bactesting finds exceptions that are in consistent with the confidence intervals, we should consider the calibration of the model to have failed.
D. The confidence interval for $\text{VaR}$ and that for the test of $\text{VaR}$ must be consistent to ensure the accuracy of the backtesting.

Answer: A

---

Question 7

Risk analysts at an investment bank are looking at the assumptions the bank uses in its foreign exchange option pricing models. They find that the implied volatility of at-the-money options is relatively low, and it gets progressively higher as the option goes in-the-money or out-of-the-money. Which of the following is correct?

A. The volatility of the asset is constant and the price of the asset changes smoothly with no jumps, asset prices have lognormal distribution as long as one of these two conditions is met.
B. The volatility smile of the foreign exchange option becomes more pronounced as option maturity increases.
C. In the implied distribution figure, the implied distribution has a heavier left tail and a heavier right tail.
D. In the implied distribution figure, the implied distribution has a heavier left tail and a thinner right tail.

Answer: C

---

Question 8

Mike and Sue are discussing the appropriateness of $\text{ES}$and $\text{VaR}$ used in current risk model in the bank. Mike believes that $\text{ES}$ would be a better measure, especially under financial distress. Sue disagrees with it and suggests using $\text{VaR}$ instead. With respect to $\text{ES}$ and $\text{VaR}$, which of the following statements is correct?

A. VaR is more complicated to calculate compared to $\text{ES}$, and $\text{VaR}$, like $\text{ES}$, meets the coherent risk measures criteria.
B. The method of calculating $\text{ES}$ is to slice the tail into a large number n of slices, each of which has the different probability mass, estimate the $\text{VaR}$ associated with each slice, and take the $\text{ES}$ as the average of these $\text{VaRs}$.
C. $\text{ES}$ is a probability-weighted average of tail losses, so we can estimate $\text{ES}$ as an average of ‘tail $\text{VaRs}$’.
D. $\text{ES}$ is not the easiest estimator in coherent risk measures, it could be less than the $\text{VaR}$.

Answer: C

---

Question 9

A risk analyst is validating a 1-day $99\%$ $\text{VaR}$ model by collecting trading data for $4$ years. Assuming $250$ trading days in a year, what is the maximum number of days, daily losses exceeding the 1-day $99\%$ $\text{VaR}$, under a $95\%$ confidence level, to ensure that the model can be accepted through backtesting?

A. $8$
B. $10$
C. $12$
D. $16$

Answer: D

---

Question 10

A company’s Chief Risk Management Officer, at the general shareholders" meeting, discussed the relevant risks the company facing in areas including investments, trading, risk management, the global financial crisis and regulation. At the meeting, he presented the following conclusions, which of these conclusions is correct?

A. The higher the degree of diversification in the portfolio, the higher the correlation.
B. Correlations between the tranches of the CDOs also increased during the crisis from 2007 to 2009, but the senior tranches were safe since it consisted of AAA-rated and they were protected by the lower tranches.
C. Lower correlation coefficients reduce the worst-case scenario for creditors, i.e. the joint probability of debtor defaults.
D. For bonds in distress, the term structure of default probabilities increases in time, while for most investment grade bonds, the term structure of default probabilities decreases in time.

Answer: C

---

Question 11

A risk manager is analyzing a firm’s equity option pricing assumptions and noticed that the volatility used to price a low-strike-price option is significantly higher than that used to price a high-strike-price option. Assuming the same mean and standard deviation, how does the implied distribution of equity option prices differ when it is compared to lognormal distribution?

A. It has a heavier left tail and a heavier right tail.
B. It has a heavier left tail and a lighter right tail.
C. It has a lighter left tail and a heavier right tail.
D. It has a lighter left tail and a lighter right tail.

Answer: B

---

Question 12

A market risk analyst at a hedge fund is considering how to model low probability, high impact extreme events using extreme value theory ($\text{EVT}$). The analyst decides whether to use generalized extreme value theory (KaTeX parse error: Unexpected end of input in a macro argument, expected '}' at end of input: …ver-threshold (\text{POT) approach to model the tails of P/L distributions. Which of the following is correct?

A. $\text{POT}$ provides a threshold $u$ which determines the number of observations $N_u$, the larger $u$ the better, because the larger u, the more representative of the extremes will be selected.
B. In the $\text{POT}$ approach, if the original distribution is a $t$, then we should choose the Gumbel distribution.
C. The $\text{POT}$ has only two parameters, one is a scale parameter $\beta$ which can be positive or negative, the other is a shape or tail index parameter $\xi$.
D. $\text{GEV}$ method involves three parameters, based on the tail index $\xi$, the $\text{GEV}$ equation has three special cases, when $\xi >0$, it’s useful for financial returns because the distribution is typically heavy-tailed.

Answer: D

---

Question 13

A risk analyst wants to predict the interest rate by constructing a fitting model. The analyst assumes that the economy tends toward some equilibrium based on such fundamental factors as the productivity of capital, long-term monetary policy, and so on. So, he decides to use Vasicek model. The analyst gathers the following information:

• Current short-term interest rate: $4.18\%$
• Long-run value of short-term interest rate: $3.24\%$
• Mean reversion rate: $0.07$
• Annual basis-point volatility: $180$ bps

The analyst then constructs an interest rate tree, determines the expected short-term interest rate in $5$ years, and calculates how long it will take for the short-term interest rate to return to half its long-term value. Which of the following statements made by the analyst is correct?

A. The expected short-term interest rate is $3.9\%$ and the half-life is $9.90$ years.
B. The expected short-term interest rate is $5.23\%$ and the half-life is $15.70$ years.
C. The expected short-term interest rate is $3.9\%$ and the half-life is $15.70$ years.
D. The expected short-term interest rate is $5.23\%$ and the half-life is $9.90$ years.

Answer: A

---

Question 14

During the next year, an internal model to measure the market risk developed by an investment bank predicts a $98.5\%$ probability of no loss occurring and a $1.5\%$ probability of a single loss occurring. If the single loss occurs, the severity is presented by three possible results: $ $12.0$ million loss with $15\%$ probability, $ $20.0$ million loss with $50\%$ probability, and $ $28.0$ million loss with $35\%$ probability. What is the model’s 10-day $97.5\%$ expected shortfall ($\text{ES}$)?

A. $ $12.96$ million
B. $ $9.89$ million
C. $ $10.88$ million
D. $ $12.50$ million

Answer: A

---

Question 15

The non-parametric approach which attempts to estimate risk measures without making strong assumptions about the relevant distribution. The essence of these methods is to let the P/L data speak for itself and to use the most recent empirical distribution of P/L rather than some assumed theoretical distribution to estimate the risk measure. One of the most popular non-parametric methods is historical simulation. Which of the following is correct?

A. For non-parametric methods, there are problems dealing with variance-covariance matrices, the curse of dimensionality, etc., and it is often the least natural choice for high-dimensional problems.
B. Bootstrapped historical simulations use bootstrap procedure to resample from the existing dataset with replacement, and are often more accurate but are not useful for gauging the precision of the estimates.
C. Non-parametric are usually quick to reflect major events, if there is a permanent change in exchange rate risk, the approach will react to it quickly.
D. Non-parametric approaches can be greatly improved and potentially enhanced if we combine them with parametric “add-ons” to make them semi-parametric.

Answer: D

---

Question 16

Jim has just taken the position of CRO of a listed company, he intends to select a risk measure that is more suitable for the company’s risk management, after examining the actual situation of the company and understanding the characteristics of each risk measure, he has come to the following conclusions, which of the following conclusions is the correct one?

A. ${\text{VaR}}_{\alpha }(L)$ is defined as the quantile of $L$ at the probability a, which measures quantiles of losses, and concurrently considers any loss beyond.
B. $\text{ES}$ is always sub additive and coherent but it does not mitigate the impact that the particular choice of a single confidence level may have on risk management decisions.
C. The spectral risk measures ($\text{SRM}$) require that larger losses are taken more seriously than smaller losses and requirements.
thus the corresponding weighting function does not respond to risk aversion.
D. $\text{VaR}$ is criticized for not being a coherent risk measure with sub-additivity not satisfied, leading to a company that had the incentive to legally break up into various subsidiaries in order to reduce its regulatory capital requirements.

Answer: D

---

Question 17

Backtesting $\text{VaR}$ is a formal statistical framework that involves verifving whether actual losses are consistentl with projected losses. James, FRM, is a researcher at a risk consulting firm who is planning to conduct $\text{VaR}$ backtesting, and in anticipation of doing so, he makes a number of observations or conclusions. Which of the following is correct?

A. Ideally, the failure rate should give an unbiased measure of $p$, that is, it should be less than $p$ as the sample size increases.
B. In backtesting, we must make assumption about the return distribution, if the distribution is skewed, or with heavy tails, the backtesting cannot work properly.
C. It is possible to set a low $\text{Type I}$ error rate and have a test that creates a very low $\text{Type II}$ error rate, in which case the test is said to be powerful.
D. The confidence level of $\text{VaR}$ model should be high to increase the effectiveness, or power, of the statistical tests.

Answer: C

---

Question 18

An investment manager is worried about adverse movement of treasury bond’s yield, which takes $10\%$ share in the current portfolio. The investment manager plans to use $\text{TIPS}$ to hedge the risk, but has difficulty in selecting either regression-based hedging or $\text{DV}01$ neutral hedge. Which of the following is most likely correct regarding these two hedging methods?

A. When the yield on the hedged portfolio and the yield on the hedging instrument is not perfectly correlated, $\text{DV}01$ neutral hedge ensures that the trade will neither make nor lose money.
B. Regression-based hedging in practice is more approximate to use in volatile market.
C. $\text{DV}01$ neutral hedge approach based on historical data can be applied to improve regression-based hedging.
D. The best practice for regression-based hedging is to estimate hedge coefficient over different periods of time to observe if the coefficient is stable or not.

Answer: D

---

Question 19

A newly hired risk analyst who works in a large investment bank is examining how financial correlation riskl affects the bank’s portfolios. The bank holds portfolios consisting of different types of assets and enters into variousl hedging contracts with numbers of counterparties. The analyst made lots of detailed analysis about correlation risk. Which of the following statements would the analyst be correct to make?

A. The buyer of a $\text{CDS}$ faces wrong-way risk when there is a positive default correlation between the reference asset and the $\text{CDS}$ counterparty.
B. The risk-adjusted return of a portfolio typically increases when correlations of assets in the portfolio increase.
C. Dynamic correlation risk in a portfolio of pairs trades is most appropriately estimated using Gaussian copulas.
D. Correlation risk is highest during periods of relatively benign market movements when correlations are difficult to predict.

Answer: A

---

Question 20

Philip, FRM, was a risk analyst in a hedge fund, he wanted to use the copula approach to estimate the correlation of the portfolio in his firm. After research, he had come to the following conclusion. which conclusion is correct?

A. The copula technique combines the marginal loss distributions for different business lines or risk types into a joint distribution for all risk types and takes account of the interactions between different risk types based on assumptions.
B. Using a copula, only parametric marginals with different tail shapes can be combined into a joint risk distribution, but not nonparametric marginals.
C. The copula approach for $\text{CDO}$ has a limitation in that it can only measure the default correlations between a limited number of assets in the $\text{CDO}$, i.e. the number of assets in a $\text{CDO}$ cannot exceed a certain number.
D. The multivariate copulas which applied to finance is the main cause of the financial crisis, and the copulal approach allow the joining of multiple univariate distributions to a single multivariate distribution.

Answer: A

---

Question 21

A newly hired analyst is calculating the expected correlation for A-share listed stocks. The analyst has collected data showing that the average monthly correlation for all A-share listed stocks in February 2022 is $20\%$ and the long-run correlation mean of all A share listed stocks is $30\%$. The regression result run by the analyst is as follows: $Y=0.186-0.62x$ What is the expected correlation for A-share listed stocks in March 2022?

A. $0.248$
B. $0.186$
C. $0.138$
D. $0.262$

Answer: D

---

Question 22

There are a number of models for modeling the movement of interest rates, the simplest model with no drift and normally distributed rates were first studied, and then we add and discuss the implications of alternate specifications of the drift, after which we also focus on the volatility of interest rates and on models in which rates are not normally distributed. What is correct of the following statements about the term structure models?

A. Model 3 is a parallel shift model and the term structure of volatility is flat at levels that change over time while the Vasicek model is a mean-reversion one.
B. If the purpose of the model is to value and hedge fixed income securities, then a model without mean reversion might be preferred.
C. Periods of high inflation and low short-term interest rates are inherently unstable, so the basis-point volatility of the short rate tends to be high.
D. In the $\text{CIR}$ model, the o is constant, the basis-point volatility is not constant and increases with the level of the time horizon.

Answer: A

---

1-22高顿

---