The following is a review of the Market Risk Measurement and Management principles designed to address the

learning objectives set forth by GARP®. Cross-reference to GARP assigned reading—Hull, Chapter 18.

READING 18

FUNDAMENTAL REVIEW OF THE TRADING BOOK

Study Session 3

EXAM FOCUS
The new banking capital requirements, as specified in this reading, will profoundly
change the way that capital for market risk is calculated. There are several key
innovations that will cause this change. First, banks will be required to forgo using the
99% confidence level VaR measure in favor of the 97.5% confidence level expected
shortfall measure. This change will better capture the potential dollar loss (i.e., tail risk)
that a bank could sustain in a given window of time. Many risk managers have already
begun using expected shortfall in practice for internal audits. Second, risk assets will be
divided into liquidity horizons that better reflect the volatility in specific asset
categories. The third innovation is a rules-based criteria for an asset being categorized
as either a trading book asset or a banking book asset. This step will help mitigate the
potential for regulatory arbitrage.

MODULE 18.1: FUNDAMENTAL REVIEW OF THE TRADING BOOK

Market Risk Capital Calculation

LO 18.a: Describe the changes to the Basel framework for calculating market
risk capital under the Fundamental Review of the Trading Book (FRTB) and the
motivations for these changes.

In May 2012, the Basel Committee on Banking Supervision began considering the next
round of changes to market risk capital calculations for banks. This process is known as
the Fundamental Review of the Trading Book (FRTB). After receiving comments on
proposals and seeing the results of a formal study, the rules were further refined in
December 2014. It is important for risk managers to understand the nature of the
proposed changes and the new calculation methodology.
In order to properly understand the changes, it is necessary to first understand the
previous market risk requirements. The Basel I calculations for market risk capital
involved a 10-day value at risk (VaR) calculated with a 99% confidence level. This
process produced a very current result because the 10-day horizon incorporated a
recent period of time, which typically ranged from one to four years. The Basel II.5
calculations required banks to add a stressed VaR measure to the current value
captured with the 10-day VaR. The stressed VaR measures the behavior of market
variables during a 250-day period of stressed market conditions. Banks were required
to self-select a 250-day window of time that would have presented unusual difficulty
for their current portfolio.
The FRTB researched if the 10-day VaR was really the best measurement for a bank’s
true risk. The value at risk measure has been criticized for only asking the question:
“How bad can things get?” VaR communicates, with a given level of confidence, that the
bank’s losses will not exceed a certain threshold. Consider a bank that uses a 10-day
VaR with a 99% confidence level and finds that losses will only exceed $25 million in
1% of all circumstances. What if the 1% chance involves a $700 million loss? This
could be a catastrophic loss for the bank. Therefore, the FRTB has proposed an alternate
measure using expected shortfall (ES), which is a measure of the impact on the profit
and loss statement (P&L) for any given shock of varying lengths. The expected shortfall
asks the question: “If things get bad, what is the estimated loss on the bank’s P&L?”
Consider the following example that illustrates the difference between value at risk and
expected shortfall. A bank has a $950 million bond portfolio with a 2% probability of
default. The default schedule appears in Figure 18.1.
Figure 18.1: Example Default Schedule for $950 Million Bond Portfolio

At the 95% confidence level, there is still no expected loss, so the 95% VaR would imply
a $0 of loss. However, the expected shortfall measure accounts for the potential dollar
loss conditional on the loss exceeding the 95% VaR level. In this case, three out of five
times the expected loss is still $0, but two out of five times the expectation is for a total
loss of the $950 million bond portfolio’s value due to default. This means that 40% of
the tail risk would yield a loss, so the expected shortfall is $380 million (i.e., 40% ×
$950 million). This presents a very different risk perspective than using the VaR
measure alone.
Instead of using a 10-day VaR with a 99% confidence level, the FRTB is proposing the
use of expected shortfall with a 97.5% confidence level. For a normal distribution, with
mean of µ and standard deviation of σ, these two measures yield approximately the
same result. The 99% VaR formula is µ + 2.326σ, and the 97.5% expected shortfall
formula is µ + 2.338σ. However, if distributions have fatter tails than a normal
distribution, then the 97.5% expected shortfall can be considerably different from the
99% VaR.
Under this FRTB proposal, banks would be required to forgo combining a 10-day, 99%
VaR with a 250-day stressed VaR, and instead calculate capital based on expected
shortfall using a 250-day stressed period exclusively. Just as with the 250-day stressed
VaR, banks would be charged with self-selecting a 250-day window of time that would
be exceptionally difficult financially for the bank’s portfolio.

PROFESSOR’S NOTE
There are approximately 250 trading days in a 12-month time period. This is
why 250-day time windows are used. Following the same logic, a 120-day
window equates to six months, a 60-day window equates to one quarter
(three months), a 20-day window equates to one month, and a 10-day window
is essentially two weeks.

Liquidity Horizons
LO 18.b: Compare the various liquidity horizons proposed by the FRTB for
different asset classes and explain how a bank can calculate its expected
shortfall using the various horizons.

According to the Basel Committee, a liquidity horizon (LH) is “the time required to
execute transactions that extinguish an exposure to a risk factor, without moving the
price of the hedging instruments, in stressed market conditions.” The standard 10-day
LH was not deemed appropriate given the actual variations in liquidity of the
underlying transactions. Five different liquidity horizons are now in use: 10 days, 20
days, 40 days, 60 days, and 120 days. Consider the 60-day horizon, which is essentially
three months’ worth of trading days. The calculation of regulatory capital for a 60-day
horizon is intended to shelter a bank from significant risks while waiting three months
to recover from underlying price volatility.
Under FRTB proposals, every risk factor is assigned a liquidity horizon for capital
calculations. For example, investment grade sovereign credit spreads are assigned a 20-
day horizon, while non-investment grade corporate credit spreads are assigned a 60-
day horizon. See Figure 18.2 for a sample listing of liquidity horizons.
Figure 18.2: Allocation of Risk Factors to Liquidity Horizons

The Basel Committee’s original idea was to utilize overlapping time periods for stress
testing. They initially wanted to find a time period’s expected shortfall (ES) by scaling
smaller time periods up to longer time periods using a series of trials. Consider a bank
that has a 10-day risk asset, like large-cap equity, and a 60-day risk asset, like a non-
investment grade corporate credit spread. In the first trial, they would measure the
stressed P&L changes from Day 0 to Day 10 for the large-cap equity and also the value
change from Day 0 to Day 60 for the non-investment grade corporate credit spread. The
next trial would measure the change from Day 1 to Day 11 on the large-cap equity and
from Day 1 to Day 61 for the credit spread. The final simulated trial would measure Day
249 to Day 259 for the large-cap equity and Day 249 to Day 309 for the credit spread.
The ES used would then be the average loss in the lower 2.5% tail of the distribution of
the 250 trials.
After the initial idea was submitted for comments, it was revised in December 2014 to
incorporate five categories. The rationale was to reduce implementation costs. As noted
earlier, the updated categories are as follows:
Category 1 is for risk factors with 10-day horizons.
Category 2 is for risk factors with 20-day horizons.
Category 3 is for risk factors with 40-day horizons.
Category 4 is for risk factors with 60-day horizons.
Category 5 is for risk factors with 120-day horizons.
Using this revised, categorical process attempts to account for the fact that risk factor
shocks might not be correlated across liquidity horizons.
This proposed new process is formally known as the internal models-based
approach (IMA). In the internal models-based approach, expected shortfall is
measured over a base horizon of 10 days. The expected shortfall is measured through
five successive shocks to the categories in a nested pairing scheme using ES1–5. ES1 is
calculated as a 10-day shock with intense volatility in all variables from category 1–5.
ES2 is calculated as a 10-day shock in categories 2–5, holding category 1 constant. ES3
is calculated as a 10-day shock in categories 3–5, holding category 1 and 2 constant. ES4
is calculated as a 10-day shock in categories 4–5, holding categories 1–3 constant. The
final trial, ES5, is calculated as a 10-day shock in category 5, holding categories 1–4
constant. The idea is to measure the hit to the bank’s P&L for ES1–5. The overall ES is
based on a waterfall of the categories (as described previously) and is scaled to the
square root of the difference in the horizon lengths of the nested risk factors. This
relationship is shown in the following liquidity-adjusted ES formula:

To better illustrate this formula, assume all risk factors belong to either Category 1 or
Category 2. In this case, only ES1 and ES2 need to be considered. The behavior of all risk
factors during a 10-day horizon is assumed to be independent of the behavior of all
Category 2 risk factors during a future 10-day horizon. After applying the square root
rule, the liquidity-adjusted ES simplifies to:

Proposed Modifications to Basel Regulations

LO 18.c: Explain the FRTB revisions to Basel regulations in the following areas:
Classification of positions in the trading book compared to the banking book.
Backtesting, profit and loss attribution, credit risk, and securitizations.

Trading Book vs. Banking Book

The FRTB also addressed regulatory modifications. One modification is to clarify if a
risk asset should be considered part of the trading book or the banking book.
Historically, the trading book consisted of risk assets that the bank intended to trade.
Trading book assets have been periodically marked to market. The banking book has
consisted of assets that are intended to be held until maturity, and they are held on the
books at cost. Banking book assets are subject to more stringent credit risk capital
rules, while trading book assets are subject to market risk capital rules. Using different
rules has enabled a form of regulatory arbitrage where banks will hold credit-
dependent assets in the trading book to relax capital requirements.
In an attempt to mitigate this regulatory arbitrage, the FRTB makes a specific
distinction between assets held in the trading book and those held in the banking book.
To be allocated to the trading book, the bank must prove more than an intent to trade.
They must meet dual criteria of (1) being able to trade the asset and (2) physically
managing the associated risks of the underlying asset on the trading desk. If these two
criteria are met, then an asset can be allocated to the trading book, but the day-to-day
price fluctuations must also affect the bank’s equity position and pose a risk to bank
solvency.
Another important distinction was made in terms of reclassification between a banking
book asset and a trading book asset. Once an asset has been acquired and initially
assigned to either the trading book or the banking book, it cannot be reclassified except
for extraordinary circumstances. This roadblock has been established to minimize the
act of switching between categories at will, based on how capital requirements are
calculated. An example of an extraordinary circumstance is if the bank changes
accounting practices that is a firm wide shift. Another caveat is that any benefit derived
from calculating capital requirements under a post-shift category is disallowed. The
capital requirement of the original method must be retained.

Backtesting
Stressed ES measures are not backtested under FRTB. This is mainly due to difficulties
backtesting a stressed measure because extreme values used for stressed ES may not
occur with the same frequency in the future. In addition, backtesting VaR is easier than
backtesting ES. For backtesting VaR, FRTB suggests using a one-day time horizon and
the latest 12 months of data. Either a 99% or 97.5% confidence level can be applied. In
these cases, if there are more than 12 exceptions at the 99% level or more than 30
exceptions at the 97.5% level, the standardized approach must be used to compute
capital.

Profit/Loss Attribution
There are two measures banks can use to compare actual profit/loss figures to those
predicted by their own internal models. Regulators use these tests to perform profit
and loss attribution. These measures are as follows:

D represents the difference between the actual and model profit/loss on a given day,
and A represents the actual profit/loss on a given day. According to regulators, the first
measure should range between ±10%, and the second measure should be under 20%. If
the two ratios fall outside these requirements on four or more occasions during a 12-
month period, capital must be computed using the standardized approach.

Credit Risk
Basel II.5 introduced the incremental risk charge (IRC), which recognizes two
different types of risk created by credit-dependent risk assets: credit spread risk and
jump-to-default risk.
Credit spread risk is the risk that a credit risk asset’s credit spread might change, and
thus, cause the mark-to-market value of the asset to change. This risk can be addressed
by using the expected shortfall calculation process discussed earlier. The IRC process
allows banks to assume a constant level of risk. This means that it is assumed that
positions that deteriorate are replaced with other risk assets. For example, if a bank has
an A-rated bond with a three-month liquidity horizon that suffers a credit-related loss,
then it is assumed that the bank replaces this risk asset with another A-rated bond at
the end of the three-month liquidity horizon. This is clearly a simplifying assumption,
which is being replaced with incremental marking to market without assuming
replacement under the FRTB proposals.
Jump-to-default risk is the risk that there will be a default by the issuing company of
the risk asset. A default would lead to an immediate and potentially significant loss for
the bank that holds the defaulted issuer’s risk asset. This risk is subject to an
incremental default risk (IDR) charge. The IDR calculation applies to all risk assets
(including equities) that are subject to default. It is calculated based on a 99.9% VaR
with a one-year time horizon.

Securitizations
In order to address risks from securitized products (e.g., ABSs and CDOs), Basel II.5
introduced a comprehensive risk measure (CRM) charge. However, under CRM rules,
banks were allowed to use their own internal models, which created significant
variations in capital charges among banks. As a result, under FRTB, the Basel
Committee has determined that securitizations should instead utilize the standardized
approach.
MODULE QUIZ 18.1
1. Which of the following statements regarding the differences between Basel I, Basel II.5, and the
Fundamental Review of the Trading Book (FRTB) for market risk capital calculations is incorrect?
A. Both Basel I and Basel II.5 require calculation of VaR with a 99% confidence level.
B. FRTB requires the calculation of expected shortfall with a 97.5% confidence level.
C. FRTB requires adding a stressed VaR measure to complement the expected shortfall calculation.
D. The 10-day time horizon for market risk capital proposed under Basel I incorporates a recent
period of time, which typically ranges from one to four years.
2. What is the difference between using a 95% value at risk (VaR) and a 95% expected shortfall (ES) for
a bond portfolio with $825 million in assets and a probability of default of 3%?
A. Both measures will show the same result.
B. The VaR shows a loss of $495 million while the expected shortfall shows no loss.
C. The VaR shows no loss while the expected shortfall shows a $495 million loss.
D. The VaR shows no loss while the expected shortfall shows a $395 million loss.

3. Which of the following statements best describe how the internal models-based approach (IMA)
incorporates various liquidity horizons into the expected shortfall calculation?
A. A rolling 10-day approach is used over a 250-day window of time.
B. Smaller time periods are used to extrapolate into larger time periods.
C. A series of weights are applied to the various liquidity horizons along with a correlation factor
determined by the Basel Committee.
D. The expected shortfall is based on a waterfall of the liquidity horizon categories and is then
scaled to the square root of the difference in the horizon lengths of the nested risk factors.
4. Which of the following statements represents a criteria for classifying an asset into the trading
book?
 I. The bank must be able to physically trade the asset.
II. The risk of the asset must be managed by the bank’s trading desk.

A. I only.
B. II only.
C. Both I and II.
D. Neither I nor II.
5. Which of the following risks is specifically recognized by the incremental risk charge (IRC)?
A. Expected shortfall risk, because it is important to understand the amount of loss potential in
the tail.
B. Jump-to-default risk, as measured by 99% VaR, because a default could cause a significant loss
for the bank.
C. Equity price risk, because a change in market prices could materially impact mark-to-market
accounting for risk.
D. Interest rate risk, as measured by 97.5% expected shortfall, because an increase in interest
rates could cause a significant loss for the bank.

KEY CONCEPTS

LO 18.a
The Fundamental Review of the Trading Book (FRTB) is changing the historical reliance
on 10-day value at risk (VaR) with a 99% confidence level combined with a 250-day
stressed VaR. The new calculation will require the use of expected shortfall with a
97.5% confidence level. This switch will better capture the value of capital at risk
below a certain confidence level.

LO 18.b
The FRTB is establishing various liquidity horizons, which are the length of time
“required to execute transactions that extinguish an exposure to a risk factor, without
moving the price of the hedging instruments, in stressed market conditions.” The
expected shortfall will then be calculated by structuring risk assets into categories and
solving for an overall value of expected shortfall for a bank’s risk assets.

LO 18.c
Some banks have engaged in regulatory arbitrage by actively switching assets between
the trading book and the banking book depending on which category would show their
capital requirements in a more favorable light. The FRTB is mitigating this arbitrage
opportunity by deploying a rules-based standard for classification into these categories
and a roadblock for easily switching between them.
When backtesting VaR, a bank should use the standardized approach if the number of
exceptions falls outside the ranges specified by FRTB. When performing profit and loss
attribution, the bank can use ratios that account for differences between actual and
model profit/loss data on a given day. The standardized approach should be used to
compute capital if these ratios fall outside the requirements specified by regulators.
Basel II.5 introduced the incremental risk charge (IRC), which recognizes credit spread
and jump-to-default risk. For securitizations, Basel II.5 introduced the comprehensive
risk measure (CRM) charge. Due to variations in computing capital with internal
models, FRTB recommends using the standardized approach.

ANSWER KEY FOR MODULE QUIZ

Module Quiz 18.1

1. C Basel I and Basel II.5 use VaR with a 99% confidence level and the FRTB uses the
expected shortfall with a 97.5% confidence level. Basel I market risk capital
requirements produced a very current result because the 10-day horizon
incorporated a recent period of time. The FRTB does not require adding a stressed
VaR to the expected shortfall calculation. It was Basel II.5 that required the
addition of a stressed VaR. (LO 18.a)
2. C The VaR measure would show a $0 loss because the probability of default is less
than 5%. Having a 3% probability means that three out of five times, in the tail,
the portfolio will experience a total loss. The potential loss is $495 million (= 3/5
× $825 million). (LO 18.a)
3. D The expected shortfall is based on a waterfall of the liquidity horizon categories
and is then scaled to the square root of the difference in the horizon lengths of the
nested risk factors. (LO 18.b)
4. C The criteria for classification as a trading book asset are (1) the bank must be able
to physically trade the asset, and (2) the bank must manage the associated risks
on the trading desk. (LO 18.c)
5. B The two types of risk recognized by the incremental risk charge are (1) credit
spread risk and (2) jump-to-default risk. Jump-to-default risk is measured by 99%
VaR and not 97.5% expected shortfall. (LO 18.c)