Handout: Financial Risks (Part I) – Topic 3
This handout will have to be read in conjunction with the lecture notes, whiteboard from the
classroom and the supplementary Excel files available on LMS.
1. Risk vs. Return
The concept of risk versus return is a fundamental principle in finance, asserting that higher expected
returns can typically only be achieved by taking on higher risks. This trade-off is central to investment and
portfolio management [1.1].
• Expected Return and Standard Deviation of Return: Investments are characterised by their expected
return and the standard deviation of return, which quantifies the variability or risk of those returns. For
example, an equity investment might have an expected return of 10% and a standard deviation of
18.97%, compared to Treasury bills yielding 5%.
• Portfolio Theory and Efficient Frontier: Early attempts to understand this trade-off include Markowitz’s
(1952) work on portfolio selection. This led to the concept of the efficient frontier, which represents the
optimal set of portfolios offering the highest expected return for a given level of risk (standard deviation
of return), or the lowest risk for a given expected return [1.2, 45, 28]. No investment can dominate a
point on the efficient frontier by having both a higher expected return and a lower standard deviation.
When a risk-free investment is included, the efficient frontier becomes a straight line tangent to the
efficient frontier of risky investments.
• Capital Asset Pricing Model (CAPM): Sharpe (1964) extended Markowitz’s analysis by developing the
CAPM. This model describes a relationship between expected return and systematic risk.
◦ Systematic vs. Nonsystematic Risk:
▪ Nonsystematic risk (or idiosyncratic risk) is specific to an individual investment and can be largely
diversified away in a large, well-diversified portfolio. Investors should not require extra return for
bearing nonsystematic risk if it can be diversified away.
▪ Systematic risk is the market-wide risk that cannot be diversified away. A well-diversified investor
should require an expected return to compensate for this systematic risk.
◦ Beta (β): The CAPM uses beta to measure an investment’s sensitivity to the return on the market
portfolio. An investment with a beta of 0 should yield the risk-free rate, while an investment with a beta
of 1.0 (like the market portfolio) would yield the market return.
For instance, if the risk-free rate is 5% and the market return is 10%, an investment with a beta of 0.5
should have an expected return of 7.5% (5% + 0.5 * (10% - 5%)).
1
• Arbitrage Pricing Theory (APT): Developed by Ross (1976), APT is an extension of CAPM,
accommodating situations with multiple sources of systematic risk.
• Risk for Companies: While shareholders in theory focus on systematic risk, companies (especially
financial institutions) are concerned with total risks (systematic plus nonsystematic). This is due to factors
like bankruptcy costs, which are real costs to shareholders arising from the bankruptcy process itself.
Regulated financial institutions are particularly compelled to consider total risks to prevent failure, with
regulators aiming to make bankruptcy a highly unlikely event by requiring sufficient capital.
2. Risk Management in Banking
Banking involves both commercial banking (taking deposits and making loans) and investment banking
(raising capital, advising on M&A, trading securities). Banks use a combination of risk aggregation and risk
decomposition approaches to manage the diverse risks they face.
• Commercial Banking Activities: This traditional activity involves accepting deposits and originating
loans.
◦ Balance Sheet Structure: A hypothetical small commercial bank (e.g., Deposits and Loans Corporation,
DLC) might have assets primarily composed of loans (e.g., 80%), cash and marketable securities (e.g.,
15%), and fixed assets (e.g., 5%). Funding largely comes from customer deposits (e.g., 90%),
subordinated long-term debt (e.g., 5%), and equity capital (e.g., 5%).
◦ Income Statement Components: Key items include net interest income (excess of interest earned over
paid), loan losses (provisions for defaults), non-interest income (fees for services), and non-interest
expenses (salaries, overheads). Loan losses are particularly sensitive to credit risk and economic
conditions, fluctuating significantly year to year. Operational risk can lead to large non-interest expenses
from litigation, business disruption, or employee fraud.
◦ Originate-to-Distribute Model (Securitisation): Banks may originate loans and then sell them off their
balance sheets by packaging them into securities. This frees up funds for more lending and capital for
other risks. However, this model faced severe problems during the 2007–2009 Global Financial Crisis due
to relaxed lending standards and declining credit quality.
• Investment Banking Activities: These include originating, underwriting, and placing securities for
corporations or governments. They also engage in securities trading, providing brokerage services, and
market making. The Dodd-Frank Act in the U.S. restricts banks from proprietary trading. Banks offer lines
of credit, services for exports, hedging contracts (foreign exchange, commodity, interest rate), securities
research, mutual funds, and increasingly, insurance products.
◦ Banking Book vs. Trading Book: A crucial distinction is made for accounting purposes. The trading book
includes assets and liabilities from trading operations, which are marked to market daily to reflect
changes in market prices. If market prices are unavailable, marking to model may be used. The banking
book holds assets like long-term loans that are not traded.
2
3
• Approaches to Risk Management:
◦ Risk Aggregation: This approach aims to reduce nonsystematic risks through diversification. For
example, insurance companies diversify across many policies, and banks diversify their loan
portfolios. However, systematic risk still remains.
◦ Risk Decomposition: This involves identifying and managing risks one by one. For market risks in a
trading room, individual traders might be responsible for specific market variables and must keep
their risk measures within specified limits. Credit derivatives also allow for risk decomposition of
credit risks.
• Risks Facing Banks: Banks face various risks, including:
◦ Credit Risk: The risk of counterparties defaulting on loans or derivatives transactions.
Traditionally the greatest risk.
◦ Market Risk: Arises from trading operations due to potential declines in the value of instruments
in the trading book.
◦ Operational Risk: The risk of losses due to inadequate or failed internal processes, people, and
systems, or from external events (e.g., fraud, natural disasters, litigation) [71, 80, 11.1]. Regulators
increasingly view operational risk as highly significant.
• Gap Risk (Management of Net Interest Income): A key risk management activity for a bank is ensuring
that its net interest income (the difference between interest earned and interest paid) remains stable
over time [69, 14.4, 273]. This is managed by the asset-liability management (ALM) function [14.4, 273].
◦ Nature of the Risk: If a bank funds long-term fixed-rate loans with short-term floating-rate
deposits, it is exposed to significant interest rate risk. If interest rates rise, the cost of rolling over
deposits will increase, while the interest earned on fixed-rate loans remains constant, squeezing
the bank’s net interest income.
◦ Mitigation: Banks use various strategies to manage this. For example, they might enter into
interest rate swaps to transform floating-rate liabilities into fixed-rate ones, or fixed-rate assets
into floating-rate ones, to match the interest rate sensitivity of their assets and liabilities. The
“bucket deltas” approach or GAP management is often used in ALM, dividing the yield curve into
segments (buckets) and calculating the dollar impact of changing all zero rates within that bucket
by one basis point [14.8.2, 274].
• Capital Requirements (Basel): Regulators set minimum capital levels to provide a cushion against losses
and ensure confidence in the banking system.
◦ Tier 1 Capital: The most important type of capital as it absorbs losses directly while the
bank continues as a going concern. It consists of equity and noncumulative perpetual
preferred stock, with goodwill subtracted. Under Basel III, it is referred to as Common Equity
Tier 1 (CET1), including share capital and retained earnings (excluding goodwill and deferred
tax assets). Additional Tier 1 (AT1) capital includes non-cumulative preferred stock.
4
◦ Tier 2 Capital: Also known as Supplementary Capital, it includes instruments like cumulative
perpetual preferred stock, certain long-term debentures, and subordinated debt (debt
subordinate to depositors) with an original life of over five years. Tier 2 capital acts as “gone-
concern capital,” absorbing losses if the bank becomes insolvent, protecting depositors.
◦ Tier 3 Capital: Under the 1996 Amendment to Basel I, Tier 3 capital consisted of short-term
subordinated debt (at least two-year maturity) and could be used for market risk capital. Tier 3
capital was eliminated under Basel III.
Basel I, II, and III Frameworks:
▪ Basel I (1988): Introduced international standards for bank regulation, focusing on credit risk. It
required banks to keep capital equal to at least 8% of risk-weighted assets (RWA). Assets were
assigned risk weights (e.g., corporate loans 100%, OECD government bonds 0%, residential
mortgages 50%). For OTC derivatives, a credit equivalent amount was calculated, comprising
current exposure plus an add-on factor for potential future exposure.
▪ Basel II (Implemented ~2007): A major overhaul with three pillars: Minimum Capital
Requirements, Supervisory Review, and Market Discipline. It refined credit risk capital calculations
to reflect counterparty credit risk (e.g., using the Standardized Approach or Internal Ratings Based
(IRB) Approach, which employed Vasicek’s model and a one-year 99.9% VaR for credit risk). It also
introduced a capital charge for operational risk.
▪ Basel II.5 (Implemented 2011): Introduced changes to market risk capital calculation in response
to the 2007-2009 crisis. Key elements included a stressed VaR, an incremental risk charge (IRC) for
credit-sensitive trading book products, and a comprehensive risk measure (CRM) for correlation-
dependent instruments. Stressed VaR used a 250-day period of stressed market conditions to
calculate a 10-day 99% VaR.
▪ Basel III (Finalized 2017, Implemented 2023): Further increased equity capital requirements,
tightened capital definitions, and introduced regulations for liquidity risk. Key components
include: Capital Definition and Requirements (e.g., higher CET1), Capital Conservation Buffer (2.5%
CET1 buffer), Countercyclical Buffer (0-2.5% CET1 buffer), Leverage Ratio (minimum 3% Tier 1
capital to exposure), Liquidity Risk (Liquidity Coverage Ratio (LCR) and Net Stable Funding Ratio
(NSFR)), and Counterparty Credit Risk (revising CVA risk capital). Global Systemically Important
Banks (G-SIBs) are required to hold additional Tier 1 equity capital.
5
3. Risk Management in Insurance
The primary role of insurance companies is to provide protection against adverse events in exchange for
regular premium payments. Insurance is broadly classified into life insurance and non-life (property-
casualty) insurance, with health insurance as a separate category.
• Types of Life Insurance:
◦ Term life: Provides coverage for a specified period.
◦ Whole life: Provides coverage for the entire life of the insured, typically with a savings
component.
◦ Variable life: Policyholder chooses investments, which affect cash value and death benefit.
◦ Universal life: Flexible premiums and death benefits, with a cash value component that earns
interest.
◦ Endowment life: Pays a lump sum to the policyholder after a specified term, or to beneficiaries
upon earlier death.
◦ Group life: Coverage provided to a group of people (e.g., employees) under a single contract.
• Risks Facing Insurance Companies:
◦ Underwriting Risk: The risk that policy reserves are insufficient to meet claims.
◦ Investment Risk: Risks associated with the performance of their investments, especially corporate
bonds, including default and interest rate risks.
◦ Liquidity Risk: The ability to convert investments into cash to meet unexpectedly high claims.
◦ Credit Risk: Exposure to defaults by banks and reinsurance companies with which they transact.
◦ Operational Risk and Business Risk: Similar to banks, insurance companies face these risks.
• Regulation:
◦ Solvency I (Europe): The original regulatory framework in the EU, primarily focused on
underwriting risks, linking capital to policy provisions.
◦ Solvency II (Europe, Implemented 2016): Replaced Solvency I, assigning capital for a wider range
of risks including investment and operational risks. It has three pillars analogous to Basel II, with
Pillar 1 specifying a Minimum Capital Requirement (MCR) and a Solvency Capital Requirement
(SCR). The SCR involves a VaR calculation with a one-year time horizon and a 99.5% confidence
limit.
6
4. Value at Risk (VaR) and Expected Shortfall (ES)
Value at Risk (VaR) is a popular measure used to summarise the total risk in a portfolio by providing a
single number. It addresses the question: “What loss level is such that we are X% confident it will not be
exceeded in N business days?”
• Definition and Parameters: VaR is a function of two parameters: T (the time horizon) and X (the
confidence level). For example, a “99% VaR over 10 days” means there is a 99% certainty that the loss will
not exceed a specific value (V dollars) over the next 10 days. It can be calculated from either the
probability distribution of gains or losses, typically being the loss at the (100-X)th percentile of the gain
distribution or the Xth percentile of the loss distribution.
• Factors Influencing VaR: The estimated VaR depends on:
1. The estimated variance-covariance terms of the risk factors.
2. The estimated sensitivity of the assets to the risk factors.
3. The portfolio weights.
4. The desired confidence level.
5. The desired holding period.
6. The amount invested.
• Probability Distributions and VaR Calculation:
◦ Continuous Distributions:
▪ Normal Distribution: If a portfolio’s gain over six months is normally distributed with a mean
of $2 million and a standard deviation of $10 million, the one-percentile point (for 99% VaR) is
calculated as $2 million – 2.33 * $10 million = -$21.3 million. Thus, the 99% VaR is $21.3
million [Example 11.1, 168].
▪ Uniform Distribution: For a one-year project where all outcomes between a loss of $50
million and a gain of $50 million are equally likely, the loss is uniformly distributed from -$50
million to +$50 million. The 99% VaR for a one-year horizon is $49 million, as there’s a 1%
chance of a loss greater than $49 million [Example 11.2, 169, 672].
◦ Discrete Distributions: If losses are discrete, VaR might not be uniquely defined. For instance, if a
project has a 98% chance of a $2 million gain, a 1.5% chance of a $4 million loss, and a 0.5% chance
of a $10 million loss, the 99% VaR is $4 million (as the cumulative loss distribution reaches 99% at $4
million) [Example 11.3, 169, 673]. If the confidence level is 99.5%, VaR is not uniquely defined
between $4 million and $10 million; a common convention is to use the midpoint (e.g., $7 million)
[Example 11.4, 170, 673].
7
• Confidence Intervals and Accuracy:
◦ The standard error of a VaR estimate can be quite high, especially for high confidence levels. For
example, if a 99% VaR is estimated as $25 million from 500 observations, a 95% confidence interval
might be from $21.7 million to $28.3 million. The standard error decreases with sample size (square
root of sample size) and with decreasing confidence levels.
◦ Back-testing: This involves comparing actual losses to historical VaR estimates to assess the model’s
reliability. If the percentage of exceptions (actual loss exceeding VaR) is significantly different from
expected, the model may be rejected.
• Advantages of VaR:
◦ Captures an important aspect of risk in a single number.
◦ Easy to understand.
◦ Answers a simple question (“How bad can things get?”) that senior managers want to know.
• Drawbacks of VaR and Expected Shortfall (ES):
◦ VaR’s Drawback: While intuitive, VaR can lead to undesirable incentives for traders. A trader
might structure a portfolio where the VaR limit is met, but there’s a small probability of an
extremely large loss beyond the VaR level.
◦ Coherent Risk Measures: Artzner et al. Proposed properties for a “coherent” risk measure:
Monotonicity, Translation Invariance, Homogeneity, and Subadditivity. Subadditivity means that
the risk of two merged portfolios should be no greater than the sum of their individual risks,
reflecting diversification benefits. VaR does not always satisfy subadditivity, meaning it can fail to
recognise the benefits of diversification. For example, combining two independent projects, each
with a 97.5% VaR of $1 million, could result in a portfolio VaR of $11 million, violating subadditivity
[Example 11.5, 177, 206].
◦ Expected Shortfall (ES): Also known as conditional value at risk (C-VaR), conditional tail
expectation, or expected tail loss. ES asks: “If things do get bad, what is the expected loss?”. It is
the expected loss during time T, conditional on the loss being greater than the VaR.
▪ Coherence of ES: ES is a coherent risk measure, meaning it always satisfies the subadditivity
condition, unlike VaR. This provides better incentives for traders as it accounts for the magnitude
of losses beyond the VaR level.
▪ Calculation of ES: For a normally distributed loss with mean μ and standard deviation σ, ES can be
calculated using a specific formula involving the 𝑋 𝑡ℎ percentile of the standard normal
distribution. For example, if a 10-day loss is normal with zero mean and $20 million standard
deviation, the 99% ES is $53.3 million [Example 11.10, 185].
8
• Time Horizon and Scaling:
◦ The appropriate time horizon for VaR/ES depends on the application, ranging from a few days for
liquid trading positions to longer periods for pension funds or less liquid instruments.
◦ For market risks, analysts often start with a one-day VaR/ES. The common assumption for T-day VaR
(or ES) is to multiply the one-day measure by $\sqrt{T}$. This “square root of T” rule assumes daily
changes are independent and identically normally distributed with mean zero. However,
autocorrelation in daily changes means this rule can underestimate VaR/ES.
• Confidence Level: The chosen confidence level depends on the financial institution’s objectives, such as
maintaining a specific credit rating. A AA-rated bank might use a 99.98% confidence level for economic
capital, corresponding to a 0.02% probability of default.
• Marginal, Incremental, and Component Measures: For portfolios with subportfolios, analysts calculate
these measures to understand contributions to overall risk.
◦ Marginal VaR/ES: Sensitivity of the risk measure to the amount invested in a subportfolio.
◦ Incremental VaR/ES: The change in the risk measure when a subportfolio is added or removed.
𝜕 𝑉𝑎𝑅
◦ Component VaR/ES: For the 𝑖 𝑡ℎ subportfolio, it is defined as 𝑥𝑖 where 𝑥𝑖 is the amount
𝜕 𝑥𝑖
invested in the subportfolio. Euler’s theorem shows that total VaR (or ES) is the sum of its component
VaRs (or ESs), allowing for risk allocation to business units [194, 200, 28.6].
• Regulatory Context:
◦ Regulators have traditionally based market risk capital on a k times the 10-day 99% VaR, where k is
at least 3.0.
◦ Under Basel II, capital for credit risk and operational risk was based on a one-year 99.9% VaR.
◦ Basel II.5 introduced the stressed VaR.
◦ The Fundamental Review of the Trading Book (FRTB) represents a shift towards using Expected
Shortfall (ES) with a 97.5% confidence level and varying liquidity horizons (10, 20, 40, 60, 120 days)
for market risk capital [27.1, 27.3.1, 609].
