Risk Assessment Techniques — The CFA Level I Survival Guide (with jokes)

Opening: Why are we poking the risk beehive?

Imagine you're back in our Financial Markets module: emerging markets looked tempting — high returns, spicy volatility — and our Financial Crisis Analysis lecture reminded you how quickly those temptations can turn into a horror movie. Risk assessment is the toolkit that helps you decide whether you walk into that movie theater or run toward the exit with popcorn and a life-jacket.

This piece builds on what you already know (markets, crises, types of risks). Now we learn how to measure and test those risks — the practical, exam-friendly techniques you'll be grilled on at CFA Level I.

Big picture: What is “risk assessment” in practice?

Risk assessment = identify + quantify + prioritize.

• Identify the risks (market, credit, liquidity, operational, model, etc.). You already met these.
• Quantify them with numbers (volatility, VaR, PD, EL, duration, etc.).
• Prioritize using heatmaps, KRIs, and stress scenarios so scarce capital (and attention) goes where it matters.

Risk without measurement is a rumor. We measure so we can manage, hedge, or politely ignore (if justified).

Core quantitative techniques (the bread-and-butter)

1) Volatility & dispersion

• Standard deviation: the classic measure of total variability. Easy, intuitive, but treats upside and downside equally.
• Semivariance / downside deviation: focuses on bad outcomes — more useful when you care about losses.

2) Value at Risk (VaR) — the exam celebrity

VaR answers: "What is the worst loss I can expect over horizon T at confidence level α?" Example: 1-day 99% VaR = $X means losses worse than $X occur 1% of the time.

Common calculation methods:

Code-style formulas:

Parametric VaR ≈ z_{α} * σ_portfolio * portfolio_value

Where z_{α} is the standard normal critical value (e.g., 2.33 for 99%).

Limitations: VaR is not subadditive in some methods (so aggregation can be tricky); it tells you the cut-off but not the average severity beyond that cut-off.

3) Expected Shortfall (CVaR)

• Also called Conditional VaR or Expected Shortfall.
• Measures the average loss given that losses exceed the VaR threshold — a better tail-risk measure.

4) Sensitivity analysis

• Delta, gamma, vega, theta for derivatives (Level I will expect conceptual understanding).
• Duration and convexity for bonds: duration ≈ sensitivity to small yield changes; convexity adjusts for larger moves.
• DV01 / PVBP: dollar value of a basis point — tells you how much your bond position moves for 1 bp change.

5) Scenario analysis & stress testing

• Scenario analysis: build a plausible set of macro moves (e.g., 30% EM FX devaluation, 300 bps rate spike) and compute portfolio P/L.
• Stress testing: more severe, often hypothetical (e.g., 2008-like banking panic). Regulators love these.

Quick distinction:

• Scenario analysis = specific story-driven outcomes.
• Stress testing = extreme but plausible, often regulatory.

6) Backtesting and model validation

• Compare VaR predictions with realized outcomes (exceptions/violations).
• If you’re seeing more breaches than your model predicts, something’s wrong: recalibrate, change model, or admit you were optimistic.

Credit & liquidity assessment metrics (don't ignore them)

Credit: PD, LGD, EAD, Expected Loss

• PD (Probability of Default): chance borrower defaults in a period.
• LGD (Loss Given Default): fraction lost if default occurs (1 − recovery rate).
• EAD (Exposure at Default): amount exposed when default occurs.

Expected Loss (EL) = PD × LGD × EAD

Exam tip: EL is used for provisioning; unexpected loss (UL) is for capital.

Liquidity measures

• Bid-ask spread, market depth, turnover, time-to-liquidate.
• Liquidity risk can explode during crises — remember the Financial Crisis module.

Qualitative techniques & visualization

• Risk mapping / heatmaps: plot likelihood vs impact; color it red, and your boss will look worried (mission accomplished).
• Key Risk Indicators (KRIs): early-warning metrics tied to risks (e.g., rising NPL ratio for credit risk).
• Checklists and control matrices: operational risks love human errors; checklists reduce surprises.

Visualizing risk often moves stakeholders more than 100 pages of spreadsheets.

Practical step-by-step for a risk assessment (exam-friendly list)

1. Inventory exposures (market positions, credit exposures, liquidity profiles).
2. Classify risk types and link to drivers (rates, spreads, FX, commodity prices).
3. Choose measurement tools (std dev, VaR method, scenario, PD/LGD estimates).
4. Run models and scenarios; compute EL, VaR, ES, DV01, etc.
5. Backtest models where possible; compare exceptions to expectations.
6. Produce heatmaps and KRIs; prioritize top risks.
7. Recommend mitigation (hedges, limits, capital buffers, contingency funding).
8. Document assumptions and limitations.

Common exam traps & model limitations (the things they’ll ask you)

• Assuming normality for returns → underestimates tail risk.
• Confusing VaR with expected loss beyond VaR (VaR ≠ ES).
• Forgetting liquidity when stressing market moves — you might be forced to sell at worse prices.
• Overreliance on historical simulation during regime shifts (emerging markets are especially risky here).

Closing — TL;DR and the takeaways you can actually use

• Use standard deviation and VaR for quick quantification; remember Expected Shortfall is a better tail metric.
• Apply sensitivity (duration/DV01) for interest rate risk and PD × LGD × EAD for credit provisioning.
• Stress test with scenarios inspired by past crises (and plausible future shocks — yes, including weird geopolitical moves).
• Visualize and prioritize with heatmaps and KRIs — numbers alone don’t get budgets.

Final(ly useful) thought:

Measuring risk doesn’t remove it. It makes you honest about it — and in finance, honesty tends to save money (and dignity).

Version: "Risk Assessment: No-Nonsense, Slightly Dramatic Breakdown"