Skip to guide content
Desk2Quant

Market and enterprise risk

Risk Quant Interview Guide: Models, Controls, and Decisions

Prepare for market and risk quant interviews: VaR, expected shortfall, stress testing, backtesting, products, data, model risk, and communication.

Desk2Quant Editorial Team13 minute read

Risk quant interviews test more than formula knowledge. A risk number has a purpose, a scope, data and model assumptions, known blind spots, controls, and a user who must make a decision. Strong candidates connect all of those layers.

The exact syllabus depends on market risk, counterparty credit risk, capital, stress testing, or model risk. The preparation framework below covers the shared core and shows how to communicate technical findings without hiding uncertainty.

What this guide helps you do

  • Explain risk measures with assumptions and decision context.
  • Design backtests and stress tests that reveal model weaknesses.
  • Connect positions, market data, valuation, aggregation, and reporting.
  • Communicate a technical risk finding with proportionate action.

1. Start from positions and risk-factor mechanics

Before discussing VaR, identify the instruments, contractual cash flows, valuation inputs, and risk-factor mapping. Know how linear and nonlinear exposures behave under spot, curve, spread, volatility, correlation, and basis moves. Aggregation can conceal concentrations when mappings or diversification assumptions are weak.

Be ready to explain PnL from both revaluation and sensitivities. A sensitivity approximation is fast and interpretable, but higher-order effects, discontinuities, and changing market regimes can make full revaluation necessary.

  • Rates: curve nodes, basis, optionality, discounting, and day-count conventions.
  • Equity and FX: spot, volatility surface, dividends or carry, and cross-risk.
  • Credit: spreads, default, recovery, migration, liquidity, and wrong-way risk.
  • Portfolio: netting, hedges, concentration, liquidity horizon, and stale positions.

2. Compare risk measures rather than defending one number

VaR is a loss quantile for a chosen horizon and confidence level; expected shortfall averages losses beyond a quantile. Neither removes choices about data windows, weighting, liquidity horizons, mapping, valuation, or treatment of missing history. A correct definition should lead immediately to those implementation questions.

Risk measures are complementary. Sensitivities explain local exposure, scenarios examine specified moves, VaR summarizes a distribution quantile, expected shortfall describes the tail beyond it, and stress tests explore severe or structurally different conditions.

Risk metric strengths and blind spots
ToolUseful forKey limitation to mention
SensitivitiesFast attribution and local hedgingMay miss curvature, path dependence, and regime changes
VaRA consistent portfolio loss quantileDoes not describe severity beyond the quantile
Expected shortfallAverage tail severityStill depends on sparse tail data and modeling choices
Stress testingNamed severe scenarios and vulnerabilitiesCoverage depends on scenario imagination and relevance

3. Treat backtesting as diagnosis, not a pass/fail ritual

A VaR exception is an observation that realized loss exceeded the forecast threshold, but the comparison is only meaningful when PnL definition, horizon, positions, and market close are aligned. Investigate clusters, regime dependence, data breaks, valuation changes, and whether the exception came from an exposure outside the modeled factor set.

Backtesting should include coverage, independence, and attribution. A model can have the expected exception count while failing during precisely the periods when risk management matters most.

  • Reconcile hypothetical, clean, and actual PnL before interpreting exceptions.
  • Examine exception timing and clustering, not only the annual count.
  • Attribute breaks to positions, factors, pricing, data, and operational events.
  • Pair statistical results with sensitivity and scenario diagnostics.

4. Design scenarios and challenge the data

Good scenarios have a narrative, internally coherent factor moves, a transmission path to positions, and a clear use. Historical scenarios are grounded but cannot cover new structures; hypothetical scenarios explore vulnerabilities but need disciplined calibration. Reverse stress testing starts from an unacceptable outcome and asks which conditions could produce it.

Data quality is part of the model. Discuss proxies, missing observations, survivorship, stale marks, corporate actions, changing identifiers, and limited history. A proxy should be justified by economic behavior and tested in stress, not chosen only because its normal-period correlation is high.

5. Turn analysis into a controlled decision

A risk quant should distinguish a model weakness from its material impact. Frame a finding with evidence, affected portfolios or processes, severity, compensating controls, remediation options, owner, and timeline. Escalation should be proportionate and traceable.

When speaking to non-modelers, lead with the decision and exposure, then explain the mechanism. Avoid both false precision and vague disclaimers. A range with drivers and actions is often more useful than a single unqualified number.

Practise aloud

Interview drills with answer direction

Question 1

What is the difference between VaR and expected shortfall?

Answer direction: VaR is a loss quantile at a confidence level; expected shortfall is the average loss conditional on being beyond that quantile. Then discuss horizon, data, tail estimation, and the decision each supports.

Question 2

A VaR model has too many exceptions. How do you investigate?

Answer direction: First align PnL and forecast definitions, then inspect timing, positions, factor mapping, market data, valuation changes, volatility regime, and clusters. Separate model weakness from data or operational breaks.

Question 3

How would you stress a portfolio with little history?

Answer direction: Combine economic narratives, cross-asset consistency, comparable episodes, sensitivity-based construction, expert challenge, and reverse stress. Report the uncertainty caused by limited calibration evidence.

Question 4

When is a proxy risk factor acceptable?

Answer direction: When there is an economic rationale, stable behavior over relevant regimes, conservative stress behavior, transparent basis risk, monitoring, and a plan for material exposures when the proxy breaks.

Question 5

How would you explain a large risk change to senior management?

Answer direction: Lead with magnitude, affected portfolio, main drivers, whether it reflects exposure or methodology, plausible stress impact, and recommended action. Keep technical details available for challenge.

Turn reading into practice

A focused study plan

  1. Layer 1

    Products and PnL

    Map cash flows, risk factors, sensitivities, and valuation conventions for the target asset classes.

  2. Layer 2

    Metrics

    Implement and compare sensitivities, VaR, expected shortfall, and scenario losses.

  3. Layer 3

    Challenge

    Practise backtesting, stress design, proxy review, data diagnostics, and limitations.

  4. Layer 4

    Decision

    Write and present findings with materiality, controls, owners, and verification.

Self-review

Frequent mistakes to catch early

  • Giving a metric definition without horizon, confidence, data, or use.
  • Treating diversification as stable during stress.
  • Counting backtest exceptions without investigating their cause and timing.
  • Reporting a model issue without materiality or a practical control.

Continue with structured practice

Relevant Desk2Quant resources

Keep building

Common questions

Frequently asked questions

What topics are covered in a risk quant interview?

Typical topics include products and valuation, sensitivities, VaR, expected shortfall, stress testing, backtesting, time series, data quality, model risk, regulation, coding, and communication. The weighting varies by team.

Do risk quant interviews require coding?

Many do. Python, R, SQL, or another analytics language may be tested through data manipulation, risk calculations, or debugging. Some model implementation teams also expect stronger software engineering.

How should I explain VaR in an interview?

Define it as a portfolio loss quantile for a stated horizon and confidence level, then explain the methodology, data window, valuation approach, backtesting, uses, and the tail information it does not provide.

What makes a strong risk case-study answer?

A strong answer connects the exposure and decision to data, assumptions, method, validation, materiality, controls, and communication. It also states what evidence would change the conclusion.