Risk Quant
Measures how much a portfolio could lose and sets the limits that keep losses survivable.
The job
A market risk quant measures how much a portfolio could lose and sets the limits that keep losses survivable. They build volatility forecasts, VaR and expected-shortfall models, stress scenarios and limit rules, and backtest them against what happened. There is no trading P&L; the job is a risk number the firm can act on. Things go wrong when returns are fatter-tailed than the model assumes, when breaches cluster, and when a limit reacts too slowly to a new regime.
- Supports
- Market risk teams and the desks whose limits they set
- Stages
- 5 to learn, then proof of work
- Two routes
- Free resources, or SelfMadeQuant programmes and projects
The roadmap
The same stages either way. The free route uses university courses, lectures, books and papers that are free online. The SelfMadeQuant route maps each stage to our programmes and ends in projects on real market data.
Stage 01 · Foundations
Statistics and testing
The standard VaR backtests are likelihood-ratio tests.
- Likelihood ratio tests
- Goodness of fit
- Confidence intervals
- 18.05 Introduction to Probability and Statistics (opens in a new tab)
MIT OpenCourseWare (Jeremy Orloff and Jennifer French Kamrin)
Estimation, confidence intervals and hypothesis tests with short applied problem sets.
- 18.650 Statistics for Applications (opens in a new tab)
MIT OpenCourseWare (Philippe Rigollet)
Maximum likelihood, goodness-of-fit and testing at the depth risk model reviews need.
- All of Statistics: A Concise Course in Statistical Inference (opens in a new tab)
Larry Wasserman
A fast, compact tour of inference, bootstrap and testing for quantitative readers.
Stage 02 · Core
Volatility modelling
Forecasting tomorrow’s volatility from today’s data.
- EWMA
- GARCH
- Realised volatility
- GARCH 101: The Use of ARCH/GARCH Models in Applied Econometrics (opens in a new tab)
Robert Engle / Journal of Economic Perspectives
A plain introduction to GARCH, worked through a portfolio VaR example.
- RiskMetrics Technical Document (4th edition) (opens in a new tab)
J.P. Morgan and Reuters / MSCI
The original EWMA volatility and covariance method, including the 0.94 decay factor.
- A Forecast Comparison of Volatility Models: Does Anything Beat a GARCH(1,1)? (opens in a new tab)
Peter Reinhard Hansen and Asger Lunde
Compares hundreds of volatility models and shows GARCH(1,1) is hard to beat.
Stage 03 · Core
VaR, expected shortfall and tail risk
Measuring the tail, and checking the measure against what happened.
- VaR and expected shortfall
- Coherent risk measures
- Extreme value theory
- VaR backtesting
- Coherent Measures of Risk (opens in a new tab)
Artzner, Delbaen, Eber and Heath
The axioms that show why expected shortfall is coherent and VaR is not.
- Supervisory framework for the use of backtesting in conjunction with the internal models approach (opens in a new tab)
Basel Committee on Banking Supervision / BIS
The traffic-light test for counting VaR exceptions that regulators still use.
- Quantitative Risk Management tutorial slides (opens in a new tab)
Alexander McNeil, Rüdiger Frey and Paul Embrechts
Slides on risk measures, extreme value theory and backtesting from the standard QRM text.
- Modelling Extremal Events for Insurance and Finance (opens in a new tab)
Paul Embrechts, Claudia Klüppelberg and Thomas Mikosch
The reference on extreme value theory, tail estimation and peaks over threshold.
Stage 04 · Specialist
Stress testing and regulation
Scenarios beyond the data, and the rules banks must follow.
- Scenario design
- Basel market risk
- FRTB basics
- Minimum capital requirements for market risk (FRTB) (opens in a new tab)
Basel Committee on Banking Supervision / BIS
The FRTB standard: expected shortfall, liquidity horizons and the standardised approach.
- Stress testing principles (opens in a new tab)
Basel Committee on Banking Supervision / BIS
Supervisory expectations for scenario design, governance and use of stress test results.
- Stress Tests (opens in a new tab)
Federal Reserve Board
Published US stress scenarios and methodology, a worked example of severe but plausible design.
Stage 05 · Specialist
Reproducible risk code
A risk number someone else can rerun and audit.
- Testing
- Version control
- Reproducible pipelines
- 6.0001 Introduction to Computer Science and Programming in Python (opens in a new tab)
MIT OpenCourseWare
Python fundamentals, testing and debugging taught from first principles.
- Version Control with Git (opens in a new tab)
The Carpentries (Software Carpentry)
A hands-on lesson in tracking, reviewing and reverting changes to model code.
- pytest documentation (opens in a new tab)
pytest developers
How to write unit tests that pin down VaR and P&L calculations before they change.
- Good enough practices in scientific computing (opens in a new tab)
Greg Wilson et al. / PLOS Computational Biology
Practical habits for data, code and project layout that make results reproducible.
Stage 06 · Proof of work
Backtest and challenge a historical VaR model
Build historical simulation, EWMA and GARCH VaR for a US equity market portfolio, then run Kupiec, Christoffersen and Basel traffic-light tests over the full sample. Answer which model fails first in stressed periods and whether expected shortfall ranks the models differently.
Stage 01 · Foundations
Statistics and testing
The standard VaR backtests are likelihood-ratio tests.
- Likelihood ratio tests
- Goodness of fit
- Confidence intervals
- Probability, Statistics and Financial Data
- Likelihood Ratio Tests, Wilks' Theorem, and Goodness of Fit
- Confidence Intervals and Interval Estimation
Stage 02 · Core
Volatility modelling
Forecasting tomorrow’s volatility from today’s data.
- EWMA
- GARCH
- Realised volatility
- Econometrics and Financial Time Series
- Volatility Estimation Theory
- Quantitative Risk Management
- Empirical Properties of Financial Data and GARCH Volatility Models
Stage 03 · Core
VaR, expected shortfall and tail risk
Measuring the tail, and checking the measure against what happened.
- VaR and expected shortfall
- Coherent risk measures
- Extreme value theory
- VaR backtesting
- Quantitative Risk Management
- Loss Distributions, VaR, and Coherent Risk Measures
- Extreme Value Theory
- Multivariate Models and Copulas
Stage 04 · Specialist
Stress testing and regulation
Scenarios beyond the data, and the rules banks must follow.
- Scenario design
- Basel market risk
- FRTB basics
- Quantitative Risk Management
- Risk in Perspective and the Regulatory Framework
- Market Risk
Stage 05 · Specialist
Reproducible risk code
A risk number someone else can rerun and audit.
- Testing
- Version control
- Reproducible pipelines
- Python for Quantitative Finance
- Testing, Debugging, and Style
- Professional Quant Practice
- Reproducible model development
Stage 06 · Proof of work
Build the projects
Each project is built on real market data and reports what the data said, including when the idea failed.
- Q5Drawdown and Tail-Loss MonitorThe normal model understates the 99% tail by 1.47×, and a VaR that passes the Kupiec test still clusters its breaches.
- VRVolatility forecasting and risk-limit calibrationHARQ on five-minute realised variance had the lowest forecast loss in the replay year; the report shows where it fails.
Practise the interview
Questions this role gets asked, each with a hint and a worked solution.
- Are daily returns normally distributed? What would you show someone who thinks so, and what does it change for VaR?
- Explain GARCH(1,1) in words, then write it down and say what each parameter does.
- You have 30 daily closes. How do you estimate an annualised volatility, and how much would you trust it?
Start the SelfMadeQuant route free: a free account opens one complete topic from each primer and the interview question bank. No card needed.
Start free