Financial Risk Specialists

• Manual portfolio analysis + early sprea…1974–1988

  Manual portfolio analysis + early spreadsheets (pre-regulatory-capital era)

  Before 1988, financial risk management was an informal function. Banks measured credit risk through relationship lending and internal credit committees; market risk was managed by traders with discretion and intuition rather than systematic quantification. Risk analysts were generalist financial analysts who ran exposures on calculators and later on early spreadsheet programs like VisiCalc (1979) and Lotus 1-2-3 (1983). There was no standard vocabulary, no regulatory capital requirement, and no organizational norm for a risk function separate from trading or lending. The Herstatt failure in 1974 created the Basel Committee, but a decade would pass before the committee produced its first capital standard.

  Effect on the work

  The era preceded the profession. Financial risk specialists as a distinct occupational category did not exist; risk analysis was performed ad hoc by financial analysts embedded in trading and lending functions.

• Basel I capital accord + basic credit e…1988–1994

  Basel I capital accord + basic credit exposure spreadsheets

  The Basel I Accord, published in July 1988 and implemented by G10 nations from 1992, required internationally active banks to hold minimum capital equal to 8% of risk-weighted assets. For the first time, credit risk had a regulatory number attached to it, and banks needed analysts who could calculate and monitor that number across their loan and off-balance-sheet portfolios. The risk-weighting methodology was simple by later standards (sovereigns at 0%, residential mortgages at 50%, corporate loans at 100%), but implementing and monitoring it required dedicated analytical capacity that had not previously existed. Merrill Lynch, responding to the new environment, created the first dedicated risk management department at a US financial institution in 1987. Spreadsheet-based capital calculation became the core tool of early risk analysts.

  Effect on the work

  Basel I was the hiring signal that began professionalizing the field. Banks assembled small teams to compute and report regulatory capital ratios; the Basel I framework created formal demand for dedicated risk calculation capacity beyond what front-office trading desks had previously performed informally.

• J.P. Morgan RiskMetrics + Value-at-Risk1994–2004

  J.P. Morgan RiskMetrics + Value-at-Risk (VaR) as industry standard

  In October 1994, J.P. Morgan published the RiskMetrics technical document and made it freely available over the internet, along with a daily covariance matrix for several hundred key financial factors. CEO Dennis Weatherstone had famously called for a "4:15 report" that summarized the entire firm's risk in one number, available within 15 minutes of market close. RiskMetrics gave every bank in the world the methodology and the data to build the same thing. Value-at-Risk became the universal language of market risk measurement. In 1996, the Basel Committee amended its framework to allow banks to use internal VaR models to calculate market risk capital, directly embedding the J.P. Morgan methodology into global banking regulation. This created an immediate need for quantitative risk analysts who could build, validate, and interpret VaR models. The 1997 founding of GARP and the FRM certification formalized the profession's credentialing structure.

  Effect on the work

  RiskMetrics democratized quantitative risk management beyond a small group of Wall Street quants. The 1996 Basel market risk amendment made VaR a regulatory tool at every internationally active bank, creating structured demand for specialists who could implement and interpret the methodology. GARP grew from 250 founding members in 1996 to a global professional body within years.

• Credit derivatives + Basel II internal…1998–2008

  Credit derivatives + Basel II internal ratings-based approach

  The collapse of Long-Term Capital Management in 1998 exposed the limits of VaR as a model: LTCM held positions sized on the basis of historical correlations that broke down catastrophically when Russia defaulted and global investors fled risky assets simultaneously. The lesson was not that quantitative risk management was wrong but that it required specialists who could challenge model assumptions, stress-test correlation estimates, and reason about scenarios beyond the historical window. Basel II, published in 2004, added operational risk as a separately capitalized category and introduced the Internal Ratings-Based (IRB) approach for credit risk, requiring banks to develop their own probability-of-default and loss-given-default models, subject to supervisory approval. The credit derivatives market (CDOs, CDS) created by J.P. Morgan in the mid-1990s reached several trillion dollars in notional outstanding by 2006, concentrating enormous exposure in instruments that required dedicated counterparty risk specialists to understand.

  Effect on the work

  LTCM accelerated the shift from naive VaR reliance to model validation as a separate function. Basel II created formal demand for credit model specialists. The credit derivatives market created an entire sub-specialty in counterparty risk and XVA. These pressures multiplied the specialist headcount substantially through the early 2000s.

• Basel III + CCAR/DFAST stress testing +…2009–2019

  Basel III + CCAR/DFAST stress testing + SR 11-7 model risk governance

  The 2008 financial crisis triggered the most significant regulatory expansion in the profession's history. Basel III (published 2010, phased in from 2012) substantially increased capital requirements and added new liquidity ratios (LCR, NSFR). The Federal Reserve launched its Comprehensive Capital Analysis and Review (CCAR) stress-test program in 2009, requiring the largest US banks to demonstrate capital adequacy under severely adverse macroeconomic scenarios: essentially a formal demand for large, sophisticated stress-testing teams at every major bank. The Fed's SR 11-7 guidance on model risk management, issued in 2011, created independent model validation as a required function at supervised institutions: every model used for risk measurement or capital calculation had to be reviewed by a validator who was genuinely independent of the model developer. This single regulatory document created thousands of model risk analyst positions across US banking. NBER research confirms that risk management positions tripled as a share of total bank job postings between 2010 and the end of the decade at stress-tested institutions.

  Effect on the work

  The 2009-2019 decade was the fastest period of growth in the profession's history. CCAR, DFAST, Basel III, Dodd-Frank Title I and II implementation, and SR 11-7 collectively created a sustained regulatory-driven hiring wave for market risk specialists, credit risk analysts, stress-testing modelers, and model validators. Risk management roles grew from roughly 4% to over 12% of total job postings at stress-tested US banks.

• Climate risk integration + AI/ML model…2019–2024

  Climate risk integration + AI/ML model governance (NGFS, TCFD, OCC AI guidance)

  Two new regulatory frontiers reshaped the profession in the early 2020s, both arriving nearly simultaneously. Climate risk became a regulatory requirement: the Network for Greening the Financial System (NGFS) published its first climate scenarios in 2019; the ECB conducted its first supervisory climate risk assessment in 2022; the Federal Reserve began climate scenario analysis pilots with the six largest US banks in 2023. Physical climate risk analysis using tools like MSCI Climate Lab and Moody's Climate on Demand became a new specialist function, requiring knowledge of GIS-based collateral mapping, TCFD disclosure frameworks, and scenario modeling under NGFS pathways where no historical training data exists. Simultaneously, banks' large-scale deployment of ML models in credit scoring, fraud detection, and algorithmic trading triggered new model risk governance obligations: the OCC's 2025 AI guidance extended SR 11-7 principles to ML systems, requiring human specialists who understood both the regulatory framework and the internals of ML models well enough to validate their fairness, explainability, and stability.

  Effect on the work

  Climate risk and AI model governance created two entirely new sub-specialties within the occupation, adding headcount at a time when more traditional market-risk-computation functions were being partially automated by Bloomberg PORT, MSCI RiskMetrics, and similar platforms. The net employment effect was growth, because the new governance work requires human judgment that the platforms themselves cannot provide.

• AI-augmented risk analytics2024–present

  AI-augmented risk analytics: Bloomberg PORT AI, MSCI Climate Lab, Quantexa, Claude/ChatGPT

  By 2024, AI tools had automated the mechanical computation layer of financial risk management: daily VaR packs, P&L attribution reports, limit-monitoring dashboards, and first-pass counterparty exposure summaries that once required several analyst-hours are now machine-generated by Bloomberg PORT, MSCI RiskMetrics, and Wolters Kluwer OneSumX. LLMs (Claude, ChatGPT) are used to draft model validation reports, parse FRTB rule text, and generate stress-test narrative commentary. The effect is augmentation rather than displacement for senior specialists: the governance layer, which includes regulatory defense of stress scenarios, independent model validation sign-off, and climate risk judgment under conditions where no historical training data exists, requires human accountability that no AI system can supply. GARP's 2025 practitioner survey found 73% of risk professionals believe AI has expanded their capacity for complex analysis, not replaced their judgment. BLS projects the occupation will grow 6.5% by 2034, well above the all-occupations average.

  Effect on the work

  AI tools automate the computation and first-draft layer, freeing specialists for higher-judgment work and enabling smaller teams to cover larger exposure universes. The regulatory moat is structurally intact: SR 11-7 explicitly prohibits a model from validating itself, and no AI system can own the personal accountability that examiners require from model risk managers.

What's changing in your day

Three parts of your work where AI is already doing real lifting, and what stays yours.