Risk analysis (lens)

Definition

The risk lens evaluates what can go wrong and how much: downside exposure, volatility, correlations between positions, concentration, and behavior under stress scenarios. Every other lens asks whether a thesis is right; this lens asks what happens to the account when it is wrong, and whether the position size makes survival independent of any single outcome.

How it works / structure

• Inputs: position sizes and exposures, return/volatility history, correlation estimates, leverage and margin terms, and scenario definitions.
• Core operations: volatility measurement (realized, and conditional — volatility clusters, per Engle 1982), exposure aggregation across positions (risk-correlation-exposure, port-exposure-netting), drawdown accounting (risk-max-drawdown-budget), sizing rules (risk-fixed-fractional, risk-kelly-criterion, risk-volatility-targeting), and scenario analysis (risk-scenario-analysis).
• Output shape: bounded quantities — “this position risks 1% of equity at its stop”, “portfolio net delta-adjusted exposure is 120%” — that the platform’s grading and the simulation engine can check mechanically.

When it applies

To every position and every portfolio, at entry (sizing), during the hold (exposure drift, correlation change), and in review (was the realized drawdown within budget). It is instrument-agnostic; leverage instruments (margin equity, futures, short options) raise the stakes because losses can exceed the amount committed (per the SEC margin publication and the OCC disclosure document).

Risk profile & failure modes

• Estimation fragility: volatilities and correlations are estimates; correlations rise toward 1 in equity-market stress (Longin and Solnik 2001), exactly when diversification is needed.
• Tail blindness: normal-distribution assumptions understate extreme moves; scenario analysis exists because variance is not the whole story.
• Risk-measure gaming: any single number (VaR, vol target) can be satisfied while concentrating tail risk.
• Leverage nonlinearity: margin calls and forced liquidation turn paper drawdowns into realized ones at the worst time.

Evidence & limits

Markowitz (1952) formalized the core result that portfolio risk depends on covariance, not just individual volatilities — the foundation of diversification math. Engle (1982) established volatility clustering (calm and stressed periods persist), which makes volatility partly forecastable and underwrites volatility targeting. Longin and Solnik (2001) documented that equity correlations increase in bear markets — the standard citation for why static diversification overstates crisis protection. What the lens cannot do: no published method reliably predicts the timing of tail events; risk analysis bounds consequences, it does not forecast triggers.

Falsifiable-thesis examples

Illustrations only, not signals:

• “Portfolio P, run at a 10% annualized volatility target, will realize 21-day volatility between 7% and 13% annualized for at least 80% of the next 6 months” — falsified by the realized series.
• “The maximum peak-to-trough drawdown of strategy S over the next quarter will not exceed 8%” — falsified by any deeper drawdown.

Cross-references

• Pillar 7: risk-fixed-fractional, risk-kelly-criterion, risk-volatility-targeting, risk-max-drawdown-budget, risk-correlation-exposure, risk-scenario-analysis
• Regime context: regime-volatility
• Adjacent lenses: lens-portfolio (aggregation), lens-quantitative (estimation), lens-options (priced risk)