Market Risk Checklist

Pull current allocations from Black Diamond or Orion for each model — conservative through aggressive growth. List equity beta, duration, credit spread, FX, and concentration exposures by sleeve. Flag any model with single-position weight above the 5% concentration limit.

Pull the FOMC dot plot, Treasury curve shape, breakeven inflation, and high-yield OAS from YCharts or Bloomberg Terminal. Note any inversion, credit spread widening above 500 bps, or VIX print above 25 — each triggers a separate stress scenario downstream.

List non-traded REITs, interval funds, private credit, BDCs, and structured notes by client and dollar exposure. These hit liquidity coverage differently than the daily-liquid sleeve and need separate redemption-gate notes.

Cross-check against the firm's alternatives allocation cap (typically 15-20% of investable assets for accredited clients).

Walk through the identified exposures with the CIO. Capture any judgment calls — sector rotation views, duration tilts, currency hedging stance — that won't show up in the quantitative output but should drive the recommendation memo.

Run parametric and historical VaR at 95% and 99% confidence for each model portfolio. Use the rolling 3-year window; document the lookback period and confidence level on the output so the committee can compare quarter over quarter.

Apply the firm's four standard shocks: 2008 GFC, 2020 COVID drawdown, 1994 rate spike, and the in-house tail scenario (equity -30%, rates +200 bps, credit spread +400 bps). Tag any model where projected drawdown exceeds the stated max-loss tolerance.

Compare each model's VaR, beta, duration, and stress-loss against the limits in the IPS. Common gotcha: aggressive growth model has drifted above its 1.2 beta cap after a strong equity quarter and now needs trimming.

Annual methodology review — confirm correlation matrices, volatility inputs, and factor loadings are current. If the last recalibration was more than 12 months ago, schedule a full backtest before the next quarter's run.

Verify the rebalancer is set to alert on 5% absolute or 20% relative drift per asset class. Pull a sample of three accounts from each model and confirm the alert log shows expected triggers from the last 90 days.

Tie the Riskalyze / Nitrogen dashboard positions back to the Schwab or Fidelity statement totals. Breaks over $1,000 per account get a same-day investigation — usually a missed corporate action or a pending ACATS in-flight.

Walk advisors through any single-position watchlist names — concentrated low-basis stock, recently downgraded credit, or alts approaching a redemption window. Advisors flag any client conversations needed before the next quarterly review.

For each breaching model, propose a specific action: trim equity sleeve to bring beta inside the cap, shorten duration via the short-Treasury sleeve, or layer a put-spread overlay on concentrated positions. Document the Reg BI rationale for each recommendation.

Before any model-level trim hits taxable accounts, run the proposed sales through Holistiplan or the cost-basis report. Short-term gains and 30-day wash-sale windows are the two most common reasons a sound rebalance creates a client-service problem next April.

Present the memo to the IC. Vote is recorded in the minutes; any dissenting view captured by name. The CIO signs off before any cross-book trade is released to the trading desk.

One PDF: executive summary, VaR table, stress results, IPS limit dashboard, watchlist, and the IC-approved mitigation actions. Use the firm template — regulators expect consistent format quarter to quarter.

CCO reviews for any disclosure implications — material changes that may trigger a Form ADV Part 2A amendment or an updated client-facing risk disclosure. File the packet in the compliance books-and-records archive per Rule 204-2.

Drop the VaR run, stress output, IC minutes, and signed memo into NetDocuments under the quarter's folder. Five-year retention; the SEC examiner who asks for last cycle's workpapers will not accept a verbal answer.