Risk budget and sleeves — a working framework

This is a Tier 1 portfolio page: citeable on the open web, with mechanics at the framework level—not live positions.

Problem statement

Most portfolio mistakes I care about are process mistakes: wrong horizon mixed into the same sleeve, rebalancing because the calendar said so, or letting a narrative override a pre-committed risk budget. The framework below is how I keep those failures visible.

Sleeves (three, not ten)

I run three sleeves with explicit mandates:

SleeveHorizonMandate
LiquidityDays to weeksTrading and tactical risk; must stay small vs total budget
CoreQuartersStructural exposures I am willing to defend in writing
ReserveYearsDry powder and convexity; not a performance sleeve

Each sleeve has a maximum risk contribution to total portfolio volatility, not a target return. Returns are an output; the input is how much pain I am willing to tolerate in that sleeve under a defined stress scenario.

Risk budget (one number, few knobs)

The budget rolls up to:

  1. Gross exposure cap (hard ceiling)
  2. Max drawdown tolerance per sleeve (soft ceiling with forced de-risk rules)
  3. Correlation awareness — if two positions answer the same macro question, they share a sleeve budget even if they look like different tickers

Rebalancing is event-driven first (rule breach, thesis change, liquidity shock) and calendar-driven second (quarterly hygiene only).

What this page deliberately omits

How trading connects

Weekly trading reviews (Tier 2 on-site) summarize what changed in the liquidity sleeve without leaking alpha. This framework is the map those reviews point back to.

Edits and provenance

Material changes get a lastmod date and, for flagship pages, provenance: true so readers see when the framework last moved.