Skip to content
FSS Knowledge Base

Pre-Decline Deleveraging

Section titled “Pre-Decline Deleveraging”

Objective — reduce existing leverage while markets are high, for an investor who has already met their goals and for whom the remaining upside is worth less than the drawdown would cost.

Risk level — 0 — Zero risk. Risk justification — level 0 on the zero-risk test: reducing leverage creates no new downside exposure — it removes exposure — while the cost is state-dependent: if no decline arrives, the foregone levered upside is real. The strategy’s own return (interest no longer paid, risk no longer carried) is certain; its relative return is not. It is not level -1 for exactly that reason, and it cannot be level 2 or above, which require exposure to increase.

One honest qualifier that can move a specific case. Where unwinding requires selling appreciated non-registered holdings, the capital-gains toll is a certain cost, not an opportunity cost, and that case should be assessed on its own facts rather than inherited from this record. The decline itself would shrink that toll — but pre-decline, by definition, it has not yet.

Benefit justification — unverified — needs Talbot. Jurisdiction — Canada for the tax treatment of unwinding. U.S. applicability unverified.

Prerequisites

Section titled “Prerequisites”
  • Existing leverage.
  • A wealth-sufficiency test that has already been passed — this record is written for the investor who has won, not for one hoping to time a top.

Mechanism

Section titled “Mechanism”

Three benefits, in the source’s own order of importance:

  1. Preserved optionality. Capacity that is not in use is capacity available later — and the source’s whole preparation philosophy is about holding free options rather than committed positions.
  2. Deteriorating forward reward-to-risk at extreme valuations. Shiller CAPE stood near 40.6 in September 2026 — the 98.8th percentile since 1881. A 20% drop lands around 32, still above the 1929 peak.
  3. Emotional risk reduction — and the source calls this the largest.

Peace of mind is a legitimate optimization target. For the financially independent, additional wealth has low marginal utility while drawdown pain is fully felt.

That is the whole argument, and it is a utility argument rather than a forecast. The right answer for those who have already won.

Benefits

Section titled “Benefits”
  • Removes carry cost and the obligation to service it through a period when income may also be under pressure.
  • Reduces the magnitude of the drawdown that must be lived through — which, under the source’s behavioural rule, raises the probability the investor holds the rest of the plan.
  • Restores the capacity that standby-credit-capacity and post-decline-deployment both depend on.

Risks

Section titled “Risks”
  • Foregone upside if no decline arrives. The source’s own frequency work puts the probability of no ≥20% decline in ten years at roughly 21% — not negligible.
  • The capital-gains toll, where unwinding requires selling appreciated holdings. See the qualifier under Risk level.
  • Deleveraging on a valuation view rather than a sufficiency view is market timing, and inherits every objection to market timing.

Failure modes

Section titled “Failure modes”
  • Doing it on a forecast. The defensible version rests on “I no longer need this risk”; the indefensible version rests on “I think markets will fall”.
  • Ratcheting risk tolerance down at the bottom instead. That is on the source’s Harmful list — it permanently impairs future returns — and it is what happens to an investor who does not deleverage in advance and then capitulates.
  • Deleveraging and then re-levering at a worse point, which converts a risk decision into two timing decisions.

Costs

Section titled “Costs”

Any prepayment penalty on the facility, and the capital-gains tax on holdings sold to repay. Both unverified — needs Talbot in magnitude.

Tax considerations

Section titled “Tax considerations”

Unwinding may realise capital gains. Interest deductibility ceases on the repaid portion; where partial, tracing matters — see interest-tracing-hygiene and disappearing-source-continuation.

Who it may suit

Section titled “Who it may suit”

The financially independent investor who is “not optimizing for wealth; optimizing for peace of mind” — the profile the source records explicitly, and the segment a terminal-wealth ranking systematically misadvises.

Who should avoid it

Section titled “Who should avoid it”
  • Anyone still accumulating toward a goal that the leverage is there to reach.
  • Anyone doing it because of a market view rather than a sufficiency test.
  • Anyone who would immediately redeploy the freed capacity into equivalent risk elsewhere.

Implementation outline

Section titled “Implementation outline”
  1. Run the sufficiency test first: is the remaining upside needed?
  2. Quantify the drawdown that current leverage implies at −30% and −50%, in dollars, not percentages.
  3. Identify which holdings can be sold with the least tax cost, and whether any are already at a loss.
  4. Repay, and document the tracing if the repayment is partial.
  5. Record what the freed capacity is now reserved for, so it is not spent by default.

Evidence status

Section titled “Evidence status”

documented — stated in market-drop-wins-library-v2 (T3-5), with the valuation context and the peace-of-mind argument from market-drop-wins-master-log-v2 §3.2 and §2.2. Not modelled in sd-math; no dollar quantification of the leverage trade-off exists.

Valuation evidence added 2026-09-24 (from cape-timing; code ~/projects/cape-timing; report Market Timing at High Market Valuations). These are US total-return figures vs T-bills, 1926–2016 starts, with real-time CAPE percentiles:

  • From the most expensive 5% of starts, stocks trailed cash over 10 years 55% of the time, vs 14% overall. The chance of a positive real decade fell from 88% to 52%, and the median Sortino ratio from 0.76 to 0.13, with volatility unchanged. That bucket is about one independent decade of history (1929, 1998–2000).
  • A textbook (Merton) allocation at moderate risk aversion falls from about 62% stocks on history to 12–33% at today’s valuation. That supports reducing exposure for an investor who no longer needs the upside. The same logic applies more strongly to levered exposure.
  • Valuation alone is a poor timing trigger: extremes lasted years (US 1996–2000; Japan +225% after its dividend yield fell below 1% in 1985). A mechanical trend confirmation works better; see valuation-trend-rule.

Counterarguments

Section titled “Counterarguments”
  • “This is market timing with a nicer name.” The objection lands on the valuation version and misses the sufficiency version. The 2026-09-24 evidence sharpens it: deleveraging on valuation alone has a poor record (it would have exited years early in 1996 and 1985 Japan), while a pre-committed valuation-plus-trend trigger (valuation-trend-rule) held up across 100 years of US data. The source’s sizing principle is “size by loss tolerance, not by estimated edge” — a rule that needs no forecast. A deleveraging decision made on loss tolerance is not a forecast; one made on CAPE is.
  • “Expected terminal wealth falls.” Conceded, and it is the point. The source’s explicit position is that ranking by expected value “systematically misadvises the segment most worth serving.”

Variants

Section titled “Variants”
  • The source’s working label was pre-decline leverage reduction.
  • A partial version — reducing to a target ratio rather than to zero — is the more common case and is governed by conservative-leverage-ratio.
  • A rule-triggered version: deleverage when valuations are extreme and the price trend breaks, and re-lever when the trend recovers. This is the leverage-side application of valuation-trend-rule.
Section titled “Related strategies”

conservative-leverage-ratio · lifecycle-leverage-glide-path · standby-credit-capacity · non-callable-first · post-decline-deployment · disappearing-source-continuation

Sources

Section titled “Sources”

Open questions

Section titled “Open questions”
  • What the sufficiency test actually is, stated as a rule rather than a judgement. Without it this record rests on the reader’s self-assessment.
  • Whether a CAPE-conditioned version can be defended, given the source’s own bound that valuation predicts 10-year returns rather than 12-month. Partly answered 2026-09-24: a CAPE-only trigger is weak, but CAPE combined with a 10-month trend break is defensible as insurance (valuation-trend-rule). What remains open is the leverage-specific version: its borrowing-cost and re-levering rules.