Optimize Your Downturn Defenses

Sequence-of-returns risk is the one that keeps people up at night: a crash in your first few retirement years, while you’re selling to fund spending, does damage a late-career crash never would. The defensive levers against it — a cash reserve, downturn-aware withdrawals — are easy to turn on and hard to size. Two months of cash or eighteen? Trigger the defense on a 5% drop or a 25% one?

That’s what the Downturn Defense group in the Optimizer is for. It sweeps one defense parameter across its range and prices every level, so you can see what each one buys you and what it costs.

New to the Optimizer? Start with how an Optimizer run works. This page assumes you know how to pick a dimension and read a curve.

The one thing that will trip you up

Every dimension here defaults to a downside objective, and you should leave it that way.

Downside protection is insurance. It costs a little in an average market and pays you back in a bad one — that’s the deal. So if you point one of these dimensions at median or mean legacy, the optimizer will faithfully report that the best cash reserve is zero and the best trigger threshold is never. It isn’t wrong; the median simply never sees the tail the defense exists for. It’s the right answer to the wrong question — and it’s exactly how a careful person talks themselves into cancelling a policy that was working.

Each dimension below ships with the objective that measures the protection that parameter is designed to buy. Change the range if you like. Think twice before changing the objective.

Cash Reserve (months)

Searches: 0 to 48 months of portfolio withdrawals, in steps of 6.
Default objective: 10th-percentile legacy.

How large a cash cushion to hold, measured in months of your portfolio withdrawals — the amount actually leaving your accounts each month after Social Security, pension, annuity, rental, and policy-loan income are netted out, or your monthly RMD equivalent if that’s larger. If guaranteed income covers a good share of your spending, a given number of months buys a smaller dollar cushion than you might assume. A bigger reserve absorbs a longer downturn without forcing equity sales; it also drags on the median, because cash returns less than stocks over a long retirement. That drag is the premium.

Because the reserve is sized on portfolio withdrawals rather than gross spending, one month is fewer dollars for a household with large guaranteed income. So the recommended month count can come out higher than you’d expect — and higher than it would have before the 2026-09-04 re-basing — for the same or even a smaller dollar reserve.

If the answer is the top of your range, it isn’t an optimum — the objective was still improving at the highest value searched. Widen the range (for example 0–96 months) to find where it levels off.

The objective is P10 legacy because lifting your worst-case floor is the whole job of a reserve. Optimizing it for median legacy recommends zero months, every time.

Where that cash sits is a separate decision, and the optimizer doesn’t make it for you — the reserve is whichever of your accounts hold a Cash asset class, so you place it by choosing the account. See where the cash reserve actually lives. If you’re past RMD age, read what the reserve can’t protect before sizing it: required distributions can force a sale the cushion won’t stop, which makes a large reserve buy less late in the plan than the curve implies.

If your strategy has no cash reserve configured, enabling this dimension creates one — asking "how much reserve should I hold?" is a perfectly good reason to start from nothing.

Downturn Trigger Threshold

Searches: −30% to −5%, in steps of 5%.
Default objective: Success rate.

The trailing-12-month stock return at which the simulation flips into downturn mode and starts draining cash first instead of selling equities. A less negative threshold (−5%) triggers often — more protection, and more of your reserve spent on drops that would have recovered anyway. A more negative one (−25%) holds fire until something serious happens, keeping the cushion intact for a real crash but riding out the smaller ones unprotected.

The objective is success rate because the trigger’s job is to keep plans from finishing at zero.

If your strategy has no downturn-aware rule, enabling this dimension writes one. Pair it with a cash reserve — downturn mode reroutes withdrawals to cash, so without a cushion it fires and has nowhere to go.

Reserve Refill Threshold

Searches: 5% to 25%, in steps of 5%.
Default objective: 10th-percentile legacy.

Once you’ve spent the reserve in a bad stretch, when should it be rebuilt? This is the trailing-12-month stock return above which the model tops the cushion back up from non-cash accounts. A low threshold refills eagerly, so you’re rarely caught without dry powder; a high one waits for a clear recovery, leaving more invested in the meantime. Refills stay within a tax type — stocks to cash inside Taxable, or inside Traditional, or inside Roth, never across the boundary, since that would be a real distribution or rollover.

This dimension is disabled unless a cash reserve exists: there’s nothing to refill without a floor. Set one on the Strategy tab, or enable the Cash Reserve dimension above, and it unlocks.

Guardrails are a downturn defense too

They just live in a different group. Upper Guardrail % and Lower Guardrail % sit under Withdrawal Strategy, and they defend against sequence risk from the spending side: when the portfolio dips, guardrails automatically trim the withdrawal instead of forcing you to sell more shares at a bad price. Flexibility is one of the strongest sequence-risk defenses there is.

Search them alongside the three dimensions here when you’re sizing a defense. A modest cash reserve plus responsive guardrails often beats a large cash pile on its own — and the optimizer can only tell you that if you give it both.

Read it as a frontier, not a winner

Every optimizer curve is already a frontier: each parameter level with its cost and its benefit at the objective you chose. That framing matters more here than anywhere else, because the "best" answer depends entirely on how much median growth you’re willing to trade for how much floor.

The most useful thing you can do is run a dimension twice — once on 10th-percentile legacy, once on median legacy — and look at the two curves together. Where they diverge is your premium, priced in dollars, level by level. Pick the point on that trade you’re comfortable with. That’s a decision, not an optimization.

Then confirm your finalists on the Compare tab, which runs them against identical market sequences, and read the Downside Protection card on the winner.

The Optimizer searches the assumptions you provide; its rankings are modeled estimates, not recommendations. Downside metrics in particular are sensitive to your assumed return distribution — treat the shape of the curve as more reliable than any single number on it.