The Client Who Wants Out at the Bottom
Your client is 64 and retires in 12 months. His portfolio peaked at $3.40M in January and is now $2.75M. He will need $140,000 from the portfolio next year, indexed thereafter, plus $38,000 of Social Security. Inflation is 2.6%. In three escalating emails he has asked to move 'everything' into Treasury bills yielding 4.3%, saying he 'will not sit through another 20% drop.' You have replied twice with long-run return charts. Yesterday he wrote: 'If we stay put and it falls again, I'll hold you responsible.' The account is a taxable joint account; there is no liquidity event pending and no debt.
1. Diagnose what is driving the request, and what the adviser has done wrong.
Read the worked answer
Biases in play.
- Loss aversion (λ ≈ 2–2.25): the 19.1% fall requires a 23.6% gain merely to return to the emotional starting point - the asymmetry is why 'stay the course' feels like doubling down.
- Anchoring and recency: the $3.40M January peak, not $2.75M, is the reference point, and only the recent data is available to mind.
- Myopic loss aversion: a volatile portfolio evaluated daily or annually generates losses far more often than gains, so the pain is structurally over-weighted.
- Regret aversion and action bias: cash converts an uncontrollable feeling into a controllable act.
The adviser's error. Two long emails of long-run return charts answered the question 'will I be okay?' with the question 'what are average returns?' - a System-2 argument aimed at a System-1 reaction. Repetition also reads as dismissal, which intensifies the client's sense that the risk is being ignored. The unasked question was the only one that mattered: can this portfolio support $140,000 a year?
2. Do the numbers support the client's fear? Test the plan on two different lenses.
Read the worked answer
Withdrawal rate: 140,000 / 2,750,000 = 5.09%, indexed at 2.6%.
Lens A - perpetual real capital. Required nominal return = 5.09% + 2.6% = 7.7%, before fees. A 55/45 has perhaps 6.3%. A shortfall of ~1.4%/yr means eventual depletion. On this lens the client is right to be frightened.
Lens B - 30-year depletion to age 94. Real expected return = 1.063/1.026 − 1 = 3.61%. The 30-year real annuity factor is 0.0361 / (1 − 1.0361⁻³⁰) = 5.51%, which supports $151,600. The $140,000 need at 5.09% fits, with a cushion of about $11,600 a year. On this lens the plan works.
The judgment: the two lenses differ only by bequest intent, which was never established. The client's fear is not irrational - it is a rational response to a 5.09% draw rate with a thin cushion, and the real planning failure predates the drawdown. Cash at 4.3% pays $118,250 - $21,750 short, funded from principal, and only 1.7% real. It converts an uncertain problem into a certain one.
3. What do you actually recommend, and how do you handle 'I'll hold you responsible'?
Read the worked answer
Neither fully accommodate nor simply refuse.
- Carve out a three-year spending reserve: $420,000 (15.3%) in T-bills and a 1–3 year ladder, plus the next 12 months of spending. This is not a return decision - it buys 36 months in which no equity must be sold into a decline. It addresses the real fear (sequence risk), not the stated one (volatility).
- Keep the remaining $2.33M at 55/45, with a stated 6.3% expected nominal return.
- Fix the underlying problem: the plan supports ~5.5% real over 30 years but not in perpetuity. Present both, obtain a decision on horizon and bequest, and if the perpetual lens is chosen, reset spending to ~$110,000 or add income.
- Reframe in the client's language: 'We are not staying the course. We are buying three years.' Frames beat data.
- Handle the blame statement by converting it into a joint written decision memo: the two options, the arithmetic, what each costs, the client's choice, and a pre-commitment on what would change your advice (e.g. a further 25% fall with unchanged fundamentals triggers a rebalancing purchase, agreed now). Blame attaches to decisions never made together; the memo is the antidote. Set a quarterly review cadence, and say so.