How Much Do You Need to Retire? 15× Annual Spending, Not 50×

In 30 seconds: A student assumed he needed 50× annual spending (2% withdrawal) — i.e. NT$48M for NT$96k/month — to retire. James corrected him: you don’t. You can consider retiring at 15× annual spending; liquid stock beyond “15 years of annual spending” doesn’t matter to your life — the excess is just inheritance. At NT$96k/month, NT$14.4M (= 96k × 12 × 15) is enough.¹

📌 This is an important update to the four retirement withdrawal tiers: those “50× / 33×” tiers think in pure withdrawal-rate terms; this page (newer, 2026-06) uses QQQI high-dividend cash flow to slash the required retirement assets. Per “newer content wins,” treat this page as the target number.

Why not 50× / NT$48M

CLEC’s two thresholds: you can consider retiring at 15× annual spending; you’d only need 50× if you withdraw 2% without relying on dividend cash flow. But once you bring in a high-dividend cash-flow instrument like QQQI, you don’t need to pile up to 50× — liquid stock beyond 15 years of spending neither affects your life nor is necessary.¹

2.5% is the dividing line: do you need high-dividend?

00542 gives a sharper test: a 2.5% withdrawal rate (≈ 40× annual spending) is the dividing line for “do you need to add high-dividend.”²

Withdrawal ≤ 2.5% (assets ≥ 40×): pure 4-3-3 / 70-30 / 80-20 withdrawal is enough — no need to touch high-dividend. “Within a 2.5% withdrawal rate, the asset multiple alone works.”
Withdrawal > 2.5% (under 40×): a pure index allocation can’t throw off enough cash flow, so you partly substitute in QQQI high-dividend to prop up the cash flow. That’s high-dividend’s real purpose — only for retirees with a higher withdrawal rate, not for the young (see 年輕人適合買高股息嗎).

🧮 The blend method (how to size the ratio): split assets into portfolios A and B and use their “returns” to hit your target withdrawal rate. A = pure index (QQQ / 4-3-3), yielding about 2.5%; B = QQQI, conservatively 11%. To draw 4% or 5%, solve “A×2.5% + B×11% = target.” E.g. to draw 2.5% you go 100% A (no B needed); the higher the withdrawal rate, the larger B’s (high-dividend) share.²

The “3% rule” as a dynamic formula: why the money never runs out

The withdrawal rates above are “a percentage of a fixed principal”; 00335 goes further with the 3% rule as a dynamic formula — spending floats with your assets each month/year, which is the real key to “never running out”:³

Monthly spending = last month’s (or month-start) assets × 3% ÷ 12. Good market, more assets → next month you can spend a bit more; bad market, shrunken assets → spending auto-shrinks. To avoid violent month-to-month swings in living costs, in practice you reset once a year: recompute this year’s monthly allowance from the assets at the start of the year, and use it through year-end.
As long as you “never exceed 3%,” it’s very safe — because you spend a percentage of current assets, not a fixed amount, so you automatically spend less when assets fall, and the principal is never drained. James: “Don’t spend more than 3%; keep to 3% and it’s actually very safe.”
Borrowing beats selling: to fund that same 3%, pledge lending (質押借款) is safer than selling stock — the principal is never touched and keeps compounding; you’re just borrowing to spend (see 欠錢不還與資產傳承). Over the long run, “borrow to spend” preserves assets better than “sell to spend.”

This is the same thing as the withdrawal-rate tiers above, told two ways: the tiers ask statically how many times your expenses you hold; the 3% rule dynamically lets spending track your assets. Both point to the same safety line — don’t let your withdrawal rate stay above 3% for long.

How to allocate NT$14.4M (NT$96k/month example)

Split the 15 years into “10 years of high-dividend cash flow + 5 years of 70/30 growth and cash”:¹

Sleeve	Amount	Purpose
QQQI (high-dividend, via sub-brokerage)	96k × 12 × 10 = NT$9.6M	monthly dividend cash flow
00662 (QQQ)	96k × 12 × 5 × 70% = NT$3.36M	asset growth
00865B (short bonds/cash)	96k × 12 × 5 × 30% = NT$1.44M	safety cushion
Total	NT$14.4M	money for life; the excess is inheritance

Margin-of-safety update: 15× → 17×, QQQI 10 years → 12 years (00575)

In 2026-07 James made a margin-of-safety adjustment to the “10 years QQQI + 5 years 70/30 = 15×” example above. The reason: a high-dividend instrument like QQQI pays a distribution that floats with the market — if stocks fall 50%, QQQI’s dividend shrinks by about 30%. Pinning it to a bare 15× with 10 years of QQQI means that in a downturn, when the dividend shrinks, the cash flow falls short.⁴

Use 17×, with 12 years in QQQI (up from 15×, 10 years). The extra 2 years of QQQI is the margin of safety: even if the dividend shrinks 30%, you still receive the dividend you originally needed.
When the market doesn’t fall: the dividend exceeds what you need, so put the surplus dividend into 00662 (growth).
15× is still retirable — just at 15×, bearing the risk of “dividend shrinks in a downturn, spend a bit less” (a NT$50k dividend might drop to NT$40k). “If you barely have 15×, then 15× it is, you just have to take on some risk.”⁴

This is a conservative-izing of the same cash-flow logic: it doesn’t change the “live on high-dividend monthly distributions” structure, only widens the buffer from 10 to 12 years of annual spending, to absorb the “downturn dividend shrink” risk that’s specific to QQQI (mechanism in QQQI).

QQQI is a “transition product”: 15 → 20 → 30×, then sell it off (a senior member)

In the same session a senior member added a useful framing: QQQI / 3451 / GEPQ are essentially “transition products” — a 5–10-year bridge — carrying you from “retirement funds just short” to “assets grown large enough not to need them,” and eventually sold off.⁵

Start (retire at 15×): 10× in QQQI for living, 5× in 4-3-3 (see 4-3-3).
After 5 years: if that 5× of 4-3-3 doubles with luck → becomes 10×, total ~20×.
After another 5 years (year 10): that 10× of 4-3-3 doubles again, total ~30×.
At 30×: you can dispose of the transition QQQI, switch back to 4-2-4 / 80-20 / 70-30, and live on pledge withdrawals of 3–4% (see 質押提領比例能不能隨資產提高), with stable and still-growing cash flow.
Corollary: since products like QQQI get sold off in a decade or so, estate tax isn’t their main concern — whether you buy QQQI or GEPQ, it’s fine (inheritance mechanism in 欠錢不還與資產傳承).⁵

Real resilience is “15 years of cash flow,” not “15 years of cash”

With no work income in retirement, compute from a market high and ensure your allocation’s cash flow lasts at least 15 years. Note: this isn’t holding “15 years of cash” — it’s having the cash flow to support 15 years.⁶

80/20 @ 2%, or 70/30 @ 3%: the paper cash position may only cover ~10 years of spending.
A high-dividend (QQQI) allocation: the 15-year cash flow comes from ongoing dividends — a completely different concept from “holding 15 years of cash.”

James’s extreme stress test — all at once: your house is destroyed by disaster so you must rent elsewhere, the mortgage still runs, the market crashes near zero, you lose your job, you fall seriously ill, and even your pledge line gets cut — do you still have enough cash flow to live? That’s what allocation is for.⁶

Regional variants (by high-dividend yield)

The required “multiple” depends on the high-dividend yield you can get. Example: China, ¥100k/year spending, 6% high dividend:¹

High-dividend sleeve: 10 ÷ 6% ≈ ¥1.667M (produces cash flow)
Growth/cash sleeve: 10 × 5 = ¥500k (70/30 or 80/20)
Required assets ≈ ¥2.167M ≈ 21.67× annual spending

How do you count your retirement assets? Taxable accounts count only “after-tax” (~70%)

An often-overlooked detail that changes your retirement number (00553): when calculating “do I have enough to retire,” assets in a taxable brokerage account count only at their after-tax value — because to use them later, you’ll owe capital-gains tax.⁷

Rough rule: use 70%: taxable-account assets count toward your retirement total at roughly 70% (the real ratio depends on your cost basis and local tax rate).
Example: a taxable retirement account of NT$3M counts as only NT$2.1M (3M × 70%); if you have another NT$2M elsewhere, your retirement total is NT$4.1M, and retirement spending should be NT$4.1M × 2% — not 2% of NT$5M. Overstating your usable assets is the most common retirement-math error.
The one exception: pledge, and never sell. If you plan to live on stock-pledge borrowing in retirement and never sell a single share of principal (buy-borrow-die), then those assets don’t trigger capital-gains tax and count in full — which is exactly why CLEC prefers “borrow to spend” over “sell to spend” (see the 3% rule: borrowing beats selling).⁷

🔗 This complements account type: Roth withdrawals are tax-free and count in full; Pre-Tax / taxable accounts get discounted. What you can truly use in full and tax-free is Roth + the pledged-never-sold portion.

⚠️ Workers and retirees differ: a worker’s cash position is an emergency fund (living costs from job loss until the next job); the 15-year cash-flow framework is for retirees. Numbers move with withdrawal rate, dividend yield, and tax rules — education only, not personalized retirement advice. Use with the three lines of defense.

CLEC channel, post “Why you don’t need NT$48M — NT$14.4M is the golden key to retirement,” 2026-06-22. Source: raw/docs/X及YouTube的貼文/0015…. ↩ ↩² ↩³ ↩⁴
CLEC James, long session 00542 “The 2.5% Retirement-Withdrawal Key; QQQ Isn’t Tech, It’s the Strongest Companies!,” 2025-11-29, the 2.5% dividing line @05:30–07:00 and the A/B-portfolio blend method @40:00–44:00. Transcript at raw/transcripts/長篇/00542…. ↩ ↩²
CLEC James, long session 00335 “Is Your Retirement Fund Enough? How to Make It Last Forever; What Is the 3% Rule,” 2022-05-20, the dynamic 3% rule (monthly = assets × 3% ÷ 12, reset yearly), ≤3% is very safe, borrowing beats selling @00:30–33:00. No corresponding deck; transcript at raw/transcripts/長篇/00335…; backtest amounts are spoken approximations, not fixed values. ↩
CLEC James, 長篇 00575 “What truly teaches you is not the result, but the process!”, 2026-07-25, margin-of-safety update @1:52:30–1:55:00 (“stocks fall 50%, the dividend may drop 30%… I now suggest people use 17 years not 15, 12 years in QQQI… if the dividend shrinks 30% you still get the dividend you should have… if you barely have 15×, then 15× it is, you take on some risk”). Transcript in raw/transcripts/長篇/00575…; the dividend-shrink ratio is James’s spoken estimate, not a fixed value. ↩ ↩²
CLEC a senior member, 長篇 00575, 2026-07-25, QQQI transition product and 15→20→30× progression @38:00–41:00 (“QQQI / 3451 / GEPQ should all be a transitional choice, this thing will eventually be sold… retire at 15×, 10× QQQI, 5× 4-3-3; after 5 years 20×, after 10 years 30×; at 30× you can dispose of the transition QQQI and switch back to 4-2-4 or 8-2-7-3 with pledge withdrawals… so estate tax isn’t a main concern”). Transcript in raw/transcripts/長篇/00575…; the doubling and years are the member’s optimistic assumptions, not guaranteed. ↩ ↩²
CLEC channel, post “Real risk-resilient allocation… cash flow that survives the worst 15 years,” 2026-06-15. Source: raw/docs/X及YouTube的貼文/0009…. ↩ ↩²
CLEC James, 長篇 00553 “How to use QQQI + 433 allocation to build 15× annual-expense retirement cash flow?” 2026-02-14, the member Q&A on taxable accounts counting only after-tax (~70%) and pledge-never-sell counting in full. Matching deck page in raw/docs/簡報資料/00553… (“assets in a taxable account, we can only count after-tax… unless you pledge and never sell, then it counts in full”). Transcript in raw/transcripts/長篇/00553…; 70% is a rough estimate depending on cost basis and tax rate. ↩ ↩²