what the research actually says about withdrawal rates, annuities and longevity math, where the smart people disagree, and how an advisor can publish something worth forwarding.

Q1. Why is retirement income such a mess to search for?

1. Demand is enormous. LIMRA says we are in the middle of “Peak 65,” with 4.1 million Americans turning 65 every year, many without a pension (LIMRA, 2026).
2. Worry is rising. In EBRI’s 2026 Retirement Confidence Survey, only 64% of Americans feel confident they will have enough to live comfortably. Among workers it fell 6 points in one year to 61%, the lowest reading since 2017 (EBRI/Greenwald, 2026).
3. Supply is mostly sales copy. Search “best retirement withdrawal strategy” and a large share of page one will be selling an annuity, a newsletter or a free consultation. The content was written to convert, and it reads that way.
4. The search box now answers for you. Pew tracked 68,879 real Google searches. When an AI summary appeared, people clicked a traditional result 8% of the time versus 15% without one, and clicked a link inside the summary just 1% of the time. Question-style searches triggered a summary 60% of the time (Pew Research Center, July 2025).

Put those together and you get the problem in one sentence. Millions of anxious people are asking a math question, and the answer they see first is often a summary of a sales page.

Q2. So what is the “safe” withdrawal rate? Give me the number.

Happily. There are six of them, and every one is defensible.

1. William Bengen (1994) found a 60/40 portfolio survived every 30-year period back to 1926 at an initial withdrawal of 4.15%. Editors rounded it to 4%, and a legend was born. He raised it to 4.3% in 1997 and 4.5% in 2006 (Wisdom, Wealth, Abundance summary of Bengen’s work).
2. Bengen (2025) now calls 4.7% the “Universal SAFEMAX”, using seven asset classes. The kicker: across 349 historical retirees the average SAFEMAX was about 7.1%, and 6.75% worked a little more than half the time (Advisor Perspectives; CNBC).
3. Morningstar uses forward-looking return forecasts instead of history. Its number was 3.3% for 2021 retirees and is 3.9% for 2026 retirees, assuming fixed real spending, a 30-50% stock mix and a 90% chance of money left after 30 years. Retirees who flex spending can start as high as 5.7%, and a 30-year TIPS ladder supported 4.5% as of September 30 (Morningstar, Dec. 2025).

On a $2 million portfolio, 3.9% versus 5.7% is $78,000 versus $114,000 a year, as Madison Partners points out. That gap is the story. Bengen asks, “what survived the worst history we have?” Morningstar asks, “what survives 90% of plausible futures?” Neither is wrong. They are answering different questions, and a good article says which question it is answering.

Q3. Where do credible researchers genuinely disagree?

Debate	One serious camp	The other serious camp
How much stock after 65?	Anarkulova, Cederburg & O’Doherty: about 33% U.S. / 67% international stocks, zero bonds, for life. Target-date investors “need 63% more preretirement savings” to match it (SSRN).	Cliff Asness of AQR: “Equities are already 100% owned,” so not everyone can do this. Cederburg himself notes that with simpler return assumptions the optimum drops to about 40% stock at 65 (Rational Reminder, ep. 350).
Does spending really rise with inflation?	Blanchett (2014): real spending falls about 1% a year. A $100,000 budget at 65 bottoms near $74,146 at 84, which lifts the worst-case starting rate to 4.73% (Journal of Financial Planning).	Later work (Hurd & Rohwedder; Blanchett, 2026) says the curve may be a “smirk” for the median retiree, while the average is pulled up by big late-life medical bills: ~$50,000 median vs. ~$250,000 at the 95th percentile.
Should people annuitize?	Economists since Yaari (1965) have argued lifetime income is valuable longevity insurance. Framed as consumption, 72% of people prefer a life annuity (Brown, Kling, Mullainathan & Wrobel, 2008).	Framed as an investment, only 21% do. Bequest motives, fees, surrender schedules, liquidity and insurer credit risk are real costs, and a 30-year TIPS ladder offers 4.5% with no mortality pooling.

Here is the fun part of the annuity debate. Total U.S. annuity sales hit a record $464.1 billion in 2025, up 7% (LIMRA). Yet the plain lifetime-income products economists write papers about, single premium immediate and deferred income annuities, were $14.4 billion and $4.8 billion. That is about 4% of sales. Indexed products took 45% of the market, up from 24% a decade ago. So when someone says “annuities are booming,” ask which ones.

Q4. How long is “long-term,” really?

1. Using the Social Security Administration’s own 2023 life table, a 65-year-old man’s life expectancy is 18.1 more years and a woman’s is 20.7. Averages are where plans go to die, though.
2. Of men alive at 65, about 24% reach 90. For women it is about 35%. For a married couple, the chance at least one of them reaches 90 is roughly 51%, and at 95 it is still about 21% (our math from SSA Table 4c6).
3. The Actuaries Longevity Illustrator, built by the Society of Actuaries and American Academy of Actuaries, is even more sobering for healthy nonsmokers: a 50% chance at least one spouse of a 65-year-old couple reaches 92 (Indiana INPRS example).
4. Social Security is the best inflation-adjusted annuity most people own, and it is still going to be there. The 2026 Trustees Report projects the retirement trust fund runs dry in late 2032, after which 78% of scheduled benefits remain payable without reform (SSA, June 2026). Plan for a haircut.

Q5. What does this look like for a real couple?

Meet the Hendersons (illustrative): both 65, $1.5 million saved, $48,000 a year in combined Social Security.

Approach	Year-one income from savings	The catch
Morningstar fixed, 3.9%	$58,500	Highest odds of leftover money; may underspend for decades
30-year TIPS ladder, 4.5%	$67,500	Money is gone at 95, and there is a 21% chance one of them is still alive
Bengen SAFEMAX, 4.7%	$70,500	Built on U.S. history, which was a very good century
Morningstar flexible, up to 5.7%	$85,500	Requires cutting spending after bad markets, and actually doing it

The honest answer for the Hendersons is a blend. Cover essentials with Social Security (Morningstar found delaying it is “a wise decision”) plus perhaps some guaranteed income, then run a guardrails strategy on the rest. That sentence won’t go viral. It would, however, get forwarded to a brother-in-law.

Q6. How does an advisor publish something a skeptic would forward?

1. Show the range, then pick a lane. Publish the 3.9% to 5.7% spread, then explain which number you would use for which client and why.
2. Name the question your number answers. Historical worst case, 90% of simulated futures, or “guaranteed”? Say so in the first line.
3. Argue against yourself. If you sell annuities, include the 21% investment-frame finding and the TIPS ladder alternative. Readers trust writers who show the other side, and so do machines: Google’s rater guidelines call trust the most important part of E-E-A-T, noting that “a financial scam is untrustworthy, even if the content creator is a highly experienced and expert scammer.”
4. Cite primary sources, by name, in the text. Pew found .gov sites made up 6% of links in AI summaries versus 2% in standard results. Link to SSA, LIMRA and the actual paper, not to a blog quoting a blog.
5. Write for the answer box. Use real questions as headings, put a number in the first sentence of each answer, use tables, and date-stamp everything. Retirement math changes every December when Morningstar updates.
6. The forward test. Before you publish, ask: would a retired actuary roll her eyes? If yes, keep editing.

Sources

1. Morningstar, What’s a Safe Retirement Withdrawal Rate for 2026? https://www.morningstar.com/retirement/whats-safe-retirement-withdrawal-rate-2026
2. Morningstar, Finding Your Safe Withdrawal Rate (2021 = 3.3%) https://www.morningstar.com/retirement/morningstars-retirement-income-research-finding-your-safe-withdrawal-rate
3. Advisor Perspectives, Bill Bengen Boosts the 4% Rule to 4.7% https://www.advisorperspectives.com/articles/2025/08/29/bill-bengen-boosts-the-4-rule-to-4-7
4. CNBC, Bengen on inflation and the 4.7% Universal Safemax https://www.cnbc.com/2025/09/03/4percent-rule-inflation-retirement.html
5. Madison Partners, Safe Withdrawal Rate 2026 https://www.themadisonpartners.com/safe-withdrawal-rate-2026-why-right-number-matters/
6. LIMRA, Final 2025 Retail Annuity Sales $464.1 Billion https://www.limra.com/en/newsroom/news-releases/2026/limra-final-u.s.-retail-annuity-sales-set-new-sales-high-totaling-$464.1-billion-in-2025/
7. LIMRA, 2025 sales: indexed products 45% of market https://www.limra.com/en/newsroom/news-releases/2026/limra-u.s.-retail-annuity-sales-top-$460-billion-in-2025-marking-fourth-year-of-record-sales/
8. SSA, 2023 Period Life Table (2026 Trustees Report) https://www.ssa.gov/oact/STATS/table4c6.html
9. SSA, 2026 Trustees Report press release https://www.ssa.gov/news/en/press/releases/2026-06-09.html
10. Actuaries Longevity Illustrator (SOA / AAA) https://www.longevityillustrator.org/
11. Indiana INPRS, All About Life Expectancy https://www.in.gov/inprs/destinationretirement/posts/all-about-life-expectancy/
12. EBRI/Greenwald, 2026 Retirement Confidence Survey https://www.ebri.org/media/press-releases/content/2026-retirement-confidence-survey-finds-americans-less-confident-about-retirement-as-worries-grow-over-social-security–medicare-and-rising-costs
13. Pew Research Center, Google users click less with AI summaries https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
14. Brown, Kling, Mullainathan & Wrobel (2008), NBER w13748 https://www.nber.org/system/files/working_papers/w13748/w13748.pdf
15. Anarkulova, Cederburg & O’Doherty, Beyond the Status Quo (SSRN) https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4590406
16. Morningstar, Should Long-Term Investors Be 100% in Equities? (Asness quote) https://www.morningstar.com/stocks/should-long-term-investors-be-100-equities
17. Rational Reminder ep. 350, Scott Cederburg https://rationalreminder.ca/podcast/350
18. Blanchett (2014), Exploring the Retirement Consumption Puzzle https://www.financialplanningassociation.org/article/journal/MAY14-exploring-retirement-consumption-puzzle
19. Blanchett (2026), A Smile, a Smirk, or Something Else? https://onlinelibrary.wiley.com/doi/full/10.1002/cfp2.70032
20. Google Search Quality Rater Guidelines overview https://services.google.com/fh/files/misc/hsw-sqrg.pdf

Figures are for education, not individual advice. Withdrawal-rate research assumes specific portfolios, horizons and success thresholds; your mileage will vary.