How Much Do You Really Need to Retire?

The Number Nobody Wants to Hear

If you asked a group of financial planners to name the single question clients ask most often, you'd hear the same answer almost every time: How much do I need to retire? It sounds like a simple question. It isn't. The answer depends on your spending, your timeline, your portfolio composition, the sequence of returns you happen to experience, and, frankly, a fair amount of luck. But the academic research on sustainable withdrawal rates gives us a framework that's considerably more useful than guessing.

Let's walk through what the data actually says — and what it means at different income levels.

Where the 4% Rule Comes From

In 1994, financial planner William Bengen ran a series of historical simulations using real U.S. stock and bond return data going back to 1926. He wanted to find the highest withdrawal rate that would have survived every 30-year retirement period on record — including the brutal sequence of the mid-1960s through 1970s, when retirees faced a toxic combination of poor early returns and persistent inflation. His conclusion: 4% of the initial portfolio value, adjusted annually for inflation, worked in every historical scenario when a portfolio held roughly 50% to 60% equities.

That finding became the foundation of modern retirement planning. Later, the Trinity Study (1998, updated in 2011 and 2021 by researchers Philip Cooley, Carl Hubbard, and Daniel Walz) confirmed the basic finding: a 4% withdrawal rate on a portfolio of at least 50% stocks had a 95% to 100% historical success rate over 30-year periods.

The important word in those sentences is historical. The research is based on past returns. Whether it holds going forward depends on things no one can predict.

What "Safe" Actually Means in the Data

One of the most misunderstood aspects of this research is what "success rate" means. A 95% success rate doesn't mean you probably won't run out of money. It means that in 95 out of 100 historical 30-year periods, the portfolio lasted. The 5% failure scenarios weren't random — they were clustered around particularly punishing sequences, mostly retirement years starting in the late 1960s.

Morningstar's 2023 research, led by Christine Benz and John Rekenthaler, arrived at a somewhat more conservative figure for new retirees: a 3.8% starting rate, updated to account for current valuations and lower expected bond returns. Wade Pfau, a prominent retirement researcher, has argued that today's combination of high stock market valuations and low starting bond yields warrants consideration of a rate closer to 3.3% to 3.5% for those highly risk-averse about running out of money.

These aren't small differences. At a 4% rate, a $1 million portfolio generates $40,000 per year. At 3.3%, that same portfolio generates $33,000 per year. Over a 30-year retirement, that gap compounds into a very different lifestyle.

The Income Level Problem

Here's something the popular conversation around retirement savings often glosses over: the target number is not the same for everyone, and the math changes substantially across income levels. A few reasons why.

Social Security replacement rates vary inversely with income. Social Security is designed as a progressive benefit. A worker who averaged $30,000 a year in earnings replaces roughly 55% to 60% of that income from Social Security alone. A worker who averaged $100,000 replaces closer to 30% to 35%. A worker earning near or above the taxable earnings cap replaces an even smaller fraction.

This changes the calculation dramatically. If you're a lower earner, Social Security might cover most of your essential expenses. You need far less in personal savings to bridge the gap. If you're a higher earner, your savings have to carry much more weight.

Let's look at three simplified scenarios, all assuming retirement at 65, a 30-year horizon, and the 4% withdrawal rule:

Household spending of $45,000 per year: If Social Security provides $22,000 annually, you need your portfolio to generate $23,000. At 4%, that requires a portfolio of roughly $575,000.
Household spending of $90,000 per year: If Social Security provides $32,000 annually, you need your portfolio to generate $58,000. At 4%, that requires a portfolio of roughly $1.45 million.
Household spending of $180,000 per year: If Social Security provides $45,000 annually (near the maximum for a married couple), you need your portfolio to generate $135,000. At 4%, that requires approximately $3.4 million.

The middle and upper scenarios illustrate why so many higher earners feel like they never quite reach the finish line — the target keeps moving because their spending expectations are high relative to Social Security income.

The Sequence-of-Returns Problem Is Real

Most retirement calculators show you an average return. The average obscures something critical: the order of returns matters enormously when you're withdrawing money.

Here's a concrete illustration. Suppose two retirees both earn an average of 6% per year over 20 years. Retiree A experiences strong returns in the early years and poor returns late. Retiree B gets the same returns in reverse — weak early, strong late. Despite identical average returns, Retiree B may run out of money while Retiree A finishes with a healthy surplus. When you're selling assets to fund withdrawals during a market downturn, you're locking in losses and depleting shares that won't be there to capture the eventual recovery.

This is why so much research has focused not just on average withdrawal rates but on strategies to manage sequence risk. A common approach is to maintain one to three years of expenses in cash or short-term bonds so you can avoid selling equities during prolonged downturns. Another is to use a variable withdrawal strategy — slightly reducing spending in bad market years — rather than insisting on rigid inflation-adjusted withdrawals regardless of portfolio performance.

Research by financial planner Jonathan Guyton and computer scientist William Klinger found that flexible "guardrail" rules — where you reduce withdrawals when the portfolio falls to certain thresholds — allowed initial withdrawal rates of 5% to 5.5% while maintaining acceptable long-term success rates. The tradeoff is that you accept spending fluctuations in exchange for a higher starting distribution.

How Longer Retirements Change the Math

The classic research assumed a 30-year retirement. That made sense when most people retired at 65 and average life expectancy placed death in their mid-80s. But a 65-year-old today has a meaningful probability of living to 90 or beyond. A couple where both partners are 65 has roughly a 45% chance that at least one lives past 90, according to actuarial data from the Society of Actuaries.

For a 35-year retirement, the safe withdrawal rate drops. Research from Pfau and others suggests the 30-year 4% rate becomes something closer to 3.5% when you extend the horizon to 35 or 40 years. For someone retiring at 60 or earlier, the adjustments are even more significant.

Early retirees also face a longer period before Social Security begins. Every year of delay in claiming Social Security after age 62 increases the monthly benefit by 5% to 8%, and the difference between claiming at 62 versus 70 can amount to 70% to 76% more monthly income. For someone with a substantial portfolio and good health, delaying Social Security while drawing down savings — essentially using savings to "buy" a higher guaranteed income stream — is often the optimal strategy from a pure expected-value standpoint.

What Healthcare Costs Do to the Estimate

Any honest accounting of retirement needs has to grapple with healthcare. Fidelity's annual estimate of healthcare costs in retirement for a 65-year-old couple currently sits around $315,000 in present value, and that figure assumes Medicare coverage. It doesn't include long-term care.

Long-term care is the financial black hole of retirement planning. The median cost of a private room in a nursing home exceeds $100,000 per year. The median stay in a nursing facility is around 2.5 years, but distributions are highly skewed — a meaningful fraction of people require care for five or more years, and costs can approach or exceed $500,000 over a lifetime.

For middle-income retirees with assets in the $500,000 to $1.5 million range, a severe long-term care event can be genuinely catastrophic. This is the income range where the choice to self-insure or purchase a hybrid long-term care policy matters most, because high-net-worth individuals can absorb the costs while lower-income individuals may rely on Medicaid.

The Number, Made Practical

After all of this, is there a rule of thumb that works? Yes, with heavy caveats.

The traditional formula — 25 times your expected annual spending beyond Social Security — remains a reasonable starting point for someone with average longevity expectations, a 50% to 60% equity allocation, and a 30-year retirement horizon. It derives directly from the 4% rule (since 1 divided by 0.04 equals 25).

For someone who:

Expects to retire before 60
Wants to leave a meaningful legacy
Is especially risk-averse about spending cuts
Faces elevated healthcare risk factors

...a multiplier of 28 to 30 (corresponding to a 3.3% to 3.6% withdrawal rate) provides substantially more cushion, with historical success rates approaching 100% across all periods the research covers.

For someone who:

Has flexibility to reduce spending in bad markets
Has significant guaranteed income (pension, Social Security)
Has lower longevity expectations or other factors reducing the withdrawal horizon

...a multiplier closer to 20 to 22 (corresponding to a 4.5% to 5% rate with flexible spending) may be defensible, particularly using guardrail strategies.

The Research Is a Foundation, Not a Guarantee

One final point worth making: the withdrawal rate research is extraordinarily useful, but it's built on historical data from the U.S. market — arguably the single best-performing stock market in history over the 20th century. Researchers including Pfau have examined international data and found considerably lower "safe" withdrawal rates in many other countries' markets. The U.S. historical record may be a ceiling, not a floor.

That doesn't mean the research is useless. It means the responsible way to use it is as a calibration tool within a broader plan that includes flexibility, guaranteed income sources, and ongoing monitoring — not as a one-time calculation that locks in a fixed spending rate forever regardless of what the market does.

The question "how much do I need to retire?" deserves a real answer, not a made-up rule. The data gives us a real answer, within ranges and with honest uncertainty attached. That's more useful than false precision — and more honest than telling people retirement is simple.