Retirement income's 8% adoption rate hides a reserve fund problem

In a December 2025 survey of 1,486 Vanguard households between 60 and 80, all holding $100,000 to $1 million in retirement accounts, only 8% reported drawing steady income, while 53% made irregular withdrawals for specific purposes and 39% took nothing until required minimum distributions forced the question—a split that describes a product almost no one uses the way its designers intended. PLANADVISER reported the findings.

Every decumulation framework on the shelf—Monte Carlo ranges, guardrails, income floors—presumes the client wants a paycheck or can be persuaded to want one; the survey says the client's mental model is a reserve fund that gets tapped when something happens, and the distance between those two positions is where retirement income plans go to die. A paycheck plan optimizes a rate, while a reserve fund runs on access, which changes when money moves, how much sits in cash, and which account gets drawn first.

What triggers the irregular withdrawals is the more useful half of the data: debt obligations were the primary driver for 31% of sporadic withdrawers, large unforeseen expenses such as medical bills and home repairs for 23%, discretionary spending such as vacations for 22%, and only 9% were drawing mainly to cover ongoing living expenses. Debt at the top of that list means the plan was built without the balance sheet attached, since a household servicing liabilities out of a tax-deferred account is solving a cash-flow problem in the most tax-expensive wrapper available, and no withdrawal rate on a planning slide fixes that.

Then there are the 567 respondents who had not withdrawn anything: nearly half, 47%, were living solely on Social Security, and among those leaning primarily on the benefit, 52% said they were maintaining their lifestyle while 44% described themselves as intentionally frugal or actively cutting spending. An advisor would flag this group as under-planned, but the clients may simply call it fine, having parked something up to $1 million in a reserve no one has sized, priced, or sequenced.

Among 176 respondents who planned to take only their required minimum distributions, 44% misunderstood what the distributions are for: 38% believed they would not need anything beyond the minimum, 29% read RMDs as the government's recommendation for a safe withdrawal rate, and 15% did not know they could take more. Keep the 29% in view—a tax-collection mechanism has been mistaken for planning advice by nearly a third of the people who intend to lean on it, and those clients are the likeliest to under-withdraw for years from a tax-deferred account that will be taxed in larger increments later, when they have less control over the bracket.

Vanguard's own argument pushes the same direction: the paper contends that delaying withdrawals can lead to a higher lifetime tax bill than taking more consistent distributions through retirement, illustrating the point with a hypothetical 63-year-old holding roughly $360,000 in a traditional 401(k) while collecting about $34,000 a year in Social Security.

The survey also captures, without appearing to intend it, a cohort doing by accident what an advisor would charge to do deliberately: roughly six in ten respondents said they reinvested some or all of their required distributions into taxable accounts, and about three in ten reported cutting spending or being intentionally frugal while reinvesting. Paying the tax early and moving the money out of the tax-deferred bucket shrinks the balance that will generate future required distributions—the same arithmetic an advisor would run on purpose—but a client who arrives at it by default, without comparing brackets or coordinating the move with charitable and estate planning, has the right instinct and none of the sequencing.

The sample deserves one caveat before anyone quotes the 8% in a meeting: these are Vanguard's clients, aged 60 to 80, in the $100,000-to-$1 million band, and they have already chosen a low-cost provider, which likely skews the group toward households that have not bought advice. The paper does not test whether clients at advice-first practices behave differently; it is reasonable to suspect they take regular income more often, and just as reasonable to suspect the gap is narrower than the industry would like, because the reserve-fund instinct is about what the account means to the household rather than who manages it. The coverage does not say, and no one should pretend to know.

The practical move is to stop arguing with the framing and design inside it: a spending floor funded by Social Security, guaranteed income, and a cash buffer; a named line item for the debt that roughly a third of the sporadic withdrawers gave as their primary reason for tapping the account; a reserve target the client has agreed to rather than one the advisor assumes; and a tax sequence that treats the RMD age as a planning deadline instead of a surprise. The reserve-fund client is asking for a plan they will actually follow, a higher bar than most decumulation models clear.

The last mile of retirement has been framed, repeatedly, as an income and benefit-claiming problem rather than a savings problem, with the work defined as claiming strategy, health-cost budgeting, and protected-income floors; Vanguard's numbers put the last mile a step earlier, at the withdrawal decision itself—who makes it, on what schedule, and against which balance sheet. For a household taking nothing, there is no income policy to optimize, only a decision being deferred until a statute makes it.

The 39% waiting on the rulebook are the ones to watch: they will arrive at the RMD age with a balance that compounded untouched and a bracket they have never tested, and their first withdrawal will be triggered by a deadline rather than by a decision. As this publication has argued, an extra working year is a chance to fix cash flow and reset the family timeline; for a client already past that year, the fixed window is the stretch between now and the first required distribution, and the advisor who runs the bracket arithmetic inside it will not need a pitch to get the meeting.