Money leaving a retirement account does not necessarily buy a meal, a holiday or a home repair. It can settle into another savings balance. That simple distinction becomes important when renewed coverage of retirement underspending, including The Atlantic's October 12 feature, turns a household budgeting problem into a claim about unused spending capacity.
Research supports the idea that the form of retirement resources is associated with how people consume. It also contains limitations that disappear easily in a headline. An investment balance, an account distribution and actual consumption are three different measures. Understanding their relationship is more useful than assuming that a large balance represents money that can immediately or safely be spent.
A distribution can be another balance-sheet transfer
The IRS describes required minimum distributions as minimum amounts that owners of applicable retirement accounts must withdraw. These rules concern an account distribution, rather than an obligation to purchase goods and services. Tax treatment and applicability depend on the account and circumstances; an RMD cannot be rolled over into another tax-deferred account.
A household can meet a withdrawal requirement and retain the remaining proceeds as cash or in a taxable investment account. In that case, the account location changes without an equivalent increase in consumption. Taxes can reduce the resources left for other uses, and the household still decides whether the remaining money belongs in an emergency reserve, future spending or a bequest. A withdrawal statistic alone cannot answer that decision.
The same distinction matters for financial businesses. Money distributed from a retirement account can remain within the savings and investment system, although under a different account arrangement. For companies selling goods or services, the relevant flow is what the household actually spends. Treating the distribution itself as an equal amount of new consumer demand skips a step in the transmission mechanism.
The study compares resources as well as pay cheques
David Blanchett and Michael Finke's 2025 paper estimates that retirees consume approximately 80% of lifetime income on average, compared with roughly half of other available savings and income resources. The underlying Health and Retirement Study is a long-running University of Michigan survey. The paper examines historical observations, rather than a new 2026 survey of spending.
Its methods restrict the sample to households meeting financial-resource and consumption criteria, including at least $100,000 in financial assets measured in 2019 dollars. The authors convert savings into potential income using assumed withdrawal factors. The comparison therefore does not mean that researchers tracked every dollar actually withdrawn and found half of those dollars spent. The estimated income available from an asset balance depends on the model's assumptions.
The authors also note that their spending-efficiency income estimate is not adjusted for taxes. That matters when comparing potential resources with a household's usable budget. Observed relationships can support a mental-accounting explanation, but they do not establish that purchasing a particular income contract will cause every household to spend more or improve its welfare. The broad survey and the paper's selected sample should not be treated as interchangeable populations.
A stronger test of the behavioural interpretation would distinguish differences in preferences and circumstances from the effect of changing the payment form itself. Evidence that a change in framing or payment frequency alters spending while preserving comparable resources would strengthen the mechanism. Evidence that the apparent gap disappears after better allowance for taxes and household needs would weaken claims of avoidable underspending. Neither result can be presumed from the headline comparison.
Holding back can be a choice rather than an error
A regular income payment gives a household a visible budgeting unit. A portfolio requires decisions about lifespan, future costs and how much principal to use. It is plausible that a payment resembling a familiar pay cheque feels easier to spend than selling an asset. This is a mechanism consistent with the research, not a claim that all reluctance to draw down savings is irrational.
Health costs, a surviving partner's needs, housing repairs and an intended inheritance can make liquidity valuable. Personal preferences can also change after retirement: some households may want fewer purchases, rather than lack permission to make them. A model that describes spending below an assumed capacity identifies a question for investigation. It does not by itself prove that the household has made a welfare-reducing mistake.
Nor can the stock of retirement wealth be converted into an annual economic boost by applying a multiplier measured on pension payments. Wealth can finance spending across many years and belongs to households with different obligations. A calculation that ignores timing, ownership and competing uses may turn a real budgeting issue into an unsupported consumption forecast.
A predictable income stream changes the contract too
An income product can pool longevity risk or make a budget feel more predictable. It also changes the terms under which the household can access capital. The SEC's annuity guidance describes differences among contracts, fees and potential surrender charges, and explains that an insurer's obligations depend on its financial strength and ability to pay claims.
Those terms matter independently of any behavioural benefit. A stream of payments may fit recurring expenses but provide less flexibility for a large unplanned need. Features intended to support income or beneficiaries can carry their own costs and conditions. The existence of a retirement consumption gap cannot establish that those trade-offs suit a particular household, or that a product promises investment gains.
The useful outcome is a connection between sustainable resources and the life a household intends to fund, with the liquidity and obligations it values preserved. An account balance can help support that outcome; so can an income stream. Neither its size nor the fact that a distribution has occurred is a complete measure of how retirement savings are serving the household.