Four in 10 Americans Have Never Worked Out How Much They Need to Retire

Forty-two percent of Americans have never calculated how much money they will need in retirement, according to a new report from Evergreen.ai. MarketWatch reported the figure on October 23, 2026. The story also ran on Morningstar's news service that morning, credited to Dow Jones and timestamped 6:00 a.m.
The headline says these savers are heading toward retirement "blind." That word refers to having no target. It does not mean their savings have been measured and found too small.
That difference matters if you read the number as proof that people are not saving enough. Someone who has never run a retirement projection can still have a 401(k), which is a workplace savings account. They may also have a pension, or a rough idea of what Social Security, the federal retirement program, will pay.
What the 42% measures is the missing target. That target is the figure that turns a savings balance into a safe yearly withdrawal, or into a set amount to put aside each month. Without it, saving tends to happen by default.
In my view, this is the part of the number that deserves the most attention. Default saving looks like saving from the outside. A person can deposit money every month and still have no idea whether it is enough.
Other surveys point the same way, though they ask different questions. The 2026 Retirement Confidence Survey from the Employee Benefit Research Institute was conducted online from January 2 through January 28, 2026. It found that 64% of Americans feel confident they have enough money to live comfortably through retirement. Confidence fell among both workers and retirees.
The same survey found that 46% of workers have calculated how much they are likely to need for health costs in retirement. Health care is one of the most concrete items in a retirement budget. So nearly half of workers have a figure for that one line, while 42% of Americans have no overall figure at all.
The Retirement Insecurity 2026 report looks at savings balances. It reports that 80% of Americans said the nation faces a retirement crisis in 2026, up from 67% in 2020. It also finds that 47% of respondents have less than $100,000 saved for retirement, and 18% have no retirement savings at all.
Access is another hurdle. Reuters, citing research from Economic Innovation Group, reported in February 2026 that nearly 54 million American workers have no access to any employer-provided retirement plan. Workplace plans are often where people first see a projection tool or a default savings rate. A missing plan can mean the calculation never happens at all.
That link is my inference, not something these sources measure directly. It is the mechanism most worth testing against the data.
The Employee Benefit Research Institute also runs a Retirement Security Projection Model. It simulates the share of Americans at risk of falling short of the retirement income they will need. That is different from asking people what they believe. A simulation relies on assumed incomes, savings paths, and benefit formulas, so its result depends heavily on those assumptions.
When you see a retirement statistic, ask how it was produced. Was it self-reported confidence, self-reported calculations, a survey of account balances, or a simulation? Each one answers a different question.
Views about the future add another layer. In the 23rd Annual survey cited by the Transamerica Center, 41% of workers said future generations of retirees will be worse off than people retiring now. Goldman Sachs has also published its Retirement Survey and Insights Report 2026, based on a survey of working and retired Americans about the financial strain reshaping retirement.
For savers, the lesson from the Evergreen.ai figure is narrow. Forty-two percent of Americans have no stated retirement target, and that gap can be measured separately from the balance in any account. Whether it leads to worse outcomes depends on who the 42% are, what they have saved, and what benefits they expect.
If you are building a plan from these numbers, start with the target. The math that settles the question still has to be done household by household.


