Life’s Persistent Questions #7: Is Retirement Income Really Pass or Fail?
By: Eric Sondergeld | July 19, 2026
One of the persistent challenges in retirement income planning is that we often describe outcomes in overly simple terms. Chief among them is the industry’s long‑standing reliance on “probability of success” as a primary measure of retirement readiness. That metric can be useful, but it is also blunt. It implicitly treats retirement as a pass‑or‑fail proposition, when in reality it is nothing of the sort.
Recent critiques of probability‑of‑success metrics, including work by David Blanchett, have pointed out that success rates alone say little about the magnitude of a potential shortfall. That observation moves the conversation in the right direction. But I think the issue can be framed even more clearly by borrowing a concept that is fundamental to how insurance works.
Incidence is Not the Same as Severity
In insurance, risk is never evaluated solely by asking whether an adverse event might occur. The event may trigger a claim, but the economic consequence depends on the severity of the loss. A minor fender‑bender and a total loss are both accidents, but they are not remotely equivalent outcomes.
Retirement income planning works the same way. Failing to meet an income goal is a risk event, but the more important question is how large the resulting shortfall is, when it occurs, and relative to what benchmark. Is the gap measured as a percentage of total desired income? Or as a percentage of the portion that savings were expected to fund? Those distinctions matter because they determine how damaging a shortfall really is.
Guaranteed Income Lowers the Cliff
Consider two retirees.
The first has a relatively low probability of falling short of an income goal, but if that shortfall occurs, the income drop is significant because much of the plan depends on portfolio withdrawals. The cliff is high.
The second has a higher probability of some shortfall, but the gap is likely to be much smaller because essential expenses are largely covered by guaranteed lifetime income. The cliff is lower.
A simple probability‑of‑success metric does not tell us which scenario is “better,” nor should it. The point is that interpreting probability is fundamentally different in these two cases because the severity of failure is fundamentally different.
This perspective helps clarify the role of guaranteed lifetime income. Existing sources such as Social Security, defined benefit pensions, and lifetime annuities materially reduce the severity of a retirement income shortfall.
If part of a household’s spending need is already covered for life, the downside is limited. Even if markets disappoint or portfolio withdrawals must be adjusted, the household is not falling from its full income target to zero. Guaranteed income lowers the cliff.
Viewed this way, guaranteed lifetime income functions much like insurance against severe retirement income loss. It may not eliminate every possible gap, but it absorbs part of the shock and limits how damaging adverse outcomes can be.
A Deterministic Way to Frame Part of the Risk
This also suggests a more deterministic way to think about part of the problem. To the extent a household’s essential income needs are fully covered by guaranteed lifetime income sources, the probability of failing to meet that portion of the goal is effectively 0%. No stochastic model is required to establish that.
The real question begins where guaranteed income ends. If guaranteed lifetime income falls short of what is needed to pay the bills, risk is introduced. At that point, the relevant analysis is not simply whether a shortfall might occur, but how large the uncovered gap is, how dependent it is on savings, and how severe the shortfall could become under adverse conditions.
Framed this way, the issue is not whether everything in retirement must be modeled probabilistically, but which portion of income is already secured and which portion remains exposed.
The Industry is Already Moving in This Direction
The industry has already begun moving toward this way of thinking. Capital Group’s Portfolio Reliance Calculator, Protective’s reliance‑rate framework, and Edward Jones’s work on reliance rate all emphasize how much of a retiree’s income plan depends on the portfolio versus guaranteed sources, such as Social Security or pensions.
That is an important step because it identifies where exposure begins. But it should not be the last step. Exposure alone does not tell us how damaging a shortfall might be. The next question is severity.
People Don’t Blindly Drive Off Cliffs
This framing also puts the common fear of “running out of money” in better context. It is a powerful phrase, but often an imprecise one. A retiree with Social Security, a pension, or other guaranteed lifetime income may run out of savings, but does not run out of income altogether. Nearly all people have at least one of these.
More importantly, most retirees are not passive passengers headed toward a cliff. When conditions deteriorate (e.g., markets decline, balances fall, withdrawals feel uncomfortable), people normally respond. They reduce discretionary spending. They slow down. They adapt.
That does not eliminate risk, but it does mean the relevant issue is rarely whether income falls to zero. It is how large an adjustment is required and whether essential needs remain covered.
Toward Better Retirement Income Metrics
All of this points to the need for better ways to describe retirement outcomes. Rather than a single probability score, a more useful framework would consider multiple dimensions: the probability of falling short, the expected severity of any shortfall, the likelihood of failing to cover essential expenses, and the share of total income that is guaranteed for life.
Retirement income is not a binary outcome. Severity matters in addition to probability. Ignoring that distinction leaves us with an incomplete picture of the risks retirees actually face, and the protection already in place when things do not go exactly as planned.
