Losses early in retirement can disrupt the journey—even when long-run returns meet expectations.
Large investment losses early in retirement often have a large impact on retirement security, an impact that can be especially pronounced for individuals who stop savings and coast to retirement. Sequence-of-returns risk: the timing of returns matters, not just their long-run average.
Introduction
A recent Wall Street Journal article by Oyin Adedoyin, Young People Are Obsessed With This Simple Retirement-Savings Formula, describes Coast FI, an approach in which people accumulate enough early in life for investment growth to carry them toward retirement without further contributions. Coasting means stopping retirement contributions, while continuing to work to cover current expenses.
Coasting is often presented as a deliberate choice: a worker reaches a savings threshold and decides that further contributions are unnecessary. In practice, life events can produce a similar outcome without any carefully calculated decision to coast. Not everyone who starts at a major law firm makes partner or can sustain the hours and demands of that career. Having children, caring for relatives, illness, or losing a job can lead people to accept lower pay or reduce their working hours. They may continue covering current expenses while contributing little or nothing to retirement savings. For these workers, coasting is less a lifestyle preference than an adjustment to changed circumstances. The question is still whether their existing savings can support retirement—and how vulnerable that plan is to investment losses when withdrawals begin.
The basic calculation discounts a retirement savings target by an assumed compounded return over the years remaining until retirement:
Savings needed today = Retirement target ÷ (1 + annual return)ʸᵉᵃʳˢ
For example, a person seeking $1 million at age 65 and assuming a 7 percent nominal annual return would need $184,249.18 at age 40: $1,000,000 ÷ (1.07)²⁵. The experiment below considers a person with $165,824.26 at age 40 who stops contributing and retires at 62. With a 7 percent annual return, those savings grow to approximately $734,668 at retirement. The question is how long that portfolio can support spending that begins at $50,000 annually—and how losses early in retirement affect the answer.
The formula uses a single assumed compounded return even though annual returns often vary by a lot. This paper does not challenge the common planning assumption that long-run average market returns will be solidly positive.
Instead, I examine one additional risk: the sequence of returns once retirement withdrawals begin. Strong long-run returns do not necessarily prevent losses early in retirement or at an end of a career from exhausting savings sooner than expected.
Previous Research on sequence risk explains why an average return alone is insufficient for retirement planning:
Wade Pfau’s regression-based analysis of lifetime returns shows that returns late in a career become particularly important when workers are still contributing, while early-retirement returns have disproportionate effects on sustainable spending.
Michael Kitces examines sequence risk for savers and retirees and the vulnerability surrounding retirement. His research with Pfau also finds that starting retirement with a more conservative portfolio and gradually increasing equity exposure can improve sustainability under unfavorable return sequences.
My earlier article, The Sequence of Returns Puzzle, illustrates why losses late in a worker’s saving career and early in retirement can be especially damaging. The experiment below applies the withdrawal side of that problem to a person who has stopped contributing at 40.
My earlier article, When Higher Withdrawal Rates Backfire, examines how higher initial withdrawals amplify the damage from early retirement losses. Flexible withdrawal rules can help preserve savings but may require substantial spending cuts—leaving retirees financially solvent without maintaining an acceptable standard of living.
The Experiment
The experiment compares three investment scenarios: constant returns, losses at the beginning of retirement, and the same losses ten years later. All three have the same compounded investment return over the 52 years from age 40 to age 92.
Scenario 1 — Constant returns: 7 percent nominal annually during both coasting and retirement.
Scenario 2 — Early losses: a 10 percent loss in each of the years beginning at ages 62 and 63, with 7.7431 percent annually in every other year, including the coasting years.
Scenario 3 — Later losses: a 10 percent loss in each of the years beginning at ages 72 and 73, with 7.7431 percent annually in every other year, including the coasting years.
Scenarios 2 and 3 each have two loss years and 50 normal years within the comparison period. Their normal-year return g is calculated from:
(0.90)² × (1 + g)⁵⁰ = (1.07)⁵²
Thus, a portfolio receiving no contributions or withdrawals would have the same ending wealth under every scenario at age 92. The stronger normal-year return compensates for the two loss years. Annual returns are converted into equivalent monthly compound rates, and calculations use the unrounded value of g.
The three scenarios share the following assumptions:
Savings and coasting. Each person has $165,824.26 at age 40, makes no further retirement contributions, and takes no withdrawals before age 62. Scenario 1 earns 7 percent annually during these 22 years and reaches retirement with $734,668.09. Scenarios 2 and 3 earn 7.7431 percent annually during coasting and each reaches retirement with $855,491.09. Calculations use unrounded balances and returns.
Consumption. Final-year salary is $150,000. Initial annual retirement consumption is one-third of that salary, or $50,000, regardless of wealth. Consumption rises 3 percent on each retirement anniversary.
Monthly withdrawals. The first withdrawal occurs at age 62. Spending is $4,166.67 per month in the first year, calculated without rounding, and remains constant within each retirement year. Withdrawals occur at the beginning of each month, before investment returns.
Other cash flows. The portfolio funds all modeled consumption. There is no additional income, saving, or reduction in spending following losses. Returns are net of fees; taxes are omitted.
Comparison criterion. Count the full monthly withdrawals the portfolio can finance before the next scheduled withdrawal exceeds remaining assets.
Scenarios 2 and 3 enter retirement with identical wealth, approximately $120,823 more than scenario 1. Their comparison isolates the effect of moving the same two loss years from the beginning of retirement to ten years later.
Comparisons with Scenario 1 reflect the fact that Scenarios 2 and 3 require higher normal-year returns to offset their two loss years. The comparison between Scenarios 2 and 3 isolates the effect of the timing of those losses.
Results with Monthly Withdrawals
Scenario 1 — Constant returns: 263 full monthly withdrawals, or 21 years and 11 months of consumption. The first shortfall occurs at age 83 years and 11 months.
Scenario 2 — Early losses: 217 full monthly withdrawals, or 18 years and 1 month of consumption. The first shortfall occurs at age 80 years and 1 month.
Scenario 3 — Later losses: 289 full monthly withdrawals, or 24 years and 1 month of consumption. The first shortfall occurs at age 86 years and 1 month.
Moving the two loss years from the beginning of retirement to ages 72 and 73 adds 72 full monthly withdrawals, or six years of consumption. Both scenarios start retirement with the same wealth and experience the same investment returns in a different order.
Despite entering retirement with approximately $120,823 more than scenario 1, the early-loss portfolio supports spending for three years and ten months less. Withdrawals during the downturn reduce the capital available to benefit from subsequent growth.
Scenario 3 benefits from ten years of 7.7431 percent annual returns before its losses occur. Its portfolio supports spending for two years and two months longer than the constant-return portfolio.
This result illustrates that delaying losses allows a portfolio to build a larger cushion before withdrawals are affected.
Implications for the Coasting Calculation
Equal compounded returns over the full comparison period do not produce equal retirement spending outcomes. The early-loss portfolio begins retirement with more wealth than the constant-return portfolio but reaches a shortfall sooner. Moving those losses ten years later extends the funding period by six years.
The results illustrate why a coasting calculation cannot rely solely on an assumed long-run return. A worker may accumulate the projected savings balance and still face an insecure retirement if substantial losses occur when withdrawals begin. Under the early-loss path, these savings cannot fully fund the next scheduled monthly withdrawal shortly after age 80, leaving a considerable spending gap for someone who lives into their nineties.
This paper examines one set of savings, retirement age, and spending assumptions. Other combinations could produce substantially different outcomes and deserve further study. Monte Carlo simulations could extend the analysis by generating many possible investment return paths and estimating how often people who stop saving early can sustain their planned retirement spending. That research could help identify how much savings—and what margin for adverse returns—a worker needs before coasting becomes a reasonable retirement strategy.
Authors Note: This article is free to read and share. If you found it useful, consider supporting Economic and Political Insights with a paid subscription. Your support helps me continue researching retirement, health care, economics, and public policy. Use this coupon to subscribe for $30 for one year.
The appendix below explains the model and calculations for readers who want to examine the results.
Appendix: How the Retirement Simulation Was Calculated
This appendix describes the calculations used in the retirement experiment and verifies the results reported in the paper. The purpose is to make the assumptions transparent and allow readers to reproduce the analysis.
Starting Conditions and Coasting Period
The model begins with a worker age 40 who has saved $165,824.26 and makes no additional retirement contributions. The worker continues working but allows the portfolio to grow until retirement at age 62.
Scenario 1 assumes a 7 percent nominal annual return during the 22-year coasting period:
$165,824.26 × (1.07)^22 = $734,668.09
Thus, the worker reaches retirement with $734,668.09.
Scenarios 2 and 3 are designed so that all three scenarios have the same compounded investment return from age 40 through age 92. These scenarios contain two years with a 10 percent loss and 50 years with a higher normal return.
The normal-year return is calculated so that:
(0.90)^2 × (1 + normal return)^50 = (1.07)^52
Solving this equation produces a normal-year return of approximately 7.7431 percent.
Applying that return during the coasting years gives:
$165,824.26 × (1.077431)^22 = $855,491.09
Therefore:
Scenario 1 begins retirement with $734,668.09.
Scenario 2 begins retirement with $855,491.09.
Scenario 3 begins retirement with $855,491.09.
The higher retirement balances in Scenarios 2 and 3 reflect the fact that those scenarios require higher normal-year returns to offset their two future loss years.
Retirement Withdrawal Assumptions
The first withdrawal occurs at age 62.
Initial retirement spending is based on one-third of final salary:
$150,000 × 1/3 = $50,000 per year
The first monthly withdrawal is therefore:
$50,000 ÷ 12 = $4,166.67
Spending increases by 3 percent on each retirement anniversary. Withdrawals remain constant within each retirement year.
Withdrawals occur at the beginning of each month. The portfolio balance is reduced by the withdrawal before that month’s investment return is applied.
The monthly return is calculated from the annual return:
Monthly return = (1 + annual return)^(1/12) − 1
Investment Scenarios
The retirement return paths are:
Scenario 1: Constant returns
7 percent annual return every year.
Scenario 2: Early losses
10 percent losses at ages 62 and 63.
7.7431 percent annual returns in all other years.
Scenario 3: Later losses
7.7431 percent annual returns through age 71.
10 percent losses at ages 72 and 73.
7.7431 percent annual returns thereafter.
The only difference between Scenarios 2 and 3 is the timing of the two loss years.
Verification of Results
The monthly withdrawal model was recalculated using the assumptions above.
The results are:
Scenario 1: Constant returns
263 full monthly withdrawals.
Spending supported for 21 years and 11 months.
First insufficient withdrawal occurs at approximately age 83 years and 11 months.
Scenario 2: Early losses
217 full monthly withdrawals.
Spending supported for 18 years and 1 month.
First insufficient withdrawal occurs at approximately age 80 years and 1 month.
Scenario 3: Later losses
289 full monthly withdrawals.
Spending supported for 24 years and 1 month.
First insufficient withdrawal occurs at approximately age 86 years and 1 month.
The calculations reproduce the results reported in the paper.
Limitations and Extensions
This experiment is designed to isolate one issue: the effect of the timing of investment losses after retirement begins. It does not estimate the probability of experiencing any particular return sequence.
A more complete analysis would use Monte Carlo simulations to generate many possible investment paths and examine how often different combinations of savings, retirement ages, spending levels, and return sequences lead to successful or unsuccessful outcomes.
The purpose of this exercise is narrower: to demonstrate that even when long-run compounded returns are equal, the timing of losses can materially change how long retirement savings support planned spending.
Additional Readings:
Pfau, Wade D. The Lifetime Sequence of Returns: A Retirement Planning Conundrum (2013).
Kitces, Michael. Valuation-Based Tactical Asset Allocation in Retirement (2014). Discussion of joint research with Pfau on rising equity glidepaths.
Bernstein, David. The Sequence of Returns Puzzle: Why Timing Hurts Workers and Retirees in Opposite Ways (2025). Examines why losses late in a worker’s saving career and early in retirement can be particularly damaging.
Bernstein, David. When Higher Withdrawal Rates Backfire (2026). Examines how higher initial withdrawals amplify exposure to early losses and why flexible withdrawal rules may preserve savings at the cost of substantial spending cuts.