pillar · 401k · fundamentals
Why your 401(k) projection is probably too optimistic
Brokerage retirement projections quietly skip three forms of risk. Here's what they leave out, and why your real-world result usually undershoots the line on the chart.
By The Yearfold team · April 30, 2026 · Last reviewed May 4, 2026
You log into Fidelity, Vanguard, Schwab, or your employer 401(k) portal. There's a chart. A green line slopes upward gracefully toward your retirement age. A confident number sits at the right edge: "$1,847,000."
Three problems with that chart. None of them are the brokerage's fault — they're consequences of how single-line projections work. But every one of them pushes the displayed number higher than your actual future portfolio will land.
Problem 1: A single line ignores variance
The chart shows you the median outcome — or worse, the expected outcome assuming a smooth average return. In real markets, returns aren't smooth. A 50/50 stocks/bonds portfolio has a one-year standard deviation of roughly 9-12%. Across a 30-year horizon, the spread of plausible outcomes is enormous: the difference between the 10th-percentile and 90th-percentile ending balance can easily be 4× or more.
The "expected" line sits in the middle of that spread. About half of real-world outcomes will land below it. A green line slopes upward confidently; the actual distribution is a fan that gets wider every year.
The Yearfold calculator shows the fan. Your "single number" projection shows the line.
Problem 2: Sequence-of-returns is invisible
Two retirees with identical 30-year returns can have wildly different outcomes if the order of those returns is different.
- Retiree A: starts retirement with three good years (2010-2012). Withdraws modestly. Portfolio grows. By year 5, balance is up 35%. They survive the rest of the retirement easily.
- Retiree B: starts retirement with three bad years (2000-2002). Withdraws the same dollar amount. Portfolio drops. By year 5, balance is down 30%. They never recover the lost ground.
Same average return over 30 years. Different lived experience. The single- line projection treats average returns as the answer; the historical record shows the ORDER of returns is what kills plans.
This matters most in the first 5-10 years of retirement, when withdrawals compound losses. It's why traditional planners talk about a "bond tent" (more conservative entering retirement, then re-risking) — to minimise sequence-of-returns exposure during the danger window.
Single-line projections don't tell you which side of this risk you're on. A Monte Carlo does — explicitly, by reporting the 10th-percentile outcome alongside the 50th.
Problem 3: Inflation gets glossed over
Most 401(k) projections show nominal dollars — the number you'll see on your statement in 30 years. That number includes inflation.
A "$1.8 million" projected balance in 30 years, at the typical 3% inflation that's baked into most projections, is worth roughly $740,000 in today's dollars. That's the real spending power. Most retirees mentally substitute the nominal number for the real number, which over-states their future buying power by 50-60%.
Yearfold runs internally in nominal dollars (so growth, contributions, withdrawals, and Social Security all interact with the same inflation trajectory) but displays everything to you in real dollars — today's dollars — so the numbers actually mean what you'd think they mean.
The hidden assumption: continuous contributions
Your projection assumes you'll keep contributing $X/month every month for the next 30 years. The Department of Labor's longitudinal data (BLS Number of Jobs survey) shows the median US worker holds 12 jobs across their career, with average unemployment spells of several months between jobs. Contribution gaps during job changes — and the temptation to cash out a 401(k) at the new employer — quietly subtract years of compounding.
The honest version of "what does my projection look like" includes a realistic distribution of career interruptions. Most projections don't model them at all.
What about employer match changes?
Your projection probably assumes your current employer match continues indefinitely. In recessions, employers routinely cut or suspend matching (Vanguard's "How America Saves" 2024 documents this in detail). The 2008-2009 recession saw ~20% of large employers suspend matching contributions for at least one year.
A 6% match on a $80K salary is $4,800/year. A two-year suspension is $9,600 of contributions you're projecting that you may not actually receive.
So what should I do?
Three concrete moves to get an honest projection:
1. Run the math against a probability distribution, not a line
The real question isn't "what's my projected balance?" — it's "what's the probability my plan succeeds, and what's the worst-case I should plan against?" Yearfold's headline output is a success probability with the 10th and 90th percentile balance bands explicit. Your brokerage's projection is a single line. Use both.
2. Always look at your number in TODAY'S dollars
If a calculator shows you a future balance, mentally divide by 1.03^N (where N is years to retirement). For a 30-year horizon, that's roughly half. Your $1.8M nominal projection is more like $750K of real spending power. That matters a lot for planning.
3. Stress-test against a 20% market drop in retirement year 1
Most plans look fine in the median case. The honest test is: does the plan still work if the market drops 20% the year you retire? If yes, you're durable. If no, you need a buffer (cash, bonds, part-time income, or a flexible spending plan).
The brokerage's chart isn't lying
The chart shows you the answer to a specific, well-defined math question: "if returns and inflation hit their long-run averages perfectly, here's where the line goes." That's a useful sanity check. It's also not the question you should be asking.
The question you should be asking is: "across the range of futures that have actually happened in US economic history, what fraction of them work for my plan?" That's a Monte Carlo simulation. Run yours against your specific inputs, then look at the percentile band — not the median line — to decide how robust your plan really is.
Related reading: