This is Part 5 of our Middle Class Trap Series
Saving for early retirement is actually pretty forgiving.
If you save 20% or more of your income and put it in low-cost index funds, the big decisions come infrequently. You might periodically rebalance, switch between Roth and pre-tax contributions as your income rises, or increase savings in a taxable brokerage as early retirement gets closer. You may only need to check in once or twice a year without making any changes. The math mostly takes care of itself.
Planning withdrawing your hard-earned savings can feel a lot more complicated. Over the first four parts of this series, we built a detailed picture of how Ben and Leslie’s withdrawal strategy could work. We modeled three bridge strategies, walked through ACA optimization, and stress-tested their pre-tax balance at different return assumptions. At 6% returns, their plan holds up well.
But that plan is built on a specific set of assumptions, and over a 30-year retirement, a lot of those assumptions will drift.
The withdrawal phase rewards people who check in and adjust. This mindset shift can be hard for FIRE optimizers who are used to setting a single plan and executing it. In fact financial research has shown that accounts belonging to inactive investors outperform those of active traders.
But saving is fundamentally different than a withdrawal strategy where tax laws, market returns, spending needs, and family circumstances will all change in ways no spreadsheet can fully anticipate.
Tax Law Changes More Than People Expect
Tax rules are likely to be different in 15 or 20 years. That’s not a reason to stop planning, but it might help take the pressure off trying to perfectly optimize the next 30 years.
A few examples worth thinking about (also check out a longer post on tax uncertainty)
The ACA and its premium tax credits are only 12 years old. They didn’t exist when many current early retirees were doing their planning. Much of Ben and Leslie’s healthcare strategy depends on a program that wasn’t part of the retirement planning conversation in 2012.
The 0% capital gains bracket didn’t exist until 2008. It came out of Bush-era tax legislation and has been extended repeatedly since. A couple retiring in 2000 and building a 30-year plan couldn’t have counted on it.
Estate tax thresholds have moved dramatically. The exemption was $675,000 in 2001. It climbed, fell, climbed again, and is now over $15 million per person. Planning assumptions that were irrelevant for most families in 2000 became critical a decade later, then receded again.
The Net Investment Income Tax threshold hasn’t kept up with inflation. The $250,000 threshold for married filers has been fixed since 2013. As incomes rise over time, more retirees get pulled into an extra 3.8% tax on investment income without any change to the law.
None of these are obscure edge cases. They are all features that sophisticated early retirement plans are built around today.
What about tax rates?
A lot of Roth conversion advice assumes ordinary income rates will eventually rise. The long-term fiscal picture, including entitlement spending and federal debt, makes tax increases a reasonable risk to plan around. But tax increases haven’t happened yet, and nobody knows when or whether they will. You could make a case for converting pre-tax accounts at today’s rates you don’t want to over-convert if rates stay flat. The math changes a lot depending on which way you assume tax rates are headed.
Capital gains treatment is worth watching too. Preferential taxation of long-term gains has existed in some form since the 1920s, but the specific rates and structure have changed significantly over that period. Someone who retired in 1975 and planned around favorable capital gains treatment would have watched the top rate climb to 28% by the late 1970s.
Ben and Leslie’s brokerage strategy relies heavily on qualified dividends and long-term gains staying in the 0% bracket. Planning around preferential treatment compared to ordinary income is a reasonable medium-term bet. Planning around the specific current structure holding for 30 to 50 years is a more tenuous one.
There’s a long list of recent tax changes that would have been impossible to model ten years ago. The standard deduction nearly doubled under the 2017 Tax Cuts and Jobs Act. The state and local taxes (SALT) deduction cap upended planning for high-income households in expensive states. The new senior tax credit is a meaningful benefit for retirees right now, and it’s set to expire.
No spreadsheet built today can account for the equivalent changes that will happen between now and 2050, because nobody knows what they are yet. Your goal should be to build enough flexibility into the plan to absorb them when they arrive.
Return Assumptions Drive Bigger Swings Than the Optimal Conversion Amount
Ben and Leslie’s plan was built on a 6% nominal return assumption. That’s reasonable. It’s also the most important assumption in their plan, and one figure that we know won’t be smooth or predictable.
Here’s what their plan looks like across a range of return assumptions:
| Nominal Return | Brokerage Survives To | Pre-Tax Balance at 75 | First RMD |
| 4% | 62 | $1.1 million | $44,000 |
| 6% | 63 | $3.2 million | $130,000 |
| 8% | 66 | $6.7 million | $273,000 |
These average return assumptions actually understate the uncertainty. A specific sequence of positive or negative returns can skew results more than the average. A bad first decade in retirement does more damage than a bad decade at 70, because early withdrawals lock in losses before the recovery arrives.
In contrast, a great first decade followed by a difficult second one is a fundamentally different retirement, even with the same average return on paper.
On RMDs: Keep Them in Perspective
RMD optimization has become something of an obsession in early retirement planning circles, and it’s worth stepping back for some perspective. A large RMD at 75 means Ben and Leslie’s investments compounded well for 25 years. That’s the outcome everyone is hoping for.
That doesn’t mean early retirees should leave obvious money on the table during their low-income years. Making Roth conversions to fill the standard deduction and low brackets is straightforward good planning. On the other hand, sacrificing real dollars today to solve a hypothetical problem 25 years out, which depends on both return assumptions and tax law staying roughly as-is, deserves some scrutiny.
As we showed in Part 4, at a 6% return assumption, Ben and Leslie’s first RMD is roughly in line with their annual spending need, and Social Security covers the rest. The RMD problem only emerges at 8% or 10% returns. Some of you are thinking “but that’s the historical return of the stock market,” and that’s true. Even so, sequence of returns is a much more likely plan-killer than an oversized RMD.
The Assumptions Nobody Models
Most retirement projections model one household moving through time in a straight line. Real life rarely works that way. Fair warning: we’re about to get a little dark.
Death of a spouse. For our happy couple, one spouse is likely to predecease the other. When that happens, filing status flips from married filing jointly to single and bracket thresholds nearly halve. A surviving spouse can face a significantly higher tax bill on the same income, right when they’re least equipped to make adjustments. For couples doing long-term tax planning, this is one of the most consequential variables and one of the least discussed.
A Social Security haircut. The Social Security trust fund is projected to face a shortfall around 2033-2035 if Congress takes no action. Most retirement projections assume full benefits. Whether to apply a discount on that assumption is a judgment call, but treating current benefit levels as guaranteed over a 30-year retirement is optimistic at best.
Spending doesn’t inflate in a straight line. Research consistently shows retirement spending follows more of a smile curve: higher in the early active years, lower in the quieter middle years, then potentially rising again with late-life healthcare costs. A flat 2.5% inflation assumption smooths over a shape that actually matters for how long money lasts.
IRMAA creep. Medicare Part B and D surcharges kick in at income thresholds that haven’t been adjusted for inflation at the same rate as ordinary brackets. Over time, more retirees get pulled into higher premium tiers without any deliberate policy change.
The Plan Is a Starting Point
The withdrawal phase isn’t harder than accumulation because it requires more precision. It’s harder because it requires more flexibility.
A good withdrawal plan for Ben and Leslie starts at age 50 and sets up a framework that can be monitored as market returns and circumstances change. Tax law will change and returns will be uneven. Spending will shift over time too.
The retirees who navigate this phase well use their plan as a living document rather than a final answer.
Coming Up in Part 6
In the final post of this series, I’ll share the spreadsheet that underlies all of this analysis and walk through what software forecasting tools get right and where they can fall short. Retirement software almost always recommends aggressive early Roth conversions, and we’ll dive into why that is and what assumptions are driving it. Understanding those assumptions is the key difference between blindly following a tool and understanding why you’re executing its recommendations.
Read Post 6: All Models Are Wrong, Some Are Useful
