I Asked 3 CPAs the Same Retirement Question and Got 3 Answers — Here’s Why

Getting this wrong costs either years of work I did not need to do or money that runs out at 84, and I was not willing to guess at a number that expensive.

So I sent one identical page to three CPAs and had three 20-minute calls in eight days. Back came: keep working to 67; retire at 64 earning $30,000 part-time for three years; retire at 64 only once $130,000 sits in cash. I had budgeted neither the extra working years nor the reserve.

Same inputs every time: age 58, $110,000 salary, $620,000 saved, $250,000 on the mortgage, out at 64, Social Security at 67, $70,000 of spending. Nobody disputed a figure. The split came from judgment and from what each practice looks at first.

The three answers only became comparable once return, spending, tax and Social Security assumptions sat side by side on one sheet. Until then they were three opinions and I was grading them by how confident the voice sounded.

My job was never to crown the smartest expert. It was to make those assumptions line up as neatly as the facts already did, then pick the answer whose conditions I can actually meet — and I have written it all down on one page to carry into the next meeting.

CPA's lens	Main recommendation	Key assumption	Best suited to
Tax practice (CPA A)	Work to 67	5% return; Roth conversions at 64-67	Someone who values tax savings over early retirement
Planning practice (CPA B)	Retire at 64 with $30k part-time for 3 years	4% and 6% return scenarios; part-time income $30k	Someone willing to work part-time and flexible on spending
Retirement-plan practice (CPA C)	Retire at 64 only with $130k cash reserve	Spending cut to $65k; annual reassessment	Someone who can build cash and revisit the plan yearly

All three CPAs got the same snapshot — did they actually answer the same question?

They answered the same question, all right. They just didn't hear it the same way.

The page I sent was one typed sheet: age 58, salary $110,000, $620,000 across the retirement accounts, $250,000 still owed on the house, a stated target of 64, Social Security at 67, spending $70,000. I attached nothing else, because there was nothing else at that point. Each CPA got one 20-minute call, then silence on my end — no tax returns, no account statements, no follow-up meeting, no chance to revise. What each of them worked from was that sheet plus whatever I happened to say out loud in those twenty minutes.

The verdicts came back three different ways. One said stay at the $110,000 job until 67. One said leave at 64 and earn $30,000 a year part-time for three years. One said leave at 64 only with $130,000 parked in cash first.

Here is what took me a week to see. Not one of them disputed a single figure I sent. Nobody said the savings looked low or the spending looked high or the mortgage number was wrong. The numbers sat there untouched while the answers walked off in three directions.

I spent those seven days treating the split as a fight about my arithmetic — as if one of them had caught something the others missed. There was nothing to catch. The arithmetic was the same entering all three calls; what changed was what each practice looked at first and what each quietly filled in where the sheet was blank. That is an assumption problem wearing a number problem's clothes, and until I named it that way I was grading the answers on how confident the voice sounded.

The tax CPA said wait until 67

The tax CPA's answer was to keep the $110,000 salary running to 67 and give up the plan to leave at 64.

Her reasoning hung on one stretch of calendar: the three years from 64 to 67. In that window I would have no wage income and no Social Security check yet, which makes it the cheapest stretch of my life to move money out of a traditional retirement account into a Roth, where it grows and comes out tax-free later. Pay tax now at a low rate, never again after. The rules that govern that move live in IRS publications, not in a 20-minute phone call, and she said so.

She modeled one return number for the whole plan — 5% on the $620,000 — and then leaned on the salary, not the portfolio, for the three years in question. No withdrawals from 64 to 67. Contributions still going in. The whole bridge paid for out of a job I would still be showing up to.

The trade-off is not hidden, it is just enormous. Three years of mornings. Three years of the commute and the meetings, in exchange for a tax window that may or may not be worth more to me than those years are.

That is the test I have to run on this answer: whether the conversion room is worth three more years of work, because this one is settled on tax, not on cash flow. Run the numbers against your own bracket before you take it as gospel — a fee-only fiduciary or a CPA who can see your actual returns is the right person for that, not a stranger on a phone.

The planner said retire at 64 with part-time

CPA B, the planning practice, spent the same three years a different way: retire at 64, earn $30,000 a year part-time until 67, and claim Social Security at 67 like the other two.

Here is the arithmetic that made it work. $30,000 of wages against $70,000 of spending leaves about $40,000 a year to come out of the $620,000 portfolio during the bridge. Retire at 64 with no income at all and that draw roughly doubles, so the part-time work cuts the pull on savings by a little over half.

CPA B did not give me one return number. The plan was run at 4% and at 6%, and the two cases do different jobs. At 4% the part-time income carries the plan; at 6% it is cushion, nice to have but not load-bearing.

So I asked the question that matters to me: what happens if the part-time work stops after the first year? That is where the $30,000 showed itself to be structural, not a bonus. Pull it out and the 4% case stops working, which is exactly the case I should plan around rather than the 6% one.

For CPA B's answer to hold, two things have to be true in my life: that I can find and keep $30,000 of part-time work starting at 64, and that the portfolio can carry about $40,000 a year while I do.

That is a bet on me, not on the market. CPA A bet on the tax code and a 5% return; CPA B bet that I can still earn at 64. I know which of those bets I can actually control.

The third CPA demanded a cash reserve first

The retirement-plan practice put a number in front of me that neither of the others had: $130,000 in cash, sitting in an account I could reach without selling anything, before I walk out of the building at 64. The call ran the full 20 minutes and that figure landed about halfway through. He called it two years of runway at his spending number, which brings me to the part I almost missed.

He set my spending at $65,000, not the $70,000 I told him I spend now. That was a decision he made on my behalf, a trim he wrote into the plan and then said out loud so I could argue with it. I didn't, that night. I sat with it for two days because it is the difference between a plan that assumes I stay the same person and one that assumes I will cut back once the paycheck stops.

What I keep turning over is the timeline. $130,000 out of a $110,000 salary, on top of a $250,000 mortgage I am still paying, inside six years. That is a little over $21,000 a year of saving that does not currently exist in my budget. And he did not hand me a single return number to lean on. His instruction was to rerun the whole thing every year as the market and my actual spending show up, which means the plan is not a document I file. It is a chore I repeat.

So the condition is real and it is mine to meet: can I build that cash and still pay the mortgage, and will I actually redo the arithmetic each spring instead of putting the folder in a drawer?

Each practice sees a different first problem

Each practice walked in looking at a different part of my page, and the part they looked at first decided the answer they gave me. None of them disputed a figure. They just opened the file at different pages.

The tax practice opened at the three-year stretch between 64 and 67, when I would have no wage income and Social Security would not have started. In his reading that window is the cheapest place I will ever convert retirement money at a low tax rate, and the way to keep it open is to keep earning. His answer is not really about retirement age. It is about not wasting a tax window. The planning practice opened somewhere else entirely: I earn $110,000 and one day I won't, so his question was what plugs that hole. Part-time work of $30,000 plus Social Security at 67 fills part of it, and the $620,000 carries the rest at roughly $40,000 a year in the gap. The retirement-plan practice opened at the first withdrawal. If my first two years of selling land in a bad market, the loss gets locked in, and cash is what lets me wait out a bad stretch without selling.

Same page, three entry points, and I could not find one of them wrong. Each was answering the problem that practice sees every week.

Which leaves me with the uncomfortable part: the practice predicts the answer. If I go to a tax practice I will get the tax-shaped plan, delivered with complete confidence, and confidence is not the same as fit. So I have to name the problem I am actually afraid of before I book the next call, because that fear, not the CPA's certainty, is what should pick the lens.

Different return assumptions change the answer

Of all the assumptions sitting behind those three lenses, the return number moved my numbers more than anything else I compared. And the three of them did not even use the same kind of number.

The tax practice ran one rate, 5%, applied across the whole plan. The planning practice refused a single figure and gave me two, 4% and 6%, which is why his answer had a load-bearing condition in it. The retirement-plan practice would not name a rate at all and told me to reassess each year against whatever the market actually did.

Drop to 4% and the arithmetic turns against leaving at 64. The distance between retiring at 64 and waiting to 67 stretches out, and the $30,000 of part-time income stops being a nice cushion and becomes the thing holding the plan up. Run the same spending at 6% and the identical plan reads as comfortable, even roomy. Nothing else changed between those two pictures except the setting. Three people can add up the same figures correctly and still land in three different places, because the assumption underneath the addition is doing the work.

I want to be clear that none of these is a forecast and none of them is a promise. They are scenario settings, and every one of those CPAs would tell me so if I asked directly. The real question is which one I could defend to myself through a bad decade, rather than the one that makes 64 work on paper. That is a harder question than it sounds when 64 is the number I have been aiming at.

Social Security timing was a bridge-cost question

All three CPAs left my Social Security claim sitting at 67, which surprised me at first, because I had assumed that was where they would argue. Nobody did. The real disagreement was over who pays for the three years between 64 and 67, and each man answered it with a different pocket.

On CPA A's sheet that stretch costs $70,000 a year, and his fix was elegant in a way that annoyed me: I never retire at 64, so there is no gap to fund. CPA B shrank the hole by putting $30,000 of part-time wages into it. CPA C never trimmed the years at all, he just stacked $130,000 in cash beside them, which at his $65,000 base covers two bad years.

So the question stopped being when to claim and became what fills the space before the claim starts.

What I had been doing, for weeks, was treating the 20-minute call as a source. Delayed credits, spousal rules, how a claim interacts with work income before full retirement age – those live at the Social Security Administration, in writing, and not in anything a CPA says off the top of his head on a Tuesday. I should not have been storing any of it in memory.

Here is the test I apply now. Can I name the thing that pays for 64 to 67, and is it something I hold – wages I can actually get, cash I can actually park – rather than a portfolio that has to behave for exactly three years? If the bridge leans on the market cooperating, it isn't a bridge. It's a hope with a date on it.

Spending assumptions were not the same

The hole can only be as deep as the spending poured into it, and the three of them did not pour the same amount. I had sent $70,000 of current spending and assumed it landed as a neutral fact. It didn't. It landed as a choice each practice made about what I was willing to believe about myself.

CPA A took the $70,000 whole, no edit, no question. CPA C did something I found mildly insulting until I thought about it for a week: he set the base at $65,000, told me plainly it was a trim rather than a prediction, and ran everything off the lower number. CPA B went further and assumed I could drop to $60,000 in weak markets. Five to ten thousand a year is not a rounding difference. It moves the retirement date before a single return assumption gets applied.

What I had never done, in any of this, was separate what I spend from what I would be willing to spend.

Ten thousand dollars of willing gave CPA C room his plan needed and CPA A's plan never had. That is the part that stung – not that he cut my lifestyle, but that I couldn't have told you which $10,000 he cut.

Commitments have to survive a bad year. If I claim $60,000 is doable when markets fall, then in the year markets fall I have to actually do it, which means knowing today which lines in the $70,000 are fixed and which are mine to move. Otherwise the trimmed budget is a nicer spreadsheet, not a plan.

Risk tolerance was read in opposite directions

The last input in the whole exercise was never written on the page at all, and it may have moved the answers more than any return figure did. I said one sentence on all three calls: I don't want to run out. That is the entire disclosure. Everything else was each man deciding what he had heard.

CPA A took it as a preference for safety and put me on the $110,000 salary for three more years. CPA B took it as a wish to be free at 64 and built a plan where $30,000 of part-time work does the work of a bigger portfolio. CPA C took it as fear of a market drop and answered with $130,000 in cash, which is what a man does when he thinks the danger is a bad first year.

Three readings of one sentence, and each man answered the version of risk his own practice was trained to see.

I cannot be talked out of this by a better spreadsheet. I have to pick which risk I am actually reducing. Outliving the money is one. Working three years I had never planned to work is the other. They are not the same fear, and the plan that guards against one is not the plan that guards against the other.

The tell, for me, was looking at which answer made my shoulders drop. That is not a financial test. But it is the only way I found to choose between three defensible answers that all assumed they knew what I was afraid of.

The first five years carry the most weight

Whatever risk I decide to reduce, every version of the plan has the same soft spot in time: the first two years after I leave work. That is where I put the pressure first.

Withdrawals in a falling market lock the loss in. Sell $70,000 of investments in a down year and those shares are gone – they don't come back when the market recovers, because I already spent them. Money taken out in a flat or rising year costs me far less. Same spending, same portfolio, wildly different damage, and the damage lands early.

So I ran CPA B's plan without its safety net. No part-time wages at all, $70,000 a year pulled straight from the $620,000 for the three years from 64 to 67, and returns at 4% instead of 6%. Then I ran CPA C's $130,000 reserve against two years of $65,000 spending. It covers exactly two years. A three-year bad stretch outlasts it, and I wanted to see that number before I trusted it.

Then the same spending against all three rates. At 6% the retire-at-64 plan holds without flinching. At 4% it needs the $30,000 of part-time work to stay upright. CPA A's 5% is the close call in the middle, where a small shift either way tips it. The math didn't change between those columns. Only the assumption did.

The habit I took away costs nothing: test the worst case of the answer I like, not the average case of the answer I want.

Make the CPAs show their assumptions

A stress test is only worth running on inputs I can line up side by side, and these three arrived in three different shapes. One gave me a single return figure, one gave me a range, one refused a number altogether. Comparing them meant asking for the assumptions in writing.

The ask is one page, four lines: return or return range, annual spending level, how the 64-67 years are taxed, and the Social Security claim age. Four lines I can lay next to each other, which was never possible with three phone calls in my head.

Then one more question per CPA, and it turned out to be the most useful question I asked: what single change flips your recommendation? For CPA B, it was losing the $30,000 of part-time work. For CPA C, it was the $130,000 not being there. For CPA A, it was the conversion window not being worth three more years of salary. Each of them answered inside ten seconds.

That tells me where the real weight sits in each plan. It also tells me what to check against my own life: how much cash I can actually hold, whether I will genuinely work part-time at 64, and which risk I want reduced.

Here is what unsettled me. Only one of those three answers was arithmetic. The other two were judgment calls – a choice about tax, a choice about fear – delivered in the same steady voice as the first, and on the phone they sounded identical. A page of written assumptions was the only way I could tell which was which.

Frequently Asked Questions

Is it too late to start planning for retirement at 58?

No. At 58 you still have years to adjust contributions, spending, and your retirement date. The three CPAs I spoke with all said the same thing: the plan matters more than the starting point. A $620,000 balance is a real foundation, not a lost cause.

How much should I have in cash before I retire?

CPA C recommended $130,000 for my situation — about two years of $65,000 base spending. That's not one-size-fits-all. The right number depends on your spending, how long you expect a bad market to last, and how much cash you can build while still working. Ask your CPA to model it.

Should I claim Social Security at 67 or wait until 70?

All three CPAs I spoke with assumed claiming at 67 for my plan, but they focused on the funding gap between retirement and claiming, not on claim age itself. The SSA website has the rules on delayed credits and spousal benefits. Run the numbers on your own earnings record before deciding.

Can I retire at 64 if I still have a mortgage?

Yes, but it complicates the cash flow. I have $250,000 left on my mortgage. CPA C's plan required building $130,000 in cash while carrying that mortgage. That's a heavy lift in six years. If the mortgage payment is large, it may push the retirement date later or require more part-time income.

What if I can't find part-time work after I retire?

Then CPA B's plan falls apart, because the $30,000 was load-bearing, not a bonus. If part-time work is uncertain, ask your CPA to model the plan with zero part-time income. That forces you to either save more cash or delay retirement, which is exactly what CPA A recommended.

Which answer fits your numbers, not your hopes

Here is my rule, stated plainly. If I can hold $130,000 in cash and earn $30,000 part-time for three years, the retire-at-64 answer is the one I test first. If I cannot do both, waiting to 67 wins — not because it is smarter, but because I can fund it.

Take the one-page assumptions sheet to every CPA you sit with. Use the IRS for tax rules and the Social Security Administration for claiming rules, not a 20-minute call.

Three calls, one page, three answers. The page never changed. My job was to make the assumptions as identical as the facts were.

The answer I act on at 64 will be the one whose assumptions I have written down and can live with, not the one I liked best on the phone. The snapshot is still sitting on my desk. This week it gets a second page stapled to it.