10:26 PM, a Thursday in April, dining table buried in printouts, two laptops open, a calculator with a cracked screen, and my brother-in-law Marcus three glasses into arguing with me about a pension I hadn't even told most family members I had. The separation package gave me thirty days to choose: $612,000 as a lump sum rollover, or $3,400 every month for as long as I live, with half continuing to my wife after im gone. Marcus kept pointing at the lump sum like it owed him money. I kept looking at the survivor column.
Stakes: this decision, once made, is irreversible, and it covers the exact years when I'll be least able to fix it. Choose wrong at 61 and there is no second attempt, no market recovery that refunds a decision. That asymmetry is what kept me at the table past midnight for six straight weeks.
The twist nobody expects: the spreadsheet war ended in a draw, and I still chose the annuity with total confidence. The math tied. The math was never the whole decision. Here's how both halves played out.
the numbers, stated fairly on both sides
Break-even is the honest starting point. At $3,400 monthly, the annuity pays $40,800 a year. Collect to age 80 and that's $776,000 received, more than the lump sum with zero market dependence. Die at 68, the lump sum wins catastrophically. Marcus, who plans everything assuming he lives to 95, kept building models where the lump sum invested at 6 percent crushed the annuity by his mid-seventies. His models were correct. They were also conditional on a 6 percent market over his specific retirement, a variable he treated like a constant.
where the lump sum argument genuinely won
Credit where due. The annuity has no inflation adjustment, so $3,400 in year twenty buys maybe half a $3,400 shopping trip. The lump sum, rolled into an IRA, keeps upside, let's me stage Roth IRA conversions in low-income years, and leaves principal my kids could inherit. An annuity dies with you except for the survivor slice; a portfolio doesnt. Marcus called this "the flexibility premium," and he's right that it has value. My dispute was with the price he implicitly put on it.
the number the spreadsheets couldn't hold
My mother had a guaranteed pension of $2,100 a month. Every other financial thing in her life — a modest IRA, a house with equity, whatever the market was doing — rose and fell. The pension never moved. Watching her eighties, I spotted what that flat line actually snagged: she never once hesitated over a heater repair, a prescription, a flight. Her confident spending came from the check that couldn't disappoint her. Behavioral research calls spending from guaranteed income easier; I watched it happen at one kitchen table for a decade, which is a sample size of one and the strongest data I own.
the sequence risk that decided the tie
A bad market decade in your first retirement years is the specific way lump sums die. Retire into 25 percent down and withdrawals from a shrunken base lock in the damage. The annuity is immune to sequence by construction. I ran Marcus's own 6 percent model with one edit: start the first five years at negative 2 percent instead. The lump sum's lead evaporated and never fully returned. He objected that this was cherry-picking. I said retiring into a bear market is cherry-picking that happens to real people constantly.
how I actually split the difference
The final structure: I took the annuity, and it covers our nondiscretionary baseline — housing, food, insurance, about 70 percent of our core budget. My 401k rollover and brokerage money, approximately $700,000 outside the pension, covers travel, the Roth conversions, the inheritance, every flexible dollar. My wife keeps her own income and Social Security coming later. Nothing about our lifestyle depends on a market we can't control, and nothing about our upside depends on a check that can't grow.
the survivor benefit, priced honestly
One number sealed it. Half of $3,400 continuing to my wife for her remaining life, however long, was worth something the lump sum models kept rounding toward zero. Marcus's spreadsheet valued it at a cold actuarial figure. I valued it as: if I go first at 72, she never once hasta learn investing from grief. We argued that line item longer than every other line item combined, and it was the only one where neither of us budged, because we were both right and only one of us will ever be proven so.
signing, and the 30 days I used to the hour
I submitted the election on day 29, after making Marcus print his final model so I could keep it. He signed it as a witness to my stubbornness, his phrase. My wife and I celebrated with takeout and a spreadsheet-free evening.
Some nights I still open that April folder at 10-something PM, the same hour this all began, and look at the two final columns: $612,000 versus $3,400, arrows and asterisks all the way down. The cracked calculator sits in my desk drawer, retired on a pension of its own. Marcus and I now argue about other things, loudly, happily. But when friends hit their own pension-choice week, I send them one line that took me six weeks and one mother to learn: run the spreadsheet til it ties, then let the guaranteed floor decide.
