Tomorrow is the last day of August. Everything I've written about this week was observation. A consumer confidence survey you can't influence. A GDP revision you can't affect. A Fed chair's speech you can't negotiate with. All of it useful to understand, none of it responsive to anything you do.
Tuesday you walk into the only four months of the year where that flips completely.
Taxes are the largest single expense most households will ever carry. Over a working lifetime the number dwarfs the mortgage, the cars, the tuition, all of it. And unlike market returns, unlike inflation, unlike the labor market, the tax outcome is substantially a function of decisions you make on a calendar you already know.
The problem is that almost everyone runs this calendar backwards. They start thinking about taxes somewhere around the third week of December, at which point most of the good options have already expired or gotten expensive. September decisions are cheap. December decisions are expensive. That gap is entirely about mechanics, and it's worth understanding why.
Payroll systems need a cycle or two to process a change. Custodians get slammed in late December and processing times stretch. Markets are where they are, so a loss you wanted to harvest may not exist anymore by the time you look. And every dollar of deferral has to come out of a paycheck that hasn't been issued yet, which means the fewer paychecks remaining, the larger a bite each one has to take.
So here's the sequence, in the order it should actually happen.
September belongs to payroll, and it's the item with the hardest deadline mechanics of anything on this list. The 2026 employee deferral limit for a 401(k), 403(b), or most 457 plans is $24,500. If you're 50 or older, you can add $8,000 on top of that. If you turn 60, 61, 62, or 63 during the calendar year, the catch-up is $11,250 instead, provided your plan permits it. Combined employee and employer contributions cap at $72,000.
Now do the arithmetic that matters. If you're paid twice a month, you have roughly eight pay periods left after September closes. Whatever gap exists between your year-to-date contributions and the limit has to be divided across those remaining checks. Every week you wait, that required percentage climbs, and it climbs steeply. Somebody who's $8,000 short in early September is deferring an extra thousand dollars per check. The same person in mid-November is trying to defer four thousand per check, which for most people simply isn't possible because the paycheck isn't big enough.
There's a second trap here that costs people real money. Some employer plans match per pay period without a year-end true-up. If you front-load your deferrals aggressively and hit the annual cap in November, those December paychecks have no deferral for the employer to match against, and the match is just gone. Whether your plan trues up is written in the plan document, and it takes one email to your benefits contact to find out. Ask before you accelerate.
September and October belong to the tax projection, and this is the step people skip that makes every subsequent step guesswork.
You cannot make an intelligent decision about capital gains, Roth conversions, charitable timing, or bracket management without knowing where your taxable income is going to land. Not your gross income. Your taxable income, after deductions. The 2026 standard deduction is $16,100 for single filers, $32,200 for married filing jointly, and $24,150 for head of household. Subtract that from your projected income and you have the number that drives everything else.
Why does it drive everything else? Because of the single most underused provision in the tax code, which almost nobody outside of financial planning circles knows exists.
Long-term capital gains in 2026 are taxed at zero percent for single filers with taxable income up to $49,450, and for married couples filing jointly up to $98,900. Not a deduction. Not a credit. A rate of zero.
Capital gains stack on top of ordinary income, so the calculation is about how much room you have left underneath that ceiling. Say you're single, your ordinary income after the standard deduction lands at $43,900, and you're holding an appreciated position. You have $5,550 of space inside the zero percent bracket. You can sell that much gain, pay nothing, immediately repurchase the same security, and you've permanently raised your cost basis for free. There's no wash sale problem because the wash sale rule applies to losses, not gains.
That's gain harvesting, and it's the mirror image of the loss harvesting everyone talks about. It matters enormously in specific years: a sabbatical, a career transition, the early retirement window before Social Security starts, a year the business took a loss, the first year after a spouse stops working. In those years the zero percent bracket is a gift with an expiration date of December 31, and the only way to use it is to know your projected taxable income by October.
October and November bring an item that catches people who did everything else right, and it only applies to taxable brokerage accounts.
Mutual funds are required to distribute realized capital gains to shareholders, and they typically announce estimated distributions in October and November and pay them in December. If you buy a fund in November and it distributes in mid-December, you receive a taxable capital gain distribution on appreciation that occurred before you owned a single share. You pay tax on somebody else's gain.
The fix is simply to check the fund company's estimated distribution page before making any Q4 purchase in a taxable account, and to wait until after the record date if the estimate is meaningful. Exchange-traded funds have a structural advantage here and generally distribute far less, which is one of the underappreciated reasons they've taken over taxable accounts. None of this applies inside an IRA or 401(k), where distributions are invisible to you.
November and December are when the two heavyweight moves happen, and they need to happen in that order.
Loss harvesting first. If you're carrying positions below your cost basis, selling them realizes a loss that offsets capital gains dollar for dollar without limit. If losses exceed gains, you can apply up to $3,000 against ordinary income, and anything left over carries forward indefinitely. The constraint is the wash sale rule, which disallows the loss if you buy a substantially identical security within thirty days before or after the sale. Thirty days on both sides, sixty-one days total, and it applies across all your accounts including your IRA and your spouse's accounts. Buying a similar but not identical fund keeps you invested without triggering it.
Roth conversions second, because you need to know your realized gains and losses before you can size a conversion properly. A conversion moves money from a traditional IRA to a Roth, you pay ordinary income tax on the converted amount this year, and everything after that grows and comes out tax-free with no required distributions.
The technique that makes conversions worth the trouble is filling a bracket rather than converting an arbitrary round number. If your projected taxable income puts you in the 12 percent bracket with $30,000 of room before the 22 percent bracket begins, converting $30,000 costs you 12 cents on the dollar, and you've moved that money permanently out of the taxable system. Converting $50,000 in the same year drags $20,000 through a rate nearly twice as high.
Two things to know about conversion timing. The deadline is genuinely December 31, with no extension and no grace period. This is different from IRA contributions, and people confuse the two constantly. And if you're anywhere near Medicare age, understand that Medicare premium surcharges use a two-year income lookback, so income you create in 2026 shows up in your 2028 premiums. That's not a reason to avoid conversions, it's a reason to size them with the surcharge thresholds in view.
December is for the items with hard deadlines and no flexibility.
Charitable giving is where most of the remaining leverage sits, and the mechanism most people get wrong is what they donate. Writing a check is the least efficient version. Donating appreciated securities you've held more than a year lets you deduct the full fair market value while never realizing the capital gain at all. The charity, being tax-exempt, sells with no tax consequence. You've eliminated a gain and taken a deduction on the same dollars. If you want the deduction this year but haven't decided which charities to support, a donor-advised fund lets you contribute now and distribute later.
If you're 70 and a half or older, a qualified charitable distribution sends money directly from your IRA to a charity, keeps it entirely out of your adjusted gross income, and counts toward your required minimum distribution. Keeping income off the top line rather than taking a deduction against it is often worth more, because adjusted gross income drives so many other thresholds.
Required minimum distributions themselves have to be out by December 31, and the penalty for missing one is severe, currently 25 percent of the shortfall and reduced to 10 percent if you correct it promptly. Custodians get backed up in late December. Take it in early December.
And then there are the two deadlines that aren't December 31, which is where people leave money on the table by assuming everything expires with the calendar.
Traditional and Roth IRA contributions for tax year 2026 can be made until the April 2027 filing deadline. The limit is $7,500, with an additional $1,100 if you're 50 or older. Health savings account contributions follow the same extended deadline, with 2026 limits of $4,400 for self-only coverage and $8,750 for family coverage, and the HSA remains the only account in the code with a triple tax advantage: deductible going in, tax-free growth, tax-free out for qualified medical costs.
The other one is the fourth quarter estimated tax payment, due in mid-January 2027. If you have self-employment income, significant investment income, or you converted a Roth in December, withholding alone may not cover you. The safe harbor is worth knowing precisely: you generally avoid an underpayment penalty by paying at least 100 percent of your prior year's total tax, or 110 percent if your adjusted gross income exceeded $150,000, regardless of what you actually end up owing. That's a rule you can satisfy with arithmetic rather than prediction.
Let me put a number on why the sequence matters. Someone in the 24 percent bracket who intended to increase deferrals by $6,000 and ran out of paychecks pays about $1,440 more than they needed to. Someone who had $10,000 of room in the zero percent capital gains bracket and never checked pays $1,500 when those shares eventually sell at 15 percent. Someone who bought a mutual fund in a taxable account in mid-November and caught a 6 percent distribution on a $50,000 position picked up a $3,000 taxable event for nothing. None of those are exotic scenarios, and all three are prevented by a calendar.
Which is the actual point. Every item above is a date with a decision attached, and the entire failure mode is that the dates arrive while you're busy. Put them in your calendar this week with reminders two weeks ahead of each one, and you've converted a subject people find intimidating into a series of appointments.
The prerequisite is being able to see your own position. You need projected income, current year-to-date deferrals, unrealized gains and losses by lot, and a running view of cash. Most people can't assemble that quickly, which is exactly why the December scramble happens. Connecting your accounts to a free tracker like Empower gives you net worth, holdings, and cash flow in one view, and its retirement and tax planning tools are the fastest way to get from "I should look at this" to an actual projection.
For the execution side, the fewer manual steps between decision and action, the more likely the decision survives contact with a busy December. Having contributions and allocations run on a schedule at a platform like M1 Finance means the only thing left for you to decide in Q4 is the tax question itself, not the mechanics underneath it.
And the calendar deserves to be automated rather than remembered. I run mine through Make, with each deadline firing a reminder and a short checklist of what has to be decided before that date passes. It took about forty minutes to build and it has saved me from the December version of every mistake above.
One honest caveat. I'm not your accountant, and several of the moves above interact with each other and with state tax rules in ways that depend entirely on your situation. Roth conversion sizing in particular is a genuine planning exercise, not a rule of thumb. The value of doing this work in September is partly that your CPA still has capacity to talk to you, which stops being true sometime in November.
September through December is roughly 17 percent of the year and a wildly disproportionate share of the financial outcome you control. Everything else this newsletter covers is weather. This is the part where you get to build the house.
I've mapped the entire fourth-quarter sequence into a single calendar: every deadline, what has to be decided before it, the 2026 limits and thresholds in one table, and the order the decisions have to happen in so that each one has the information it needs from the one before it.
Reply with the word WINDOW and I'll send it over.
Somebody in your life is going to do all of this in the last week of December. Send them this now instead. Three referrals unlocks the Money Systems Lab playbook library. Ten unlocks lifetime premium access.
Taylor Voss
Money Systems Lab