Taxed on Money They Never Received

Most of the conversation about taxing Social Security is about the percentage. Is it fifty percent? Is it eighty-five? That part is well covered, and most advisors can recite the provisional income thresholds from memory.

Almost nobody checks what the percentage is applied to.

Medicare premiums come out of a Social Security benefit before the check is issued. They do not come out of the amount the IRS taxes. Form SSA-1099 reports net benefits in Box 5, and Box 5 is Box 3 minus Box 4 — where Box 4 is benefits repaid to Social Security, not premiums. The premiums appear only in the description of Box 3, itemizing where the gross benefit went. Nothing anywhere subtracts them from the taxability base.

The result is a retiree paying income tax on money that was routed to Medicare before it ever reached them.

In the scenario modeled below, a single filer entitled to $46,800 in benefits at 65 receives $38,017 after Medicare — and is taxed on $39,780. The taxed figure is $1,763 higher than the amount that arrived, in the first year of benefits, before any inflation assumption enters the picture.

The lower panel is where this ends up. By 81 the benefit nets to zero, Medicare is billed directly rather than withheld, and the IRS still taxes $54,609 of a benefit the client no longer receives. The rest of this piece is how it gets from the top panel to the bottom one — and the mechanism is not the one most projections assume.

The obvious objection is that Medicare premiums are deductible. They are — as a medical expense, on Schedule A, and only to the extent total medical expenses exceed 7.5% of adjusted gross income.

Two conditions have to hold before that deduction is worth anything: the client has to itemize, and their medical spending has to clear the floor. Most retirees take the standard deduction. For them the premium is taxed as income and never recovered anywhere on the return.

This is the same shape as the charitable-giving problem: a deduction exists, the client cannot reach it, and the planning conversation has to happen before the money moves rather than after.

The interesting part of this scenario is not that the gap exists at 65. It is what happens at 77.

Required minimum distributions begin at 75. Modified adjusted gross income more than doubles that year, from $103,430 to $234,773. The Medicare bill does not move — because the income-related surcharge runs on MAGI from two years earlier.

It lands at 77. Total Medicare cost goes from $18,451 to $47,368 in a single year, and net Social Security falls from $39,739 to $11,986.

Nothing about the client's behavior changed. They took a withdrawal they were legally required to take, and the consequence arrived two years later, in a different line item, attached to a different program. That two-year delay is precisely why it is almost never connected back to its cause — and it is the window in which planning is still possible.

Premiums keep compounding. The benefit keeps growing at a 2% cost-of-living adjustment. Those two lines cross.

At 81, Medicare exceeds the entire Social Security benefit. The net benefit is zero, the premium stops being something withheld from a check and becomes something billed directly, and the difference has to be funded from somewhere else. By 85 that shortfall is $30,430 a year.

And the benefit is still taxed. At 85 the scenario reports $59,111 of taxable Social Security in a year when the client receives none of it and writes a check for the privilege.

Every figure past 65 depends on the scenario's inflation assumptions: Part B at 7.3% a year — the sixty-year historical average — and Part D at 6%. A lower assumption moves the crossover later. It does not remove it, because the gap at 65 is already there without any inflation at all.

None of this is a prescription. It is a projection problem, and the fix is to run it while the client can still act.

• Model Social Security net of Medicare, not gross. A plan that spends the statement figure is spending money that never arrives.
• Project the MAGI that required distributions will create, and look at the surcharge two years forward from the year it first spikes.
• Model whether the client will ever itemize in retirement. If not, the medical-expense deduction is not a mitigation and should not be treated as one.
• Run the crossover: the year Medicare exceeds the benefit. It is a cash-flow event that has to be funded from elsewhere, and most projections never surface it because they stop at the gross benefit line.

The premium is the part everyone models. The tax on the premium is the part nobody does.