HSA in Stocks or Cash When Mortgages Sit Above 7% and CPI Jumps 0.4%? The 20-Year Math on $8,750 (September 2026)
Picture Jordan and Priya. Both are 45, on a family high-deductible plan, in the 24% federal bracket, spending about $3,000 a year on real medical costs, with roughly 20 years until Medicare. (They're a worked example, not real people.) It's September 30, 2026, and they have the same tabs open that you probably do: a jobs report, a mortgage-rate article, and a column asking whether the stock market is about to wreck their retirement.
They have three HSA questions: how much to contribute, stocks or cash inside the account, and what changes at 65. None of the headlines answers those directly, but they do shift the trade-offs. Here is the math, with the caveats left in.
What the September 30 headlines actually say
| Data point | Latest figure | Source | What it changes for an HSA |
|---|---|---|---|
| Consumer Price Index | +0.4% in Aug 2026 | Bureau of Labor Statistics, "Major Economic Indicators Latest Numbers" | Makes cash sitting in an HSA more expensive to hold |
| Unemployment rate | 4.1% | BLS | A liquidity question, not a tax question |
| Payroll employment | +162,000 (preliminary) | BLS | Little direct HSA effect |
| Average hourly earnings | +$0.10 (preliminary) | BLS | About $208 a year if it stuck (2,080 hours), against an $8,750 family limit |
| Mortgage rates | "Steadily above 7%" | NerdWallet, "Mortgage Rates Today, Wednesday, September 30" | Sets the guaranteed return you give up by choosing the HSA |
Two cautions on reading this table. First, 0.4% is one month. If it repeated twelve times it would compound to about 4.9% a year (1.004¹² ≈ 1.049), but one print is not a forecast. NerdWallet's own summary says rates are in a holding pattern while inflation is "still running hot," which is a description of now, not of next year. Second, nothing here changes the deduction. Your HSA tax break is set by your bracket, not by the CPI release.
Dial 1: How much to contribute
The 2026 limits are $4,400 self-only and $8,750 family, plus a $1,000 catch-up at 55 and older. Here is what the first leg of the triple-tax advantage is worth on the family limit:
| Federal bracket | Income-tax savings on $8,750 | Plus 7.65% payroll tax (payroll route only) | Year-one total |
|---|---|---|---|
| 22% | $1,925 | $669 | $2,594 |
| 24% | $2,100 | $669 | $2,769 |
| 32% | $2,800 | $669 | $3,469 |
The payroll-tax savings apply only when contributions run through an employer cafeteria plan. Direct contributions and self-employed contributions get the income-tax deduction but not the FICA savings. A few states, including California and New Jersey, don't follow the federal treatment, so your state line could be zero.
Against a mortgage above 7%: $8,750 of extra principal saves about $612.50 in first-year interest (8,750 × 7%). Jordan and Priya's HSA deduction is $2,100 on day one, about 3.4 times larger, and it doesn't depend on any market. The mortgage side is also guaranteed, though, and it never asks whether you have medical expenses. Both sides are real. The long-run comparison depends on a break-even return that I walked through in the 5.0% break-even math for a 7% mortgage.
Against a rewards card: NerdWallet's sponsored pieces on the IHG Premier card (fourth night free) and Bilt's new launch are useful for what they are, which is marketing. Suppose a fourth night is worth $200 (my assumption, not IHG pricing). Jordan and Priya's $2,100 deduction equals ten of those nights, earned before any growth. I broke this down in the hotel-redemption vs. HSA comparison. A card perk can still win if you would have booked the stay anyway and the fee is covered. It usually just isn't competing for the same dollar.
When you should contribute less than the max:
- You have no emergency cash and a surprise bill would land on a 25% credit card.
- Your payroll cash flow is tight. HSA contributions can be changed mid-year, so this isn't all-or-nothing.
- Unemployment at 4.1% and an uncertain job means liquidity matters more to you than to someone with a stable paycheck.
This is the kind of analysis Trivexano runs for you, so you don't have to build the bracket-by-bracket spreadsheet yourself.
Dial 2: Stocks or cash inside the HSA
The contribution is the easy part. What the money does after it lands is where two people with identical contributions end up $100,000 apart.
Here are 20 years of $8,750 annual contributions ($175,000 total, contributed at year-end) at a range of example returns. These are illustrations, not forecasts. The last column discounts by an assumed 3% inflation, so it's in today's dollars:
| Assumed annual return | Balance after 20 years | In today's dollars (3% inflation) |
|---|---|---|
| 2% (cash-like) | $212,602 | $117,713 |
| 4% | $260,558 | $144,265 |
| 6% | $321,874 | $178,214 |
| 7% | $358,708 | $198,608 |
| 8% | $400,418 | $221,702 |
Cash at 2% has a real return of roughly −1% a year under a 3% inflation assumption. If the August pace of 0.4% a month were sustained, it would be about −2.8%. That's the quiet cost of a "safe" HSA. I covered the full cost of that choice in the AI-bubble cash move math.
The 7% vs. 2% gap depends on your time horizon
| Years to spend | 7% invested | 2% cash | Gap |
|---|---|---|---|
| 10 | $120,894 | $95,372 | $25,522 |
| 20 | $358,708 | $212,602 | $146,106 |
| 30 | $826,534 | $354,971 | $471,563 |
If you're 10 years out, the gap is $25,522. At 30 years it's $471,563. A 45-year-old and a 60-year-old can both say "I have an HSA" and need different allocations.
The AI-bubble stress test
Mr. Money Mustache's "Will the AI Bubble Destroy our Retirement?" starts from the observation that the market keeps surprising us. Crashes scare people, and so do record highs. I won't predict which comes next, so I ran a blunt test on Jordan and Priya's 7% path: one 30% crash in a single year, and 7% in every other year.
| Scenario (20 years, $8,750/year) | Ending balance | vs. cash at 2% |
|---|---|---|
| 7% every year, no crash | $358,708 | +$146,106 |
| 30% crash in year 10 | $282,427 | +$69,825 |
| 30% crash in year 19 | $240,933 | +$28,331 |
| All cash at 2% | $212,602 | baseline |
A year-10 crash costs about $76,000 against the smooth path but still leaves them $69,825 ahead of cash, because later contributions buy at lower prices. A year-19 crash costs about $117,775 and shrinks the lead to $28,331. The same crash does much more damage when it lands just before you spend.
Two lessons. Staying fully in cash from the start gives up the most. Staying fully invested up to the spending date carries the most timing risk. That's why many people split the account: a cash buffer for near-term medical bills (for Jordan and Priya, call it $3,000, about one year of their spending) and the rest invested, stepping more toward cash as the spending date nears.
There's a behavioral caveat too. The 7% path only works if you don't sell after a 30% drop. If you know you would, your realistic return is lower and your allocation should be more conservative. That's a valid input, not a failing.
Bond yields matter here as well, and I looked at a 5-gate checklist for allocation and Medicare in a different rate environment.
Dial 3: Medicare coordination at 65
Once you're enrolled in any part of Medicare, you can't contribute to an HSA. The trap is that Part A can backdate up to six months when you enroll after 65. If you claim Social Security or apply late, contributions in the backdated months can become excess contributions.
Worked example: at the family limit, six months of contributions is about $4,375 (729.17 × 6). At 24%, that's $1,050 of deduction you don't get, plus a 6% excise tax of $262.50 if the excess stays in the account. The worst case is about $1,313 for a single year. My longer write-up of the 6-month lookback rule covers the edge cases.
After 65, the account's second life begins:
- The balance keeps growing tax-free even though contributions stop.
- Medicare Part B, Part D, and Medicare Advantage premiums are qualified expenses. Medigap premiums are not.
- Non-medical withdrawals after 65 are taxed as ordinary income but carry no 20% penalty.
The key allocation insight is that your spending date isn't 65. Premiums and out-of-pocket costs flow out over decades. So only the near-term slice needs the cash buffer, and the rest can reasonably keep a long-horizon allocation. That's a judgment call, and a retiree with a pension and a retiree with no other assets would make it differently.
When the answer flips
- Horizon under 10 years: the gap to cash is $25,522, so a cash-heavy mix is defensible.
- Non-conforming state or self-employed: the tax savings shrink. Drop the payroll-tax line for self-employment, and use your actual state rate.
- Bracket change at retirement: the deduction is worth more at 32% than at 22% ($2,800 vs. $1,925 on $8,750), which favors contributing in peak-earning years.
- Fees: an account with a 1% fund expense ratio and a flat platform charge behaves differently from a low-cost index fund. Another post compares the HSA to a 1% advisor fee.
One more rough number. Over 20 years, Jordan and Priya's $175,000 of contributions at 24% saves $42,000 in income tax. At 7%, the $183,708 of growth would face roughly $27,556 at a 15% capital-gains rate in a taxable account, a simplification that ignores dividends. That's about $69,556 of the triple-tax advantage, and about $13,388 more if the payroll route applies. It's a rough figure, not a projection. The 7%/20-year result is also in the 5-input calculator walkthrough.
Your numbers will differ
Everything above uses one couple, one bracket, one return assumption, and one crash. Yours will differ based on your:
- Federal and state bracket
- Payroll vs. direct contribution route
- Years until you spend the money (and whether it's one date or a stream)
- Current cash/invested split
- Mortgage rate, emergency fund, and other debt
- Comfort with a 30% drop
- Medicare start date
The headlines will keep moving. The deduction, the tax-free growth, and the tax-free qualified withdrawals won't, but how much of each you capture depends on decisions only you can size. You can model this for your specific situation at Trivexano, so the next decision comes from your inputs rather than a rule of thumb. The math should speak for itself, and it only speaks clearly with your numbers in it.
Sources
- Major Economic Indicators Latest Numbers — Bureau of Labor Statistics
- Mortgage Rates Today, Wednesday, September 30: Steadily Above 7% — NerdWallet
- Will the AI Bubble Destroy our Retirement? — Mr. Money Mustache
- 3 Reasons This Hotel Credit Card Deserves a Spot in My Wallet — NerdWallet
- Why Bilt’s New Launch Could Be the Most Rewarding Card to Rule Them All — NerdWallet