The Unrecoverable-Cost Framework

To compare renting and buying fairly, you must compare like with like. Folk wisdom says rent is “money thrown away” while a mortgage “builds equity.” Half of that is correct: rent is indeed an unrecoverable cost. The misleading half is the unspoken implication that owning has no unrecoverable costs. It has many, and most buyers never add them up.

Each year, ownership burns money you will never see again — money that builds no equity. A widely cited rule of thumb, the 5% rule, estimates this annual unrecoverable cost at roughly 5% of the property’s value. The rule was built with three terms; add insurance as a fourth, because it has become a major expense:

Property tax –- about 1%

Varies sharply by state. California’s Proposition 13 pins the base rate near 1.1–1.2% of assessed value, with annual assessment growth capped at 2%. The SALT deduction relief written into OBBBA (Sec. 70120, IRC §164(b)(7)) — a headline cap of $40,400 in 2026, growing 1% per year through 2029 before snapping back to $10,000 in 2030 — looks generous on the marquee, but for the readers of this book it largely is not: the cap is phased down by 30 cents on every dollar of MAGI above $505,000 in 2026 (the threshold is indexed 1% a year too), with a $10,000 floor, so anyone clearing roughly $606,300 of MAGI is back at the old limit. On top of that, OBBBA’s new “2/37” rule (section “Tax Deductions of The Mortgage Interest”) caps the tax value of itemized deductions at about 35 cents per dollar for top-bracket filers. Net effect: on a property-tax bill of any size, most of the outflow is genuinely unrecoverable.

Maintenance –- about 1%

A conservative long-run estimate of what it costs simply to keep the asset from degrading: roof, systems, paint, and the slow grind of entropy. Older or larger homes run higher.

Insurance –- 0.3% and climbing

The term the classic 5% rule folds into maintenance and then forgets, which was defensible when homeowners policies were cheap, but is indefensible now. Insure the dwelling for full replacement cost instead of market value or loan balance (section “Homeowner’s and Renter’s Insurance”), and budget the premium as its own line. In a catastrophe-exposed market the number stops being small: admitted carriers have withdrawn from wildfire and coastal wind zones, and the surplus-lines or state-plan-plus-wrapper replacement routinely costs three to five times the policy it replaced (section “Wildfire, Hurricane, and the California Insurance Crisis”). A $2M home in a California high-severity fire zone can carry $20,000 of annual premium — 1% of value by itself, and rising faster than any other term in this formula. Underwrite the premium you will pay in year five, not the initial quote at closing.

Cost of capital –- $c_k$

The largest, most overlooked, and most rate-regime-sensitive piece. It has two halves: the mortgage interest you pay on the borrowed portion, and the opportunity cost on your own equity — the return your down payment could have earned in a diversified portfolio instead of sitting in drywall. Express it as a weighted average so it tracks your actual capital structure:

$$c_k=\frac{D}{V}\left[i_m(1-{\tau }_{\text{eff}})-\pi \right]+\frac{E}{V}{r}_e$$

where $D∕V$ is the loan-to-value ratio, $i_m$ the nominal mortgage rate, ${\tau }_{\text{eff}}$ the effective marginal tax benefit on interest (zero if you no longer itemize, capped at 35% for top-bracket itemizers under the 2/37 rule, and zero on interest attributable to acquisition debt above $750,000), $\pi$ expected inflation, $E∕V$ the equity share, and $r_e$ the real long-run expected return on a diversified portfolio. Subtracting $\pi$ from the debt leg is not an embellishment: the rent you are comparing against escalates with inflation while a fixed mortgage payment does not, so the two legs have to be stated on the same real basis or the comparison is rigged. In a 3–4% mortgage regime $c_k$ runs 1–2% and the full unrecoverable cost sits near 4–4.5%, under the advertised 5%; at 6–7% mortgages $c_k$ is 3.5–4% and the total nears 6%; with no deductible interest at all it passes 6.5%, and the 5% rule is then optimistic by a third. Plug in today’s actual mortgage rate, your real effective tax position, and a realistic forward equity premium; do not trust the historical number printed in any rule of thumb, including this one.

Write the annual unrecoverable cost of owning, $C_h$, as:

where $V$ is the property value, $t$ the property-tax rate, $m$ the maintenance rate, $s$ the insurance rate, $c_k$ the blended cost of capital from the formula above, and $g$ the expected real appreciation of the property. The baseline rule sets $g=0$, which is close to the long-run U.S. record — real house prices have historically tracked inflation with only a slim premium — so if you want to plug in a positive $g$, you are making a forecast and should say so out loud. The homeowner’s tax benefits already live inside $c_k$ through ${\tau }_{\text{eff}}$; do not count them twice.

The decision rule follows immediately. Compare $C_h$ with the annual rent on a genuinely comparable property. If rent is meaningfully below $C_h$, the mathematically optimal move is to rent and invest the difference — that gap is real money you can compound elsewhere. If rent runs well above $C_h$, buying is the cheaper way to consume that shelter. Note what is not in this comparison: mortgage principal. Principal repayment is a transfer from one of your pockets to another — forced savings, not a true cost — so it does not belong in $C_h$.

Worked example. A $2,000,000 house, $1,500,000 mortgage at 6.5% nominal, $500,000 of your own equity, expected inflation 2.5%, expected real portfolio return 5%, and a top-bracket itemizer whose deduction is capped by the 2/37 rule — but only on the first $750,000 of acquisition debt, so half the loan generates no benefit and ${\tau }_{\text{eff}}\approx 0.175$:

Set property tax at the Prop 13 base of 1.1%, maintenance at 1%, insurance at 0.35%, and $g=0$:

The example is federal-only. California allows interest on the first $1,000,000 of the loan against its own 9.3–13.3% rates, which lifts ${\tau }_{\text{eff}}$ to roughly 0.25 and trims $c_k$ to about 3.1%; the chapter keeps the federal figure because it is the conservative one and because California’s own itemized-deduction phase-out shaves the state leg for exactly the incomes that reach it. That is roughly $9,750 a month of pure burn before a single dollar of principal. If the comparable rental is $7,500 a month, renting and investing the $2,250 difference wins on arithmetic, and the only question left is how much you are willing to pay for the things arithmetic does not capture. Move the same house into a high-severity fire zone — $s=1\%$ — and $C_h$ clears $130,000, or $10,800 a month, without changing anything else about the property. Insurance is now large enough to flip a marginal buy-versus-rent decision on its own, which is exactly why it belongs on its own line instead of buried inside a maintenance assumption from a gentler era.

One benefit does sit on the ownership side of the ledger: imputed rent. By living in a home you own, you collect the rent you would otherwise pay a landlord — effectively a dividend you pay yourself, and one the IRS does not tax. That is a genuine advantage. But it is the return that must be measured against the $C_h$ computed above; it is not a license to ignore the cost. The rigorous, dollars-and-cents version of this comparison — built on net present value — follows immediately below.