How Much House Can You Actually Afford?

Three constraints, and only one of them is the bank's

Ask what you can afford and almost every answer you get is a ratio: some percentage of your income, applied to a payment, run backwards into a price. That is one of three constraints, and for a great many buyers it is not the one that decides anything. The other two are the cash you have and the payment you are willing to live with. All three produce a maximum price. The one that produces the smallest maximum is the answer, and which one that is varies from household to household with no relation at all to income.

Set the case up once and reuse it. Household income of $90,000 a year, so $7,500 a month gross. A thirty-year loan at 6.5 percent. Property tax at 1.2 percent of the purchase price, home insurance at $1,200 a year, no association dues. Closing costs of 3 percent of the price, paid from savings. Minimum deposit of 10 percent. Everything the buyer has saved goes in, so the loan is 1.03 times the price minus savings — the extra 3 percent being the closing costs the savings must also cover. That single equation is what makes the three constraints comparable: each of them caps the same price, from a different direction.

Two buyers, same income, and the constraint switches

Buyer A has saved $30,000. The deposit constraint is arithmetic and brutal: the savings must cover the 10 percent deposit and the 3 percent of closing costs, so the price cannot exceed $30,000 divided by 0.13, which is $230,769. At that price the total monthly payment — principal, interest, tax and insurance — is $1,644, or 21.9 percent of gross income. The lender's 43 percent test would have allowed $441,343. The comfort test at $2,200 a month would have allowed $304,864. Buyer A is not constrained by income at all. Buyer A is constrained by having $30,000 instead of $60,000, and no amount of shopping for a better rate changes that.

Buyer B has saved $90,000 on the same income, and everything inverts. The deposit now permits $692,308 — a number no one on this income should look at twice. The lender's test permits $491,839. But Buyer B has decided that $2,200 a month is the most they will hand over, and that decision caps the price at $355,360. The binding constraint is the one Buyer B chose. That is not a weakness of the method; it is the method working. The buyer with savings has bought their way out of the two external constraints and is now negotiating only with themselves.

Buyer C completes the set. Same $90,000 income, same $90,000 saved, but a $450 car payment and a tolerance for $3,200 a month. Now the deposit permits $692,308, the comfort constraint permits $488,511, and the lender's 43 percent test — reduced by the car to a housing allowance of $2,775 — binds at $431,922. Three buyers, three binding constraints, one income. If a calculator gives you a single number without telling you which of the three produced it, it has hidden the only fact you needed.

Front-end and back-end are two different tests

The American convention uses two ratios and they measure different things. The front-end ratio is housing alone — principal, interest, taxes, insurance, and any association dues — over gross income. The back-end ratio adds every other recurring obligation: car loans, student loans, minimum card payments, court-ordered support. FHA's published benchmarks are 31 percent front and 43 percent back, both exceedable with documented compensating factors, and the two do not bind together. On our buyer with $90,000 of income and $90,000 saved, the 28 percent front-end convention permits $342,045 while the 43 percent back-end permits $491,839 — a gap of $149,794 depending on which test you were shown.

The back-end test is where existing debt does its damage, and the damage is larger than it looks because the ratio is a subtraction on the payment and the payment is a multiplier on the price. On our buyer, a $450 monthly car payment reduces the housing allowance from $3,225 to $2,775 and the maximum price from $491,839 to $431,922 — a loss of $59,918 of purchasing power. Per unit that is roughly $133 of price for every $1 of monthly obligation, or $13,315 for every $100 a month. It is the single most useful number in this article: before you borrow to buy anything with a monthly payment in the year before a mortgage application, multiply that payment by 133 and ask whether the thing is worth that much house.

The ratio is not the same rule in six markets

France is the strictest and the most explicit. The Haut Conseil de stabilité financière fixes an effort rate of at most 35 percent of income, and — this is the part that surprises people who translate the American ratio — loan insurance counts inside the 35 percent. Maturity is capped at 25 years, extendable to 27 where a deferral is tied to an off-plan purchase or to works worth at least 10 percent of the operation. Lenders may breach both criteria within a flexibility margin of up to 20 percent of new lending each calendar quarter, and that margin is itself directed: the great majority is reserved for buyers of a principal residence, with a defined share for first-time buyers.

The United States has moved in the opposite direction. The 43 percent debt-to-income limit that used to define a General Qualified Mortgage was removed by the CFPB's 2020 final rule and replaced with price-based thresholds tied to how far the loan's annual percentage rate sits above the average prime offer rate, with mandatory compliance from 1 October 2022. What remains at 31 and 43 percent are FHA's underwriting benchmarks, which are programme policy and openly exceedable with compensating factors. So the number that everyone quotes as the American rule is no longer the American rule; it is one programme's starting point.

Germany and Italy have no binding income ratio at all. In Germany the Ausschuss für Finanzstabilität and the Bundesbank now collect loan-level data on new lending, which for the fourth quarter of 2025 showed an average loan-to-value of 83 percent, an average debt-service-to-income ratio of just under 38 percent and household debt averaging 6.3 times annual income — figures published as monitoring, not as limits, with income-based instruments recommended rather than enacted. In Italy the Banca d'Italia holds the macroprudential toolkit but has not activated borrower-based measures; the well-known 80 percent figure comes from the CICR deliberation implementing article 38 of the banking act, which defines credito fondiario and can be raised to 100 percent with supplementary guarantees. It is a rule about the loan against the property, not against the income.

Spain and Portugal sit between the two poles and differ from each other. Spain has no statutory ratio: Ley 5/2019 obliges the lender to assess and document creditworthiness, the Banco de España's consumer guidance points at a payment of roughly 30 to 35 percent of net income, and 80 percent of appraised value is the practical financing ceiling because of how mortgage funding is structured rather than because a rule forbids more. Portugal is the one market here with an explicit numerical recommendation and it was tightened this year: under the Recomendação Macroprudencial n.º 1/2026, applying to solvency assessments from 1 August 2026, debt service to income is limited to 45 percent computed after an interest-rate shock and on expected income, loan-to-value runs to 90 percent for own permanent housing and 80 percent otherwise, and maturity runs to 40 years. It is a comply-or-explain recommendation, not a statute, which in practice banks follow.

What the French rule does to our buyer

Run Buyer B through the French rule rather than the American one and two things change at once: the ceiling rises to 35 percent of income, and the maturity falls from thirty years to twenty-five. Add loan insurance at an assumed 0.36 percent a year of the original capital — an assumption, since the premium depends on age, health and cover — and the whole 35 percent, which is $2,625 a month, must contain the repayment and the insurance together. The answer is a maximum price of $448,768, of which the insurance line is $112 a month. The shorter term costs purchasing power and the higher ceiling more than gives it back; the insurance inside the ratio quietly takes some of it away again.

The wider lesson is that a ratio is only meaningful together with three definitions: what counts as income, what counts as a charge, and over how many years the payment is computed. Change any one and the same headline percentage produces a different price. Gross or net income; twelve or thirteen or fourteen monthly instalments; insurance in or out; a stressed rate or the contract rate; expected income or current income. Portugal's 45 percent is computed after an interest-rate shock, which makes it stricter than a 45 percent computed on today's rate. France's 35 percent includes insurance, which makes it stricter than a 35 percent that does not. Comparing the numbers without the definitions is how people conclude that one country's lenders are generous when in fact they are measuring a different thing.

The constraint you set yourself is the one that survives

The deposit constraint moves on the timescale of saving, which is years. The lender's ratio moves on the timescale of rate changes and rule revisions, which is quarters. The comfort constraint moves when you change your mind, which is never, if you write it down before you start looking. That asymmetry is why the third constraint deserves more respect than it usually gets: it is the only one of the three that is a statement about the life you want rather than about the money that exists.

There is a practical order of operations that falls out of all this. First compute the deposit ceiling, because it is a division and takes ten seconds. Second compute the payment you would be content with in a bad year — not a good one — and turn it into a price. Third, and only third, ask a lender what they will allow, and treat their answer as an upper bound you have no obligation to reach. If the lender's number is the smallest of the three you have a debt problem to solve before you have a housing problem to solve, and the fastest fix is usually clearing the obligation with the largest monthly payment rather than the largest balance. That distinction — payment, not balance — is the whole of the back-end ratio in one sentence.

Frequently asked questions

Is the deposit really the binding constraint for most first-time buyers?

Very often, yes, and the arithmetic shows why. The deposit constraint is a division: savings over the sum of the minimum deposit percentage and the closing-cost percentage. Because the divisor is small — 0.13 in our case — the answer is highly sensitive to the numerator. Doubling savings doubles the ceiling, exactly. Meanwhile the income constraint barely moves. Buyer A in this article could raise their income by half and still not change their maximum price at all, because income was never what was stopping them. That is why programmes aimed at first-time buyers almost always attack the deposit rather than the ratio.

Should I clear my car loan before applying?

It depends on what clearing it costs. In our case a $450 monthly payment costs $59,918 of purchasing power, so if the remaining balance is well under that, repaying it converts cash into borrowing capacity at a favourable rate — but only if the cash was not part of the deposit, because removing it from savings tightens the deposit constraint by roughly $7.69 of price per dollar removed. The right test is to compute both constraints again with the money moved and see which ceiling falls. If the deposit constraint is already binding, paying off other debt buys you nothing at all.

Why does a longer term not always help?

Because it only relaxes the payment constraints, and there are three constraints. Stretching from twenty-five to thirty years lowers the monthly payment and therefore raises the ceiling set by the ratio and by comfort — but it does nothing whatever to the deposit ceiling, which is pure division. For Buyer A, whose deposit binds, a longer term changes the maximum price by exactly zero. It is also capped by rule in some markets: France limits maturity to 25 years, Portugal to 40. And in every market it buys the lower payment with more total interest, which our article on paying a loan off early works through in detail.

Do the property tax and insurance assumptions matter much?

More than most people expect, because they scale with the price and therefore compete with the loan for the same monthly budget. At 1.2 percent of price, the tax on a $400,000 home is $400 a month before a single unit of interest is paid — roughly 15 percent of a $2,625 allowance. Local rates vary by a factor of several between jurisdictions, and in some markets the recurring property charge is far smaller than the transaction taxes on the way in. Put your own figures in: a calculator using a default rate for a jurisdiction that charges half or double will be wrong by tens of thousands on the ceiling.

Can a lender exceed the published limit?

In every market described here, yes, but through a different door. In France the excess is quantitative: lenders have a flexibility margin of up to 20 percent of new lending each quarter, steered towards principal residences and first-time buyers, so a file above 35 percent is possible but competes for a scarce allocation. In Portugal the recommendation is comply-or-explain and includes its own tolerance. In the United States exceeding a benchmark is a documentation exercise using compensating factors. In Germany, Spain and Italy there is no binding ratio to exceed, only the lender's own policy and the general duty to assess creditworthiness. Being inside a limit is never the same as being able to afford it.