What is a good debt-to-income ratio?

Two ratios, and lenders look at both

Debt-to-income is a single division — monthly debt payments over gross monthly income — but it is quoted as two different numbers, and confusing them is the most common way people arrive at a figure that does not match the one an underwriter produces.

The front-end ratio counts housing only: the mortgage payment, property tax, buildings insurance, and any service charge or association fee. The back-end ratio counts housing plus every other required monthly payment. When someone says "my DTI is 31%" without qualifying it, they almost always mean the back-end figure, because that is the one that decides applications.

The traditional guideline pairs them: 28% front-end, 36% back-end. That is where the "28/36 rule" comes from, and it is old enough that it long predates the underwriting software now doing the actual work. It survives because it encodes something durable — that housing should be the largest single claim on your income but not so large that nothing else fits behind it.

The gap between the two is the part worth attention. A front-end of 26% with a back-end of 44% describes someone whose house is affordable and whose car and cards are not. A front-end of 30% with a back-end of 33% describes someone stretched on housing and clean everywhere else. Those are different problems with different fixes, and a single blended number hides which one you have.

Gross income, not what arrives

The ratio uses gross monthly income — before tax, before pension contributions, before health premiums and before anything else deducted at source. This is not a technicality. It is the single largest source of disagreement between a self-calculated ratio and a lender-calculated one, and it always runs the same way: using take-home pay makes your ratio look considerably worse than the one being underwritten.

The reason is consistency rather than generosity. Deductions vary enormously between two people earning identical salaries — different tax codes, different pension rates, different benefit elections — so a ratio built on net pay would compare borrowers against a moving baseline. Gross income is the same measurement everywhere, which is what makes a threshold like 43% mean anything at all.

For salaried employment the figure is simply annual salary divided by twelve. Everything else needs care. Variable pay — bonus, commission, overtime — is normally averaged over two years and only counted where there is a documented history of receiving it, so a first strong bonus year often contributes nothing. Self-employed income is typically taken from filed accounts, again averaged, and again after allowable expenses rather than as turnover.

The awkward consequence is that people whose income rose recently are underwritten on the income they used to have. That is a real constraint rather than an oversight, and the practical response is to know which number a lender will use before you build a plan on the one in your bank account.

What counts as debt, and what surprisingly does not

The rule is narrower than most people expect: DTI counts recurring obligations that appear on a credit file or a court order. It does not count the cost of living.

So groceries, utilities, fuel, phone contracts, streaming subscriptions, childcare, health insurance premiums and pension contributions are all excluded — however large they are and however unavoidable they feel. A household spending significant sums on childcare has exactly the same DTI as one spending nothing, which is one of several reasons a good ratio is not the same thing as an affordable payment.

What does count: the proposed housing payment including tax and insurance, car loans and leases, student loans, personal loans, credit card minimum payments, any instalment agreement, and court-ordered obligations such as child support or alimony. Cards are counted at the minimum due rather than at what you actually pay or the full balance, which means a large balance you clear every month still adds its minimum to the ratio.

Two cases catch people out. A student loan in deferment or on an income-driven plan is frequently not counted at zero — many lenders substitute a percentage of the outstanding balance, so a loan currently costing nothing per month can arrive in the calculation as several hundred. And a car lease is counted for its full remaining term regardless of how few payments are left, because the obligation exists until it does not.

The other direction is worth knowing too. A loan with a small number of payments remaining — commonly ten or fewer — is often excluded entirely, which is occasionally the cheapest way to move a ratio that is marginally over a threshold.

The thresholds that decide an application

The 36% guideline is advice. The number that functions as a gate is 43%, and it became conventional because it was written into mortgage regulation as the ceiling for a category of loan carrying particular legal protections for the lender. The rules have since moved toward a pricing-based test, but the figure outlived the regulation that created it and remains the point at which most lenders start asking for a reason.

Above 43%, approval is usually still possible and increasingly depends on what else is true about the application. Automated underwriting at the major agencies commonly extends to around 50% where there are compensating factors: substantial reserves, a large deposit, a long and clean credit history, or income documented as stable over years rather than months. Government-backed programmes are frequently more permissive still, sometimes well past 50% on a manual underwrite.

Which is why the honest framing is not a pass mark. Below roughly 36% the ratio is simply not the constraint on your application. Between 36% and 43% it is one factor among several. Past 43% it becomes the thing that has to be argued around, and the strength of the rest of the file decides the outcome.

It is also worth separating what you can borrow from what you should. A 45% back-end ratio can be approved and still leave a household with very little room, because — as the previous section describes — the ratio is blind to childcare, commuting and every other real cost. The threshold protects the lender against default. It was never designed to tell you whether the payment is comfortable.

Moving the number, and the trap in doing it

DTI is a fraction, so there are two levers, and one of them is far more responsive than people assume.

The counterintuitive part is that paying down a large loan usually does nothing. An instalment loan has a fixed monthly payment; reducing the balance of a car loan from eighteen thousand to twelve thousand leaves the payment exactly where it was, and the ratio therefore unchanged. Only clearing it outright removes the payment from the numerator. Directing a windfall at the biggest debt is the instinct, and for DTI purposes it is often the least effective thing you can do with the money.

What moves the ratio quickly is eliminating small obligations with disproportionate monthly payments. A card with a two-thousand balance and a sixty-a-month minimum, and a personal loan with three thousand left at three hundred a month, cost five thousand to clear and remove three hundred and sixty from the numerator. On a seventy-five-hundred gross monthly income that is nearly five percentage points — the difference between arguing about a threshold and being comfortably inside it. Sort candidates by monthly payment relative to payoff cost, not by balance.

Two things to avoid while an application is live. Do not open new credit: a new card or car loan adds a payment immediately and the ratio is re-checked before completion, which is a genuinely common way a conditional approval falls over at the last step. And be careful about closing cards you have just cleared, since that affects credit utilisation and history without improving DTI at all — a paid-off card contributes nothing to the numerator whether it stays open or not.

On the income side, only documented and durable increases help. A raise with a payslip behind it counts; an expected bonus generally does not. Adding a co-borrower brings their income and all of their debts, so it improves the ratio only when their own ratio is better than yours — which is worth checking before assuming a second income helps.

Questions