A mortgage lender computes two debt-to-income ratios on every file, and they routinely disagree. One borrower in this guide has a front-end ratio of exactly 28.00% — dead on the classic housing guideline — and a back-end ratio of 43.85% that fails the line lenders watch most. Another runs a 33.00% front-end that breaks the housing guideline while carrying one of the healthiest back-end ratios of the three. This guide defines both numbers precisely, works each one step by step on real households, and shows which programs judge which ratio — so you can read your own file the way an underwriter will.

Two ratios, two different questions

The front end dti — the front end debt to income ratio — asks one narrow question: how much of this paycheck does the roof take? Its numerator is the housing payment alone: rent, or the full mortgage payment with principal, interest, property tax, homeowners insurance, HOA dues and mortgage insurance combined. The back end dti asks the wider question: how much of the paycheck is already spoken for by everything? Its numerator is the housing payment plus every other contractual debt payment — car loans and leases, credit card minimums, student loans, personal loans, support payments. Both divide by the same figure: gross monthly income, before tax. When a lender quotes a single “DTI”, they almost always mean the back-end; the front-end lives in the mortgage file specifically.

The two numbers catch different failure modes. A household can have a modest house and a dangerous total load — three car payments and a stack of card minimums behind a reasonable rent. Another can have no debts at all and a housing payment that eats a third of gross pay in an expensive city. Neither ratio alone sees both households clearly, which is why mortgage underwriting keeps both.

How to calculate the front end ratio (the housing ratio)

The front end ratio and the housing ratio are two names for one calculation, so how to calculate front end ratio and how to calculate housing ratio are the same three steps. Take the household this guide returns to — call it Household B: gross income of $6,500 a month, buying with a housing payment of $1,820.

Step 1 — build the true housing payment. Add every component, not just principal and interest: principal and interest $1,430, property tax $240, homeowners insurance $110, HOA dues $40, mortgage insurance $0 — a total of $1,820. Leaving the small lines out is the most common front-end error; underwriters rebuild the payment from the tax certificate and the insurance quote, and their version is the one that counts.

Step 2 — divide by gross monthly income. $1,820 ÷ $6,500 = 0.28.

Step 3 — express it as a percentage. 28.00%. Household B’s housing ratio sits exactly on the classic 28% guideline — the old rule of thumb that housing should take no more than 28 cents of each gross dollar. By this ratio alone, the file looks tidy.

Two refinements separate a borrower’s estimate from an underwriter’s figure. First, the swap: the front-end ratio on an application uses the proposed housing payment, not the current one. A renter paying $1,450 on $5,200 of gross income has a current housing ratio of 27.88% — but if the proposed mortgage payment, fully loaded with tax, insurance and mortgage insurance, is $1,690, the front-end ratio on the file is 32.50%, and the 27.88% appears nowhere in it. Second, the escrow lines move: property tax and insurance are estimated at application and trued up afterwards, which is why a front-end ratio computed to two decimals in January can drift a few tenths of a point by closing without anyone making an error. Compute with the fullest version of the payment you have, and treat the second decimal as an estimate, not a measurement.

The same household, back-end

Now the wider lens. Household B’s remaining payments:

Household B: $2,850 of debts on $6,500 of gross income is a back-end ratio of 43.85% — against a front-end ratio of 28.00% on the same file.

Total debts of $2,850 against the same $6,500 of income: $2,850 ÷ $6,500 = 43.85%. The back-end ratio is nearly sixteen points above the front-end, and it is the back-end that crosses the famous 43% line. Nothing about the house changed — the disagreement is the debt load behind it. This is the single most common DTI surprise in mortgage applications: a borrower who has carefully kept housing at 28% discovers the file was being judged on 43.85% all along.

Pass one, fail the other: both directions

All three households computed with the Debt-to-Income Ratio Calculator; every figure is reproducible in the tool.

Household C runs the disagreement in reverse. In a high-cost market its housing payment is $2,970 on $9,000 of income — a 33.00% front-end that no classic guideline blesses. But behind the roof there is almost nothing: one $60 card minimum. Its back-end ratio is 33.67%, healthier than Household B’s and healthier than the worked household A’s 36.00%. Programs read these files differently precisely because the two ratios separate: a strict housing-ratio test punishes C and waves B through; a total-debt test does the opposite. Knowing which test your program applies tells you which number to work on.

A five-minute drill: both ratios from your own statements

Run the pair on a generic file to fix the method. Gross income $5,000 a month. Proposed housing payment, fully loaded, $1,300: front-end = 1,300 ÷ 5,000 = 26.00%. Now the recurring payments: the same $1,300 housing, a $289 car payment, a $75 card minimum and a $210 student loan payment total $1,874: back-end = 1,874 ÷ 5,000 = 37.48%. Read the pair: the housing cost sits comfortably under the classic 28% line, but the total load runs past the classic 36% — this file’s binding constraint is the non-housing payments, $574 a month of them, not the house. Your own drill needs four figures: gross monthly income (before tax — the number lenders use), the fullest housing total you can assemble, and the minimum or contractual payment from each statement. If a payment on your list is small, that is exactly the cheap fix an underwriter hopes to see: a $75 minimum retired is a point and a half of back-end ratio on this income, bought for the price of one balance.

Which ratio does each lender actually use?

Mortgage programs use both — with one famous exception. FHA judges both ratios against its tiers (31% housing and 43% total in base manual underwriting). USDA judges both, as a 29% PITI ratio and a 41% total debt ratio. Conventional underwriting classically quotes both — the 28/36 guideline — although Fannie Mae’s automated system weighs the back-end ratio and sets no separate front-end cap. VA is the exception: it sets no front-end ratio at all. A VA file is judged on the back-end figure as a guideline (41%) and, more decisively, on residual income. Auto and personal lenders use the back-end only, with the new payment you are applying for added to the numerator. And when no mortgage is involved at all — a card application, a personal loan — the front-end ratio simply does not exist; there is only the total load.

Reading the pair like an underwriter

Put the three households in this guide in front of each program and the pair’s division of labour becomes concrete. Household B (28.00% / 43.85%) passes FHA’s 31% housing tier without discussion, then needs one documented compensating factor to lift its back-end tier from 43% to 47% — the same file, waved through by the front-end test and workshopped by the back-end test. Under USDA it fails the 41% total debt standard and needs the automated system or a waiver. Under VA its front-end ratio is never even computed; the whole decision rides on the 43.85% against the 41% guideline and the residual income beneath it. Household C (33.00% / 33.67%) inverts the pattern: USDA’s 29% PITI ratio and FHA’s 31% housing tier both push back on the housing cost itself, while its total load would satisfy every back-end standard in this guide at a glance. The practical lesson is diagnostic. If your binding constraint is the front-end, the lever is the housing payment — a cheaper property, a larger down payment shrinking the loan, or waiting for tax and insurance estimates to firm up. If the binding constraint is the back-end, the lever is the other payments — and a smaller house bought while keeping three car payments does nothing for the ratio that is actually failing. Borrowers routinely fix the wrong ratio for months; the pair, computed honestly, tells you which one is yours in a minute. Run both of your numbers in the Debt-to-Income Ratio Calculator and read them against the program you actually intend to use — the pair, not either one alone, is your file.