Why Every Home Affordability Index Gives a Different Answer

In its October 2026 research update, the National Association of Realtors put its Housing Affordability Index at 105.4, meaning a family earning the median family income had about 5% more income than NAR requires to buy the median-priced existing home.² The Federal Reserve Bank of Atlanta’s Home Ownership Affordability Monitor, run on August 2026 data, found the median household would need 43.9% of its income to own the median home, far above the 30% line it treats as affordable.³ The National Association of Home Builders put the figure at 36% for the second quarter.⁴

All three are calculated correctly. They disagree because they divide different payments by different incomes and compare the result with different thresholds. Once you can see those three choices, the headline numbers stop contradicting each other, and you can tell which one, if any, describes your own purchase.

Three published indices, three answers

These are the most recent national readings available on October 25, 2026. Each comes from a different month or quarter, so treat them as a snapshot of each method rather than a precise comparison.

Measure	Latest reading	What the reading says
NAR Housing Affordability Index	105.4 (October 2026 update)	Above 100: the median family qualifies under NAR’s rule, with principal and interest at about 24% of income
Atlanta Fed HOAM	43.9% of income, index 68.3 (August 2026)	Below 100: the median household falls about $40,000 a year short of the income needed
NAHB/Wells Fargo Cost of Housing Index	36% of income, existing home (Q2 2026)	In NAHB’s “cost-burdened” band of 31% to 50%; 71% for a family earning half the median
Harvard JCHS required income	$120,800 (end of 2025)	Income needed to carry the median home at a 31% ratio; 32% of households earn that much

NAR’s implied payment share is 25% divided by 1.054. The HOAM income shortfall is from the Atlanta Fed’s national data file. Sources: NAR, Federal Reserve Bank of Atlanta, NAHB, Harvard Joint Center for Housing Studies.

What each index counts

Every payment-based affordability index makes the same four choices: how big a down payment the buyer makes, which monthly costs count, whose income goes in the denominator, and where the line between affordable and unaffordable sits.

	NAR HAI	Atlanta Fed HOAM	NAHB CHI	JCHS
Down payment	20%	10%	10%	3.5%
Costs counted	Principal and interest	P&I, property tax, insurance, mortgage insurance (0.558% of the loan)	P&I, property tax, insurance, mortgage insurance	P&I, 1.15% property tax, 0.35% insurance, 0.55% mortgage insurance
Income	Median family	Median household	Median family (HUD)	Household income distribution
Threshold	P&I at 25% of income	30% of income	30% or less; 31–50% burdened; over 50% severely	31% debt-to-income
Output	Index, 100 = just qualifies	Payment share and index	Payment share	Required income and share of households who earn it

NAR’s formula is the simplest. It finds the income a family would need for the principal-and-interest payment on the median home to equal 25% of income, then compares that with the median family income:¹

\text{HAI} = \frac{\text{median family income}}{12 \times M / 0.25} \times 100

where $M$ is the monthly payment on a loan of $L$ dollars at monthly rate $r$ over $n = 360$ months:

M = L\,\frac{r(1+r)^n}{(1+r)^n - 1}

The Atlanta Fed uses the same structure with a 10% down payment, a fuller payment, household income, and a 30% threshold drawn from the HUD standard.³ NAHB’s index, which replaced its long-running Housing Opportunity Index in 2024, reports the payment share directly instead of converting it to an index.⁴ JCHS asks a different question: how many households earn enough to carry the median home at the 31% ratio it describes as typical for lenders.⁵

The same home under two sets of rules

To separate method from data, take one synthetic case: a $500,000 home, a household earning $120,000, a 6.5% 30-year fixed rate, property tax of 1.1% of the price, and insurance of $1,440 a year. Apply both formulas to that one household and that one home.

NAR-style: $400,000 loan, $2,528 a month in principal and interest, 25.3% of gross income. Index about 99.
HOAM-style: $450,000 loan, $2,844 in principal and interest, $458 property tax, $120 insurance, $209 mortgage insurance. Total $3,632 a month, 36.3% of gross income. Index about 83.

For this synthetic household, 11 points of income separate the two answers. In order: the smaller down payment adds 3.2 points, property tax 4.6, insurance 1.2, and mortgage insurance 2.1. The threshold then moves the index on top of the payment. Scored against NAR’s 25% line instead of the Atlanta Fed’s 30%, the same $3,632 payment would produce an index of about 69 instead of 83.

An index of 100 has a narrow meaning. It says the representative household exactly meets that index’s rule under its assumptions. Raise the threshold from 25% to 30% and more markets cross 100 without any change in prices, rates, or incomes.

Family income and household income

NAR and NAHB divide by median family income. A Census family is two or more related people living together, so the measure leaves out people who live alone or with roommates. The Atlanta Fed chose household income because single-person households, which family income leaves out, account for a large share of housing activity.³

The gap is large. NAHB used a median family income of $106,800 for the second quarter of 2026.⁴ The Atlanta Fed’s August 2026 data imply a median household income of about $86,700.³ The dates and sources differ, but the family figure is roughly 23% higher, and dividing the same payment by the smaller number raises the burden by the same proportion. The synthetic $3,632 payment above is 41% of the NAHB family income and 50% of the HOAM household income.

Rates move payment measures more than price-to-income

The price-to-income ratio divides the median home price by median income. It contains no interest rate, which makes it useful for long historical comparisons and poor at describing what buyers pay each month. On a $400,000 loan, principal and interest cost $1,686 a month at 3%, $2,147 at 5%, and $2,797 at 7.5%. Moving from 3% to 7.5% raises the payment 66% with no change in price.

The national data show the same split. JCHS reports that the median existing single-family home cost 4.1 times the median household income in 2019, peaked at 5.0 in 2022, and eased to 4.7 in 2025.⁵ Over the same years, the Atlanta Fed’s annual average payment share went from 29.0% to 38.1% to 42.8%.³ Price-to-income rose 15% from 2019 to 2025 and the payment share rose 48%, and the payment measure kept climbing after the price ratio had turned down. Freddie Mac’s weekly survey put the average 30-year fixed rate at 7.32% on October 22, 2026, against 6.19% a year earlier.⁶

Neither measure is wrong. Price-to-income describes how expensive houses are relative to earnings. A payment index describes what it costs to finance one at today’s rate. A buyer who expects to refinance later cares about both; a buyer paying cash cares mostly about the first.

Price-to-rent measures valuation

Price-to-rent divides a home’s price by a year of rent on a comparable home. A ratio of 15 is a gross rental yield of 6.7%; a ratio of 20 is 5%. The ratio says nothing about whether a household can make the payment. It asks whether owning is expensive relative to renting the same housing, which is a valuation question.

Economists frame that question with the user cost of owning: what a homeowner gives up each year, expressed as a share of the price. Himmelberg, Mayer, and Sinai write it as¹¹

u = r_{f} + \omega - \tau\,(r_{m} + \omega) + \delta - g + \gamma

where $r_f$ is the risk-free rate (the return the owner’s money could earn elsewhere), $\omega$ the property tax rate, $\tau$ the income tax rate that applies to deductible mortgage interest $r_m$ and property tax, $\delta$ maintenance and depreciation, $g$ expected appreciation, and $\gamma$ a risk premium for owning. If owning and renting cost the same, annual rent equals price times $u$ , so the price-to-rent ratio that counts as fair value is $1/u$ . It changes whenever rates, taxes, or expected growth change. The approach builds on James Poterba’s 1984 asset-market model of how taxes and inflation affect the cost of owning.¹²

Using this method across 46 metro areas, Himmelberg, Mayer, and Sinai argued that price-to-rent and price-to-income ratios can mislead when real interest rates or expected growth differ across cities and years, and that by their measure most markets in 2004 were not overvalued.¹¹ Prices fell nationally from 2006, which shows how much the conclusion depends on the expected-appreciation term. User cost is the right framework, and its answer is only as good as the growth assumption you feed it.

Renting carries its own risk. Sinai and Souleles show that owning hedges against future rent increases, and that the hedge is worth more to households that expect to stay longer and that live where rents are volatile.¹³ A fixed-rate mortgage locks in principal and interest; taxes, insurance, and repairs still move. Buying and renting are both risky ways to secure housing. Our rent vs. buy analysis works through that comparison.

Every threshold is a convention

The cut-offs used to call housing affordable come from different institutions doing different jobs.

Threshold	Who uses it	What it measures
25%	NAR	Principal and interest only, median family
30%	HUD, Atlanta Fed, NAHB, JCHS cost burden	Housing costs relative to pre-tax income; JCHS applies it to actual costs reported in the Census survey
28% / 36%	Traditional lender rule of thumb	Housing payment / all debt payments, gross income
36% to 45%	Fannie Mae, manually underwritten loans	Total debt-to-income; above 36% requires credit-score and reserve conditions
41% plus residual income	VA home loans	Debt-to-income and the money left after housing and debts
43%	Former General Qualified Mortgage limit	Removed by the CFPB; mandatory compliance with the price-based replacement began October 1, 2022
50%	Fannie Mae Desktop Underwriter; JCHS severe burden	Automated-underwriting ceiling; also the severe cost-burden line

Many articles still describe 43% as the federal limit. The CFPB’s 2020 final rule removed the 43% debt-to-income cap from the General Qualified Mortgage definition and replaced it with limits on how far the loan’s APR can exceed the average prime offer rate.⁷ Fannie Mae allows a total debt-to-income ratio of up to 50% on loans run through Desktop Underwriter.⁸ Lenders can set stricter limits of their own.

We found no study that tests 25%, 30%, 36%, or 43% as the line that best predicts delinquency, depleted savings, or financial hardship. These numbers are administrative standards and descriptive conventions. They are useful for comparing markets or years under one fixed rule, and much less useful for deciding whether one household should buy one house.

Residual income: what is left after the house

Every ratio above divides by gross income, which is not money a household can spend. Two households with the same income and the same mortgage can have very different budgets once taxes, childcare, other debts, and retirement contributions come out. Michael Stone argued that affordability should be measured by what remains for other necessities after paying for housing, which he called the residual income approach.¹⁰

\text{Residual income} = \text{take-home pay} - \text{housing costs} - \text{essential non-housing costs}

The VA home loan program is the main U.S. underwriting system that requires it. VA borrowers ordinarily must meet both a 41% debt-to-income standard and a minimum monthly residual income that varies by family size, region, and loan amount. The table in the regulation sets $1,117 a month for a family of four in the West borrowing $80,000 or more. A borrower above 41% can still qualify without extra review if residual income beats the guideline by at least 20%.⁹ The hard part is defining essential spending. A household’s actual spending includes choices, and a fixed national floor ignores local costs. A useful residual-income test uses your own essential spending and your own savings target.

Translate the indices yourself

The calculator applies NAR’s and the Atlanta Fed’s formulas to a home and income you choose, then walks from one burden to the other one assumption at a time. Raise the property tax rate to see how much of the gap is tax; drop the rate to 3% to see how much a rate regime changes both measures.

What we recommend

Use published indices for what they do well: comparing markets and tracking one market over time under fixed rules. When you read one, check which costs it counts, whose income it uses, and which threshold it applies, because those choices can move the answer more than a year of price changes.

For your own purchase, skip the search for the right percentage. Test the house against a short list of constraints, in this order, and let the tightest one set your price:

Cash to close. Down payment, closing costs, and moving costs from money you can spend without selling at a bad time.
Lender eligibility. What a specific loan program will approve. This is a ceiling.
Monthly budget. Full housing cost, including maintenance, against take-home pay after essential spending and your savings target.
Reserves after closing. Enough liquid savings left to cover several months of expenses at the new housing cost.
Stress. A job loss, a lower bonus, or a major repair in the first few years.

Do not average the answers or trade one off against another. A household that fails the reserves test cannot make up for it with an attractive price-to-rent ratio. Our guides on house-poor risk and the full cost of ownership work through steps 3 to 5 in detail.

A rule of thumb is enough when your situation looks like the one it was built for. A salaried household with ordinary expenses, modest other debt, and a solid emergency fund will get a sensible first answer from 28/36. The rules fit badly for households with large variable pay, high childcare or medical costs, substantial assets but modest income such as retirees, or homes with unusual tax and insurance costs. For high earners, our guide to how much house high earners can afford compares three personal frameworks on one household.

Frequently asked questions

Is 30% of income still a good rule for housing?

It is a reasonable first screen and a poor final answer. The 30% line comes from federal housing policy and is used to describe burdens across many households. It is applied to gross income, ignores other debts and essential spending, and was not chosen by testing which share best predicts financial trouble.

Why does my lender approve more than affordability calculators suggest?

A lender asks whether you are likely to repay under a specific loan program. Fannie Mae’s automated underwriting accepts total debt-to-income ratios up to 50%. A lender does not check whether the payment leaves room for retirement saving, childcare, or a repair fund.

Why do price-to-income ratios ignore interest rates?

By design. The ratio compares prices with earnings, which makes it stable across decades with very different mortgage markets. It shows how expensive houses are; a payment index shows how expensive they are to finance right now.

Does a NAR index above 100 mean homes are affordable?

It means a family with the median family income could cover principal and interest on the median home at 25% of income with 20% down. It leaves out property tax, insurance, mortgage insurance, the down payment itself, and households that are not families.

Why does the Atlanta Fed show worse affordability than NAR?

It assumes a smaller down payment, counts taxes, insurance, and mortgage insurance, and divides by median household income, which is lower than median family income. Each choice raises the measured burden, and the 30% threshold only partly offsets them.

Key takeaways

Indices differ on four choices: down payment, costs counted, the income measure, and the threshold. In late 2026 the national market scored 105 on NAR’s index and 68 on the Atlanta Fed’s.
For one synthetic $500,000 home and $120,000 household at 6.5%, method alone moves the burden by 11 points of income: 25.3% under NAR-style rules and 36.3% under HOAM-style rules.
Rates drive payment measures. From 2019 to 2025, price-to-income rose 15% while the Atlanta Fed’s payment share rose 48%.
Thresholds are conventions. The 43% federal cap was removed; Fannie Mae’s automated limit is 50%; no study validates any single percentage as the safe line.
For your own purchase, find the binding constraint. Cash to close, lender eligibility, monthly budget, reserves after closing, and stress. The tightest one sets the price.

Related guides

How Much House Can High Earners Really Afford? runs lender, White Coat Investor, and balance-sheet rules on one household.
Are You About to Be House-Poor? stress-tests a purchase against RSU drawdowns, job loss, and repairs.
The True Cost of Owning a Home adds maintenance, closing, and selling costs that no index counts.
Rent vs. Buy: A Financial Analysis compares the two over a holding period, including opportunity cost.
How Much to Put Down on a House weighs PMI against keeping cash in reserve.
How Much of a Ten-Year Treasury Yield Is Risk Compensation? explains the long-rate premium that feeds into mortgage rates.
Why a 50-Year Mortgage Won’t Fix Housing Affordability shows what a lower payment does and does not change.

Sources

National Association of Realtors. Methodology: Housing Affordability Index. 20% down payment, 25% qualifying ratio on principal and interest, median family income; 100 means the median family has exactly enough income to qualify.
National Association of Realtors. October 2026 Monthly Research Update (October 20, 2026). “The Housing Affordability Index registered 105.4, down from 106.7 one year ago.”
Federal Reserve Bank of Atlanta. Home Ownership Affordability Monitor, methodology and US Affordability Index data file, accessed October 25, 2026. August 2026: annual payment share 43.9%, index 68.3, qualifying-income gap $40,194. Annual averages computed from the monthly file.
NAHB. NAHB/Wells Fargo Cost of Housing Index, and Housing Affordability Worsens on Higher Mortgage Rates (Eye on Housing, August 20, 2026). Q2 2026: median family income $106,800; 36% for an existing home and 34% for a new home; 71% and 67% for a family earning half the median.
Joint Center for Housing Studies of Harvard University. The State of the Nation’s Housing 2026. Price-to-income ratios (4.1 in 2019, 5.0 in 2022, 4.7 in 2025); $3,120 monthly cost and $120,800 required income at a 31% ratio; payment assumptions in Figure 18.
Freddie Mac. Primary Mortgage Market Survey. 30-year fixed average of 7.32% on October 22, 2026; 6.19% a year earlier.
Consumer Financial Protection Bureau. Qualified Mortgage Definition under the Truth in Lending Act (Regulation Z): General QM Loan Definition. Final rule published December 29, 2020; mandatory compliance date later delayed to October 1, 2022.
Fannie Mae. Selling Guide B3-6-02, Debt-to-Income Ratios. 36% for manual underwriting, up to 45% with credit-score and reserve requirements; 50% for Desktop Underwriter.
38 CFR § 36.4340, Underwriting standards, processing procedures, lender responsibility, and lender certification. Debt-to-income and residual income standards for VA-guaranteed loans.
Stone, M. E. (2006). What Is Housing Affordability? The Case for the Residual Income Approach. Housing Policy Debate 17(1), 151–184.
Himmelberg, C., Mayer, C., & Sinai, T. (2005). Assessing High House Prices: Bubbles, Fundamentals and Misperceptions. Journal of Economic Perspectives 19(4), 67–92; NBER Working Paper 11643.
Poterba, J. M. (1984). Tax Subsidies to Owner-Occupied Housing: An Asset-Market Approach. Quarterly Journal of Economics 99(4), 729–752.
Sinai, T., & Souleles, N. S. (2005). Owner-Occupied Housing as a Hedge Against Rent Risk. Quarterly Journal of Economics 120(2), 763–789; NBER Working Paper 9462.

Editor’s note

Educational content, not lending or financial advice. The $500,000 home and $120,000 household are synthetic and apply each index’s published formula to one set of inputs; they are not readings of any index. Index readings and mortgage rates are as of October 25, 2026 and change monthly.