Two well-known indices, both anchored at 100, both counting different things. The National Association of Realtors publishes a Housing Affordability Index whose own methodology states that "a value of 100 means that a family with the median income has exactly enough income to qualify for a mortgage on a median-priced home." The Federal Reserve Bank of Atlanta publishes a Home Ownership Affordability Monitor whose explainer states that "an affordability index score below 100 indicates home ownership is unaffordable, while a score above 100 means a median-priced home is affordable for a median-income household." The two sentences sound interchangeable. The arithmetic behind them is not.
Where they diverge, in the four places that matter. NAR uses median family income reported by the Census Bureau; the Atlanta Fed uses median household income, a broader population that includes people living alone. NAR's test is whether the monthly payment fits inside a qualifying ratio of 25 percent of monthly income; the Atlanta Fed's is whether annual ownership cost stays under 30 percent of annual income. NAR counts principal and interest only, and its methodology states that "the calculation assumes a down payment of 20 percent of the home price". The Atlanta Fed counts "principal and interest costs in addition to homeowner's insurance, property taxes, and other components of home ownership cost". So one index is measuring whether a family could get approved, and the other is measuring whether a household could carry the whole cost of ownership.
The consequence is that the two can point in different directions, and neither is wrong. An area where property taxes and insurance are unusually expensive will look better on a measure that excludes them than on one that includes them. An area with many single-person households will look different depending on whether the income used is a family's or a household's. When two affordability figures for the same city disagree, the first question is not which is accurate but which question each is answering.
A mortgage rate sits inside both, which is why affordability and price move apart. The cost being compared with income is a monthly payment, and a payment is a function of the loan amount and the interest rate together. A market can therefore become less affordable while prices are flat, and can become less affordable even while prices fall, if borrowing costs rise faster than prices come down. Reporting that treats an affordability index as a proxy for house prices misses the variable doing most of the work in any period when rates are moving.
The price-to-income ratio is the simpler cousin and has a different weakness. Dividing the median home price by median annual income gives a figure with no interest rate in it at all, which makes it a cleaner measure of how expensive housing is relative to earnings and a poor measure of whether anyone can currently afford the payment. It is the more useful of the two for comparing decades, because it is not distorted by whatever rates happened to prevail, and the less useful for describing this month.
What an index built on medians cannot do. The median-income household and the median-priced home are statistical constructs, and the household is not necessarily the one buying that home. First-time buyers earn less than the median and buy cheaper homes than the median; existing owners bring equity that no index models. So a reading below 100 does not mean nobody in the area can buy, and a reading above 100 does not mean a particular family can. The index is a good instrument for tracking a market over time and comparing markets with each other, and a poor instrument for answering a question about one household.
