Anyone who has looked for an apartment to rent or a home to buy has probably heard the rule of thumb that you shouldn't spend more than 30% of your income on housing.
This seems like a reasonable benchmark. After all, you wouldn't want to spend so much money on housing that you are left with nothing for the rest of life's necessities. The problem is, many Americans aren't abiding by this rule of thumb, not by a long shot. In places like New York City, for instance, the average family spends 40% of its income on housing, while a recent report by Harvard University’s Joint Center for Housing Studies shows that the share of renters who are "cost-burdened" rose from 40% to 46% from 2003 to 2013.
The data underscore the fact that in an America with stagnant incomes and a convalescent home-building industry, housing is becoming more expensive. But how expensive is too expensive? And is the 30%-of-income benchmark useful?
The short answer is no. Here's why:
1. The standard is an arbitrary holdover from the New Deal era: The United States National Housing Act of 1937 marked the beginning of federal public housing subsidies as we know them today. The law set income limits, where recipients were eligible for housing subsidies only if their income wasn't more than five or six times the amount of rent they owed each month. According to the Census Bureau:
By 1969, the escalation of rents by public housing authorities struggling to meet spiraling operation and maintenance costs nearly nullified the purpose of the public housing program established in 1937 to serve the nation’s neediest. To reverse this, the Brooke Amendment (1969) to the 1968 Housing and Urban Development Act, established the rent threshold of 25 percent of family income; that is a family would be required to pay one-quarter of its income in rent. By 1981, this threshold had been raised to 30 percent, which today remains the rent standard for most rental housing programs.
In other words, the idea stems from a policy crafted 80 years ago, and it has fluctuated based on the vagaries of available public funding. While government programs might need arbitrary standards to decide who is eligible for welfare, it doesn't make sense to then use that standard to decide whether or not a particular individual is spending too much on housing, or whether or not housing in general is too expensive.
2. Housing affordability should take other costs into account: David Bieri, an economist at the University of Michigan, recently published a paper analyzing housing affordability measures, arguing that an accurate measure would have to take into account the quality of housing, including what benefits residents gain from living where they live. For instance, the average New York family that spends 40% of its income on housing is also deriving benefits, like superior public transportation, from living in the city. Quality public education is another good that also drives up housing prices, and it might make perfect sense for a family to spend more money on a home if it means they can then spend less on education.
3. We should expect housing to take up an increasingly large part of our incomes: As the economy grows, the cost of many goods are going to fall. For instance, in 1936 the average family spent about one-third of its income on food, compared with roughly 13% in 2003. A person only needs to consume so many calories per day, and as food production becomes more efficient, we should expect to spend a smaller portion of our incomes on it.
But some goods, like housing, won't follow this trend. That's because much of the cost of housing is made up of the land on which the housing sits, and the cost of land is largely the result of its proximity to economically productive areas. This makes housing somewhat similar to healthcare and education, where the main driver in the good's cost (skilled labor in the cases of healthcare and education) won't be made much cheaper via technological advances. So, we shouldn't expect these goods to account for the same portion of an individual's budget as time goes on.