As recently as 1990, Americans earned only 0.5 per cent of their personal income from rents. Now they earn 4.5 per cent of their income from this source.
Since February 2007, more than 12 per cent of the total growth of personal income can be attributed to rising rental income, even though rents were only responsible for 1.5 per cent of personal income ten years ago. Put another way, rental income has more than tripled in the past decade during a period when employee compensation only grew by 31 per cent.
On the surface, these data seem to corroborate the increasingly common view that the increase in American wealth inequality can be blamed almost exclusively on rising house prices.
However, an analysis of what the Bureau of Economic Analysis is actually measuring suggests that’s not quite right. It’s more accurate to say the booming returns to home ownership have mostly been driven by the fall in mortgage interest rates.
The BEA defines “rental income of persons” as: The net current production income of persons (except those primarily engaged in the real estate business) from the rental of real property, the imputed net rental income of owner occupants of farm and nonfarm dwellings, and the royalties received by persons from patents, copyrights, and rights to natural resources.
It’s easiest to start by explaining the “net” part. Just as net income for a business is sales minus costs, net rental income is what landlords keep after covering their property taxes, depreciation, maintenance expenses, and mortgage interest. The headline income number is therefore affected both by the rent landlords (including homeowners) collect and by their costs.
But people who live in their own homes don’t actually pay — or get paid — any rent. They either own their dwelling outright or they have a mortgage. Mortgage down payments and ongoing principal payments are counted as saving because they either involve buying assets or reducing indebtedness, both of which boost net worth. Mortgage interest payments are considered part of the cost of owning a home rather than a cost of living in it.
So what is being measured here?
Statistical agencies around the world “impute” what homeowners would pay were they renters “based on the rents charged for similar tenant-occupied housing”. The thinking is that this makes measures of GDP “invariant when housing units shift between tenant occupancy and owner occupancy.” This imputation is one of the single biggest “expenses” for the average American — about 12 per cent of personal consumption expenditures.
This is actually quite strange.
If you buy beer at a bar, your consumption is the price you pay, which covers the cost of the drink plus the service of having people serve you in a dedicated space. If you buy beer at a grocery store and drink it at home you spend far less money because you’re providing the services of pouring the drink and hosting the venue yourself. Yet nobody argues the statisticians should “impute” the value of these services in the national accounts as a supplement to what you actually spent at the grocery store.
That said, imputing rental payments of home owners — and nothing else — leads to perverse mismeasurements of labour output in the home, which in practice tends to be sexist. If a progressive stay-at-home husband devotes himself full-time to caring for his children, preparing food, cleaning the house, driving everyone around, etc, no economic activity is recorded. If the same man enters the workforce and pays other people to perform those services, however, the economy gets bigger by the amount the amount he earns and by the amount he spends, because it’s income for others. That’s double-counting, however, since the labour involved in cooking and cleaning and childcare was always occuring even if it wasn’t always measured. Refusing to count labour performed in the home is inconsistent with the principle animating the imputation of rent “paid” by owner-occupiers.
The other argument for imputing housing rents “paid” by owners is that a home is an asset providing benefits for many years, unlike, for example, a bottle of beer, which is consumed once and (hopefully) never experienced again. As the BEA puts it:
Unlike money spent on items that do not hold their value, money that is spent to acquire a residence is not “gone” in the sense that it cannot be reclaimed. Money spent on a residence usually can be recovered—often with a profit—by selling.
That was written in 2007, which seems in poor taste, but the basic idea makes sense. You spend a lot upfront, finance the purchase with debt, and get benefits over time as you gradually repay what you owe. Unless the house burns down and the land it sits on is covered in radioactive waste, there is probably some value in selling it at some point in the future.
However, this logic applies to many purchases besides homes.
Cars, for example, are not replaced particularly often. Auto leases nowadays can stretch as long as six or seven years, comparable to the typical length of time people live in the same home before selling. Those who don’t lease typically borrow for a similar length of time. The age of the American vehicle fleet is more than 11 years, and even in a bumper year like 2016, new vehicle sales account for less than 10 per cent of the total number of cars and trucks on the road. Yet spending on vehicles is counted up front, rather than spread out over their expected lives.
You could even extend this logic further to include anything you buy but don’t immediately destroy after a single use. An owned tuxedo is an asset that substitutes for rentals. It may require maintenance to offset wear and tear but that’s no different from a house. Even a good pair of socks can last for years. None of these things are treated the way owner-occupied housing is, however.
As it happens, the BEA looked into these issues a few years ago and estimated that a few methodological tweaks would have boosted GDP by about 40 per cent in 1965 and by more than a quarter in 2010. Put another way, growth has been overstated because the gains from women entering the workforce were partly offset by the decline in unmeasured services performed at home. Interestingly, the BEA’s estimates are strongly affected by the collapse in the relative wages of household workers. Assuming the value of childcare provided by a parent hasn’t plunged over time, for example, the mismeasurement could be even larger.
After adjusting for the loss of this labour and the long life of durable goods, total spending on services grew only two-thirds as much. Moreover, the dramatic increase in the share of consumption spending going to services — from 48 per cent in 1965 to 67 per cent in 2010 — almost completely disappears with this new methodology (75 per cent to 82 per cent).
The notion that homeowners “pay” themselves rent that must be imputed is odd and inconsistent with the methodology in the rest of the national accounts, but it is what it is. The good news is the trend of this fictitious income/cost stream has tended to track actual rents paid by people who deal with actual landlords, although there have been some temporary discrepancies based on changes in the home ownership rate and the geography of where people rent vs own.
The reason is that owner-occupiers are much more indebted and are therefore extremely sensitive to changes in mortgage interest rates. People who lived in houses they owned were losing money throughout the 1980s, according to the BEA, because debt burdens were so high. The other costs — maintenance, depreciation, and net taxes — have consistently hovered between 40 and 50 per cent of total imputed rents paid.
The mortgage interest burden has ranged from 21 per cent of rents “collected” by home owners to more than 50 per cent. Since the financial crisis, this figure has dropped by more than 20 percentage points. The absolute amount of mortgage interest paid each year by homeowners has dropped by nearly $200 billion (37 per cent) since 2008, which explains more than half the increase in net (imputed) rental income earned by home owners.
Given who owns (and owes) what, there might be distributional consequences of these shifts in the composition of asset income. But it’s tough to argue there’s anything unique about housing supply and demand that explains these data. Instead, what seems to matter is that many people finance their home purchases with lots of debt.
Source: https://ftalphaville.ft.com/2017/01/27/2183181/americans-are-making-more-money-from-renting-out-homes-than-ever-before-sort-of/