RDP 2016-04: Housing Prices, Mortgage Interest Rates and the Rising Share of Capital Income in the United States 2. The Measurement of Housing Services

A very useful feature of the national accounts produced by the Bureau of Economic Analysis (BEA) is the ability to decompose the aggregate GDP estimates along several dimensions. For example, it is possible to decompose total housing output by the type of housing (owner-occupied and tenant-occupied) and by the type of geographic region (e.g. states and metropolitan statistical areas). To understand the drivers of the rise in housing capital income, it is important to discuss some key measurement issues (for more details on the measurement of rents and housing services in the United States, see, for example, Diewert (2003), Mayerhauser and Reinsdorf (2007), Short, O'Hara and Susin (2007) and McCarthy, Peach and Ploenzke (2010).

Housing services are a component of both household income and consumption in the national accounts. Housing services consist of two main components: 1) the rent paid by tenants to landlords (or ‘market rent’); and 2) the rent paid by home owners to themselves (or ‘imputed rent’). For the home owner, the value of housing services is measured as the income the home owner could have received if the house had been rented to a tenant. In essence, it is assumed that a home owner is paying rent to themselves as they represent both the tenant and landlord of that property. The rents imputed on owner-occupied housing are the largest component of personal consumption expenditure, valued at US$1.4 trillion in 2014 (or about 12 per cent of total personal consumption and 8 per cent of GDP). The imputation is needed for GDP to be invariant to changes in the rate of home ownership.^[3]

To estimate the value for the services of owner-occupied housing, the BEA uses information on the rents charged for similar tenant-occupied housing. The imputation is based on the ‘rental equivalence’ method – the ‘gross imputed rent’ is valued by the estimated rent that a tenant would pay for the same home, taking into account factors such as dwelling size, dwelling quality, and location (Mayerhauser and Reinsdorf 2007). Importantly, this method implies that the prices (and rents) of owner-occupied and tenant-occupied housing will essentially move together over time for a given location.

As Rognlie (2015) highlights, there are important differences between ‘gross’ and ‘net’ housing services. The ‘gross output’ of housing services is the amount earned and spent on housing services, which is the same as the amount of rent paid (whether imputed or not). Gross housing output has gradually increased as a share of total net domestic income from 7.8 per cent in 1950 to 12.3 per cent in 2014. The ‘gross value added’ of the housing sector is equal to the total rent paid less expenditures on home maintenance (or ‘intermediate housing consumption’). Gross value added can be further divided into gross operating surplus (capital income), compensation of employees (labour income) and property taxes.

There are two different ways to estimate net housing income: 1) net housing capital income (or ‘net operating surplus’), which is equal to gross operating surplus less depreciation; and 2) net housing profit income (or ‘rental income of persons’), which is equal to gross operating surplus less depreciation and interest payments. Existing research tends to focus on the former measure of net housing income (e.g. Piketty and Zucman 2014; Rognlie 2015). Measured on this basis, the share of the economy going to housing capital has effectively doubled since 1950 to stand at 6.4 per cent in 2014.

I mainly focus on the long-run trends in housing capital income (κ^R) to be consistent with the existing research. But, I also consider housing profit income (π^R) as this provides an important cross-check on the results and sheds additional light on the ultimate ‘winners’ and ‘losers’ of the secular rise in the housing share of the economy. The net profit measure has risen from 2.4 per cent of the total economy in 1950 to 3.9 per cent in 2014. It has also followed a much more pronounced U-shaped pattern over recent decades, gradually declining in the period between the Second World War and 1980 from 2.4 to 0.4 and rising thereafter.

The following accounting definition summarises the differences between the various housing income measures:

According to the BEA, net housing profit income (π^R) is equal to housing gross output (i.e. the rental price, P^R, multiplied by real housing services, Y^R) less housing maintenance (M), compensation of employees in the housing sector (COE), property taxes (T), housing depreciation (D) and mortgage interest payments (IP). Also, note that housing gross operating surplus (κ^R) is equal to gross value added (GVA^R = P^RY^R − M) less compensation of employees and property taxes. The decomposition of housing services is shown in Table 1.

Footnote

In 2014, the total value of the housing stock was estimated at US$18.9 trillion. The owner-occupied housing stock made up US$14.9 trillion (or 78.8 per cent of the total housing stock) while the tenant-occupied stock made up US$3.9 trillion (20.8 per cent). The tenant-occupied housing stock can be further divided by the type of owner into households (9.2 per cent of the total stock), sole proprietorships and partnerships (7.6 per cent), non-profit institutions (1.2 per cent) and corporates (1.0 per cent). The government-owned stock made up US$0.4 trillion (2.1 per cent). [3]