Economics: Tight housing market
admin | May 24, 2019
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.
Homebuilders since the housing bust of the late-2000s have operated as if once burned, twice shy. Housing starts and new home sales 10 years after the bust, are running far below even pre-boom levels. It took several years for housing demand to absorb the significant supply created during the boom, but evidence is accumulating that the housing market is getting tight, which should support strength in building activity going forward.
Housing stock arithmetic
The basic concept for thinking about the housing stock is that the number of housing units should roughly track the growth of the population, or, more precisely, the number of households, over time. Homes age or otherwise become uninhabitable, too, so new construction should modestly outpace the household growth to account for both population growth and turnover of the housing stock. Various metrics can provide insight into how the current situation compares on a historical basis.
Vacant units and occupancy rates
The Census Bureau reports each quarter on housing inventory, occupancy, home ownership rates and other aspects of housing. Estimates of housing inventory include a breakdown of units that are occupied versus those that are vacant, either year-round or seasonally. The number of vacant housing units fell sharply in 2017 and 2018 (Exhibit 1).
Exhibit 1: Vacant housing units
The same data implies an occupancy rate—the percentage of housing units that are occupied year-round. This gauge similarly shows that the housing supply has tightened considerably over the past two years (Exhibit 2). In fact, the occupancy rate is the highest in nearly 20 years; this measure actually understates occupancy, since some vacant homes are vacation homes occupied only part of the year.
Exhibit 2: Housing occupancy rate
The Census Bureau also separately calculates vacancy rates for homeowners and renters. These data show that the homeowner vacancy rate has fallen to levels last seen in the early 1980s after an unprecedented spike during the housing bust.
Exhibit 3: Homeowner vacancy rate
Similarly, the rental vacancy rate has dropped to 1.4%, matching the lowest reading since 1981.
Exhibit 4: Rental vacancy rate
Using occupied housing units as a proxy for household formation
For the purposes of this discussion, a household” could be defined as a family, group, or individual (of any sort) that occupies its own housing unit. Thus, we can use the Census Bureau series on occupied housing units as a proxy for household formation (the Census Bureau definition of “number of households” does not quite conform to our housing-centric concept). Thus, the annual change in the number of occupied units is a somewhat choppy, but rough proxy for the increase in the number of households, or annual household formation. This gauge of household formation dropped sharply and persistently after the Great Recession, as families and friends doubled up due to the deterioration in household finances—the primary example being the stories of millennials living in their parents’ basements (Exhibit 5). With the economy growing robustly, household formation has risen sharply in the past few years.
Exhibit 5: Occupied housing units as a proxy for household formation
Armed with a proxy for household formation, we can now compare the pace of housing demand (household formation) with the growth of housing supply (as measured by housing starts). Housing starts outpaced household formation during the Housing Boom years, creating an inventory overhang (Exhibit 6). The two series recovered more or less in tandem through much of the Bust and subsequent recovery, which actually constitutes a period when housing inventories were slowly being pared (remember that there is a certain amount of turnover/destruction in the housing stock over time). However, in the last two years, the pace of household formation has surged while housing starts have failed to accelerate. This chart provides another independent confirmation that the supply of housing has tightened considerably over the past few years.
Exhibit 6: Household formation outstripping housing starts
In a sense, the housing market can be thought of similarly to any goods market. When inventories accumulate, manufacturers tend to slow down the pace of production, and vice versa. In the case of the housing market, these data make a compelling case that housing starts will need to pick up going forward (unless demand deteriorates substantially). This supports my contention that the housing sector should begin to contribute positively to real GDP growth again soon after a string of five consecutive negative quarters.
A tight housing market also suggests that home prices are likely to remain firm. We have seen a significant moderation in home price appreciation recently after several years of unsustainably rapid gains. Some of the cooling reflects a general sense among potential buyers that homes had gotten too pricey, and some portion of the slowdown may be a one-off adjustment in high-end markets and high-tax states to the tax changes (SALT deduction and mortgage interest deduction limits). In any case, these data would suggest that home prices will not weaken very much or for very long as long as demand holds up.
Of course, home prices do not figure directly into key measures of consumer prices, but rent does. The tightness in rental markets, as indicated in Exhibit 4, is likely to keep shelter costs firm. Indeed, even as home price appreciation has cooled dramatically over the last few quarters, rent and owners’ equivalent rent have remained firm. In the CPI, rent has risen by 3.8% over the past 12 months, while OER has increased by 3.4%. In both cases, these gains are nearly identical to the corresponding 12-month advances through April 2017 and 2018, even though measures of home price appreciation have slipped from the past few years.