Leading vs Lagging Indicators: Rental Vacancies
December 19th, 2011 by PotatoOne problem with trying to pick apart brewing bubbles in real estate is that there’s a real shortage of leading indicators.
A leading indicator is just what it sounds like: some measure that identifies a problem (or conversely, an upward move) before that is upon you. So for house prices, a leading indicator would be something that might signal the future direction of house prices. If there was something that reliably foreshadowed prices, you could point to that and some history, and maybe help people see what’s coming.
There are a few that might work, my favourite of course being the price-to-rent multiple. Unfortunately, while I believe that one is both reliable and particularly relevant to the down-to-earth decision of how to go about getting a roof over your head in your own life, it’s not particularly timely. You can use it to say that there may be over-valuation, but it gives no information on when that over-valuation may come to an end, or how: lower costs (like interest rates), higher rents, or lower prices could all interact to change it. Like all indicators, it’s a bit beyond imperfect.
Other indicators might be the health of the economy: if unemployment is rising, people may be unable to afford their mortgages (or may not be in a home-buying mood), so prices may fall, and vice-versa when unemployment falls. Unfortunately, sometimes that works in reverse, as happened in the US: house prices dropped first, which then sank the economy. It could be that only after prices fall does the rampant over-building slow, and after that all the construction workers and real estate agents start looking for new jobs.
Lots of people like to point at these lagging indicators — which we know don’t really tell us anything about the future of the housing market — and say that since they are fine, so too must be the housing market. Mortgage arrears and defaults, declining interest rates, GDP growth, etc.
Recently David Fleming looked at vacancy rates, saying that since they’re low in Toronto, then the housing market must be fine. Now I can totally see the logic: vacancy rates should be a leading indicator. Yet again, history doesn’t seem to bear that out: Toronto had a real estate crash in 1989, with many looking back and saying that there was over-building occurring at that time. Yet the vacancy rate stayed persistently low — less than 1% — all through the late 80’s. It wasn’t until the prices started coming down that the vacancy rate started going up in 1990. There wasn’t as much of a lag: it was a better indicator than mortgage defaults. You potentially could have looked at that data and realized the drop in prices (still modest at that point) wasn’t a temporary opportunity. We saw the same thing in the states: the vacancy rate was flat through most of the 2000’s until after the peak: about 2007 in many cities, after which it spiked up (in the comments, I used Miami’s numbers in particular, since that was the comparison DF made in his post).
This is a little lot counter-intuitive: if there’s a building boom and over-supply, then there should be vacancies. If tenants are being taken out of the rental pool and becoming owners, and speculators are taking on multiple units and adding to the supply, there should be vacancies. Yet we’re not seeing them now, and that doesn’t appear to be a contrary indicator: the vacancies weren’t leading indicators in past bubbles, either.
How can this be? I have a few speculations below, but the short answer is that I honestly don’t know. I even went out and asked Ben Rabidoux. Anyway, some speculations:
- Shadow inventory: with prices going up 10%/year in Toronto, and condos being cash-flow negative, it’s almost not worth the hassle of renting your place out and having to deal with a tenant when you’re making all your money from appreciation anyway. Some speculators may be holding vacant inventory; as soon as prices stop going up, they may become reluctant landlords, and this will increase the supply shortly after the peak, leading to rising vacancy rates after prices peak. This is a bit nutty, even for real estate speculators, but it only takes a few thousand to increase the vacancy rate by ~1% in Toronto.
- The way the stats are collected: CMHC collects data exclusively from multi-unit properties, largely professionally-run apartment buildings. These are not in the business of having vacancies, so there may be many vacancies in detached houses, basement apartments, duplexes, and condos, while the professional buildings find ways to get their spots filled. This may also be why the average rent figure always appears to be lower than you expect.
- Booms attract people too. Once the construction jobs dry up, vacancies may increase from the demand side as those workers leave the city; and for those who own, they may put their units up for rent, increasing supply. If the economy is good, then perhaps people are renting apartments on their own rather than getting roommates.
- Construction time: it takes a few years to get a condo tower completed, and if the building boom reaches a crescendo just before the end, those units won’t be completed and on the market until after the peak. For example, if construction reached the nutty stage in 1987 or 1988, it wouldn’t be until 1990 that the over-supply would be evident on the rental market.
- Household formation can diverge from population growth. When trying to say that there’s an over-supply of housing stock we may focus on how population growth compared to housing starts over some period (e.g., saying that housing construction outpaced population growth for much of the past decade). But household formation is not quite the same as population growth if household size is changing. So if easy lending is borrowing demand from the future, leading to new household formations with smaller family sizes (e.g., single people buying condos), then that can help soak up some of the over-supply. Does that translate into the rental market? I have no idea, but it could be that the rental market stays insulated from all this insanity: as renters are pulled over to owners, rental units also undergo condo conversions.