Percent unoccupied dwellings compared to percent apartments and percent rented in selected Sydney suburbs (source data: ABS 2011 Census)
Percent unoccupied dwellings compared to percent apartments and percent rented in selected Sydney suburbs (source data: ABS 2011 Census)

The Sydney Morning Herald (SMH) reported yesterday that researchers at UNSW’s City Futures Research Centre reckon “Sydney’s housing affordability crisis is being artificially inflated by up to 90,000 properties standing empty in some of the city’s most desirable suburbs” (Thousands of empty homes adding to Sydney’s housing crisis, experts say).

The alleged culprit is negative gearing:

“Leaving housing empty is both profitable and subsidised by government,” researchers Bill Randolph and Laurence Troy said. “This is taxation lunacy and a national scandal.”…The results suggested property investors in some of Sydney’s most sought-after areas were focusing on growing the value of their properties, with losses offset by tax incentives such as negative gearing.

The SMH links to an opinion piece by the two researchers which it also published yesterday; it reinforces the theme that negative gearing is the bad guy here (Negative gearing has created empty houses and artificial scarcity).

Naturally, the SMH doesn’t provide a link to the study the researchers did (it seems the SMH still thinks print) and I can find no trace of it on the UNSW web site.

Nevertheless, based on the two SMH articles, the argument seems to go like this:

  • The proportion of dwellings that were vacant on Census night 2011 was highest in up-market suburbs like Manly and Potts Point where the capital gain is high and rental yield is low; and lowest in less salubrious suburbs like Green Valley and Erskine Park where the capital gain is low and rental yield is high.
  • This indicates a very large number of investors in up-market suburbs don’t rent out their dwellings because the combination of negative gearing and strong capital gains means they don’t need to bother.

Is this a reasonable interpretation of the data? Does the proffered explanation make sense? While I don’t doubt some dwellings are left unoccupied for long periods for no “good” reason, I suspect the scale of the problem – supposedly “up to 90,000” – is exaggerated.

The first point to make is that negative gearing has no role in increasing the number of dwellings that are unoccupied. If an investor doesn’t earn income from a property then the negative gearing benefit can’t be applied. Obviously there’s no rental income if the dwelling is left empty. So the negative gearing argument can be scratched outright; it’s nonsense.

The second point is the numbers don’t seem to support the proposition. The 2011 Census shows 6.5% of properties in the Sydney urban area (94,282 out of 1,442,089) were unoccupied on Census night. That’s actually a lot lower than the figures for NSW (9.7%) and Australia (10.7%)! It doesn’t seem consistent with the idea that Sydney’s a capital gains and negative gearing driven hot-house.

Third, there’s Occam’s Razor. There are other simpler, more plausible reasons why a property might have been unoccupied for a single night in August 2011:

  • Families go away on holidays, often for weeks at a time. The Census is conducted in winter so it’s no surprise well-heeled people take the opportunity to move to warmer places like the northern hemisphere, Bali or Noosa.
  • Some proportion of properties are always going to be vacant for a period while new owners or new tenants move in. This period will usually be longer for newly constructed dwellings.
  • Some properties are vacant because they’re being renovated for occupation or sale. Some are awaiting demolition by a developer.
  • Some professional and business people in regional centres own dwellings in the city they use for weekend visits and often plan to retire to.

Another factor is the significant increase in the proportion of one person households; in their case there’s no one else to “occupy” the dwelling if they’re away for the night:

  • They go away for work, perhaps overnight to see a client interstate or for a week to a conference overseas. Some might work on a FIFO arrangement or others on a months-long contract in Asia.
  • They might stay regularly at their lover’s place (who’s never done that for long periods?). When someone moves in to their partner’s place permanently they often hang on to their home for an extended period before selling it.
  • They might be in hospital, in care, or interstate to attend a wedding, a funeral, etc.

The probability that a dwelling will be unoccupied for a single night is also amplified by the increase in the number of couple households where both work. It’s quite plausible there were some where both members were away on Census night.

It doesn’t strike me as necessary to resort to arguments like “taxation lunacy and a national scandal” in order to explain why as many as 16% of dwellings in suburbs like Manly and Potts Point were unoccupied for a single night compared to 2% in Green Valley and 3% in Erskine Park.

A possible alternative explanation is people on higher incomes are much more likely to be away because they can afford to; and because they have more (usually work-related) reasons to.

Another alternative explanation is suburbs like Manly and Potts Point have very high proportions of apartments and renters; so we should expect more small households, as well as a high turnover of occupants with consequently more unoccupied nights.

In Manly, 73% of the dwelling stock is a flat, unit or apartment and 52% of private occupied dwellings are rented. The corresponding numbers for Potts Point are even higher: 95% and 63%.

The average for the Sydney urban area, on the other hand, is 28% flat, unit or apartment and 32% rented. The corresponding figures for Green Valley are 1% and 23%; in Erskine Park it’s 0% and 14% (see exhibit).

Of course there are bound to be some owners who let their property sit idle for a very long period for no apparent reason other than they can afford to avoid the bother of renting it. Better data is needed to understand the scale of this issue but on the face of it I think it’s doubtful it’s anything remotely like “up to 90,000” dwellings (which would be almost all dwellings recorded as unoccupied on Census night in the Sydney urban area!).

Nevertheless, there’s still a case for investigating whether there are ways to reduce the number of long-term unoccupied dwellings. There’s anecdotal evidence that some developers let properties slated for redevelopment sit empty for unnecessarily long periods (see Are too many houses empty?).

Overall though, I suspect this problem is much exaggerated. There are simpler and more plausible explanations for the great bulk of “empty homes” on Census night that must be considered before pointing the metaphorical gun at greedy investors/owners.