Senin, 28 Januari 2013

The case against the case for negative gearing

Negative gearing is an issue that is rather close to our hearts. We're not the biggest fans of this tax break for landlords, and we've talked about it a fair bit over the last couple of years. We understand that many landlords don't feel the same way about it - about 1.12 million of them, in fact. Whenever the question of its efficacy gets a run in the press, we start to see many of its supporters pointing to the calamity of 1985-1987, after the Hawke government trimmed it a little.

(Bob Hawke and Paul Keating
had a bit of a go at negative gearing in 1985)

During that time, rents went on a bit of a rampage. It is often claimed that this is because investor sentiment was so damaged by the restrictions to negative gearing (losses could only be offset against rental income, not other income) that landlords simply started to abandon the market. In the face of serious pressure from the propertied classes, the government capitulated, and negative gearing was reinstated in full.

We've always thought that this was a bit of old hokum, and we're not the only ones. In April 2011, while the world was gearing up for the National Tax Forum, Saul Eslake pointed out that it was extremely low vacancy rates in Sydney and Perth, and not a nationwide landlord strike, that caused all this trouble with rents. More recently Tim Lawless, head of RP Data's research and analytics team and friend to all Australian property speculators, has published a spiel that appears to support the idea that the great rent rise of '85 was not the result of changes to negative gearing.

Lawless refers to this graph:

Source: Tim Lawless, via Michael Yardney's Property Update

... which shows that rents were already well and truly on the rise before any changes to negative gearing were made in 1985. It also shows sharp increases during the mid-nineties and mid- to late- noughties.

He also refers to data from the Australian Bureau of Statistics, on loans for property investment over the period in question. He says:
The total value of investment finance commitments in September 1987 was 41.5% higher than in September 1985.  These figures seem to suggest that at that time there was no weakness in demand for investment housing ...
... The reason why negative gearing was reinstated in September 1987 was that it was proclaimed that rents rose sharply on the back of a fall in housing market investment.  However, it doesn’t look as if investment in the housing market dried up throughout this period. Rents clearly did rise quite sharply throughout as demonstrated.
Lawless offers no alternative ideas as to why rents rose so sharply during that time, but we're inclined to agree with Eslake. Steady increases in rents are more likely to be linked to vacancy rates than to the way we (un)tax the nation's landlords. Well might you argue that disinvestment by landlords en masse could have an adverse effect on vacancy rates, but the reality is likely to be otherwise. Landlords are, in the main, not a particularly riotous bunch. They tend not to burn down their houses as they exit the market - it's far more orderly than that. When they opt to bail out, they sell, and some other body usually moves in in the end...

But the main point of Lawless' article was not to exonerate those who have decried the myth of disinvestment, and to join the chorus of voices calling for reasonable tax reform. Instead, it was to put an alternative argument in support of negative gearing. The argument is: if not for negative gearing, landlords would not invest in new houses, so there wouldn't be enough to go around. This would be a disaster for taxpayers because it would fall to the government to pick up the shortfall.

Let's have a closer look at this argument.

Lawless looked at the ABS dwelling approvals numbers, and saw that in the 12 months to October 2012 there were 145,515 dwelling approvals granted to private interests, compared to 2,065 granted to public housing providers.

He then checked the census data, and saw that 29.6% of Australians live in rented accommodation.

He concluded that the private sector must have built and let 43,684 new homes, because that's 29.6% of the houses built in the period (adjusted to allow for the 2,065 new public housing tenancies).

He checked the median home price for October 2012, which was $386,000. He multiplied 43,684 by $386,000 and got a little over $16.86 billion. This is how much it would cost the government to build all those homes for tenants.

He compared this to the tax foregone by negatively geared landlords: estimated at $4.81 billion over the 2009-10 financial year - as good a figure as any to rely on.

He concluded that the government would not be able to use the increase in tax revenue to build enough houses to make up for a projected shortfall, if negative gearing was abolished.

This logic may appear sound, but it relies on a heck of an assumption - that almost 30% of newly built housing is commissioned by landlords, to be rented out to tenants. Quite simply, this is wrong. As Chris pointed out some time ago in his analysis that negative gearing is not your friend, only a small proportion of the money borrowed by landlords is used to fund new construction. Eslake has spotted this, too, and he puts this figure at 8%.

(RBA Table D06)

That means Lawless is going to have to review his argument.

Let's make a start for him:

If 8% of privately commissioned newly constructed housing goes straight into the rental market, and we apply that to the 2012 dwelling approvals numbers Lawless has used,* we end up with a figure of 11,641 dwellings built by private landlords. If we multiply this by his median house price, we find that the government would have to spend a shade above $4.49 billion in order to keep pace. This falls in just below the $4.81 billion of foregone revenue that we can attribute to negative gearing.

But we can take this further. A recent independent report into the Social Housing Initiative of the Nation Building Economic Stimulas Plan tells us that the government can build houses somewhat more cheaply than private landlords can. In fact, the figures suggest that the government can build houses at a cost of around $266,000 per dwelling - having procured 19,699 new dwellings out of the $5.238 billion spent on social housing across the stimulus package. In order to build those 11,641 new houses for tenants, the government would have to spend a little under $3.1 billion. Perhaps we could have houses, and a budget surplus, too?

That report tells us that there are many other economic benefits when the government invests in social housing, and we've spoken before about some of the social benefits, too. Spending tax dollars on social housing is very clearly a good thing to do.

But we're mindful of the fact that there are still 133,874 new dwellings that we're yet to account for in our example here - those properties that are built by/for owner/occupiers. Of course these purchases would not, we'd assume, contribute to any negatively geared property portfolios. But does the construction of these properties have any bearing on vacancy rates in the private rental market? And what, if anything, is the impact of negative gearing on the cost or availability of those old houses that are left behind?

We're not going to be able to answer these question in any definitive kind of a way, but we can certainly explore them a little. Let's see how far we can get...

We'll look at the ABS' feature article First Home Buyers In Australia for some data and analysis. First, it tells us that, in the three years before 2009/10, there were 429,000 first home buyer (FHB) households in Australia. We must note that changes to the First Home Buyer Grants scheme that came with and since the Nation Building Economic Stimulus Plan will cause this figure to fluctuate over time - the previous three year window saw 318,000 FHBs enter the market, and we're likely to see a drop in the next period following the winding back of FHB grants schemes in some parts of the country (late in 2012). We must also expect fluctuations to have occurred within each three year period, so it would be wildly inaccurate to simply divide the latest number by three in order to get a yearly uptake of FHBs. But, for the sake of the discussion, let's do it anyway... 429,000/3 = 143,000. So let's say there are roughly 143,000 Australian first home buyers in any given year.

Next, we can see that of Australia's FHBs, 18% of them bought newly constructed homes in 2009/10. Again, this number is hardly indicative of long-running trends - the 2007/08 period saw 9% of FHBs in newly constructed homes. Changes to grants schemes - and a wealth of other factors affecting Australian housing markets - will ensure these numbers continue to move around from year to year. But let's take the latest figure again, just because it's there. 18% of 143,000 is 25,740. So let's say roughly 25,740 new homes are purchased and occupied by first home buyers in any given year.

Now, we can also see that, in 2007/08, 63% of Australia's FHBs moved out of a dwelling rented from a private landlord, in order to take up in their new digs. We'll have to make some more bold assumptions here: a) that this number is fair enough to compare to other data from the 2009/10 period, b) that this number applies across purchases of both new and existing dwellings and c) that every FHB household to exit the rental market is equal to one household of tenants. These assumptions are clearly wrong, but let's make them anyway. 63% of 25,740 is 16,216. So let's say roughly 16,216 rental properties are vacated each year due to FHBs building new homes and exiting the rental market.

16,216 new rental vacancies that negative gearing had nothing to do with? Imagine that! No doubt that's a foolishly optimistic conclusion, but it's a train of thought that's worth pursuing...


But what of the 108,134 new properties that are still unaccounted for in our example? We'll just have to assume that most of these are purchased by established mortgagors - upgraders who are looking to move into a nicer place. This means that, as part of the day-to-day cycle of second-hand house trading (and continuing with our rather clumsy simplification of the numbers), about 108,134 established dwellings will be added to the market by vendors each year. These houses will be bought by either investors or owner/occupiers - and it's fair to assume that many first home buyers would be looking at getting a hold of one of these if they could. But with 92% of landlords - the vast majority of them negatively geared and willing to borrow up big - competing with 82% of first home buyers, it's not hard to see why FHBs might struggle to keep up...

So, after all that, we're left with a combination of FHBs and upgraders paying overs for a new home, and negatively geared landlords gobbling up second-hand properties to rent out to those who've missed out. We refer again to earlier analysis from Chris, and quote:
  • negative gearing does not cause an individual landlord to charge less rent;
  • negative gearing does not create net additional rental housing;
  • negative gearing has contributed to more higher-income households renting, which both pushes rents up, and pushes lower-income households out of lower rent properties;
  • negative gearing has contributed to low-value proprties dropping out of the rental market, which pushes up the rent for those that remain in rental; so therefore
  • negative gearing is not your friend.
Policies that encourage the construction of new dwellings - whether by governments or by private interests (indeed, can't we have both?) - do a great deal more for the rental market than negative gearing ever will.

*Nb - Lawless uses the number of "dwelling approvals", which include renovations and alterations as well as new builds. These numbers must all be taken with a grain of salt.