Gavin R. Putland,  BE PhD

Wednesday, November 11, 2015 (Comment)

Negative gearing in context

My address to the meeting of the ALP Northcote branch, Tuesday, 20 October 2015 — invited by Ian Rogers, publisher & editor of Banking Day (I am not a member of the ALP).

NEGATIVE GEARING has two meanings: an investment strategy, and a tax lurk.

As an investment strategy, of course it means borrowing to buy an asset, and paying such an exorbitant price that the interest and other outgoings exceed the income from the asset.

In this context, gearing is apparently another name for operating leverage, which is a ratio in which the numerator is

REVENUE − VARIABLE COST

and the denominator is

REVENUE − VARIABLE COST − FIXED COST.

Now the variable cost is the cost that rises and falls with output. For a typical property investment, that doesn't seem applicable. So we treat all costs as fixed, in which case the numerator becomes simply REVENUE, and the denominator becomes

REVENUE − COST.

And if the cost exceeds the revenue, the denominator becomes negative; hence the whole ratio becomes negative; hence the term.

The interesting thing about that is that if the ratio is still positive, then, as costs approach revenue, the denominator approaches zero and the ratio approaches infinity. Hence a negative ratio is to be understood not as less than zero, but as greater than infinity! So if you thought negative gearing was a paradoxical concept, you were right — for more reasons than you probably realized.

Nevertheless, the investment strategy is the more literal meaning of the term. Negative gearing as a tax lurk means full deductibility for the investment strategy — that is, deducting all the interest and other outgoings although they exceed the income, or, equivalently, deducting the current loss on the investment against unrelated income.

The tax lurk is said to be inequitable because it's available only to investors, so that it helps investors to outbid prospective owner-occupants: if you buy a home to live in and the interest exceeds the rental value, you can't deduct the excess interest against your wage or salary; but if you let it to tenants, you can. Adding insult to injury, you can't deduct the cost of commuting to work against your wage or salary. The cost of commuting is deemed to be a “personal” expense, not a work-related expense, as if you always had the option of living where you work, or working where you live. The law doesn't admit that you live where you can afford it and get work where you can find it. But if you run a loss on an investment property, sure, you can claim that against your wage or salary.

Of course, negative gearing as an investment strategy isn't possible unless you have enough unrelated income — perhaps augmented by in-kind support — to be able to carry the current loss. Hence about half the benefit of the NG tax lurk goes to the top 20 percent of households, ranked by household income, and about a third of it goes to the top 10 percent (according to figures from NATSEM, quoted on the Macro Business blog and elsewhere). This obviously undermines the progressiveness of the income-tax system, leading to further criticism of the tax lurk on equity grounds.

Equally obviously, negative gearing as an investment strategy doesn't make sense unless you expect the income from the asset to grow and eventually exceed the interest paid or forgone on the purchase price, or (more likely) you intend to resell the property for a so-called capital gain, which anticipates the higher rent. (Speaking of which, I believe that we're on the traditional ground of the Wurundjeri people, who might have trouble buying it back at current prices.) In the main, this rise in value reflects greater effective demand for your property's location — that is, greater effective demand for the land — due to economic growth and the provision of amenities and infrastructure by agents other than yourself. In other words, in the main, the rise in value is an unearned windfall, or an unearned increment, or an economic rent — not a reward for productive activity.

That's inequitable in two ways:

  • First, those who create the capital gains don't get the full reward for their work and investment, because some of it is creamed off by property owners who just happen to own land in the right places; and
  • Second, it seems that capital gains are even more skewed towards high-income households than NG benefits — as we should expect, because some people can buy appreciating assets without going into debt. According to the discussion paper that preceded the Henry review, in 2005-6 about 2/3 of capital gains went to the top 10 percent of tax filers, ranked by taxable income. And remember that taxable income, by definition, includes only half of discounted capital gains, and doesn't include non-taxable capital gains on the principal residence. If we only consider the benefit of capital-gains discounting, then (according to NATSEM) about 73 percent of that benefit goes to the top 10 percent of households, ranked by household income — further undermining the progressiveness of the income-tax system.

Well, after all that pontificating about equity, you may be surprised to learn that I'm a social conservative and an economic dry, with a distinct Austrian libertarian streak. Far be it from me to complain about how we divide up the economic pie, while ignoring the size of the pie! But, if we want the nation as a whole to get richer, we want people to be creating wealth — not merely looking for ways to appropriate wealth created by others. Hence we don't want to encourage land speculation. And while it's possible for land speculators to be productive on the side — by building, or rebuilding, on the land that they buy and sell — they don't have to do those things in order to get their capital gains. Or to discount their capital gains! And under present rules, they don't have to do those things in order to claim the NG deduction! That invites criticism of the NG tax lurk, together with the capital-gains tax lurk, on efficiency grounds.

So, if the major political parties were true to their principles, the Labor Party would be criticizing the NG lurk because it's inequitable, and the Tories would be criticizing it because it's inefficient, and we'd have bipartisan agreement that something about it needs to be changed. Dream on!


When people speak of abolishing negative gearing, they obviously mean abolishing the tax lurk, not banning the investment strategy. But that doesn't resolve the ambiguity. It's often said that the Hawke-Keating government “abolished” negative gearing between 1985 and 1987; but what actually happened was that the investment strategy, for future investments only, was quarantined for tax purposes. Properties acquired up to 17 July '85 were grandfathered (that is, excluded from the changes), and for properties bought after that date, NG losses could be carried forward and offset against subsequent positive cash flow from properties bought after that date, whether the cash flow was in the form of rent or capital gains. In those days, the cost base for capital-gains tax was adjusted for inflation, but there was no discounting, so there was no need to specify whether carried-forward losses should be discounted if they were claimed against capital gains rather than current income. Nowadays, of course, we would need to specify that.

But in the current proposals for reforming NG, there's an embarrassing lack of specificity about the carrying of losses.

The recent report by Richard Holden for the McKell Institute, with the title “Switching Gears: Reforming negative gearing to solve our housing affordability crisis”, looks at 5 scenarios. “Scenario 1” is business as usual — a continuation of the present regime — and it produces an estimated tax expenditure (or revenue loss) of $51 billion over 10 years. Scenarios 2 to 5 involve some restriction on negative gearing, and all except “Scenario 5” involve grandfathering of past acquisitions. But nowhere is it specified whether the losses that can't be claimed against unrelated income would be carried forward. So I have to try to infer that from the stated impacts on revenue.

“Scenario 5” is the complete disallowance of negative gearing, with no grandfathering. It sounds simple, but its revenue impact isn't quantified because it would involve an unknown number of forced sales by established investors facing a sudden increase in their tax bills, and those sales would presumably change the tax status of the affected properties.

“Scenario 2” is similar to what the Greens are proposing — to disallow NG on future acquisitions while grandfathering past acquisitions. The effect on revenue is a bit rubbery because it depends on how fast existing debts would be paid down, which in turn might be influenced by the policy change: investors might pull their horns in and pay down their debts faster, and we don't know the magnitude of this effect. But according to the report (p.26), a faster debt reduction means a bigger boost in revenue; and if the debt is eliminated over 10 years, most of the tax expenditure is also eliminated over 10 years. Now it seems to me that if losses were carried forward, this wouldn't happen, because a faster reduction in debt means earlier offsetting of losses against gains and more offsetting of losses within the 10-year period, hence a smaller boost in revenue.

Hence I infer that losses are not carried forward. Hence, where the author says (p.24) that the calculated budget impacts are abstracted from other tax implications, I take that to mean that they're also abstracted from any carrying-forward of losses — although they apparently allow for growth dividends, as in “Scenario 4”.

“Scenario 4”, which is the preferred option [in the report], is to disallow NG on future purchases of established homes — in other words, allow negative gearing for new homes only, while grandfathering past acquisitions. This doesn't restrict NG as much as “Scenario 2”, but it's expected to increase revenue by a greater margin, because it encourages construction and its multiplier effects, and therefore increases economic growth, and increases revenue relative to “business as usual”. That's what I mean by the “growth dividend”.

In the Greens' proposal to disallow NG, it's clear that past acquisitions would be grandfathered, but again I haven't been able to find any explicit statement on carrying of losses. What the Greens do say is that the Parliamentary Budget Office has crunched the numbers and concluded that their proposal would increase revenue by $42.5 billion over 10 years. That's above the range of “Scenario 2” in the McKell report. Hence I'm inclined to think that if the McKell figures don't allow for the carry-forward of losses, neither do the Greens' figures. In both cases I'm ready to be corrected by better-placed sources, but that's the way it appears to me.

So, that explanation was a bit wonkish, but the take-away from it is clear: If you're going to propose any restriction on negative gearing, then, in the circumstances where you're not going to allow current losses to be deducted against unrelated income, you need to specify whether you're going to allow those losses to be carried forward. And if you're still going to discount capital gains, you also need to specify whether carried-forward losses must be likewise discounted before they can be offset against capital gains.


Defenders of the NG tax lurk love to pretend that it reduces rents by increasing the supply of rental housing.

But, to claim NG under current rules, you don't have to build a dwelling, or buy one off the plan, or even buy a newly built dwelling. You only have to outbid a prospective owner-occupant for an existing dwelling, thereby turning an otherwise owner-occupied property into a rental property, and an otherwise owner-occupant household into a tenant household, increasing both the supply of and the demand for rental accommodation and therefore doing nothing to reduce rents.

Furthermore, to claim NG, you don't have to get a tenant. You only have to claim that the property is available to let, while you hold out for a higher rent, or for a tenant household that you regard as more desirable. Double income and no kids, perhaps. Sure, it's illegal to discriminate against children; but every time your agent recommends tenants with children, you can reject them for some other reason, or no reason, until the agent gets the message.

The ability to claim your current loss against other income makes it easier to hold out. And anything that makes it easier for landlords to hold out will raise rents rather than lower them, because holding out reduces the effective supply; it means that some of the so-called vacancy rate is not a measure of supply, but rather a measure of holding back supply.

“Scenario 4”, the preferred option in the McKell report, addresses the first problem — by allowing future investors to claim NG for new homes only, with the proviso that “new homes” are defined as construction that increases the supply of dwellings and doesn't merely replace demolished dwellings. Past investors are not affected. Future investors who intend to build or buy new homes are not affected. Future investors who intend to positively gear established homes are not affected. The only investors affected by the changes are future investors who intend to negatively gear established homes. They have to choose whether they will proceed without the tax benefit, or not proceed, or retain the tax benefit by switching their attention to new homes. To the extent that they choose the last option, they will add to the supply of housing — either by directly causing new construction, or by paying out the builders who will then be free to move to the next project.

So the aim of “Scenario 4” is to make negative gearing work as advertised — to make it do what its defenders claim that it does.

This idea isn't new, by the way. It was suggested by the Acting Premier of WA, Mal Bryce, back in 1987, and considered by Paul Keating, who eventually took a different proposal to Cabinet, which eventually decided to reinstate full deductibility of NG. I didn't know that story until I read the McKell report. But I had been independently pushing the same idea — limiting NG to new homes, with grandfathering — since 2003. The McKell Institute floated the same idea, but without calculating the revenue impact, in its inaugural report in 2012. In April 2014, the same idea was reported to be under consideration by Treasury and the Parliamentary Budget Office, although it didn't find its way into the Budget; and it's been in the news, on and off, since then.

So, faced with this sensible and increasingly mainstream proposal, which would make negative gearing do what its defenders say it does, what did the Property Council and the Real Estate Institute of Australia do about it? They commissioned “independent” consultants ACIL Allen to produce a report on NG and the capital-gains discount and the potential impacts of reform. And that report, dated 12 June 2015, just happens to contain a section, numbered 5.5, with the title “Limiting negative gearing and the 50 per cent discount on CGT to new dwellings” (CGT meaning capital-gains tax). In that section, the “independent” consultants draw the conclusion that limiting tax concessions to new homes would raise both prices and rents. Their argument relies largely on confusing a lower-than-expected increase in construction with a decrease in construction. At one point they even allege that “a negative shock to the market... would likely result in upward pressure in prices.” But I guess that's what you have to do if the conclusion is predetermined and your job is to find arguments to support it.

I wrote a rebuttal of that section, under the title “Squaring the circle: The contradictory arguments against limiting negative gearing to new homes”, which was posted on the Land Values Research Group blog (lvrg.org.au) on 8 July, and reposted at Macro Business on the same day. I suspect it was this article that brought me to the attention of Mr Rogers, in which case he won't be pleased to hear that I've had to make a correction. Our “independent” consultants defined new properties to include vacant residential lots. To this I replied (and I quote):

...if you want to claim negative gearing on a residential property under the current regime, you need to claim that the property is rented or available for rent. That could be difficult if the property doesn't include a dwelling! The purpose of limiting negative gearing to new dwellings is to increase the incentive to build. Allowing negative gearing for vacant lots would reduce it. That's not the proposal.

(Unquote.) Well, in fact, you can claim negative gearing on a vacant lot that doesn't earn any income, provided that you buy it with the intention of producing income from it — for example, by building a house and letting it — and provided that you pursue that intention with due diligence. There was a case called Steele v. Federal Commissioner of Taxation (1999), in which the High Court found that there was sufficient intent, and another one called Temelli v. FC of T (1997), with different facts, in which the Federal Court found that there wasn't. And it must be conceded that if you could never deduct the interest for the land acquisition while waiting for the building to be built, this could be a barrier to construction, whereas if you could always deduct the interest for the land acquisition, you might never get around to building anything.

So it's conceivable that a policy of allowing negative gearing for new homes would retain the present regime concerning vacant lots. But it would not treat a vacant lot as equivalent to a finished home as our “independent” consultants seem to be suggesting. So I've annotated my article accordingly.

You can find that article by searching on the title: “Squaring the circle: The contradictory arguments against limiting negative gearing to new homes”. So tonight, rather than further discussing the economics of the issue, I thought I should say something about the politics of it.


When you're trying to sell desirable economic reform in the teeth of vested interests, you come up against what has become known as Brandolini's law — although I'm sure he wasn't the first to state it — to the effect that it takes 10 times as much energy to refute bull$#!t as it takes to produce it. Actually Brandolini used the term order of magnitude; and because he's a computer programmer, that might mean a factor other than 10.

An example is the claim that land tax is passed on in rents. To make that statement takes 7 words. To refute it, I've written a blog post of more than 2000 words, and I might get around to adding a few hundred more.

But negative gearing, at least to me, looks like an exception, provided that we maintain focus on the desired reform. When the usual suspects say negative gearing reduces rents, we can say: only if the negative gearers build more homes. When the usual suspects say negative gearing increases the supply of rental housing, we can say it also increases the demand as negative gearers outbid first home buyers and keep them on the rental market — unless, of course, the negative gearers have to build new homes. It each case the response takes a few more words, but it's still only one soundbite.

And if we strike first, the boot is on the other foot. The proposition is: if you have to build or buy a new home to claim negative gearing, more rental homes will be built. That fits into a 9-second soundbite. It fits into a 6-second TV grab. It fits into 140 characters. It passes the laugh test. It passes the sniff test. It passes the online comment test when I'm outnumbered 10-to-1 by Catallaxy readers. Which suggests to me that it probably also passes the pub test. ACIL Allen's attempt to prove the contrary proposition takes more than two pages, and fails to land a punch.

Now what about the urban myth that the quarantining of negative gearing between '85 and '87 caused skyrocketing rents? If you're going to adopt “Scenario 4”, all you have to say is: “What happened between '85 and '87 is irrelevant because that's not our policy. It's more like the Greens' policy. Ask them about it! Our policy is to increase the supply of housing by requiring future negative gearers to build or buy new homes.”

That said, even the Greens might do well to distance themselves from the Hawke-Keating policy, on the ground that they're proposing to spend the extra revenue on social housing and thereby increase supply.

What about the Tories? The Abbott cabinet publicly ruled out any changes to negative gearing. But on 29 June the Herald Sun reported, apparently reliably, that the government was considering disallowing NG for foreign investors in Australian property. And foreign investors are supposed to invest in new dwellings, not established ones. So disallowing NG for foreign investors does the opposite of “Scenario 4”: it restricts NG for new homes but not for established homes and thereby reduces construction — ingeniously harnessing xenophobia to restrict competition and prop up rents for the benefit of established investors. That should be easy enough to call out.

There is, of course, a countervailing political consideration. If Labor promises to limit NG to new homes, claiming that it will reduce or contain rents, and if the Tories attack the policy, claiming that it will increase rents, there is a class of voters who will say: “Well of course the Labor Party is telling the truth, and that's why we're not going to vote for it.” Your internal pollsters would know more about that than I do. But my impression, for what it's worth, is that this class is shrinking, because its offspring are taking too long to leave the nest, and then too long to commit themselves to home ownership.


There is also a countervailing economic consideration. Suppose that between now and the next election, it becomes undeniable that the bubble has burst and we've built too many high-rise dog-boxes. If we then change the NG rules so as to stimulate even more construction, won't that accelerate the crash in prices? Not necessarily. For one thing, if everyone know the crash is on, construction will stall anyway, and the problem might be how to avoid an overshoot and a shortage down the track. For another, continued overbuilding would certainly reduce or moderate rents, but that doesn't mean it would accelerate the crash in prices, because more building means more activity not only in the building industry but also in upstream and downstream industries, hence more jobs, hence more borrowing power, which props up prices for a bit longer, and buys a bit more time in which to get the economy onto a more productive, less speculative footing.

That raises three questions. First, what do we have to do to reduce prices, as distinct from rents? Second, do we really need to accept the political and economic risks of falling prices in order to improve affordability of home ownership? And third — the big one — what sort of policy would make the economy more productive and less speculative? Well, again I'm afraid that my response is a bit wonkish, but I hope it's redeemed somewhat by the fact that the questions are intertwined.

To the first question: In a rational market, the price of a property is the net present value of the future cash flows attributable to the property, including any eventual resale. But if you derive a formula for the price/rent ratio in terms of a appreciation rate, a discount rate, a holding time, and a set of tax parameters, you get a denominator with a subtraction in it (or several, depending on what simplifications you make), and there doesn't seem to be anything to stop that denominator from falling to zero, giving an infinite price, or even going negative, causing that “greater than infinity” problem that I mentioned in the introduction. And even if you plug in realistic values of the parameters, you can easily get price/rent ratios that obviously exceed the capacity to service loans. In that situation, the price of a property will be decided by whatever you can borrow against it; and in the absence of regulatory constraints, prices will be bid up until the financial system breaks.

Which brings me to a shameless plug: in the July 2015 issue of World Economic Review, which you can find via Google, I have a paper with the title “The Price Cannot be Right: Taxation, Sub-Intrinsic-Value Housing Bubbles, and Financial Instability”, which deals with that problem. It also shows that there's a very simple method of bringing price/rent ratios within sustainable bounds, and thereby stabilizing the financial system. That method is to impose a whopping great capital-gains tax with no exemption for the family home. The CGT can do the job by itself, without any stamp duty or rates or land tax, provided that the CGT rate is increased to compensate for any property taxes that it replaces.

To reduce price/rent ratios, the CGT only needs to apply to future purchases. But in that case the revenue stream starts at zero and rises gradually. If you want an immediate revenue stream to replace some other tax, then unfortunately you can't have any grandfathering of past acquisitions under the CGT — unless that “other tax” is the other property-transfer tax, namely stamp duty, in which case you can have a sort of partial grandfathering, in which you pay the stamp duty that would have been payable at the transition date, plus the CGT on the rise in value since then. But if the CGT only raises enough revenue to replace stamp duty, then it doesn't have the desired effect of reducing price/rent ratios. If we want that effect, we must either renounce any sort of grandfathering of past acquisitions, or be content with a gradual rise in revenue from the CGT, hence a gradual replacement of other taxes besides stamp duty. If we pick the first option — no grandfathering — then we have to argue that the abolition of some other taxes is the compensation for the CGT.

To the second question: If we want to make rents and prices more affordable without being crucified by the property lobby or the financial lobby, we need a period of time — the longer the better — during which rents and prices continue to rise, but spending power rises faster.

One ingredient is to improve supply, so that the rise in spending power isn't entirely competed away. And of course one way to improve supply is to limit negative gearing to new homes. Another way is capital-gains tax, because it increases the attractiveness of current income relative to capital gains, and therefore encourages land owners to build accommodation and seek tenants. And if you want that incentive to work with current landowners, then again you need to apply the CGT to the increase in value from this time forward.

Another ingredient is infrastructure, which raises rents and prices in the serviced locations. This is compatible with affordability, in the sense that the higher rents and prices are for improved amenity, not the same amenity. Furthermore, that amenity tends to translate into spending power — for example, because it lets you get by with fewer cars or reduces your travel distances.

Speaking of infrastructure, if the capital-gains tax on each property is refunded to the State in which the property is located, or imposed by the State itself, then the State has the incentive and the ability to invest in infrastructure which causes such capital gains. This benefit of CGT is in addition to its ability to replace stamp duty, because the infrastructure funding comes from the consequent uplift in property values whereas the stamp-duty replacement comes from the property values as they would have been without the infrastructure. Again this is compatible with the partial grandfathering arrangement in which you pay the stamp duty on the value at the transition date plus the CGT on the increase in value since then. But again, if you want to replace other taxes besides stamp duty, you can either renounce grandfathering or wait for the CGT revenue to rise.

To the third question, one answer is now obvious: One way to discourage speculation and encourage production is to raise capital-gains tax and cut taxes on productive activities. But if you want the rise in capital-gains tax to be grandfathered, the only other tax that you can abolish immediately is stamp duty. The rest have to wait for the capital-gains revenue to build up.

If that wait isn't politically acceptable — and indeed it probably isn't — then you need to consider what abolitions of existing taxes might be acceptable to the voters as compensation for an ungrandfathered capital-gains tax on the family home. I'll offer two hints, and then wind up, leaving the hints hanging in the air.

First hint: By my rough calculations, an ungrandfathered capital-gains tax on property, levied on real gains, not nominal gains, at a rate of about 45 percent, would raise enough revenue to replace conveyancing stamp duty, payroll tax, State insurance taxes, and most of the existing land tax, so that land tax could become an unoccupied land tax, which would increase the incentive for landlords to find tenants and pull the rug under almost all of the arguments against land tax, be they valid or invalid. (Feel free to throw those arguments at me as soon as I finish!)

Second hint: In the latest Tax Expenditure Statement from Treasury, for 2014-15, it seems that the revenue sacrificed by the Commonwealth through capital-gains tax concessions and superannuation concessions is greater than the revenue raised by the... GST! $46 billion sacrificed from capital gains on the family home, plus about $30 billion from super, makes about $76 billion in forgone revenue — compared with a budgeted $57 billion for the GST.

I thank you for inviting me.


Creative Commons License         Return to Contents
comments powered by Disqus