Why Demand Rules the Housing Market

One boundary first. This is about the owner-occupier market, the price of buying a home to live in. Rents run on a different mechanism, the balance between demand for housing services and the dwellings available, and are a separate story. And, while both markets operate over the same total housing stock, what follows is about the price of buying.

Most goods compete for a share of your budget, and when they get expensive, someone makes more of them and the price eases. Housing competes far less than most, and in the places people most want to live it does closer to the opposite. It climbs toward whatever buyers can pay, because the supply that would hold the price down cannot easily be added there. In every market both sides - supply and demand - set the price. This essay is about why, in housing, the demand side rules, and why the supply side is so unusually unable to answer it. Take them in turn.

Two goods in one house

A house is really two goods sold together. One is shelter, a normal need. You want a roof, and one roof is enough, so demand for shelter is satiable like demand for anything else. The land underneath is different. It is not a need but an asset, and while one roof is enough, the value of the land position under it has no such limit. There is always a better-located, scarcer piece of ground to want, from Blacktown to Parramatta to Point Piper. And the two legs move in opposite directions. The building depreciates like a car and stays standing only through maintenance. The land is not manufactured, so where a location is wanted its scarcity value can rise without limit. That is not automatic. Land in a dying town falls like anything else. But in the places people compete to live, the gain sits in the land, not the building. When people say their house tripled, the house did nothing of the sort. The land under it did. So to buy a home is mostly to buy land, and that is where the trouble starts, because you cannot secure the shelter without being pushed to buy the asset.

The national accounts show the same split in aggregate. The value of Australia's dwelling structures has held roughly steady as a share of the economy for decades, because the stock is constantly repaired, renovated and replaced, so it neither wastes away nor runs ahead. Land has done all the moving. Dwellings flat, land tripled, and the gap between them is the whole story.

Owning is irresistible

Look at the asset you are pushed into, because for an ordinary household it is the most heavily privileged one in the economy, and not by accident.

Some of the pull is real and would exist under any system. Owning gives you secure tenure, which renting does not. It lets you consume housing without paying anyone rent. And it hands you, at the end, an appreciating asset you hold for life. Set against an insecure rental market, those are genuine advantages.

The rest of the pull is built by policy, and it is decisive. Four privileges do most of the work:

• The capital gain on your home is completely tax-free.

• The rent you save by owning, your imputed rent, is untaxed, unlike the return on almost any other asset.

• The family home is exempt from the pension assets test, so wealth poured into it buys a pension as well.

• Mortgage credit is uniquely accessible: cheap, long, and lent against collateral so bankable that an ordinary household can hold a large leveraged position no other asset would allow.

Two lesser tilts sit beneath them. First-home grants lift what buyers can pay at the margin, so where supply is tight they largely capitalise into the price and help the seller as much as the buyer. And some state land taxes fall on investment property while exempting the home you live in, though every house still pays rates.

Stack those together and housing becomes a leveraged, capital-gains-tax-free, appreciating asset with the softest downside in the market. Prices do fall. They just rarely fall far, or for long. The central bank has no house-price target, but it does not need one. Household mortgage debt is now so large that a serious housing downturn drags the wider economy with it, and that, when inflation allows, is exactly the condition that pulls rates lower. So easing arrives to rescue the economy and catches housing on the way down, an accidental floor rather than a deliberate one. That safety is real but unplanned, and the buyer does not need to know how it works. They only need to see that everyone who held on got richer.

Not owning is punished

The push matters as much as the pull, and it is the half most accounts miss. Not owning is not a neutral alternative. It is punished. Rent is insecure, controlled by someone else, and it rises for life. It buys shelter and nothing else, no asset at the end, and it sits on the wrong side of the very land boom enriching owners. Run it to the end of a life and the renter keeps paying for shelter long after the owner's mortgage is gone. The owner still pays rates, insurance and upkeep, but the big monthly payment stops. The retirement system is calibrated around that difference, so the pension stretches for someone with no rent to pay and strains for someone still paying market rent. For a low-income lifelong renter, the safety net can fail at the exact age they can least recover. The choice is not house versus flat. It is security against precarity, and a comfortable old age against poverty.

Why buyers stretch so far

Put the pull and the push together and owning stops being a free choice. The outside option is so unattractive that households stretch further for it than for almost anything else. A buyer straining like that does not bid what a house is worth. They bid close to what they can finance, because missing out feels nearer to ruin than to inconvenience. Not every buyer maxes out, but enough sit close enough to their borrowing limit that when the limit moves, the winning bid moves with it, and the price is set at that margin. This is why housing absorbs so much of what people can pay. Hand households more, a rate cut, a pay rise, a grant, and much of it does not stay with them. It goes into the bid, and where supply cannot answer, it capitalises into the land. The price of a home is then less about the value of the shelter than about what the winning buyer could borrow.

Why prices outran income

Prices have not merely risen. They have risen faster than incomes, and the reason sits inside the demand story rather than beside it. If housing takes whatever a household can pay, then anything that lifts what a household can pay lifts the price. Over the past fifty years three things lifted it a long way, and only one of them was income.

The first was real wages. On average across the half-century, pay grew faster than the cost of living, so households brought more in real terms to the auction. The second was the shift from one income to two. As women moved into paid work, the single-income household gave way to the dual-income household, and a couple with two salaries can borrow far more than a couple with one. The third, and probably the largest, was the long fall in interest rates, first when inflation was tamed in the early 1990s and rates came down from the double digits, then again after the global financial crisis, when policy rates sat unusually low for more than a decade. A given repayment services a far bigger loan at three per cent than at ten, so borrowing capacity climbed even when wages did not.

Housing did what it does with capacity where supply is tight. It absorbed much of it. For the first couples with two salaries, the second income was an advantage, letting them outbid the rest. But once most competing couples had two as well, two incomes became the price of entry rather than an edge, and what began as extra buying power for some hardened into the price faced by all. Cheap credit did the same: a bigger loan is an advantage until everyone has one, and then it is just the going rate. So the gains raised the bar rather than clearing it. That is why prices outran income in particular, because two of the three forces lifted household borrowing capacity faster than individual wages. And both are now close to spent, since no one can add a third income and rates cannot fall again from seventeen per cent to two, which is why the early rate rises of the 2020s pushed prices down, before other demand pressures reasserted themselves.

And because it now takes two incomes to buy, people typically buy later, after they have partnered. The single first-home buyer in their mid-twenties has given way to the couple in their thirties. The age of entry drifts upward, household formation waits on the mortgage, and anyone trying to buy on one income is quietly priced out of the market the second income built.

Ricardo and George

This is old knowledge, not a new complaint.

David Ricardo explained rent two centuries ago. As population grows and poorer land is brought into use, the best-placed land earns the difference between what it yields and what the marginal, worst-placed land still in use yields. That gap is the rent, and the owner collects it for holding the better ground, not for improving it. In his framing the price of the crop sets the rent, not the other way round, and as the margin is pushed onto worse land the gap widens and the rent climbs. Swap farm yield for urban access and the resemblance is plain. The rent of a location is its advantage over the marginal one, and it grows as the city grows around an owner who lifts not a finger.

Henry George saw what that meant for a whole society and called it the great enigma of the age. Why does poverty deepen as nations grow richer? Because part of every gain that makes a place more productive or more wanted, higher productivity, more people, better infrastructure, is captured by the scarce land around it rather than staying wholly with the people who create it. Where that capture outruns wages, the worker runs to stand still while the landowner banks the difference. George's answer was to tax the unearned value of land itself, which is why land tax remains the one lever that reaches the cause.

The residual claimant

So housing demand is not like other demand. A house is a near-necessity, a positional claim on land that is hard to reproduce, and a leveraged bet on your future income, all at once. No other good is even two of these. Together they make the land under your home behave, where supply is tight, like a residual claimant on your income: the more households can bring to the auction, the more of that surplus ends up in the site price rather than in living standards. That is why affordability tends to worsen as the country grows wealthier, which looks like a paradox and is not one. The richer we get, the more there is for scarce land to take.

All of which raises the obvious objection. If demand is this fierce and prices this high, why does supply not do what supply is supposed to do, and compete the price back down? Fat margins are meant to pull in builders who undercut each other until price meets cost. In housing that never happens, and why it never happens is the other half of the argument.

Competition works by substitution

Competition lowers prices by producing more of a good, or a close enough substitute, until rivals undercut each other toward cost. For the building, that works well. Builders compete hard, and construction is supplied close to cost. For land it is harder, because you cannot copy a location. But that does not mean competition cannot reach it. You cannot make a second Paddington, yet you can offer substitutes: the next suburb over, an apartment on the same street, more homes on the same block. Each is an imperfect copy that draws off some of the demand and shaves the scarcity rent.

The catch is that this substitution weakens exactly where it is needed most. The more singular a location's advantages, the harbour view, the ten-minute commute, the sought-after school, the fewer real substitutes exist, so the scarcity rent there is the hardest to compete away. Competition disciplines the bricks everywhere, and disciplines the land only where good substitutes can be built. Which raises the question of the one substitute that copies a location almost exactly, by building up.

Metering needs concentrated control

Even where land can be built on, the holder often has reason not to rush. Recall that a house is two assets, and in a wanted location the land is the one that can appreciate. Release a big parcel all at once and you flood the market you are selling into and push down your own price, while releasing in stages protects the price and lets the land you still hold keep gaining. So staged release can be the rational strategy, and the large estate developers do hold land and drip it out over years.

Two honest qualifications keep this from overreaching. Much slow building is not strategy at all, because developing thousands of dwellings simply takes years and gets held up by finance, infrastructure, labour and approvals. And the metering motive works only where control is concentrated, where one owner holds enough of the supply to move the price by withholding. Break the land into many separate owners who can each build as-of-right, and the logic reverses. Each races the others to market before prices soften, and none can meter. That is the difference between a market that withholds and one that competes, and it is the hinge the density argument will turn on.

Supply cannot answer in time

The deeper problem is not that supply refuses to come. It is that it cannot come in time. Financial capacity can reprice almost overnight. When rates fall, buyers can borrow more and lift their bids within weeks, well before any new home could be built. Supply cannot answer at that speed. Land has to be found, rezoned, serviced, financed, approved and built, and that takes years. So a rate cut capitalises into land long before a single extra home appears to compete it down. What supply does come tends to arrive late in the boom and dry up in the bust, so it is weakest exactly when it is needed most. The supply-siders are right that the fix is to build faster, and we should. But faster is not instant. Nothing you do to the planning system lets a house be built the week rates are cut. As long as demand can reprice overnight and supply takes years, the demand shock wins the race, and the gains settle into land before the homes arrive. Whether they stay there is a second question, and it turns on how much building eventually comes. Where supply is quick and contestable, later homes compete much of the premium away. Where it stays slow and constrained, the shock hardens into a persistently higher price. Australia has had the second kind.

Density does reproduce location

There is one serious answer to all of this, and it deserves a real hearing. The densifiers argue that you can copy a location after all, by going up. The scarce good was never land area. It was access to a place. And while you cannot make a second square kilometre of inner Sydney, you can put fifty homes where one stood, and that multiplies the supply of living there even though the ground is fixed. Density is the technology that reproduces location, vertically. It is the one supply lever that attacks the scarcity at its root instead of adding fringe land nobody wants. And it is not just theory. For the two decades to the mid-2010s, Tokyo, a rich and hugely desirable megacity, held real housing costs roughly flat while building at a rate that would be unthinkable here. Where a city genuinely lets density happen at scale, prices behave. More recently, Tokyo's prices have risen sharply while Japanese construction has weakened, a reminder that even an elastic system does not abolish scarcity when delivery slows.

The right to build is not the act

But upzoning grants permission, not delivery, and the two are easy to confuse. Two things happen the day the rezoning lands. The new development rights capitalise into the land at once, so the existing owner is handed a windfall whether or not anything is ever built. That is real, and it is why upzoning so often enriches whoever owned the dirt. But it does not by itself mean the homes fail to come. Whether the permission turns into dwellings depends on the rest: how fast approvals convert to construction, how contestable the sites are, and whether one holder controls enough of them to meter. Where sites are many and building is genuinely as-of-right, the windfall and the homes both arrive, and the homes compete the price down. Where sites consolidate into a few large holdings, or approvals stay slow and discretionary, the permission can sit as a land value while delivery trickles. So upzoning is necessary for elasticity, but not sufficient. It is the start of supply, not the act of it.

What Tokyo actually proves

So the real question is what makes density deliver, and here I will concede more than the demand-side case usually does. Tokyo shows that supply can win. Its zoning is national and permissive, building is largely as-of-right, and ownership is fragmented, so thousands of small players build continuously and none can meter. That combination genuinely overwhelms much of the land-scarcity mechanism, in one of the most desirable cities on earth. If supply is elastic, fragmented and fast enough, it works.

A weaker appreciation story helped Tokyo too, since after the 1990 crash land was no longer expected to rise, which took heat out of demand. But I will not lean on that, because it makes the theory too easy to defend. The load-bearing difference is the supply system itself. The honest claim is not that strong demand always defeats supply. It is that supply has to be genuinely elastic to overwhelm it, and Australia's is not.

Australia risks the other outcome

Which is what makes Australia's institutions the problem. Our upzoning is discretionary and negotiated rather than as-of-right, and our best locations are politically walled off by incumbent owners protecting amenity and scarcity premium alike. And where developable sites consolidate into relatively few hands, as they can in greenfield and large-project markets, strategic staging can weaken competition further, since an entrant who buys in pays the going price for land, which becomes their floor, so they tend to join the staging rather than break it. Those are the conditions under which permission sits as a windfall and delivery trickles. So mass upzoning here risks the other outcome, not Tokyo but metered projects on land whose uplift went to its owners, adding homes too slowly to move the price. It is a risk, not a certainty, and it varies by state and site. But our institutions make it likely. And the scale of the constraint is not hypothetical: the Reserve Bank estimated that by 2016 zoning had pushed Sydney detached-house prices about 73 per cent above the cost of their physical inputs, roughly 42 per cent of the final price, with large effects in the other capitals, and it notes that as demand grows those restrictions bind more tightly and push prices up further.

The verdict on density

So the densifiers are right about the mechanism and half right about the cure. Density is the one substitute that copies a location, so it belongs at the centre of any serious answer. But upzoning alone is not it. Density lowers prices only when the permission converts into enough building, which means it has to be as-of-right and contestable, so that many builders compete and none can meter. Whether the government captures the rezoning windfall is a separate question. It matters for fairness, since a public decision should not simply enrich whoever owned the land, but it is not what decides whether prices fall. Done the first way it works, as Tokyo shows. Done the Australian way, discretionary and negotiated, it is the metering with more storeys.

Why helping buyers raises prices

This is also why the policies that promise to help people buy tend not to. A grant, a deposit scheme, a tax break, a rate cut, all of them work by lifting what buyers can pay, and where supply cannot answer, what buyers can pay is the price. So demand-side help mostly ends up in the land, raising the bar it was meant to lower. A subsidy that lifts what buyers can borrow feeds the bid, not the supply. It is not that these policies do nothing. It is that where supply is tight they do close to the opposite of what they promise.

So supply is the lever

Which is the case, in the end, for leaning on supply. If demand-side help only capitalises into land, then supply is the lever that changes the outcome rather than the sticker price. Much of the bid cannot easily be switched off, since it runs on credit, tax, retirement and the plain fact that people must live somewhere. The levers that would touch demand directly, land tax, tenure reform, credit limits, are real, and some can move faster than people think. But supply has an advantage they lack. It changes the quantity of housing rather than the size of the bid, so unlike a subsidy it does not just feed the auction, and unlike a credit squeeze it improves affordability without rationing ownership by shrinking what buyers can borrow. That is the case for supply, and it does not need the other levers to be impossible.

But it has to be the right kind. Targets and blanket upzoning do not do it, because permission is not delivery and a windfall is not a home. What works is the elastic, contestable, as-of-right kind that turns permission into building quickly and lets enough builders compete that none can meter, with the rezoning windfall captured for fairness rather than for price. Supply first, but not supply naive, and not while the other hand keeps widening the bid.

There is another way to read the same failure: if the market's constitution cannot be changed, you bypass it and deliver ownership publicly, Singapore-style. That is a different model and an argument for another day, and I take the supply route here because I think the market can still be fixed.

Two blades, one cut

Set the two halves together and the picture closes. In every market both blades cut (supply and demand set the market price). In housing they cut at different speeds. Demand reprices quickly and stretches toward what buyers can borrow, so cheap credit and rising incomes are bid into land. Supply could answer, but only slowly, and only where it is fast, fragmented and built as-of-right. So two things set the price. Speed sets the opening bid, because demand moves first and supply lags. Elasticity decides how much of that bid survives, because only enough later building competes the premium back out. Australia has had both, demand that reprices fast and supply too slow and too constrained to take much of it back over thirty years.

Plenty of markets have slow supply. Oil, hotel rooms in a boom, a fixed stock of Old Masters, all can spike when demand jumps and supply cannot follow. Housing is not unusual for being one of them. It is unusual for stacking slow supply on top of everything else here: mortgage leverage that turns future income into a bid today, a positional claim on scarce location, a wall of policy privilege, and an underlying shelter need that makes walking away costly. No other good carries all five.

Which is why the demand side rules. Demand sets the bid, and supply decides how far it runs. A house is not one good but two, a shelter you need and the land you compete for, and where the land cannot be added fast enough, the bidding decides the price. Ricardo and George saw where that leads. We have built an economy that keeps proving them right.