What this is
Home purchase prices in Australia have risen a long way, faster than incomes, for at least a generation. There is a great deal of argument about why, and what to do. This piece is a tour of that argument, drawn on a common set of axes so the competing views can be compared rather than shouted past each other.
It helps to see the argument as having two levels. The first is a threshold question: is there an affordability problem at all? One serious view says no, or rather, that what looks like a problem is the market working exactly as a property market works. Prices reflect what people can pay; the distress is the market clearing. If that view is right, most of the rest of the debate is misconceived, so it has to be dealt with first. The second level opens only once you have decided there is a problem. Then a cluster of schools agree that something is wrong but disagree, often sharply, about what the problem is and what to do. Supply is too constrained. Population pressure is too high. Government has withdrawn from building. We should subsidise people to purchase. Each is a different diagnosis and a different lever.
So the tour goes: first the view that says there is no problem to solve, and why I think it is not the whole story; then the cluster of problems and solutions; then what I make of it. This is a tour of the headlines, not an exhaustive catalogue. I have picked the positions and the policies that carry the argument, and left aside the many smaller schemes and variations; a reader after the full inventory will need to look elsewhere.
I will stay, as much as I can, in the language of supply and demand. Not because a supply and demand diagram settles anything, it does not, but because it is the cleanest way to see where two careful people actually disagree. Usually they do not disagree about the facts. They disagree about which fact was doing the work.
The focus is home purchase prices, not rents and not mortgage repayments. Those are three different affordability problems and they often move against each other, but the purchase price is where the structural story is clearest and where the schools genuinely part company.
One premise sits underneath the whole debate, and it is worth stating plainly because every school works from it. Land in cities is definitionally scarce. The good locations, close to jobs, transport and amenity, are fixed; you cannot manufacture more inner-city. So at any given population, the only way to put more dwellings on the same well-located land is to build up. Densification is not one option among several; it is the thing the argument is about. Where the schools differ is on what follows from it. Does freeing density lower prices, as the supply school holds? Or does something, rising demand, developer capture, the loss of amenity that density brings, eat the gain before it reaches the buyer? The scarcity of urban land is the shared premise. Whether densification delivers cheaper homes is the contested question, and much of this piece is, in one way or another, an argument about it.
The charts, and what they quietly assume
The diagrams in this piece are Marshallian supply and demand crosses, with two adjustments that matter.
First, the vertical axis is the real price. Inflation is netted out, so a rising price means housing got more expensive relative to everything else, not just that all prices drifted up together.
Second, the horizontal axis is dwellings per person, not dwellings, which nets out population growth. Australia's population grew by roughly half over twenty-five years, and left on the horizontal axis it swamps everything else and the picture becomes unreadable. Per capita removes it. The axis also assumes, for simplicity, that compositional preferences for housing have not meaningfully changed over time, the mix of household types and the kind of dwelling each wants. They have shifted, of course, and a fuller account would allow for it; but holding them steady keeps the diagram legible and the compositional changes are second-order beside population. Per capita clears the big distortion and keeps the picture readable. We are largely assuming other changes away for the sake of argument.
With both axes adjusted, a movement in the picture means something. A real price rising while dwellings per person falls is a clean statement that housing got dearer and scarcer at the same time, with population and inflation already accounted for.
There is a practical payoff to the per-capita axis worth spelling out. Population growth pushes both curves to the right at once: more people want more dwellings (demand out) and a growing city builds more of them (supply out). On a raw dwellings axis you would have to draw both of those shifts in every diagram, and the picture would be a tangle of multiple lines before any school's actual argument even appeared. Putting dwellings on a per-capita basis absorbs that shared rightward drift into the axis itself. What is left to draw is only the movement a school is actually claiming, the steepening of supply, or the climb in capacity to pay, against a curve that can sit still. The axis does the population bookkeeping so the diagram does not have to. This is particularly important with the argument that the solution to housing affordability is to reduce immigration.
One honest caveat, stated once here so I do not have to keep repeating it. Like all models, these diagrams simplify. When I draw a single, unmoving demand curve, I am not claiming that demand stood still over twenty-five years. Of course it moved. I am claiming that, in the particular school's account of the price rise, the movement of demand was the less consequential part of the story, and the movement of supply was the consequential part. Drawing demand as a single line is a way of saying "hold this roughly still and watch what the other curve did." A different school will hold supply still and watch demand. The disagreement between the schools is, almost exactly, an argument about which curve to hold still. The diagrams make that argument visible.
These are comparative statics: deliberate, all-else-equal thought experiments that isolate one mechanism. The real history had everything moving at once, which is precisely why the schools can each find support for their story, and why the honest account, at the end, lets both curves move together.
The threshold view: the market is largely working
Before any of the schools that propose a fix, there is a view that says there is nothing here to fix. Or, more precisely, that what we are looking at is not a malfunction but a market doing exactly what a property market does. Prices are high because that is what an unsubsidised market for a scarce, highly desirable good produces. The distress is the market clearing. On this view, high prices are precisely what we should expect when government steps back from providing housing, and the breathless crisis talk mistakes normal functioning for breakdown. This school starts with the simple observation that for at least 100 years, for most people their first home purchase was always a challenge, scary and risky, consuming around a fifth of their household income. It also notes that in every decade over that period, people have complained about housing affordability. There's nothing new here.
This is worth taking seriously, because it is not complacency. The claim is sharper than "everything is fine." It is that the market is working, and a working property market is, among other things, a machine for keeping housing dear, because the people who already own it are the majority and they do not benefit from it getting cheaper. Everyone says they want affordable housing; almost no one with a stake actually wants cheap housing, which would mean their own asset falling. So the market delivers what its participants, weighted by power, actually want: stable or rising prices, dressed up as a problem nobody can quite solve.
On our axes, this view holds demand as the active force and treats supply as close to fixed. Real prices have climbed slowly, year on year, because the capacity to pay has climbed, with incomes (wages typically grew faster than inflation), with the long fall in interest rates, with the shift to two-income households, with the loosening of credit. Housing absorbs that rising capacity because shelter is wanted close to the top of the budget, ahead of almost everything else. Push more dwellings into the market and they are absorbed at a rate that protects the price; build them and the rising capacity to pay simply meets them at a level that keeps the market clearing where it always clears. So loosening zoning does not lower the price by much. It changes where building happens, not how much is paid. The lever the supply school wants to pull is, on this view, pulling on the wrong curve.
The honest force of this position is real, and two parts of it are, I think, correct. Capacity to pay genuinely is the engine behind the rising demand curve, and a great deal of the post 1990 price story is the fall in interest rates (compared with the 1970s and 80s) lifting how much a given income could borrow, compounded by the move to two earners. And the political economy is real too: a country of homeowners does not vote for cheaper homes, which is why governments reach so readily for measures that look like help without delivering the price falls that would hurt their own constituents.
In this view, the word doing the work, though, is "largely." There is one exception that matters, and it is on the supply side. The development industry is not competitive in the way the textbook market assumes. Land in the good locations is concentrated, each site is in effect a local monopoly, and developers are not benign conduits who build whatever they are allowed to and pass the savings on. Whatever the cause, deliberate release management by firms that know buyers will pay close to the most they can bear, or financing and holding costs, planning risk and delay, or simple concentration, the supply response is slower and weaker than the competitive model predicts, and the price stays above the cost that real competition would drive it to. On the strategic-release version, the rational play is not to flood the market and compete the price down to cost; it is to meter supply, releasing each tranche at the rate that keeps it scarce and sells it near the buyer's ceiling.
The point does not depend on developers being villains, only on the response being sluggish: the price-cost wedge is not competed away. It is held open and captured. But the mechanism has a floor, and this is where I part company on the dose. Metering only pays while price sits above the cost of building; a developer gains nothing by withholding stock he could only sell at or below his margin, so the wedge that can be held open is bounded by how far price stands above cost, not by the developer's appetite. Where that gap is wide, in the well-located inner suburbs, release management is real and does the work this school says it does. Where it is thin, the story has little left to explain, and what looks like withheld supply is mostly supply that was never worth building at the going price. So I take the mechanism, and grant it the inner-city ground where it bites; I do not grant it the whole shortfall, because below the margin there is no economic rent to meter.
The consequence is that prices are sticky downward. On the way up they move freely: when capacity to pay rises, the supply side captures it at once as higher prices. On the way down there is no matching force, because nothing on the supply side ever pushes prices lower. Developers do not readily cut to clear stock; they slow release and wait. So prices come down only when the demand side forces them down, that is, only when capacity to pay itself contracts, typically when interest rates rise and borrowing capacity falls, or when government changes the settings, credit rules, prudential limits, tax treatment, that determine how much a given income can pay. Prices rise on their own and fall only when pushed. A market that ratchets up easily and resists coming down is working, but it is not working competitively, and the difference is paid by the buyer.
What pushes, mostly, is the interest rate, and that is worth following through, because it drives a cycle that the threshold view rightly treats as normal. Rates fall; borrowing capacity rises; the demand curve moves out and prices rise. Higher prices widen the developer's margin, the gap between what a dwelling sells for and what it costs to build and finance, and building is a margin decision, so commencements rise. The extra building, and the activity around it, helps the economy run hot, and eventually the Reserve Bank lifts rates to cool it. Borrowing capacity falls, the demand curve retreats, prices soften, margins thin, and building slows. In time the cycle turns again. This is a self-correcting loop, and it is one of the better arguments for the threshold view: the market is not malfunctioning, it is oscillating, with the central bank leaning against it.
The thing to see is that both curves oscillate, but out of phase. Demand reprices almost at once when rates move, because borrowing capacity does. Supply moves slowly, because building takes years from approval to completion. So the new dwellings called forth by cheap money tend to arrive just as demand is already turning, which is why building so often seems mistimed, and why the market overshoots in both directions. Demand lurches; the slow building line chases it, always a step behind.
And here is the catch that keeps this from being a clean vindication of the market. The cycle oscillates, but it oscillates around a trend, and the trend rises. Capacity to pay grows secularly, with productivity, with real wages, with the long fall in rates; the supply side grows only slowly, against the constraint. So the demand trend outpaces the supply trend, and each peak is higher than the last, each trough higher than the trough before. The oscillation is two-sided; the divergence underneath it is one-sided. A working, self-correcting cycle, riding a trend it cannot escape, is exactly how you get a generation of rising real prices without anything that looks, quarter to quarter, like a breakdown.
This also explains a feature of the public debate. The complaints arrive at the expensive part of the cycle, the peak, when demand has surged and supply has not yet caught up, and they fade in the trough. The threshold view reads this, fairly, as evidence that much of the crisis talk is cyclical, the top of a normal cycle mistaken for a permanent emergency. And it is partly right: rhetoric timed to the peak should be treated with some suspicion. But the complaint at the peak sits on top of a real secular deterioration, because the peak is higher every cycle. The person whingeing at the top is feeling both the cycle and the trend at once, and dismissing them as merely cyclical misreads the moment as badly as treating it as the end of days. There is a wry corollary for later: because the noise peaks at the top, governments are pushed to act precisely when the cycle is about to turn anyway, so a demand-side handout thrown in at the peak both pumps demand into the dearest moment and then takes the credit when the cycle turns on its own.
But "the market is largely working" cannot be the end of the story, and the reason is the peculiar nature of the good. In an ordinary market, when a good gets too dear, people go without and the shortfall is visible: they do not buy, and the empty shelf or the unmade sale signals the unmet demand. Housing is different, because almost no one goes without shelter except in the extreme. Instead of going without, people compromise. The adult children who do not leave home. The share house that was meant to be temporary. The family that takes the long commute because the near suburb is out of reach. The household that never forms, the move that never happens. None of this shows up as unmet demand. It shows up as people quietly accepting less, and the market clears at a price that looks, on the surface, like equilibrium. I have argued this at greater length elsewhere, in a piece on the elasticity of housing demand; the short version is that a good no one can refuse has a demand curve that bends rather than breaks, and the bending is invisible.
So the market is working, in the narrow sense that it clears. But for a good that people cannot walk away from, clearing is not the same as adequacy. The compromises are the unmet need, and they are nearly invisible because they take the form of degraded conditions rather than absent consumption.
This is not a separate claim bolted onto the threshold view; it is the same fact the threshold view already runs on, seen from the other side. Capacity to pay says how much a buyer can bear; inelasticity says they bear close to it rather than do without. They are two descriptions of one mechanism, which is why rising capacity to pay flows into price instead of being declined, and why prices track what people can borrow so closely. If demand actually broke when housing got dear, the capacity-to-pay engine would not turn. That it plainly does is the evidence that demand bends rather than breaks. The compromises are what the bending looks like from inside the lives doing it.
There is a sharper version of this, and it sits inside the very mechanism the threshold view treats as benign. The move to two-income pricing is not just a number that lifted the demand curve. Lived forward, it changed the shape of people's lives. When a home requires two incomes, buying one is delayed in the life cycle until after pairing, so ownership becomes conditional on finding a partner, and the timing of family is pushed back to wait on it. And once a household has bought, both must keep working to service the price, which forces the trade between work and family, between paid childcare and a parent at home, onto young families who may not have chosen it.
It also resets the floor under everyone trying to buy on a single income, and makes many of them the losers in the new world. A price built for two earners is a price that many single earners cannot reach, however hard they work, because the benchmark moved out from under them. Single people are structurally disadvantaged by it. So are the separated and the widowed: a household that could carry a home on two incomes cannot carry it on one, so the loss of a partner now comes with a housing demotion it did not carry when homes were priced to a single wage. The capacity to pay that the threshold view counts as the engine of rising prices is, from inside these lives, a set of felt pressures: form a couple before you can buy, buy later, both work whether you wanted to or not, and fall out of reach entirely if you are on your own. That is not a market clearing without cost. It is a cost displaced onto the people the price reshaped.
A working market that clears by making people accept less, and hides the shortfall in the accepting, is not evidence that there is no problem. It is the mechanism by which the problem is concealed. That is why I do not stop here, and why the rest of the tour is worth taking.
The YIMBY view: supply was strangled
Start with the supply-side school. It is really an argument with two faces, and it is fairest to name both. On one side are the NIMBYs, "not in my back yard", the residents and neighbours who, over the years, asked the planning system to protect their amenity, their views, their heritage, their quiet leafy streets. On the other are the YIMBYs, "yes in my back yard", who argue that the accumulated weight of all that protection is what strangled supply and drove prices up, and that it must now be wound back. The two are the same argument seen from opposite ends: the NIMBY impulse built the constraint, the YIMBY reaction wants it removed. This is the most organised and, in its best form, the most rigorous of the positions, and the YIMBY side has been winning the public argument in recent years.
The diagnosis is that home prices rose because supply was prevented from keeping pace with demand, and that the prevention was largely a matter of planning. Demand grew, as it always does, with incomes and population and the rest. The pathology is not that demand grew; demand growth is normal and expected and would have been absorbed if supply had kept pace.
The pathology is in how the supply response was throttled, and it is worth being precise about the agency here. Supply was not strangled by some outside force. As a community, over decades, we allowed the planning system to consider more and more factors, each one reasonable on the day it was added, and each a small further brake on how readily new dwellings could appear. Amenity, heritage, overshadowing, traffic, environmental impact, the right of neighbours to object: every additional consideration we asked the system to weigh was a consideration that could delay, shrink, or stop a project. The cumulative effect of all these reasonable additions was to reduce the supply response. The growing demand then met a stock that could no longer expand at the old pace, and the price climbed. We did this to ourselves, slowly, with good intentions at every step.
On the adjusted axes, with population controlled, the story shows up as the supply curve steepening and shifting to the left over time. The supply curve of the past, drawn to the right and relatively shallow, is a world in which more dwellings could be brought to market without much more cost. The supply curve of today is steep and sits to the left: each additional dwelling per person is harder and dearer to deliver, and fewer of them appear. Equilibrium climbs from A to B, almost straight up. Price rises a great deal; dwellings per person fall. The demand curve barely needs to move for this to happen, which is the point.
It is worth being detailed about what "supply was strangled" actually means, because the strength of the YIMBY case rests on the accretion being real and specific, not a slogan. Over decades, a great many separate constraints have piled up, each defensible on its own day, that together steepened the supply curve:
Zoning that fixed low-density envelopes over established suburbs, so that the land best served by existing infrastructure and closest to jobs cannot be built on at any real intensity. Height limits and floor-space ratios that cap what a given parcel can yield. Heritage overlays and view-sharing and overshadowing rules that constrain individual sites. Minimum lot sizes and car-parking minimums that raise the cost and lower the yield of every project. Development assessment processes that have lengthened from months to years, with each additional referral, objection right, and appeal stage adding holding cost and risk. Infrastructure and developer contributions that load a large fixed cost onto each new dwelling before a brick is laid. Environmental and servicing requirements that gate land release on water, sewerage and transport capacity that governments have been slow to fund.
The YIMBY argument is that this accumulated thicket is the steepening of the supply curve, and that most of it is not load-bearing. Much of it could be wound back without anyone being harmed, and the winding-back is the solution.
But the honest version of the case has to concede three things, and the better advocates concede at least the first two.
The first is that not all of this regulation is unnecessary, and some of it is necessary in a way that no reform should touch. Flood and fire zoning exists because people die otherwise. Transport corridors and the reservations for water, sewerage and power are land you must hold idle now so that a city of the future can function; build over them and the loss is permanent and ruinous to reverse. Parks and open space, once built on, are gone, and a denser city needs them more, not less. Contamination, structural safety, the things that protect the people who will actually live in the building. These are not amenity niceties. They are the part of the steepened supply curve that should stay steep, and a serious supply agenda has to separate the constraints that earn their cost from the ones that do not.
The supply school has a test for this, and it is a good one. A restriction is justified, on this view, only if it offsets an externality, a cost imposed on others, at least as large as the cost the restriction itself imposes by holding back supply. Where the externality is real and large, the constraint earns its place; where it is small or imagined, the constraint is just suppressing housing and should go. As a principle this is hard to fault. The difficulty is in the application, because the test is only as good as your ability to value the externality, and the load-bearing constraints are often the ones whose externalities are hardest to price. What is the value, today, of a transport corridor that a metro line might use in 2050? What is the cost of the lost light, the lost quiet, the changed character of a low-rise street? These are real, but they do not come with numbers. So the externalities test does not remove the judgement; it relocates it into the valuation of things the market never priced. Whether we have "enough" housing, on this test, turns out to depend on how you value externalities that cannot be valued with any confidence. That is where the supply school's certainty sometimes outruns its evidence.
The second is that not all of the cost is planning. Some of the steepening is in the building industry itself: a construction sector with flat productivity for decades, concentrated land holdings, and labour arrangements that raise costs and slow delivery. This is where the conduct of the construction unions belongs, the recent CFMEU troubles being the sharp end of it, alongside the developer concentration on the other side. Freeing the planning system does not, on its own, fix a building sector that is slow and dear for reasons that have nothing to do with zoning. So even granting the diagnosis in full, the supply curve has a component that planning reform cannot reach.
The third concession is the one the supply school is least comfortable with, and it is not really a factual matter at all. The standard reply to the amenity objection is that density is efficient: dense cities are more productive, have shorter commutes, lower per-capita infrastructure costs and emissions, and Hong Kong, Tokyo and Singapore show that very large populations can be housed well at very high density. All of that is true, and it is measurable. But the same density that one person reads as efficient and alive, another reads as an urban jungle: towers where there were houses, canyons where there was light, crowding where there was quiet, the loss of the suburban form they deliberately chose. That loss is real to the people who feel it, and it is precisely the kind of thing the market cannot price. So the disagreement about how much density is too much is not a dispute one side can win with data. It is a clash between a measurable gain in efficiency and an unmeasurable loss of amenity, and weighing the two against each other is a judgement, not a calculation. The supply school's confidence comes partly from counting only what can be counted. The resistance it meets comes from people defending something that does not show up in the count. Neither side is being stupid, and no supply and demand diagram adjudicates between them.
The solution is the diagnosis run backwards. If accumulated constraint steepened the supply curve and pushed it left, then removing the constraint that is not load-bearing flattens it and pushes it back to the right. On the diagram, supply moves from its steep present position back out, and equilibrium slides down the demand curve from B toward A: price falls, dwellings per person rise. That is the whole YIMBY prescription in one move, and where the constraint being removed is genuinely discretionary, it is hard to argue with.
The questions to hold, as we leave this school and turn to the next, are the ones its own diagram raises. How much of the steepening was discretionary regulation that can be wound back, and how much was the necessary floor and the building sector itself, which cannot? And what stops a freed-up supply from being captured as developer margin rather than delivered as homes? Those are the questions the next schools, in their different ways, are really answering.
Reduced immigration
Some argue that the answer to Australia's housing affordability is to reduce immigration. The logic is immediate and intuitive: fewer people arriving means less demand for housing, and less demand means lower prices. Cut the inflow and the pressure eases.
This is the school where the per-capita axis earns its keep, because it changes what the lever looks like. Immigration changes the number of people. On a dwellings-per-person axis, the number of people is exactly what we have already netted out. So to a first approximation a change in immigration does not move the demand curve: each person still wants roughly the same housing, and per capita, demand is largely unchanged. The lever that everyone files under "demand" turns out, on the axis that matters, to be doing much less to demand than the headline suggests.
That "largely" is doing some work, and it is worth being honest about it. Migrants are not an average slice of the population. They cluster in Sydney and Melbourne, skew younger, rent before they buy, and may form smaller households. So a migration surge can lift housing demand per person a little, and can lift it sharply in particular places, not just add to the population count. The per-capita axis nets out the numbers; it does not fully net out the composition. So immigration is not quite the pure population story the axis would make it. But the bulk of the effect is still the population effect, and that is what the axis lets us see clearly.
What it actually touches is supply, and only for a while. Picture the recent surge rather than a cut, because that is what actually happened and the data is clean. When migration jumped after 2022, the number of dwellings did not suddenly fall. But each dwelling now had to house more people, so dwellings per person dropped: a short-run negative on supply, not because we lost any housing, but because the same stock was spread across more people. A migration cut is the mirror image: a short-run positive on supply, as the same stock is spread across fewer people, helped along briefly by a construction pipeline calibrated to the old, higher growth, and by any dwellings vacated when people leave. Both are temporary. Both are the stock being stretched or loosened by a sudden change in the rate of population growth, and both fade as building and the population path re-settle.
So the immigration advocates are right that the recent gap is real and that it is migration-driven. You can see it in the chart: the breakeven line jumps above the building line precisely when migration surges. And the national chart, if anything, understates the local pressure, because migrants concentrate. A national dwellings-per-capita ratio that looks tolerable can sit on top of acute shortage in inner Sydney, Melbourne and Brisbane, where the arrivals actually land. Immigration can worsen that spatial mismatch even as the national ratio later normalises. What the chart also shows is why that gap is a transitional thing and not a structural one. It is the product of a slow-moving rate of building meeting a fast-moving population, and the mismatch is a matter of timing, not of dynamics. Building does catch up; it is just slow to turn. The volatility is already subsiding, the breakeven line has peaked and is coming down toward the building line, and building is drifting up to meet it. Cut migration and you lower the breakeven line faster, closing the gap sooner. But the gap was always going to close as the surge normalised and building caught up, and nothing in this adjustment touches the things that set the price. It is a within-market adjustment to a change in the rate of population growth, working itself out at the speed the construction pipeline allows.
Two honest qualifications, before this sounds too reassuring. "Transitional" is a word about models, not about lives. A stock-adjustment that takes several years to work through is, for the tenant living through it, a multi-year rental squeeze, and there is nothing trivial about that. Temporary for the model is real for the renter. And the catch-up is not guaranteed to be quick. The argument that building eventually adjusts assumes building can adjust; where planning, finance, labour and serviced land are all constrained at once, the adjustment can be slow enough that "temporary" stretches into the better part of a decade. The transition is real, and it can be long and painful.
There is a sting in the tail of cutting, too. If a sustained cut persuades builders that demand has permanently softened, the one part of the system that was at least stable, the steady flow of building, will eventually shrink. Fewer firms, fewer trades, less capacity. When population growth resumes, as it tends to, the supply side is smaller and slower than before. The lever that bought a brief dip can leave the system worse placed for the next upswing.
Here is the deeper reason immigration sits awkwardly in this tour. For every other school, I can draw the problem on the supply-and-demand axes, because every other school is making a claim about the structural equilibrium: supply is too constrained, or capacity to pay is the engine. Immigration is harder to draw that way, and the difficulty is the point. A change in the rate of population growth produces an adjustment, the stock stretches or loosens, building eventually recalibrates, the ratio re-settles, but it is an adjustment to a change in the rate, not a change in the dynamics that set the price. The capacity to pay, the weak competition on the supply side, the inelastic stock, the fact that shelter cannot be refused: these operate at any level of population and any rate of growth. I should not overstate this. There is one channel by which sustained migration does reach the structural picture: expected future population growth is capitalised into land values today, because developers and investors price in the trajectory, so a credible long-run migration program lifts land prices now, ahead of the people arriving. That is real, but it is an indirect effect that works through expectations, and it is second-order beside the capacity-to-pay engine and the supply constraint. So immigration mainly affects the transition, and only indirectly touches the deeper pricing mechanism. It moves the economy along the path, and nudges the path through expectations; it does not set the shape of the path.
So I would put the claim carefully, because the careful version is the one that holds. Immigration can intensify housing stress when population growth jumps faster than the construction pipeline can respond, and it can do real damage in the places and years where the surge and the lag collide. The mismatch runs both ways: a surge makes building too little for the population, and a drought, like the border closure, briefly makes it too much, which is why rents softened for a while when the borders shut. Both are the same thing, a sudden change in the rate of population growth that the slow building line takes years to meet, and both unwind as the two re-converge. But that is mainly a stock-adjustment problem, not the deep pricing mechanism. It worsens the squeeze; it does not explain why the squeeze exists. Slow the arrivals and you ease the adjustment; you do not touch the things that set the price, and those are what make Australian housing structurally expensive once population growth steadies.
There is a sharper version of the immigration argument that has to be met, because it is the strongest one. It says: migration is not a one-off shock but a persistent, high inflow, and building has long lags, so the adjustment never actually finishes; the temporary mismatch becomes effectively permanent, and the transition is the equilibrium. Grant the premise, because it may well be true. It still does not rescue the immigration diagnosis; it converts it into the supply diagnosis. If a steady inflow keeps outrunning the building response year after year, the thing that is failing is the building response. Migration is then not the cause of the structural problem but the load that reveals it: a supply side elastic enough to keep pace would absorb a persistent inflow without a permanent squeeze, and a supply side this steep would produce rising prices under any sustained source of growing demand, migration or otherwise. So even in its strongest form, the persistent-migration story points back to supply responsiveness as the binding constraint. It tells you to fix the supply curve, not the inflow.
None of this is to say the level of immigration is beyond legitimate debate. It bears on wages, on services, on infrastructure, on inflation, on a great many things, and reasonable people weigh those differently. The breakeven chart makes the narrower point better than any argument: the line that lurches is population, and building, slow to turn, follows it at its own lagging pace. The gap between them is a stock-flow mismatch, painful while it lasts, that the market grinds closed. It is real. But it is also temporary. And it is not the underlying reason that housing is expensive.
Negative gearing and the capital gains discount
Two tax settings have been blamed, together, for a great deal: negative gearing, which lets an investor deduct a rental loss against other income, and the capital gains tax discount, which has taxed only half the gain on sale. The pair made geared residential investment attractive, sheltering income along the way and taxing the profit lightly at the end. The long-standing argument is that this tilts the field toward investors and away from owner-occupiers, bids up the price of existing homes, and that winding it back would help affordability.
As of the 2026-27 Budget this is no longer a hypothetical. On 12 May 2026 the government announced it would reform both, with effect from 1 July 2027 (the measures are announced, not yet law). The design matters more than the headline, and it is cannier than the blanket change debated for a decade. Negative gearing against salary income will be limited to new builds: buy an existing dwelling after budget night and you can deduct rental losses only against residential property income, not your wage, though losses carry forward; buy a new build and full negative gearing survives. The fifty per cent CGT discount is replaced by inflation indexation plus a minimum thirty per cent rate on realised gains, again with new builds able to choose the old treatment. Everything already owned is grandfathered.
The carve-out for new builds is the tell, and it is to the government's credit. The complaint about negative gearing was never really that investors exist; it was that investors bidding for the existing stock add nothing to supply while adding to price. This design removes the concession for precisely that, buying established homes, while keeping it for the thing that adds dwellings. It is an attempt to redirect investor capital out of price competition for existing homes and into financing new ones. Read in the language of this piece, that is a tacit admission that supply is the binding constraint: the policy is built around the very point the whole debate keeps circling.
So what does it do? On the market for existing dwellings, it lowers what investors can pay, by removing the tax shelter that topped up their return. Investor demand for established homes eases, and the purchase price softens. Their capacity to borrow is reduced. That is the affordability benefit, and it is real. But here the analysis has to go further than the advocates usually take it, because the concession was doing something specific: it was substituting for yield. It let investors accept a low rental yield, because the tax treatment made up the rest of the return. Take it away and yield has to do that work itself. So rents drift up, not as a one-off, but until the rental yield rises far enough that property clears an investor's required return without the tax break. At that point investors come back, on a different proposition, paid by rent rather than by tax and expected capital gain.
This is why I am ambivalent about the policy rather than for or against it, and why the published numbers, read carelessly, mislead. Treasury's estimates are all small: house prices perhaps two per cent lower over a couple of years, about nineteen thousand dollars on a median home; rents up by less than two dollars a week; thirty-five thousand fewer dwellings over ten years, roughly three and a half thousand a year, much of it offset by the new-build carve-out and an infrastructure fund, so most market economists read the net supply effect as close to neutral. None of these is a large number. And the shape matters as much as the size. Almost every private forecast shows the same pattern: a brief shortfall, prices growing more slowly or dipping for a year or two, and then a return to normal growth from a slightly lower base. The two per cent is the depth of a transient dip, not a permanent discount.
That shape is exactly what the mechanism predicts. Investor demand for established homes eases, so prices step below trend for a while. Then rents drift up, yields recover, investors return on a yield basis, demand comes back, and the price path rejoins its trend. So the purchase-price benefit is real but temporary, a shortfall that the market grows out of, and in the long run it is close to a zero-sum game for prices. The durable residue is not cheaper houses; it is a small, permanent rise in rents, the level at which yield-driven investors were drawn back in. There is a fair case that the rent effect is larger and more lasting than the headline two dollars, because as population grows against a rental pool the change has made slower to expand, the pressure compounds. The concession had been holding rents down and prices up at once; removing it does the reverse on both, then mostly unwinds on the price side and persists on the rent side.
Whether that is a good trade is a genuine question, not a slogan. It shifts cost from buyers, who get transitional price relief and a less speculative market, onto renters, who bear a rent rise that compounds. The new-build carve-out softens the rental hit, by keeping investor money flowing into new supply, which is the design's real virtue. And the grandfathering that makes it politically survivable also makes it gentle to the point of being slight for years, since only new purchases of existing stock are touched and the investor base turns over slowly. So it is a smarter lever than the one everyone fought about, deliberately calibrated to be mild, whose real effects will be small and slow, and which still does not touch the structural drivers, capacity to pay and the supply constraint, that set the price in the first place. Reasonable people will weigh the buyer-renter trade differently. I find I genuinely do not know whether it is worth doing, which is, I think, the honest position once you follow the mechanism all the way through.
Please note: A narrower verdict on this reform is not the same as a verdict on the capital gains discount itself. My ambivalence here is about what it does to house prices and rents, where the effect is small and two-sided. The broader case that the change to the discount is poor tax policy, for reasons that have little to do with housing, I think is stronger, and I have set it out separately.
The 5% Deposit Scheme, and demand-side subsidies in general
Then there are the demand-side subsidies: the first-home grants, the stamp-duty concessions, the shared-equity schemes, the deposit guarantees. These are the most popular interventions of all, because they are visible, immediate, and feel generous. They are also, on the economics, the closest thing in this whole tour to a policy that does the opposite of what it claims. I am not ambivalent about these.
The live example is the Australian Government 5% Deposit Scheme, expanded on 1 October 2025. It lets a first home buyer purchase with a five per cent deposit, two per cent for a single parent, with the government guaranteeing the gap to twenty per cent so the buyer avoids lenders mortgage insurance. The October expansion removed the income caps and the limit on places and lifted the price thresholds, turning a targeted, rationed scheme into a broad, uncapped boost to first-home-buyer borrowing power. It now stands behind better than one in three first home buyers.
The mechanism is the one we have now seen several times, and it is unforgiving. Removing the deposit hurdle and the mortgage-insurance cost lets a whole cohort buy sooner and borrow more against their savings. That is a rise in what they can pay. Against a steep, slow-moving supply of the entry-level dwellings these buyers compete for, more capacity to pay does not summon many more homes; it raises the price. The help is capitalised into the price of the starter stock. The seller captures it. The buyer hands over the saved deposit and a little more, and ends up no better off, often worse, because the price rose and they are now carrying a ninety-five per cent loan into it. The government's own promotion gives the game away: it boasts that a Brisbane buyer can purchase a one-million-dollar home on a fifty-thousand-dollar deposit. Enabling larger purchases on smaller savings is exactly how you bid up the price of the homes being purchased.
There is no solution chart for this section, because the policy is not a solution to draw. The single diagram says it all: demand shifts right, supply barely moves, price rises, and the subsidy ends up in the vendor's pocket. Worse, it is self-perpetuating. By lifting entry-level prices, it raises the very bar it is helping people over, so each cohort of buyers needs the scheme more than the last, because the last cohort's use of it nudged prices up. The subsidy manufactures the need for the subsidy. And it loads first home buyers into thin-equity, high-leverage mortgages at those elevated prices, which transfers risk squarely onto the people it claims to help.
I will not be neutral about this one. Demand-side subsidies of this kind are, for the most part, populist nonsense. They are chosen precisely because they are visible and because they threaten no incumbent's asset, indeed they gently inflate it, and their failure to improve affordability is not an unfortunate side effect but the very reason they are politically safe. A policy that actually lowered prices would hurt the owners who vote. A policy that raises prices while appearing to help buyers offends no one who matters and earns a grateful headline. The uncapped 5% Deposit Scheme is what a government reaches for when it wants the appearance of action without the substance, and it works, as politics, precisely because it fails as economics.
All we need is more publicly funded supply
The last position is the one most often offered as the obvious answer: build more public and community housing. If the market will not deliver enough affordable homes, the state should. It is the prescription I am least sure about, not because the impulse is wrong, but because the argument for it is usually made without the step that would justify it.
Start with the question an analyst has to ask of any intervention: what market failure makes it necessary? This is the discipline I applied to the planning constraints earlier, and it cuts the same way here. Most of the popular case does not survive it. "There are not enough homes" is not a market failure; it is a market outcome, and if the problem is quantity, the remedy is to let the market build more, not to have the state do the building. "Homes are too dear" is not a market failure either; it is the market clearing against a high capacity to pay. Neither complaint, on its own, tells you the government should become a builder.
There are genuine candidates, but each points somewhere more specific than "build lots of public housing." Weak competition on the supply side is a real failure, but its clean remedy is to fix the competition, not to add a public builder alongside the private oligopoly. The private capture of publicly created land value is real, but it argues for value capture, betterment levies, public land development, not for public housing as such. The non-optional nature of shelter is the closest thing to a true failure, because it means the market clears by forcing compromise rather than by people walking away, but even that argues for housing the people being squeezed out, not for a program to move the general price.
What is left, when the efficiency arguments have been worked through, is a merit case, and it is a real one. Shelter is a good a society may decide everyone should have to a decent standard, regardless of capacity to pay, like basic healthcare or schooling. That is a normative judgement, not an efficiency one, and it is perfectly legitimate. But notice what it justifies: a targeted floor for those the market will not house, the very low income, the vulnerable, the people for whom no amount of market functioning produces a home. It justifies a safety net. It does not, by itself, justify public housing as a tool to lower the price for everyone. And this is where the public-supply argument usually goes quiet.
Grant the merit case, though, and you still have not finished, because you then have to choose the instrument. There are two: help people pay for housing in the market, or build housing outside it. Rent assistance, the first, is fast, flexible and portable; it follows the person to where they want to live, and it can be raised tomorrow. Commonwealth Rent Assistance does real, quiet good, keeping a large number of low-income renters out of the worst stress. But it has the flaw we have now seen several times: in a tight market it capitalises into rent. Give renters more capacity to pay against an inelastic supply and the landlord captures much of it. So it helps the individual recipient, who is invisible to their own landlord, while in aggregate leaking into the very rents it was meant to relieve.
Public housing, the second instrument, does the opposite on the diagram, and that is its real virtue. It does not add to demand; it adds to supply. On our axes it shifts the supply curve to the right, and because it adds stock rather than capacity to pay, it lowers the equilibrium price a little for everyone, even though the dwellings are let to low-income tenants. A demand subsidy pushes price up; public building pushes it down. In a supply-constrained market, that is the difference between a policy that backfires and one that helps. So if the question is which instrument moves price the right way, public building wins.
But, and it is a large but, public building does not change the slope of the supply curve. It moves the steep curve a touch to the right; it does not make it less steep. The structural inelasticity, the weak competition, the slow pipeline, all remain. So the relief is only ever as large as the quantity built, and to move the price meaningfully against a steep curve you would have to build a great deal, continuously, year after year, because population growth keeps pushing the requirement back out. And here the walls appear. The first is fiscal, and it is about permanence rather than size: a one-off injection is politically easy, but the sustained, growing capital commitment needed to build at scale forever is the line that loses every budget contest, which is why the last forty years show occasional bursts of public building and long stretches of neglect, and why the public share of the housing stock has shrunk rather than grown. The second wall is the supply constraint itself: public building faces the same planning system, the same construction sector, the same concentrated land and limited labour as private building. Government does not escape the steep curve; it queues in front of it like everyone else. And the evidence is in the delivery. The flagship fund was set a modest enough target, forty thousand social and affordable homes over five years, around eight thousand a year. Yet roughly seventeen months into that sixty-month window it had completed about eight hundred and ninety homes, with some nine and a half thousand more under construction and the rest committed on paper or still in planning. Two per cent finished; the bulk of the promise still a pipeline. The money was there and the contracts were signed; the homes came slowly, because building them ran into the same constraints that slow everyone else. Promised acceleration may yet lift the numbers, but the lesson of the first stretch of the program is plain: even a funded, flagship effort with a modest target delivers public housing years behind the promise.
Put the numbers beside the need and the limit is plain. Around ten thousand a year, even delivered in full, sits against a shortfall that runs into the tens of thousands of dwellings a year. It is the right kind of lever, pointed the right way, at perhaps a seventh of the scale the gap would require, and undeliverable even at that. So public supply is legitimate as a targeted floor, the merit case holds, and it is the one instrument that moves the price the right way. But it is oversold as a general solution: it does not change the structural slope, it cannot be built at the scale that would matter, and the fiscal commitment it needs is precisely the one governments have shown, over four decades, they will not sustain. Even a reasonable answer, it turns out, is trapped behind the same supply constraint as everything else.
What I make of it
A tour like this is supposed to end with the answer. I do not have one, and I have come to think that anyone who does is telling you more about their certainty than about housing. What I have instead is a way of holding the whole thing at once, and a few conclusions that survive the holding.
Start with the picture that ties the sections together. Two things have been moving for a generation. Capacity to pay has risen, with productivity and real wages, with the long fall in interest rates, with the shift to two incomes and the loosening of credit. And the supply curve has steepened, building has become less and less responsive to all that capacity. The clearest single piece of evidence is the plainest chart in the set: completions per thousand people, the supply flow with population netted out. It moves in great cycles, three big building booms across forty years, but each boom peaks lower than the one before, and the latest reading is the lowest in the record, even as real prices reached the highest. Price up, building per person down, sustained over decades. That is what a steepening supply curve looks like in the data, and it is the structural fact the whole tour keeps circling.
So who is right? I take much of the market-largely-working machinery: capacity to pay is the engine, competition on the supply side is weak, the supply response is slower and weaker than the competitive model predicts, prices ratchet up and fall only when something pushes capacity to pay down. Those observations are sound, and I have leaned on them throughout. My quarrel with that school is narrower than disagreement: I take its mechanisms but I reject the conclusion some draw, that because the market is working in its own terms, supply-side intervention is futile. The very facts it gets right, weak competition, land-banking, a steep and unresponsive supply side, are not the case against intervention. They are the case for it. They just point to different interventions.
The supply school I resist differently. Its case is real, and the piece has granted it at length: supply does matter, the steepening is real, and it is a large part of the story. My disagreement is with the confidence, not the substance. Its proponents press a sound observation past what it will bear, into a clean headline number and a clean policy conclusion, and the cleanness is what I distrust. The constraints that steepened the curve are not all of a kind. Some are load-bearing, the flood zoning, the corridors, the utility infrastructure, and cannot be removed. Others are defended by an amenity loss that is real even when self-interested. Sorting one from the other is slow, contestable, parcel-by-parcel work, and the externalities that decide it are precisely the ones no one can price with confidence. So the certainty with which the supply school advances its conclusion is more than the evidence supports. Winding back the constraint is also far harder than the optimists allow.
The supply school also offers what it presents as a clean test for "enough": keep building until granting another development right no longer moves the land price. The trouble is that this test can never read "enough," because there are always rights we withhold, for flood, fire, contamination, a transport corridor a future city will need, and relaxing any of them would move the price. Pressed on this, the test narrows to a fairer claim: enough once supply is held back only by justified externalities. But that hands the whole question to the valuation of the externality, and the load-bearing constraints are precisely the ones that cannot be priced. When should land be held for a corridor a railway might use in 2050? The cost of getting it wrong is enormous and unknowable today, and the decision is about preserving an option over decades, something a static supply-and-demand cross has no axis for. So the test that looks precise is precise only because it prices the benefits of restriction at close to zero. Admit a single unpriceable constraint and it becomes the same judgement everyone else is making, dressed as a measurement. The headline figure that falls out of it, the great stock of "missing" homes (some put it as high as 1.9 million dwellings), is best read as a ceiling, not an estimate of what reform could actually recover. There is a death or a complaint behind almost every rule, a fire, a flood, an overshadowed yard, a street someone fought to keep, and much of that cost is buying the thing the rule was made for. Some of it is genuine waste, the parking minimum that protects nothing, the limit drawn for a view long gone, and that should go. But sorting the waste from the load-bearing is slow, contestable, rule-by-rule work, and neither the headline number nor its critics can yet say how much is truly recoverable. A ceiling is not an estimate, and the work that would turn it into one has no agreed method, only judgement.
Hold both and the levers fall into place, and the verdict on each is sobering. Reduced immigration is a sugar hit: it eases the squeeze on the stock for a while, then the relief unwinds, and the dose is paid for with consequences across the rest of the economy, in wages, in skills, in inflation, in the workforce, that have nothing to do with housing. Trimming negative gearing and the capital gains discount is, in the long run, close to a wash for prices: a transient dip that reverts as investors return on yield, with the durable effect landing on rents rather than on purchase prices. It is better designed than the old blanket version and unlikely to do much good or much harm for house purchase affordability. The deposit scheme is simply bad policy: it pours capacity to pay into a steep, uncompetitive curve, the help capitalises into the price, the vendor pockets it, and the buyer is left carrying more debt for the same home. And public housing has real benefit, it is the one lever that moves the price the right way, but it is hard to sustain, because the commitment it needs is permanent and the building runs into the same constraints as everything else.
So what is left to do, once the magic cures are set aside? More than nothing, and less than anyone is selling. There are legitimate supply-side reforms: freeing the constraint where it is genuinely discretionary rather than load-bearing, releasing well-located land where the amenity cost is low, applying the honest test of whether a given rule offsets an externality at least as large as the cost it imposes. There are real microeconomic reforms, the ones the largely-working diagnosis actually points to: penalties on holding land idle so that banking it stops paying, measures against the concentration of land and development, attention to the building sector's flat productivity and the frictions in its labour market, reforms that make the supply side behave more like a competitive market and less like the sluggish, concentrated one it is. And there is a place for a small, sustainable public housing program, not the occasional grand splash that the budget cannot keep, but a modest, continuous one that builds steadily and compounds over decades, which is the opposite of how we have done it and probably the only way that works.
None of these is a cure. Each is partial, slow, and modest, and they work, to the extent they work, together and sustained, not singly and in a hurry. That is the unglamorous truth, and it is why the politics defeats it. The reforms that would actually help have no natural constituency and threaten interests on every side: the constraint that protects the leafy suburb is defended by the owners one party depends on, the building sector and its labour frictions are defended by the unions the other party depends on, and the land-bankers and concentrated developers like the system as it is. So each party can pull only the lever that hurts the other's base, which means neither pulls a real one, and the system defaults, again and again, to the demand-side gesture that offends no one and quietly makes things worse. The clamour peaks at the top of the cycle, the government reaches for the visible handout just as the cycle is about to turn, and the handout takes the credit while the steepened curve and the rising capacity to pay grind on underneath.
Which leaves one piece of advice, and it is the through-line of everything above. Be wary of the one-trick pony. This debate is run by advocates, each holding a single lever and selling it as the answer: just build, just cut immigration, just scrap the tax breaks, just help buyers, just have the government build. Each is right about something real and wrong to promote that something into the whole story. The confident single cause and the single cure are not the marks of insight; they are the marks of advocacy, and the more certain the pitch, the more it should be distrusted. Housing is expensive for several reasons at once, the reasons interact, and the honest response is a patient portfolio of modest reforms, held together over a long time, by governments that mostly will not.
I keep coming back to where the working market hides its failure: not in empty shelves, but in the compromises people make because they cannot refuse shelter. The household that does not form. The family that takes the long commute. The single income locked out of a market priced for two. The move that never happens. Those are the stakes, and they do not show up in the clearing price. Whatever we choose to do, or more likely not do, that is what it is finally about: whether the next generation can form a household and a home with something less than the whole of two lives spent paying for it.
No comments:
Post a Comment