The supply flood is not arriving. This is the big one

Market analysis · Eastern Suburbs

Sydney housing affordability: what a unit actually cost in 1990, 2000 and today

A Sydney unit took 4.7 years of income to buy in 1990. Today it takes 8.2. But the repayment figures tell a different story to the one most people expect — and it explains the empty open homes.

Alan Weiss · Weiss Real Estate

There is a particular quality to a Saturday morning open home with nobody in it. The agent standing in the kitchen. The register with two names on it, one of them a neighbour.

I have been selling property in the Eastern Suburbs since 1990, and I have felt that quiet twice before — once in the early nineties, once in the years either side of 2000. It has come back. So I went and pulled the numbers to work out whether it has come back for the same reasons.

It has not. And the difference matters more than almost anything else being written about this market.

How many years of income does a Sydney unit cost?

Start with the simplest question. How long would you have to work — every dollar, before tax — to buy the median Sydney unit?

The price figures for 1990 and 2000 come from Abelson and Chung’s Sydney median unit series, built from Valuer-General and Real Estate Institute data. The income figures are the Australian Bureau of Statistics measure of average weekly ordinary time earnings for a full-time adult, annualised. Same definition, same source, all three periods.

Since 1990 the median Sydney unit has multiplied 6.4 times. The average full-time wage has multiplied 3.7 times. The gap between those two numbers is the entire story.

What surprised me is how evenly it happened. Affordability deteriorated by a factor of 1.31 across the 1990s, and by a factor of 1.33 across the twenty-six years since. Roughly the same again, spread over nearly three times as long. This was not a shock. It is a grind, and it has been running the whole time I have been in this business.

1990: when the barrier was the price of money

In January 1990 the Reserve Bank cash rate hit 17.5 per cent. Standard variable mortgage rates sat around 17 per cent.

What followed was the recession Paul Keating called the one we had to have. On the national accounts, GDP fell 1.5 per cent between the June quarters of 1990 and 1991. Unemployment went from 5.7 per cent in the December quarter of 1989 to a peak of 10.6 per cent three years later.

And here is what did not happen. On Nigel Stapledon’s long-run series, Sydney house prices fell about 9 per cent in real terms from their 1989 peak to the 1991 trough. Across all capitals, 8 per cent. The worst recession since the war, mortgage rates near 17 per cent, unemployment in double digits — and residential values gave up less than a tenth.

The damage was somewhere else. Sydney office property fell roughly 40 per cent over the same cycle. The banking crisis of the early nineties was a commercial property event that happened to coincide with a housing cycle.

1990 was not a housing crash. It was a period when houses were cheap relative to income and money was impossibly expensive.

Buyers stayed home because they could not service a loan — not because they could not reach the asset.

2000: the decade the composition changed

Rates came down and Sydney ran. From its 1991 trough to the 2004 peak, Sydney rose 91 per cent in real terms, finishing 73 per cent above the 1989 high. Note the date: Sydney peaked in 2004, four years ahead of the rest of the country.

But the number that matters from this period is not the price. It is what you were buying.

Stapledon separates land from structure. In the first half of the 1990s, land made up 37.5 per cent of Sydney’s median house price. By the first half of the 2000s it was 69.8 per cent. Land went from roughly a third of what you were paying for to better than two-thirds, in a decade.

That is the structural break. Construction costs track general inflation reasonably closely; land does not. Stapledon attributes about two-thirds of the rise in Sydney’s median house price from the mid-1970s to the mid-2000s to fringe land prices alone — and land supply is a policy variable, which he traces back to state legislation from the 1940s and the State Planning Authority Act of 1963.

What Sydney mortgage repayments actually cost in each period

Now the calculation that changed my mind.

Take the median unit in each period, assume a 20 per cent deposit and a twenty-five-year principal-and-interest loan at the prevailing variable rate, and express the repayment as a share of gross average full-time earnings.

Look at the first two rows. The monthly repayment in 1990 and 2000 is almost identical in dollar terms — about $1,560 either way. But wages had risen 45 per cent in between, so the same payment fell from two-thirds of income to under half.

And then look at today against 1990. On servicing, 1990 was worse. Substantially worse. Sixty-five per cent of gross income against fifty-one.

I want to be honest about that, because it cuts against the argument I have been making in conversation for a year. If servicing were the whole story, this market would be easier now than it was when I started.

So why can’t anyone buy?

Because the barrier moved.

In 1990 the wall was the repayment, and a repayment is a price — prices move. When rates fell, the buyer who had been sitting out came straight back in, and the market cleared. The recovery mechanism was built in.

A deposit is a stock, not a flow. It does not fall when the Reserve Bank cuts. It rises with the asset.

Every rate cut that lifts values pushes the entry price further away from the person trying to save for it — which is why the two most recent easing cycles improved affordability for people who already owned and worsened it for everyone else.

There is a second mechanism working the same way. Under APRA’s serviceability buffer a new borrower is not assessed at the rate they will pay. They are assessed roughly three percentage points above it. So the market is quietly running a version of 1990 at the assessment desk, even while the actual repayment is comfortably below the 1990 burden.

The rate that decides whether someone walks into my open home is not the one on the front page.

What I think happens from here

What follows is my own assessment rather than anything I can source.

I do not think this resolves the way 1990 resolved, and the numbers above are why. In 1990 we were waiting on a rate cut, and the cut came, and it worked — because the barrier was a flow that monetary policy could move. The barrier now is a stock, and it sits in the land market rather than the credit market. A cut lifts borrowing capacity at the margin, but it lifts values first, which pushes the deposit further out of reach for the very person it was meant to help.

One caution, and it is a real one. Stapledon’s series shows Australian peak-to-trough falls becoming less severe over time — averaging 23 per cent before the war and about 13 per cent after it. Residential prices here are sticky. Anyone reading this as a forecast of a crash is reading it wrong. The historical pattern is years of grinding sideways in real terms, not a collapse.

Which is the uncomfortable part. A crash would at least reset the entry price. Grinding sideways does not — it holds the door shut for longer, and it does so quietly enough that nobody is obliged to call it a crisis.

So when I stand in an empty open home on a Saturday morning, I am not looking at a market waiting for a rate cut. I am looking at one where the arithmetic that once let an ordinary income reach an ordinary home has been withdrawn by degrees, over thirty-five years, and has not been replaced with anything.