News stories about Australia's overheating real estate market have been missing a key part of the story. Rents in Sydney and Melbourne are far too expensive. On the other hand, they're likely to increase even further. Unless we take action now, the poorest households could bear most of the pain of the property market rebalancing. As if that wasn't reason enough, the state of the rental market will give the RBA a severe headache in the next few years, leading to tighter interest rates and an economy kept below its potential.
By now we're familiar with charts of capital city house prices shooting up. This chart hasn't gotten the same airtime, but I think it's just as interesting:
Typical rents in the centre of Sydney and Melbourne have roughly doubled in the last decade, an average increase of 7% a year. I've focused on the inner-ring suburbs here, but the story is pretty similar across the greater metro regions. You can find more detail in the replication files for this blogpost, along with the underlying code and data for the other graphs.
Wages have been growing at annual rates of roughly 3% over that period, so you'd expect to find that housing has become less and less affordable for renters. One way to measure that is by plotting rent as a percentage of the average wage. Here I've done that for rents in greater Sydney, and the entire Melbourne metro area, using statewide average salaries. (See the replication files for more details.)
A rule of thumb that people often use is to define ‘housing stress’ as having to spend more than 30% of your income on housing. On that basis, it seems that an average wage-earner renting the median Sydney house would have been in housing stress since 2007. Things are a bit easier for a typical wage-earner in Melbourne, but not by much.
These numbers are consistent with a snapshot of the Sydney real estate market at the start of the year that was put together by Anglicare. They looked across the Sydney basin for rental properties that people on pensions or income support could reasonably hope to afford. In most regions, the availability was close to nil.
As well as being hard to find and expensive, rental properties are often poor quality. Eleanor Robertson elaborated on this point recently in a great piece for the Guardian. She's right. The state of rental housing should be a scandal, and we should be spending a lot more time thinking about how to fix it.
The trouble is that rental yields are currently too low. The rental yield is equal to a year's worth of rent, divided by the price of the flat or house. It's the return you can get on money invested in real estate, before any taxes or subsidies. Right now, rental yields are not as low as they were during the previous real estate frenzy, in 2003-04, but they've been getting close to that level.
TODO: upload this chart again
I've only plotted Sydney yields here, because there doesn't seem to be a publicly available house price series for Melbourne, but based on recent comments by RP Data I think it's probably a similar story. The most recent numbers have houses in both cities delivering about 3%, which is also the return you get from a term deposit, except that real estate is much more risky. Let's assume, purely for the sake of argument, that landlords are greedy. Why would they accept such a low return for their money? Partly it's cyclical—with the economy weak, yields are low across the board—but the main reason is that landlords are expecting to be able to sell their assets for a much higher price than they paid.
However, with APRA cracking down on loose bank lending to investors, auction clearance rates dropping, and (in a year or two's time) interest rates beginning to tighten, those expectations are likely to change. In which case, real estate markets will only reach equilibrium if rental yields increase. Increase by how much? It's difficult to pin down a precise value for what the long-run rental yield should be in equilibrium: the economics of housing investment is quite complex, and, in Australia, renters subsidise homeowners in several ways, making it impossible to draw direct comparisons to other asset markets. For more details on the calculations involved, see the recent RBA paper by Tulip and Fox.
Looking at the movement of rental yields since the early 1990s, particularly in the wake of the 2003-04 period, it seems reasonable to expect that they will head back towards 4½ to 5% in a few years’ time. And this is a rare case where percentages of percentages actually make sense, so we're talking about an increase of about 50%.
The rental yield is a ratio, so the maths is pretty simple: there are two ways to make it get bigger. The numerator (rents) could increase, or the denominator (house prices) could fall. Either of those, or both, could happen now. So far, the data is looking very similar to the end of the real estate mania 12 years ago. Back then, house prices fell a little and then stagnated for a while. Meanwhile, rents increased at a rapid pace that was sustained for five years. It's entirely possible that the misalignment in rental yields will be solved by a house price crash, but my money's on a replay of last time.
Interest rate pain
Having chunky rent increases on the horizon is obviously less than great news for anyone who cares about housing affordability, but it's also terrible for the RBA. Rents are a major contributor to the ‘nontradables’ part of inflation, which is what monetary policy really focuses on. In fact, rent increases were the largest contributor to the inflation breakout of 2007-08. In retrospect, the central bank was too slow to raise interest rates during that period, so it wouldn't be surprising if the staff there are determined not to repeat that mistake. If rents start to move in 2016, then interest rate tightening may be brought forward, even if the rest of the economy is still weak.
An obvious difference from the earlier period is that ten years ago the commodities boom was in full effect and wages were growing strongly. So perhaps the adjustment in rental yields will be slower and gentler this time. I hope that's the case, but it's also true that the rental inflation problem happened in eastern cities, which were left behind by the mining boom; at the time, people complained about a ‘two-speed economy’. If that pattern extends to the whole country, then there's very little wiggle room for monetary policy. In the face of an adverse supply shock such as a rent breakout, then an inflation targeting central bank has to raise interest rates, even if it increases unemployment further.
Another way out
There is a more benign solution to the problem of capital city rents. Setting aside the ups and downs caused by speculative activity in Sydney and Melbourne, it seems clear that the fundamental price of housing—to buy or to rent—is too high. The most effective solution to that is to build more houses. I'll elaborate on that in a later post.