Wednesday 19 September 2018

The Australian housing bust: Why this time is different

In recent months, media articles have begun to surface highlighting the fact that Australian property prices have started to fall, and indeed in some suburbs of Sydney, have already fallen about 5-10% from their peaks. The response from the establishment has been predictable - there is no need to worry; prices will moderate somewhat, but a crash is far fetched. We have heard all this doom-mongering before, they argue, and the doubters have always been proven wrong. This time will be the same.

However, there is a strong argument to be made that this time legitimately is different. Not different in the sense commonly argued - that Australia's property fundamentals are unique and so prices will forever rise - but rather that calls for a crash this time could well prove spot on, because as I hope to explain, things really have changed in a material way. Price declines of 20%+ are not only possible, but arguably even quite likely, and bearish scenarios could witness prices declines that vastly exceed that (50-70% in a crisis is possible, although far, far less likely). Indeed, a 30% decline will merely return prices to 2012-13 levels - levels that were already amongst the most expensive in the world.

The next few years is potentially going to pose unprecedented challenges to the Australian financial system and economy, and a crisis is quite possible. Indeed, the outright insolvency of the Australian banking system is not that far fetched (no doubt the Australian government will bail them out). The idea that Australian banks are "unquestionably strong" is ridiculous - they might be ok, or they might not be, but the idea that there are not major life-threatening downside scenarios for the banks is sheer, unadulterated delusion. Remarkably, financial markets are reacting to recent events with an extraorindary degree of complacency, with Australian bank share prices continuing to trade at elevated valuations of more than 2x book, and low-teen PE multiples despite very low credit costs (c25bp), and share prices have barely declined over the past six months.


So why might this time be different?

The most fundamental thing that has changed in the past year or so has been the Royal Commission investigation into banking and financial sector practices which - amongst other things - has revealed a lot of the things previously predicted on this blog: widespread mortgage application fraud, with complicity from both banks and mortgage brokers in overstating borrower income and understating realistic living costs. The extent of the revelations make an absolute mockery of the common claim that Australia's property market is different to the pre-GFC US because lending standards have been more conservative - liar loans are estimated to have represented as much as one third of mortgage origination in recent years. The extent of this mortgage fraud is made more worrying still in light of the fact that about 50% of mortgages are interest-only in nature (i.e. so inflated income has been used to assess the affordability of interest payments only, not interest-plus-principal).

This malfeasance should not have been difficult to predict. 50% of mortgage loans in Australia have been originated by mortgage brokers in recent years, whose incentives centre around loan volumes rather than quality, and who have every incentive to help borrowers misrepresent the extent of their ability to repay. The flawed incentives associated with external mortgage origination was a key contributor to the GFC - something to which Australia seems content to ignore, on account of excessive faith in Australian exceptionalism ('we are different and better'). Australian borrowers - eager to get in on the seeming one-way-bet that was the Australian housing boom, will have actively sought out mortgage brokers capable of navigating around whatever residue of lending standards remained at the banks, while banks have been wilfully blind to these practices. They have been complicit in this deception in order to boost loan volumes, and hence profits and executive bonuses, believing that ever-rising house prices would continue to underwrite the downside.

One of the most inevitable outcomes of these RC revelations will be a serious regulatory clampdown on mortgage lending standards (indeed it has already started). This will force banks to restrict the flow of credit to the housing market to levels borrowers can actually afford, rather than what they previously fraudulently claimed they could afford. The problem is that house prices have already escalated to the point where they have capitalised the marginal borrowers' ability to borrow using a fraudulent rather than realistic loan application. This means that as lending standards are tightened, the difference between what people can actually afford to pay and what they were previously claiming they could afford to pay will have be reflected almost immediately in house prices, and we are about to find out how large that gap was. Given the degree to which Australian houses have far outrun realistic affordability metrics, that could be - at the very least - something in the order 10-20%. This is not hyperbole - this should be a base case expectation.

One of the single worst delusions in the housing market is the idea that real estate prices are determined by the physical demand and supply of housing. The physical demand and supply of residential real estate for underlying use drives the level of rents, not the level of capital values. Cap rates, by contrast, are driven entirely by the availability and cost of credit. Using a tight physical real estate market as a justification for low cap rates is therefore fundamentally an exercise in double counting, because tight supply is already reflected in elevated rents. This sort of fuzzy thinking is always prevalent at the top of bubbles, as people grasp at any rationalisation possible to justify prices.

This is fundamental to understand, because it explains why a regulatorily-mandated tightening in credit must impact house prices proportionately. The level of complacency amongst the Australian establishment reflects, to a large degree, failure to understand the above, and to ascribe high house prices to strong net migration and limited supply, and a strong economy, amongst other spurious justifications. If they were blaming those factors for high rents, I would agree, but they are using them to justify 2.5% cap rates on those elevated rents! The 2.5% cap rates have everything to do with credit, and nothing else (other than peoples' speculative instincts).

Furthermore, this is not even mentioning the fact that mortgage rates have in recent years been struck on the lowest RBA benchmark interest rates in history (1.50%); some of the lowest international wholesale financing costs for Australian banks in history; and with 50% of loans being on interest-only terms - the latter of which is also already being clamped down on. All these factors have combined to create the lowest monthly mortgage payments relative to the size of mortgage loans in history as well, which have also been capitalised into house prices. This also means that any increase in mortgage repayments (whether from higher interest rates or higher principal payments) will also have to be reflected in house prices, and the difference could be significant.

This is a key difference between today and immediately prior to the GFC, where RBA interest rates were closer to 7%, and only reached a trough of just over 4% in the early 2000s. This meant that the RBA had significant room to lower interest rates to new lows to juice the housing boom up into one last blow-off phase in the post GFC years. Meanwhile, in the US, the Fed cut rates from about 6% to 1% in the early 2000s after the dot.com bust, which was (for practical purposes) about as low as they could go. Australia, from a mortgage interest rate perspective, therefore only achieved the final capitalisation of extraordinarily low interest rates into house prices this cycle, not the last cycle, and there is not too much further that rates can be cut. Banks have been using both inflated incomes and the lowest interest rates in history to calculate mortgage affordability - what could go wrong?

Unlike in the US pre-GFC, the risk of a major RBA rate hike cycle from here is negligible, barring political upheaval that results in left wing policies that drive up inflation. I think RBA rates will hit 0.0% before they hit 2.5%. However, internationally, funding costs are rising, and the credit spreads on lending to Australian banks are also likely to rise. This means Australian banks are likely to see rising wholesale funding costs (Australian bank LDRs are 120-130%, another warning sign). This means that it is unlikely mortgage rates can fall materially further, even if RBA rates are further cut.

Furthermore, more pointedly, it is possible RC findings precipitate other more onerous regulatory restrictions on the banks, including a requirement that they stress test floating rate mortgages for a significant rise in interest rates, and include a buffer in their lending standards for that risk as well. If this is required of the banks, the practical effect could be to tighten mortgage standards on par with an effective 100-200bp increase in mortgage rates anyway (if you have to prove your borrowers could afford such an increase to lend to them). If this occurs in combination with more stringent income verification requirements, it will be a complete and utter disaster for house prices.


But wait, there's more

If the above weren't enough, the housing market is also facing four more significant headwinds. First, a significant amount of new supply is hitting the market. Indeed, as I discussed in a blog post last week, on a per capita basis, Australia is actually constructing twice as much new residential real estate (by sqm) than China. Secondly, China has clamped down on capital outflows, and domestic politicians in Australia have also moved to restrict foreign purchases of real estate as well. Together, these policies have cooled the 'China bid'. Thirdly, anti-immigration sentiment is growing in Australia (as it is in many parts of the world), which could slow population growth. And finally, the rising unaffordability of housing to new home buyers is now starting to trigger a political backlash, which has recently given rise to a discussion about abolishing negative gearing tax deductions.

I have long thought that the ultimate trigger for a housing crash would be political - albeit still some time away. The truth is, high house prices are purely a political decision. You could crash the market tomorrow simply by announcing a new annual property tax on homes (or second homes) equal to 2% of the market value of the property. It would literally be that simple. The idea that house prices are some uncontrollable force of nature is nonsense.

In the past, there has been no political impetus for such a move, because the vast majority of households have owned a house. However, as home ownership levels continue to decline on account of unaffordable housing and growth in wealth inequality, it is only a matter of time before that calculus changes. The trigger point, in my view, would be if home ownership rates fell below 50%. However, we may reach that political tipping point sooner than expected (even if it is not preempted by a regulatorily-mandated tightening in lending standards), and in this respect it is highly notable that discussions around abolishing negative gearing tax allowances in Australia have recently gained some momentum. It will be interesting to see what the popular appetite is for this move, but a protest vote against a growing 'rentier class' in Australia is entirely possible, and if negative gearing tax allowances are indeed curtailed, the carry costs of investment property will sharply rise, and also contribute to materially lower prices.

If all of these things happen together (tighter lending standards, higher interest rates, and a disallowance of tax deductions or other policies designed to reduce house prices), it is not only possible, but highly likely that Australian property prices would fall by 30-50%. A 30% decline would only take prices back to 2012-13 levels, and if people believe that is impossible, they are kidding themselves. Day-to-day life has not changed that much since 2013, and the increase in prices between 2013-17 was driven almost entirely by low interest rates and a relaxation of lending standards, and the elevated prices that already existed in 2013 reflected generous tax deductions. If you take these props away, the market is virtually guaranteed to crash.

Regulators are doing the right thing in instituting these clampdowns, but unfortunately, they are doing it far, far too late. These restrictions should have been put in place at least 10 years ago, and now the cake is already baked. Over the past decade, Australian politicians and regulators have seriously dropped the ball, and many Australians are going to pay for it dearly (and learn some very old investment lessons).


All this is before mentioning the psychological turning point

All of the above is before mentioning the most powerful force of all that could eventually turn the boom to bust - a shift in investor psychology from greed and FOMO (fear of missing out), to fear. Before we get to fear, denial will persist for quite some time, before progressing through creeping anxiety, outright concern, fear, and then panic, followed by a long period of disinterest. That's how markets work - always have, always will.

A large number of Australians are geared heavily into property, running property portfolios with LVRs of say 80%. With the negative carry on properties running at perhaps 1% of the property value per annum (interest costs on the mortgage, less rents net of property expenses and allowable negative gearing tax deductions), even a 5% fall in capital values will represent an annual 30% loss of equity for the investor (6/20). Leverage cuts both ways, and losses can accumulate extremely fast on the downside. They may shrug off property price weakness at first, believing that prices will inevitably rebound in time. But as prices continue to fall and equity starts to evaporate, disquiet can turn to concern, and then morph into outright fear and panic - particularly if mortgages are moving from interest only to interest-plus-principle, and negative gearing deductions are disallowed, which result in a serious cash flow crunch.

Furthermore, on the demand side, one of the fundamental drivers of a property bubble is the emotional conviction buyers develop that prices can only go up. Because bubbles can go on for a long, long, loooong time, this conviction can become extremely strong over time, and is reinforced every time warnings of a decline prove unfounded. The strength of this emotional conviction is strong enough to overcome rational, abstract, data-driven analysis. This is why everyone ignores all the charts which all demonstrate quite clearly that Australian house prices are off-the-charts expensive on virtually any metric you care to look at. Emotional convictions trump abstract analysis.

The problem is, that emotional conviction will buckle in the face of a protracted period of falling prices. This is what drives long drawn out bear markets in asset prices. Owning a property, with all the maintenance obligations it entails, while taking a 1% cash flow carry loss each year, is fine when price are rising 10% every year and you're building equity, but it's not so much fun when property prices are stagnant to falling year after year after year. People will eventually become disillusioned and throw in the towel. By the end of a long drawn out down market, the inverse situation prevails where all the data clearly highlights that the asset is cheap, but no one is willing to buy it, because by this point people have formed an emotional conviction that prices never go up. 

If and when the psychological tipping point is reached in Australia, there is a long, long, loooong way down, and it will probably take a decade or more for prices to bottom out. A lot of people are going to lose a lot of money, and all of this is before mentioning the likely catastrophic impact of a housing bust on the Australian economy, either.


The inflated numerator

Affordability metrics (as well as mortgage repayability metrics) always use current levels of income and unemployment, but this methodology overlooks the fact that during a housing bubble, the level of economy-wide income and spending is itself inflated by the housing and credit boom. This means the numerator in many affordability ratios is also overstated.

In a housing bust, it is entirely reasonable to expect unemployment to rise to something in the order of 10-15%, and for many people who retain their jobs to suffer much reduced incomes. Spending levels across the board will also decline as people attempt to deleverage. Indeed, as mortgage stress increases, the first thing people will do is cut discretionary spending. When they do, businesses such as retailers and restaurants will feel it, and see a decline in sales and profits, and so they will start to lay people off, who will then spend less and start defaulting on their mortgages. A vicious cycle is triggered. That is how a recession occurs (or if particularly severe, a crisis).

The problem for Australia is that they have not had a recession in 27 years, and so a whole generation of Australians has never experienced one; don't know what it feels like; and haven't planned for one. This is classic Hyman Minsky - stability creates instability. Australians would never have recklessly taken on so much leverage and financial risk if they had lived through a recession in the past 10-15 years, and the length of time since the last recession simply promises that when the next recession does hit, it will be so much worse, as 27 years of cumulative imbalances are redressed in one go, instead of broken down into smaller bites (smaller shorter booms, with smaller busts). A 27 period of calm has lulled people into a false sense of security, and allowed the imbalances to build up to exceedingly dangerous levels.


Conclusion

So is the 'lucky country' about to experience a turn in fortunes? Time will tell. Crises are difficult to predict, and there are graveyards full of people who tried to call an end to the Australian housing bubble. The above must therefore be heavily caveated by that fact - economies are complex and hard to predict. Nevertheless, Australia strikes me as perhaps the riskiest developed economy in the world at present, and its asset prices don't reflect it. Its banks are some of the most expensive in the world, and the stock market generally is also dear. The AUD, although having fallen somewhat, is also still fairly expensive and the country has a 3% of GDP current account deficit, despite the recent somewhat elevated nature of commodity prices.

What can be said with assurance is that if Australia does in fact have a property crash and resultant recession/financial crisis, it ought to surprise no one. Any regulator, central banker, policymaker, or banking executive that tries to claim, after the fact, that such an outcome was impossible to predict, will be some combination of culpably incompetent and downright dishonest. The Australian housing market is an accident waiting to happen, and many, many people have known that and warned about it for a long time, and I find it reprehensible that the country has been so lacking in leadership that this has been allowed to happen, and so soon after the GFC, no less. It's a disgrace (this goes for NZ and Canada as well).

What can also be said with assurance is that - drawing on my recent post about the correct way to think about risk - many asset prices in Australia, including in particular the banks (commodity companies or those with offshore operations are a different animal), currently offer investors extremely poor compensation for the risks they are underwriting - merely 8% earnings yields on top-of-the-cycle earnings in the case of the banks (indeed, earnings which are significantly overstated on account of material under-provisioning for credit risk).

Despite my general aversion to short selling, I am considering shorting them, because it is hard to see a great deal of upside for the banks from here on account of high valuations and limited growth (other than the cost of carry from their high dividends), whereas the downside risks are substantial; clear catalysts are emerging for a reset in the housing market; house prices are already falling; and widespread financial market complacency exists.


LT3000