The ABS house price index data for December 2012 was released last week and confirmed what the private indices had found: that house prices rose by 2.1 per cent in nominal terms over the calendar year, with most of the rise (1.6 per cent) coming in the last quarter.
This has led several commentators – including my fellow Spectator Stephen Koukoulas – to predict that rising house prices are baaaaaaack, with Koukoulas musing that "house prices could rise by around 10 per cent this year” (All signs point to a house price hike, January 10; Ticket for a housing trifecta, February 1).
Figure 1: Real House Price Index since 1986
Koukoulas gave several reasons to look to the skies for house prices: "Housing affordability (interest rates, current house prices and wages) plus the consistently low unemployment rate, a tight rental market and a likely positive wealth effect from what has been a strong lift in the stock market. An acceleration in population growth is another factor that is likely to add to underlying demand for housing (All signs point to a house price hike, January 10).
Before I dissect these reasons, I’ll note that I too expected house prices to rise in the last quarter – but on the back of an entirely different beast to Koukoulas' list of causes above: the dynamics of mortgage debt.
My starting position on housing is the inverse of the NRA’s "guns don’t kill people – people kill people” argument against gun control. It is that "people don’t buy houses – people with mortgages buy houses”. A prime determinant of what will happen to housing prices is therefore what’s happening to mortgage debt.
On the face of it, that might appear to argue for falling house prices – because mortgage debt growth is now well below trend, at the lowest levels since records began. But it’s a tricky issue, so let me approach it slowly.
Firstly, here’s the data in simple dollar terms – and even that is enough to imply that something different is happening to the housing market than was the norm until 2008, when the global financial crisis hit.
The change in mortgage debt (shown on the right hand scale of figure 1) roughly paralleled the level of mortgage debt (shown in the left hand scale) until 2008, which also roughly means a constant exponential increase in mortgage debt over time. But then in 2008 the growth of mortgage debt broke below the level, and stayed below – so much so that the increase in mortgage debt last year was the same as in 2001, despite CPI inflation and the increases in population and incomes since then.
Figure 2: Level and rate of change of nominal mortgage debt in Australia
If we take out the impact of CPI inflation, the real value of new mortgages today is the same as in 2000 (see figure 2).
Figure 3: Mortgage debt level & change in 2013 prices
Finally, mortgage-financed spending on houses has plunged compared to expenditure on goods and services since its peak in 2005. Back then, new mortgages represented a stunning 12 per cent of GDP; now they are down to 3.5 per cent, a level it first reached in 1993.
Figure 4: Mortgage debt as a percentage of GDP
So if mortgage debt plays any role in setting house prices, and it’s been falling since 2005, why are prices rising now, and why did they rise after 2005? It’s because it isn’t the change in mortgage debt that drives change in house prices – it’s the acceleration of mortgage debt.
Explaining why this is so is complex – there’s a technical economic argument which I discuss here (it’s a proposition that most conventional economists think is obviously wrong, which of course is a strong argument in its favour). But I can give a colloquial explanation in terms of 'supply and demand' that is less formally correct but less complicated and not too misleading.
Demand for housing is monetary, and overwhelmingly sourced from mortgages: the flow of demand for housing is thus largely the flow of new mortgages. Supply for housing is physical, and the flow of new supply reflects both turnover of ownership of existing properties plus the much smaller flow of new properties. When the flow of new demand roughly equals the flow of new supply, the price level will tend to remain constant. For house prices to rise from that point of balance, the flow of new demand has to increase – it has to accelerate.
Here’s where it gets tricky, because though humans get a great thrill out of acceleration (what else explains the fascination with fast cars in a world with 110 km/hr speed limits?), it’s something we understand poorly. In particular, it’s possible to still move forward while decelerating, or move backward while accelerating.
This is what is happening to mortgage debt right now. Even though its rate of growth is generally continuing to fall (see figure 4), it is falling more slowly – and therefore accelerating. That is putting upward pressure on house prices right now – and has been doing so since early 2012.
How much acceleration would be needed to bring about Koukoulas' 10 per cent increase (working in real terms here rather than nominal – which is not the issue it used to be since inflation is now so low)? From my debt-oriented analysis, that would require the acceleration of mortgage debt hitting 2-3 per cent of GDP this year (see figure 5).
Figure 5: Mortgage Acceleration & House Price Change Relationship
What are the odds of mortgage debt acceleration hitting those levels? Not good. Figure 6 shows the same data as figure 5, but over time. There are only a couple of times that acceleration has been that high, as you can see from the line on the chart labelled 'The Kouk'.
Figure 6: Mortgage Acceleration & House Price Change since 1993
Of course the relationship is only one factor in house pricing – I haven’t even considered the supply side here for example, and a tightening of supply could add to price pressures (equally, none of the property bulls considers the debt side like I do). But I think this recent price rise, which over the year was no better than the rate of inflation, is the sign of a sucker’s rally, not a new boom.
The main reason that I doubt that the market is going to get the debt-based fuel it needs to make price change hit the levels the bulls are salivating over is that mortgage acceleration has a limit. It’s easy to accelerate that Ferrari from a standing start; much harder to do when it’s already travelling at 300 kilometres an hour. Ditto mortgage debt: acceleration was easy back in 1995, 2000 and even 2004 when mortgage debt levels were substantially lower than today. Now, since Australians have done so little deleveraging compared to the Americans, the headroom for a quick sprint in mortgages simply isn’t there.
Figure 7: Mortgage debt to GDP in Australia compared to the US
The Americans, on the other hand, have plenty of headroom for another debt-driven rally to their house prices, given the 18 per cent fall in their mortgage debt to GDP levels since 2009. If you want to see 10 per cent hikes in house prices over a year, look to the US rather than Australia.