House prices: crash postponed or crash avoided?
The Nationwide House Price Index for the UK ‘average’ property reached £165,000 in June, was up another 0.8% in July and Nationwide have just announced the same increase for August. In real terms, adjusted by general inflation, the new level is 4% above the peak immediately preceding the dip in 2004/5 but it is too early to call either the bulls or the bears back then right. London dipped earlier and rallied more, tied intimately to the stockmarket and bonuses.
Compared with its trend growth since 1957, the real house-price level over the last few years is still at a level as extreme as in 1989 and 1973. Both of those peaks were followed by price falls – in real terms – of about a third.
Real house prices cycle around a long-term trend which is close to real incomes growth – a bit over 2% pa. Let’s assume the bulls are right and this time it’s different: real prices simply level out until the trend catches up. With current values nearly 40% above trend, anyone holding, buying for the first time, trading up or running buy-to-let, on borrowed money, and who does not buy the crash story still needs to assume that their after-tax real interest rates will be below about 1% over the next 20 years if they are to avoid slow real wealth destruction. With capital (and liquidity) getting scarcer, avoiding long-term wealth destruction looks as unlikely as avoiding a near-term bear market.
If you are an owner occupier in mid-life, would never cash in for purely financial reasons and have no plans to trade up, you can ignore these cycles for now. Every one else: you might just need to ‘get real’.
The trend is not that good a friend
Homes and the land they stand on represent far and away the largest capital asset owned by households, many times greater than direct and pension-fund holdings of financial assets. It should surprise nobody that the long-term historical growth in house prices, once general inflation has been taken into account, is only a bit over 2% pa. This corresponds quite closely to the historical growth of real personal incomes, which ultimately must constrain the capital values of household asset holdings.
‘They don’t make land’ any more but they do make buildings. Low on technology and high on labour, construction prices are not likely to fall relative to other prices for long. When they do, affordability will tend to drive up the price of the land they sit on, to compensate.
When house prices do rise much faster than real incomes, particularly for quite a long period, it makes no sense to pretend that the sustainable growth rate has somehow shifted permanently higher. Unless some amazing productivity miracle has occurred unnoticed in these islands, and both incomes and house prices are together launched on a higher trajectory, much stronger house price inflation will weaken the investment underpinnings of the entire economy and so reverse itself.
For any asset class explained by some ‘equilibrium model’ long histories of real (or relative) price tell us most of what we need to know. Anyone who wants to know exactly what the future will hold, or exactly when, will be disappointed most of the time. But knowing when an asset class is exceptionally undervalued or overvalued, even if such conditions arise only occasionally in a lifetime, can make a dramatic difference to wealth outcomes.
At No Monkey Business Limited, we look at property the same way we look at equity markets. For equities, we think we understand why the sustainable trend of real total return (income plus capital growth) is so similar in different markets, at around 6% pa, regardless of the state of development of an economy. We also think we understand the common typical risk aversion, both between countries and over time, that leads the cyclical deviations from the trend, though very large, to be bounded. Like property prices, the cyclical deviations from trend, whether faster or slower, can be several decades in duration. As for property, it is then all too easy to confuse the cycle for the trend.
We are less advanced in modeling real house prices. We have less need, as most of our clients are not dynamically managing their property exposures for return reasons. We also have less data to work with. The Nationwide index of a representative or average UK house starts in 1957 and is the only long history. Because it adjusts for the rising quality of the housing stock (so it is not even notionally the same house throughout), it is approximately equivalent to an ‘internal rate of return’, reflecting the continued new investment of capital to improve the stock. But that is only a convenient approximation of the difference between property and equity returns. Moreover, the UK’s depressing history of political interference in rents means there is no meaningful equivalent of the equity market ‘total return’.
The Nationwide index, to June, is shown in Fig 1. To make successive cycles proportional, the scale is logarithmic. The slope of the regression fitted to the 49.5 years of quarterly data is 2.56% – that’s the basis of what we refer to as a ‘sustainable growth rate’. The most recent cyclical upswing has shifted the entire slope upwards, as occurred to a ‘hindsight-free’ trend during earlier upswings. Once the bull phase was followed by a bear phase, the trend shifted back down again. That serves as a reminder that even 50 years of data is not enough to be sure of the ‘true’ growth trend. Our best guess is that the true trend is about 2.2% pa.
Fig 1. - Real house prices and fitted trends
The best way to plot the deviations between actual market levels and the sustainable trend is to plot these as a ratio. Fig 2 does this. The hindsight-free ratio is included because it uses data known only at the time and so is what people could have seen. Over time, the two behave similarly although the data not common to the two (arbitrarily chosen as the first ten years) is still preventing complete convergence of the two trends.
Fig. 2. - Ratio of real house prices to trend
What can we tell about house-price cycles?
For one thing, we can note from Fig 2 that, when the fog of high inflation is rolled back, history looks very different to how it was actually experienced. How it was experienced is the explanation for widely-believed myths about property as an investment.
During the two bear markets that started in 1973 and 1989, general inflation was about the same as the fall in real terms for the duration of the fall. The popular perception was therefore that the worst that can happen in a bear market is property moves sideways. This ignores the high opportunity cost when inflation is eroding the real value of an unchanged nominal price.
Money illusion is a dangerous mindset. It is probably clouding the judgement of people who think we can now avoid a bear market and just need a period of stability to bring values back in touch with reality. In real terms, it needs about 15 years. It could happen but it would be a first.
We can also note that the lowest and highest ratios are less extreme than the cycles in real equity returns. But this may not mean much. The notion that property prices are less volatile than equities is a common trap, reflecting differences in the process for ‘discovering’ market prices. The level of liquidity is also quite different. Academic research that adjusts property prices for the lower frequency of valuations tends to reveal very similar true volatility (and also, therefore, much higher correlations between the two investments).
It is true that that houses are mainly held passively for long periods and so the proportion of the stock traded in any short period, including traded as a direct response to unusual average prices, is always likely to be lower than for equities. But buy-to-let replicates the behaviour of equity investors and that might cause the deviations to increase in future.
Can we tell anything from the profile of the ratio in Fig 2 about the possible causes of unusually strong or weak house price growth? My opinion is that these are liquidity-driven public-market mood swings rather than the result of changes in supply and demand for homes. Since new construction is a tiny fraction of existing stock, the shortage explanation has always looked implausible. For causes, look to the Bank of England and the Treasury.
Do we even need to understand the reasons? Not really. As for equities, awareness of when market conditions are, by historical standards, extreme is alone very useful information and worth making decisions on. Clearly, this is the case today for property.
This is not such a silly question. Owner-occupiers in mid-life who intend to resist any temptation to ‘lock in’ unsustainable gains can ignore these cyclical conditions – unless they are planning to trade up and have discretion over the timing.
They have mortgage considerations, of course, but at this stage in your life, and if you are not trading up, your mortgage is financing anything in your household balance sheet that can be turned into cash and not just the property. What should drive these decisions are the expected return relative to the cost of money (it costs the risk free rate plus the lender’s margin) and the risk aversion of the investor.
Some financial advisers do apply this approach but only to some marginal asset (like a maturing policy, a bonus or an inheritance) rather than to the uncertain range of returns from the combination of all financial assets held and without any quantified level of risk aversion. They cannot therefore express the difference in long-term wealth outcomes from all sources as a result of borrowing.
Those at risk
The owner-occupiers most at risk are those buying for the first time or trading up or who bought near the top with high leverage.
There is no difference in impact for these people as long as the cash flows, and risks to cash flows (particularly through loss of earnings), are the same. The impact comes in two very different forms: volatility of disposable income and foreclosure risk. In the past, bankers were happy not to interfere, whatever the asset cover for the loan, as long as you could service the debt, so cash flows dominate capital values.
Could this be about to alter? Over-extended credit-card borrowings, the hallmark of this cycle, may spill over into foreclosure as profits are maximized when a bank can collect usurious credit card interest, which has effectively been ‘capitalized’, from the net equity in a property. This possibly explains the superficially irrational behaviour of banks who are both mortgage and credit-card lenders, deliberately to draw customers into unsustainable debt (perhaps you caught the recent Panorama programme too).
The long-term cash-flow impact on household living standards of paying too much for property, when borrowing, is not very different whether property prices fall or move sideways for a long period.
Given our estimates of where we are in relation to trend and what the trend itself is, current market conditions imply a mean expected growth rate of about 1.4% over the next two decades, whatever the actual profile. Returns for shorter holding periods, however, are definitely impacted by the profile. A positive return over, say, ten years has quite a low probability. Implicitly, therefore, owners who think they can avoid erosion of real wealth by borrowing to finance a property purchase must be expecting real interest rates for a mortgage to be no more than about 1% pa. It cannot be ruled out but it is a very low-probability outcome.
It is also possible that house buyers are in fact expecting their real equity to decline, when borrowing costs are taken into account, and that they regard this as an acceptable additional cost of owning rather than renting. I doubt it.
For investors still building buy-to-let portfolios, even this rather implausible ‘utility’ argument cannot apply. Their economic reasoning must be different.
Their real wealth outcomes are based on total return, not just capital, and are comparable with total returns from financial assets. Their cost of funds is higher than an owner-occupier because of debt financing but the interest is tax-deductible. However, they also have to make new capital investments in order to sustain a given rent. The net numbers are not very different: they are highly likely to be going backwards in terms of real wealth after financing costs unless rental growth can compensate. In a bear market for capital values in an economy already over-extended due to mass popular illusions about property wealth, strong real rental growth looks a truly heroic assumption.
Buy-to-let is typically a key part of the story for those who believe a crash is unavoidable (see for example www.housepricecrash.co.uk). There are several versions of this that focus at the margin on recent buyers with less loan cover on both income and capital account. But this is not the only dynamic to watch for. In investment markets, animal spirits work across the great mass of rational and experienced holders, not just marginal investors who were unlucky or least informed. There are substantial gains in this sector that the earlier players will want to lock in once they see them being eroded. They may be back, but not before they have added to the bear-market momentum.
How low can we go?
In real terms, a correction from over 30% above trend to about 30% below trend, as occurred with quite high inflation between 1989 and 1996, must again be a possibility.
If you knew only the past history and had no views about particular dynamics that might shape the next downswing, you might assign no more than a 10% chance to this. But the bears are focused on the particular impact of a housing bear market on consumer confidence, on consumer spending, on jobs and on the banking sector. They might therefore assign a greater probability to such a loss. They could reasonably argue that these dynamics are more akin to the equity market conditions of 1974 and that the low could well be even lower than 30% below trend – by which time, by the way, the measured trend will itself drop below 2%pa.
As professional investors, we try to make realistic estimates of the range of probable outcomes for all the assets we advise on or manage, but not over short periods and we avoid timing calls. Having estimates of long-holding period returns, we can respond to improving or deteriorating ‘competitiveness’ between assets by ‘phasing’ in or out. This may work for equities but not for property. If you have only one shot at an important decision, you cannot avoid making a call. I hope these fairly technical observations help you make that call, whether you are a would-be buyer or a seller.