Are we seeing an asset price bubble in UK housing?
0Government policies designed to cheapen and improve access to mortgage finance – Funding for Lending and Help to Buy – have led to renewed focus on the UK housing market with much criticism that house prices are already at stretched and at unrealistic levels, and the policies risk fuelling an asset price bubble that would put economic and financial stability at risk.
Certainly house prices are expensive, but that looks to be largely down to fundamentals, and in particular lack of supply, and based on activity levels it’s arguable that the housing market is closer to a “bust” rather than a “boom”. Both transactions and house building are at historical lows and in effect, there hasn’t been a functional housing market since late 2008. That certainly cannot be construed as optimal and policies to improve liquidity in an otherwise illiquid market are not inherently “moronic”. Indeed, there’s plenty to suggest that a pick-up in housing market transactions should support a rise in house building. That would help the supply issue at the margin, and also mean a stronger contribution to GDP growth from construction, which has hitherto been weak.
As for the dynamic of change, UK house prices have been rising since late last year, but they are not doing so at a particularly alarming rate. Thus common measures of house price inflation suggest that short-term momentum seems to be running at an annualised rate of 5-10%. By historical standards those rates are neither unusual nor extreme, but a more significant point is that aggregate UK indices are not representative of the state of the UK housing market at present with the recent rise in house prices almost entirely down to rapidly rising prices in London, where prices are now 15% above their 2008 local high. Prices outside London have been stable since the start of 2010, and are 8% below their 2008 peak. So in effect there are two housing markets in the UK: London, where prices are high and rising rapidly; and the rest of the UK where prices are low and stagnant. That’s certainly not consistent with the concept of a generalised UK housing market “bubble”.
Looking forward and apart from idiosyncratic developments in the London housing market, there’s little to suggest a marked acceleration in house prices is in the offing. In terms of what drives prices, a simple regression suggests that lending growth is by far the most significant explanatory variable behind sustained house price growth. When this is combined with changes in the unemployment rate, which can be considered to be a proxy for confidence and economic conditions, over 60% of the movement in annual house price growth since 1985 can be explained. If we add in changes in policy rates and changes in earnings growth, we can explain another 5% of the variability. But we would be cautious of assigning too much causality here, and it seems likely that earnings growth is a substantial driver behind lending growth, and the correlation between these variables may be confused in a simple regression. Nevertheless, we can tentatively conclude that rapid house price inflation is unlikely without a material acceleration in mortgage lending and history suggests that every 1% year on year (yoy) increase in mortgage lending growth, there is a 1% yoy increase in annual house price inflation. Equally, based on historic data relationships, a fall in the unemployment rate from 7.7% to 7.0% should be expected to lead to a shift up in annual house price inflation by around 4.5% yoy, and an increase in earnings growth of 1% should be expected to stimulate house price growth by around 1%.
Taking all the factors in the round suggests that the potential for house prices to be driven substantially higher in the near term is limited, and until the labour market tightens meaningfully there’s unlikely to be a significant shift in disposable income and whilst recent policy measures should be consistent with a modest pick-up in mortgage lending, given continued pressure on banks to de-lever, we would be surprised if lending growth returned to the rapid rates seen in 2000-07.
Apart from the rate of change of house prices, there is the issue as to whether house prices are just too high relative to fundamentals. Some commentators have therefore pointed to the historically stretched relationship between prices and average earnings, although the ratio has fallen over recent years. But the premise that the dislocation of price/earnings relationships has become structural and relates to supply limitations is supported by the rental yield on housing. Low rental yields could suggest an overvalued asset price, but analysis of rental yield calculated using the value of the UK housing stock and actual and imputed rents shows that after having fallen to relatively low yields in the mid-2000s, these yields are now back above their long-term average. Equally if we look at the Bank of England’s Financial Stability Report numbers on the ratio of residential real estate to rental indices, we can see a marked improvement in rents relative to house prices since the cyclical peak, albeit that the ratio has only reversed to levels seen in 2003 and in absolute terms is still 20% above its 1987-2006 average. Neither of these metrics is perfect, but both suggest that valuations are not overly stretched even if prices are relatively high.
As for transactional activity, for there to be a bubble we’d expect there to be signs of a frenzy yet actual housing transaction volumes have been stable at what are all-time lows, excepting the lows seen in late 2008.
The government will care about housing market transactional activity both because a more active market would support the recovery in economic activity levels and growth and because a liquid housing market is necessary to facilitate labour market mobility, and as a connected point new house building, like housing transactions, has been at record lows. Indeed it looks likely that this year total housing completions will be at their lowest peacetime level since 1923, and relative to the level of population, the numbers are even more extreme. Thus with the exception of a few years on either side of the First World War, house building completions per head of population are their lowest peacetime level since the series began in 1856.
As an aside, back in mid-1932, the UK had experienced a recession of a similar magnitude to that of 2008-09, was engaged in fiscal consolidation that reduced the structural budget deficit by about 4% of GDP, had short-term interest rates that were close to zero, and was in a double-dip recession. But the years from 1933 through 1936 saw a very strong recovery with growth of over 4% in every year. A key factor in the recovery was the housing market, and the number of houses built by the private sector rose from 133,000 in 1931/2 to 293,000 in 1934/5 and 279,000 in 1935/6. This building probably accounted for about a third of the increase in GDP between 1932 and 1934, and the house building room reflected growth in the supply of mortgage finance, with Building Society mortgage debt rising from £316mn with 720,000 borrowers in 1930 to £636mn with 1,392,000 borrowers in 1937 when about 18% of non-agricultural working-class households were buying or owned their own homes. In these years, deposits fell in some cases to 5% and repayment terms were extended from around 20 to 25 or even 30 years, reducing weekly outgoings by 15%. In addition, houses were affordable to an increasing number of potential buyers. 85% of new houses sold for less than £750 (£45,000 in today’s money). Terraced houses in the London area could be bought for £395 in the mid-1930s when average earnings were about £165 per year. Houses were cheap because the supply of land for housing was very elastic which in turn meant that there was no incentive for developers to sit on large land banks. Underpinning the availability of land for house-building was an almost complete absence of land-use planning restrictions which applied to only about 75,000 acres in 1932.
But back to today. We can reasonably expect that there will a relationship between the number of transactions in the housing market and the scale of house building, and the data on housing transactions and private sector house building reveal that quite reasonably private house builders won’t and don’t build houses unless there is a reasonably liquid market. It follows that policies designed to increase liquidity in the housing market, such as reducing the size of deposits required from first-time buyers, can be viewed as sensible counter-cyclical measures at this stage of the housing market cycle, and absent a vibrant public sector-driven house building programme, the Help to Buy scheme can be seen to be a reasonable policy with some chance of success. Furthermore the scheme can be scaled back or reversed if the housing market does show signs of over-heating, by reducing the size of the government guarantees on LTV ratios for new mortgages.
The recent housing market indicators from the RICS survey suggest that activity in the housing market has recently picked up with improvements in both demand and supply. The upshot of that is that a reliable gauge of housing market tightness – the ratio of sales to stocks – has not risen significantly. This tends to correlate well with conventional measures of house price inflation and suggests that as the rise in activity has been driven by a pick-up in both demand and supply, there has not yet been a broad-based rise in house prices. It points to annual house price inflation of just above zero.
James Bevan is chief investment officer of CCLA, specialist fund manager for charities and the public sector. CCLA launched The Public Sector Deposit Fund in 2011 to meet the needs of local authorities and other public sector organisations. You can follow James on twitter @jamesbevan_ccla