Synopsis
Recent data covering the US housing sector has shown a mixed picture, with some indicating renewed weakness in new home sales and existing home sales. In this insight, we examine what is driving some of this renewed softness and detail our longer-term outlook for this key sector of the US economy.
Detailed Insight
Recent data covering the US housing sector has shown a mixed picture, with some indicating renewed weakness in new home sales and existing home sales. The chart below shows that new home sales for a single family declined substantially in March 2014, and for Q1 2014 are likely to record a lower figure than for the comparable period a year ago.
The deleveraging undertaken by the US household sector since 2007 has resulted in the financial obligations ratio of households, namely debt repayments & interest as a % of household disposable income, declining to near its lowest levels over the past three decades. This is illustrated in the chart below.
With the US population continuing to grow by roughly 1% per annum and overall employment having grown in excess of 1% over the past two years, the question being asked is why is the US housing sector not showing a more robust recovery? There are number of reasons at present, each of which we address in this market insight.
Firstly, despite continuing population growth, household formation rates remain low, with the total number of households having remained largely unchanged over the past year.
There are various reasons being given for the recent ‘stall’ in new household formation, including the indebtedness of younger adults due to the large amount of outstanding student loan indebtedness in the US. New household formation is an important driver for new home or housing demand. The chart below shows that, following the recovery in new residential home construction since 2011, and coupled with a lack of new household formation, the ratio of total housing stock relative to the number of households has returned towards the upper end of its range over the past two decades. This suggests that there exists adequate housing stock relative to the number of US households at present.
Further analysis, however, suggests that the number of households should be much higher than the current figure of roughly 115mn. By some estimates, based on the natural growth of the population, the number of households in the US should be at least 2mn higher. Indeed, if we consider the total number of households relative to the male population over the age of 20, the ratio stands at the lowest level in more than three decades.
The chart below illustrates that this ratio has in fact trended persistently lower over the past three decades, which also reflects structural changes in the nature of the average US household. Increased numbers of single mothers, and the increased participation of women in the workforce have led to an increasing number of households which are not headed by a male.
The sharp decline in the ratio over the past few years is nonetheless steeper than the natural declining trend observed since 1990. This provides some support for the view that the current number of US households is too low, possibly by as much as 2mn. The natural population growth of around 1% per annum which is the base case for the US would point to the annual household formation of between 1mn and 1.5mn. Given the lack of household formation in recent years, should employment growth average around 1.5% per annum, average annual household formation could easily average between 1.5mn and 2mn for a number of years as the demographic ‘imbalances’ are addressed. Such a dynamic would require a similar number of new homes every year.
What are the key factors constraining housing demand apart from household formation?
We believe that two other factors may also be playing a more pertinent role at this time, namely new credit regulations and a sharp rise in the cost of newly built homes and their prices. The chart below illustrates that the percentage (47%) of new homes costing above $300,000 has risen to the highest levels on record and substantially higher than was the case just three years ago.
The sharp rise in new homes’ prices and the indebtedness of younger adults to student loans means that the affordability level of your average first time buyer is additionally not comparable to that of the average US adult at present.
This lack of affordability is also in part driving more restrictive lending standards, thus creating a negative feedback loop. Tighter lending standards and higher prices for newly built homes have predictably made many new homes unaffordable for younger families.
Tighter lending standards are fuelled in part by the Frank-Dodd regulations that came into effect on 1 January 2014. The effect of these regulations is clearly visible in the weekly mortgage applications data, which reflects that new mortgage applications approved has declined substantially back to the lowest levels since the 2008 financial crisis.
As employment grows in tandem with the population, however, the inevitable demand for housing will have to be satisfied one way or another. As we have shown, US household formation could easily recover to around 1.5mn per annum, taking into account natural population growth. The chart below shows that 1.5mn also represents the median level for new housing starts in the US since at least 1960. As such, we believe that it is almost inevitable that new home construction will eventually return to a level of around 1.5mn per annum, or 50% higher than present levels.
In a capitalist society, economic participants will furthermore find eventual alternative ways of meeting demand for a product or service. Recent evidence has pointed to a marked rise in demand for rental units, suggesting that these affordability constraints are already having an impact on the structure of the economy.
As such, the latent demand for housing is likely to be resolved eventually. Firstly, demand for rental units will increase, with some demand for new housing being diverted to available rental stock. This will have the effect of increasing rental rates and the rental return on housing. This will inevitably act as another stimulant for the construction of new housing. Secondly, the size of new homes being built could be reduced, enabling home builders to offer new homes at more affordable prices. The chart below shows that this dynamic is already well entrenched, with the number of ‘multi-family’ units being built already back to the levels reported in 2007, prior to the recession.
Finally, as the unemployment rate declines and financial institutions become more confident about the economy’s future prospects, we believe lending standards will decline or become less stringent as banks become more comfortable with the new credit regulations.
These three factors combined, coupled with the existing low housing inventory, will eventually lead to two major trends which are likely to remain intact for the rest of the decade. New housing starts are likely to eventually return to their long-term median level of around 1.5mn and move 50% higher. US residential house prices and rental rates will also keep moving upwards. This is a positive for those companies and sectors exposed to the US housing market. Companies that have purchased or are holding large inventories of ‘rentable’ housing stock will be able to generate higher than average yields, while also enjoying capital appreciation on the housing stock they hold. Finally, as US home construction of single and multi-family units returns to pre-recession levels, homebuilding companies and suppliers of construction materials etc will also benefit.
This dynamic will also inevitably lead to a period of elevated housing inflation, however, when coupled with an anticipated acceleration in US wage growth, and this could place further upward pressure on core inflation in the US and, by implication, longer-term US rates over the next few years. We believe the expansion of employment and concomitant wage growth will be sufficient to offset the increase in interest rates for some time. It is only when interest rates approach or exceed 4%, and in the context of a further rise in property prices, that affordability ratios would once again return to a level that would eventually dampen housing demand.
What about the ageing of the US population?
There is some concern that a rise in the dependency ratio over the next 10 years will lessen the natural demand for housing as the number of persons over the age of 60 increases relative to the total population. The charts below show that a rising dependency ratio indeed appears to have some correlation with regard to demand for housing, and therefore real house prices. This is self-explanatory: although the total population grows, older people will vacate their homes and move to retirement villages, suggesting that overall housing stock relative to the active working population would not need to grow as much as the number of new households formed.
Indeed, this dynamic is and has been a key factor in many countries that are now experiencing an ‘ageing’ of their population and the US will not be exempt. We have shown, however, that the current level of households appears too low as a starting point. In contrast to some of the countries reflected below, the US population is still expected to grow in aggregate. The data depicted below illustrates how the dependency ratio in the US is anticipated to remain at a higher level in comparison to most other developed countries facing a secular ‘ageing’ population.
Real house prices are additionally lower on a relative basis when measured relative to median income or rent. This suggests less scope for decline in house prices, importantly when compared to countries with elevated house prices and valuations, such as the UK and Australia. Demographics suggest that these countries will also face a much sharper decline in the dependency ratio over the next 10 to 15 years.
In conclusion, many workers are retiring at a more advanced age due to better healthcare and in some cases as a result of financial constraints, This suggests that household contraction due to mortality and/or shift to retirement clusters may only happen much later than has been the case in the past. This all suggests that, although the long-term median level of new household formation in the US will be lower than it has over the past 50 years, it should still be significantly positive and more than likely between 1mn and 1.5mn. Coupled with ‘pent-up’ household formation, new housing demand should still average around 1.5mn per annum, a level largely similar to the long-term median level since 1960.