Over the years there has been lots of discussion about the “fundamentals” of the housing market. Our “strong demographics” were often cited as contributing to the buoyant demand for housing. By this was meant that the rapid growth of the numbers in the household-forming age groups – relative to the number of units being vacated by deaths etc – translated into a firm demand for additions to the housing stock. On this site, Colm McCarthy looked at the impact of demographics on the demand for housing yesterday, referring to recent evidence on the resumption of net emigration. Some of the ensuing discussion tended to get bogged down in trying to interpret very short-term indicators. I thought it would be helpful to provide a medium-term perspective on this aspect of the housing market.
The point of departure is that over the past ten or fifteen years Ireland’s population has been growing faster than that of any other OECD country. For example, our population grew by 14 per cent between 2002 and 2008 when the population of EU15 managed only a 4 per cent increase.
Both components of population change – natural increase and net immigration – contributed to Ireland’s exceptional growth. The rate of net immigration was extraordinarily high in 2006 and 2007, when it reached 1.8 per cent of the population and accounted for two-thirds of total population growth. But the CSO figures released yesterday show that there was a resumption of net emigration between April 2008 and April 2009, which could be a harbinger of higher net outflows in the short to medium term. Ireland’s rate of natural increase (the excess of births over deaths) has also been exceptionally high in the recent past. It rose from 0.4% in 1996 to 1.0% in 2008 at a time when the rate in several European countries – including Germany and Italy – turned negative as deaths outnumbered births.
In light of the exceptional rate of net immigration in the middle of the present decade, and the continuing high rate of natural increase, the talk of strong demographic underpinnings of the housing market was understandable, although economists never lost sight of the responsiveness of migration to the performance of the Irish economy and the unusually open nature of our labour market. However, we should also be wary of extrapolating the recent evidence of resumed net emigration.
To get a better fix on the medium-term influence of demographic factors on the housing market over the medium term, I have used the population projections for 2011-2041 published by the CSO in 2008. I confine attention to the years 2011-2021 and to the M0 or zero net migration assumption. (This allows for some net inflows and outflows in individual age groups that sum to zero for the whole population.) This seems a reasonable compromise between rushing to assume that there will be an upsurge in emigration on the basis of this year’s figure and the more bullish net immigration scenarios published by the CSO two years ago.
My approach was simply to apply 2006 household male and female headship rates by ten age groups to the projected population in 2011, 2016, and 2021. This yields projections of the number of households on the assumption of no change in average household size within each demographic group. Thus, no allowance is made for the probable continuing trend towards smaller households, especially at older ages or for any effects of rising unemployment/falling disposable income on rates of household formation. The projection of the number of households is entirely driven by the underlying population projection.
Before discussing the results, it is worth looking at how the projected population for 2011 compares with the estimated 2009 population published by the CSO this week. On the assumption of zero net migration the projected 2011 population aged 20 and over was 3,232.5 thousand, while the estimated 2009 population is 3,243.4 thousand. On the face of it, the discrepancy of 10.7 thousand or 0.3 per cent is trivial but when the numbers in individual age groups are compared bigger discrepancies come to light. In particular, the large fall that was projected for the population aged 20-29 does not seem to be happening on cue. The estimated 2009 numbers in the 20-24 and 25-29 age groups are 12 per cent and 18 per cent ahead of the projected 2011 numbers. On the other hand, the estimated numbers aged 30 and over in 2009 are 3.3 per cent below the 2011 projection. This highlights the uncertainty surrounding Irish population projections. None the less, the detailed projections published in 2008 are the best available for preparing estimates of the likely evolution of the number of households in the country.
The projected number of households in the country in 2011, 2016, and 2021 obtained from applying the methodology described above is shown in Figure 1.
The actual change in the number of households over the four-year period 2002-2006 is also shown. All figures are annualised for comparative purposes.
The projections show the growth in the total number of households halving from 45,000 a year between 2002 and 2006 to a fairly steady 20,000+ between 2006 and 2021. This reflects the replacement of exceptionally high net immigration between 2002-06 with the assumption of zero net migration after 2006 as well as the impact of the fall in the birth rate in the 1980s on the numbers entering the 25-44 age groups after 2006.
Optimists might take comfort from the projection of a continued net increase of 20,000 – 25,000 a year in the number of households after 2006. However, if we dig deeper and look at the projected change in the number of households by age of the head of households (or what is now called the ‘reference person’) the picture becomes less assuring (Figure 2).
What we now see is that the change in the number of households headed by individuals aged under 45 years (what might loosely be identified as potential ‘first time buyers’) collapses from 24,000 a year in 2002-06 to minus 6,000 in 2016-21. On the other hand, the number of households headed by individuals aged 45-64 (‘trader-uppers’?) drops from 16,000 a year to around 11,000 a year after 2006. Finally, the number of households headed by an individual aged 65 or over (‘downsizers’?) trebles from 5,000 to 15,000 a year.
We might wish to concentrate on the projection of households headed by individuals aged 20-29 as being most relevant to the demand for (new) housing. The dramatic reduction in the projected number of households headed by individuals in this age group is a simple reflection of the fact that the underlying projected population falls from 715.6 thousand in 2006 to 521.8 thousand in 2021. This in turn reflects the falling number of births recorded during the 1980s and the first half of the 1990s. If the zero net migration assumption used in these projections were to be replaced by a more pessimistic assumption, the result would of course be an even sharper fall in the projected number of young households. Thus on any reasonable set of assumptions the demand for housing emanating from a key demographic group is likely to weaken considerably over the coming decade.
The accelerating increase in the number of households headed by individuals aged 65 and over will not compensate for the decline in the number of younger households. Some of the older households will presumably be interested in smaller, easier-to-maintain units while freeing up the larger houses they now occupy. This will support a demand for apartments and dwelling suited to the elderly, but its net effect on the demand for housing units will be negative.
Thus our “strong demographic fundamentals” seem to have evaporated quite rapidly. The implications for the “long run economic value” of our housing stock are negative.