Estimates that go round the houses highlight need for better data


Conflicting statistics make it hard to get a clear picture on house vacancies – or to develop coherent policy, writes PAT McARDLE

IT’S NOT so long since one of our leading banks had the temerity to suggest that we were the second wealthiest country in the world. This calculation, if it ever had any foundation, was based on the value of the housing stock and was, thus, largely illusory.

You would think, however, that a country with such aspirations would take the trouble to properly count the number of new houses built each year. Not so; we rely on ESB connections for our estimate of new housing completions. This can have occasionally hilarious consequences, such as the time a few years ago when the ESB got their manpower allocations wrong, built up a huge backlog and were unable to supply connections data for half a year. It also means that the annual figures are dubious because some new houses are not connected and others are connected long before they are finished.

Against this background it is perhaps not surprising that there was considerable confusion recently regarding the number of vacant houses, with apparently contradictory statements from three different sources.

The debate started when Dr Rob Kitchin, director of the National Institute for Regional and Spatial Analysis in Maynooth, posted some work on the Ireland after Nama blog. Kitchin, et al, using a variety of methodologies, came up with an estimate of 302,625 vacant houses, “give or take 10 per cent”. Around the same time, the Minister for Housing was quoted as telling the Cabinet that there were 100,000-140,000 houses empty; and the Construction Industry Federation (CIF) has long claimed the overhang is 35,000-40,000 units. Little wonder there was much confusion. In fact, all three are correct; or more precisely, approximately correct. They are all based on estimates and assumptions, but they purport to measure different things.

It is relatively easy to arrive at a 300,000 figure for total vacancies. CSO data, summarised in the table (right), show there were 217,000 unoccupied dwellings in April 2006. This information was contained in the Census volume on housing, published on August 16th, 2007. For some reason it received little attention. We know there were around 200,000 additional ESB connections since April 2006, so it is not too much of a stretch to assume that 40 per cent, or more than 80,000 of them, can be added to the Census number for vacancies.

A total of 300,000 may even be on the low side and there are many more houses in varying stages of completion. Kitchin’s data relies heavily on the Geo Directory, a database of addresses in Ireland, jointly established by An Post and Ordnance Survey Ireland.

Another way of looking at it is that tax on second homes was paid in respect of 283,500 residences last year.

Though this included holiday homes, it excludes unsold new homes and probably underestimates the situation by 10 to 20 per cent. This would suggest that 300,000 is, indeed, on the low side.

Which brings us to the Minister’s 100,000-140,000 range. The source of this was economic consultants DKM’s Annual Review of the Construction Industry, published in September 2009 on behalf of the Department of the Environment. The methodology is set out in some detail in a table on page 32 of the review. DKM deducted an allowance for a “normal level of vacant stock” to get “excess vacant units/overhang” as of June 2009. The range was 122,000-147,000, depending on the assumptions used. The Minister seems to have rounded down.

The key factor here is the allowance for normal vacancies which, at 7 per cent or more, is surprisingly large. However, as the accompanying table shows, this is a longstanding feature of the Irish experience and, moreover, is not out of line with other countries. In the table, I have assumed that half of the 300,000 unoccupied dwellings are “normal”, leaving an excess of 150,000 units overhanging the market, eg a 7.5 per cent excess vacancy ratio.

This excess in turn can be split into new and second-hand homes. The 50,000 figure for excess new units compares with the CIF’s 35,000-40,000, based on a survey it did almost a year ago. The CIF figures are low because they refer to new houses only.

One way or another, it is obvious that the total excess supply is enough to cater for demand over the next three to four years at least. Moreover, the excess is unevenly distributed. In Dublin, it is relatively small and could be absorbed in a year once sentiment turns, but parts of the State could be facing a wait of up to 10 times that.

One thing that is clear is the need for better statistics. The earliest results from the next Census will not be available until mid-2011. There is a strong case for an annual housing survey which would track not alone the stock of unoccupied dwellings but also provide detailed information on their precise status and distribution. How else can we even begin to fashion a housing policy?