Ireland has managed to get itself a reputation for producing statistics of questionable quality, relevance and accuracy in recent years. In the economic sphere, this view has stemmed from the inflated GDP data in 2015 and the subsequent declaration by the Nobel prize-winning US economist Paul Krugman that the data should be described as "leprechaun economics".
Unfortunately, the moniker appears to have stuck. The criticisms are highly unfair and without merit on this issue. The Central Statistics Office (CSO) was simply applying agreed international conventions.
However, an area where criticism is justified is in the compilation of house completion statistics. These are not produced by the CSO but they should be.
Housing completions data are the responsibility of the Department of Housing, Planning, Community and Local Government and are calculated by way of connections to the electricity grid. This has always been the case. The economics community has always had concerns about these data but it appears that these issues have only come to public prominence in recent times due to the scale of the housing crisis.
Unrealistic picture
There has been a growing list of evidence to suggest that the electricity connections are providing an unrealistic picture of the level of new supply coming on the market. Most prominently, this week’s census data revealed that the housing stock increased by only 8,800 units in the five-year period to April 2016. Over the same period, there were completions or, more accurately, 51,329 electricity connections. Some of the difference can be explained by homes becoming obsolete but cannot explain the full difference.
In the 12 months to February, there were 4,407 transactions for new homes according the CSO’s analysis of stamp duty returns. This compares to more than 15,000 electricity connections over the same period. Excluding an estimate for self-build (which are not included in the transactions data), it appears the level of transactions over the past year are about half of total completions. It seems implausible that more than 4,000 units have been completed by builders over the past 12 months and now remain idle, yet to be sold.
Double counting
There are a number of potential reasons for the inflated figures. For example, a home disconnected from the network for more than two years must apply for a new connection, thus double counting can occur. A separate connection may also be needed for an outhouse/shed. Related to these issues, the finishing of ghost estates over recent years should count as a completion, even though the majority of the associated building activity may have been done years earlier.
This is more than a trivial statistical point. For example, the data is used in the compilation of Irish GDP. If the data is inaccurate, it has implications for the level of output in the economy and for the calculation of important issues such as the amount of fiscal space.
More importantly, housing is the biggest policy priority for the Government. Minister for Housing Simon Coveney has set a target for 25,000 new homes per annum by 2021 under a (false) belief that output in 2016 was 15,000 units. The policy prescriptions required for a much lower level of output would be entirely different. In other words, "if you can't measure it, you can't manage it".
There is another way of measuring housing completions. Since 2004, builders have to input into the Building Control Management System (BCMS). This effectively tracks the home through the building process, from commencement (which data is published) to completion (which isn’t).
Credible housing targets
How can there be any credibility around housing targets if there is no confidence around the statistics underlying them? Moreover, it is best practice that there is a separation between a department that has set targets and the responsibility of producing statistics that would corroborate those targets.
The CSO is the appointed national statistics body for Ireland. It fully subscribes to the principles of the European Statistics Code of Practice that stresses the importance of such things as accuracy, sound methodology, quality and independence. It already has fine form in the housing area, now producing an excellent monthly database on transactions and prices.
The starting point for any policy prescription should be accurate and reliable statistics. In Ireland’s most pressing policy area, it appears we don’t have this basic requirement. Addressing this issue is well overdue.
Dermot O’Leary is chief economist with Goodbody
