State’s housebuilding numbers still do not add up

We are incapable of accurately counting our housing output with a variance between methods of 6,100 for 2022

For all sorts of reasons from planning for infrastructure to making investment decisions and the calculation of gross domestic product GDP, it is imperative to know what has been built and where. Counting houses – accurately – is a critical part of policy analysis. For Ministers, in a quantity-over-quality target-driven political system, output is also electorally important.

In 2017, along with architect Orla Hegarty, we queried the Department of Housing’s calculation of annual housing output. We suspected a considerable over-estimation of housing output every year.

At the time, the Department of Housing was using ESB connections as a proxy for housing completions. Each new connection was counted as a new house. Except that in many cases it wasn’t, as new ESB electricity connections can be triggered by work to existing buildings, or by formerly vacant units coming back on stream.

For safety reasons, for example, any new or old residential building that had been empty for more than two years needs a new ESB connection. This was then counted as a new house in that year, as well as having been counted as a new house when it was first connected.

READ MORE

In a lengthy letter to this newspaper, then minister for housing Simon Coveney roundly dismissed our suspicions, saying that the department had used the same methodology to count houses since 1970 and that it was reliable.

His successor, Eoghan Murphy, asked the Central Statistics Office (CSO) to take over, and their analysis showed that the Government had overstated housing output by 58 per cent in the period 2011-2017. House completions were about 53,000 and not the 84,000 claimed by the department and Coveney.

The CSO now has a more refined methodology to calculate housing output, still using ESB connections as the main proxy, but filtering out any misleading data.

There are, however, two other systems of collecting data on new house completions.

Since 2013, it is a legal requirement for a Building Energy Rating (Ber) to be recorded on the Sustainable Energy Authority of Ireland (SEAI) register before occupation. The SEAI should have a good idea of what is being built and where.

More usefully, since 2014, the Building Control Management System (BCMS) records certain legally required forms. Before construction, and on finishing, there is a requirement to lodge a commencement notice, to inform authorities that construction will be starting on site and a certificate of compliance on completion when built. This is a notice to say that the building is finished to the required standard and can now legally be occupied.

Builders of one-off homes can opt out of this process, but must lodge an “opt out” notice to the BCMS before construction starts. Most opt-out notices result in a new house the following year. The BCMS should also have a good idea of what is being built and where.

There are therefore two formal sources of data on house building, the SEAI and the BCMS, and one informal source of data, the use of ESB connections.

All had apparently been good on the house-counting front since 2018 until two weeks ago when Construction Information Services (CIS), a private business information company, revealed that their analysis of the BCMS showed 23,751 new houses completed in 2022 and not the 29,851 as reported by the CSO. This is a difference of 6,100, or 20 per cent less than the official statistic.

Further analysis of CIS statistics shows that the vast majority of the difference between the two sets of figures is in large urban areas. Over one-third of the difference is in Dublin city. Half the difference is in Dublin city and Dún Laoghaire-Rathdown. The four local authorities of Dublin account for 57 per cent of the difference, and nearly three-quarters of the total 6,100 difference between the CSO and the CIS figures are accounted for in the largest urban settlements in the country and their hinterlands: Galway, Cork, Limerick and Dublin.

Given the largely urban location of these differences in figures, a miscounting of one-off houses does not explain the huge gap between the two organisations’ statistics. For example, just 61 one-off houses were built in Dublin city in 2022, compared with 212 estate houses and 3,609 apartments (CSO figures). Apartment completions account for the bulk of the difference between the CSO and CIS figures.

There could well be other interpretations and the CSO and CIS can argue the toss among themselves. Indeed, they may both be right and the figures may be pointing to some other activity in the market that we haven’t yet noticed. Perhaps there is a large number of apartments that are connected to the ESB grid, but without legally required completion certificates and therefore vacant.

Interpretations aside, the most important point here is that there should be no significant variation in metrics. There should be one credible method of measurement.

In 2017, it was argued that the BCMS – the State’s official register of construction activity – was not designed to count houses. The reasoning was that as the BCMS only commenced in 2014, any construction begun before that would not be captured by it, especially developments like refurbished “ghost estates” started years before, abandoned, but then slowly being rekindled into life.

That was nine years ago now, and very little that was started before 2014 has still not been completed. The BCMS is therefore now in prime position to act as the official statistical database for housing output. It has another advantage in that only housing with a certificate of compliance on completion can be occupied so there is no danger of counting housing that can be left vacant.

Ireland is a small country so counting houses should be a straightforward process, but again there are question marks where there should not be any. The BCMS is the most valuable tool the State has that offers reliable, credible and legally required information on what is being built, and it should be used as the main measure. Otherwise housing will never add up.

Mel Reynolds is an architect and housing policy analyst. Dr Lorcan Sirr is senior lecturer in housing at the Technological University Dublin