Ghost Estate Report

From the Department of the Environment “Mr Michael Finneran, T.D., Minister for Housing and Local Services, and Mr. Ciarán Cuffe, T.D., Minister of State, with Special Responsibility for Planning, Sustainable Transport & Horticulture’, today (21 October, 2010) published a National Housing Development Survey.”

In other words, the Ghost Estate report has been published. For reasons best known to themselves, the DoE has only released a summary (link to a Word document here) so far. Is it so difficult to generate a PDF and stick it on the website?

50 replies on “Ghost Estate Report”

Fascinating and horrifying but not as bad as I had feared in view of the total vacancy figures produced by other sources (census, NIRSA, UCD). a quick canculation reveals that 12.2 per cent of the complete and vacant dwellings are in the five counties which were subject to the Rural Renewal Scheme Section 23 incentives, which is very high considering the local population. There is also a very significant overhang of vacant apartments in Dublin. I am currently completing a study on Section 23 as part of which I compared the 2006 census small area on all vacant dwellings with the electoral divisions designated under the scheme. This revealed a strong relationship between the two. I wonder is the relationship between Section 23 and ghost estates even stronger? Will the DoE make their detailed data available to researchers so we can find out? I notice the footnote to the table mentions that they want to conduct further analysis with academics and researchers.

@Karl – These are interesting numbers that would seem to contradict those put forward by NIRSA. I had a look at a county that I am very familiar with and it would seem to me that the numbers in this report are closer to the true position than the NIRSA numbers.

It is interesting to see only a synopsis. On the one hand it would have been nice to have some more detail and anlaysis but on the other hand do we need several hundred (costly) pages when the data tells its own story.

The summary (your first link) also outlines the terms of reference of the Expert Advisory Group on Unfinished Housing Developments that is to be established. The group will be chaired by Mr. John O’Connor, Chief Executive of the Housing and Sustainable Communities Agency (HSCA) and will include representatives of:

• Department of the Environment, Heritage and Local Government
• Construction Industry Federation / Irish Home Builders Association
• City and County Managers Association
• National Asset Management Agency
• Irish Banking Federation
• Health and Safety Authority
• Irish Council for Social Housing
• Royal Institute of Architects of Ireland
• Irish Planning Institute
• Society of Chartered Surveyors
• Institute of Engineers of Ireland

Given that the second bullet is “identifying the respective roles and responsibilities of key stakeholders in the resolution of unfinished housing developments” it is surprising that the list of bodies that will be invited to nominate persons to the group does not include the Law Society (note also the absence of any independent voice or god forbid an economist!).

Does anyone have a link to the terms of reference?

My understanding is that the survey only includes estates over a certain size with more than a certain number of empty houses built since 2007. If this is the case, I’m not really sure what it demonstrates? Perhaps “if you pick the right parameters, you can get the answer you want?”

The committee is the usual stakeholder, “experts” etc. nonsense I am yawning just thinking of each and everyone of them having to get their say in. The list suspiciously reads like a list of all those responsible for the mess. Are the enlightened ones now going to tell us what we must do?

A committee of 5 people is more than enough sometimes less is more and it certainly is in this case.

@Edgar Morgenroth

These are interesting numbers that would seem to contradict those put forward by NIRSA

JTO again:

Too darned right! They totally contradict those put forward by NIRSA. And they contradict even more those put forward by David McWilliams and Morgan Kelly.

When the NIRSA report came out, I ridiculed their figure of 300,000 for the number of empties. I asserted that, of their figures for the number of empties, the vast majority were old (and by old, I mean very very old) and lacked modern amenities such as bathroom/indoor toilets/central heating when built. I showed this by adding the number of houses with bathrooms/indoor toilets etc that existed in 1971 (according to the 1971 census), adding on the number of new houses built since then, and comparing it with the changes in the number of households since then. I even emailed NIRSA at the time with the results of my analysis. They were kind enough to reply. They didn’t contradicted my assertions, but simply claimed that further information was needed before reaching firm conclusions (with which I agreed). Well, now we have the further information, and it totally contradicts the wild exaggerations made by NIRSA, David McWillaims and Morgan Kelly about the number of ’empties’ in Ireland. The number of habitable-with-modern-amenities ’empties’ is but a small fraction of that claimed. If we want to add to the list of ’empties’ houses similar to my late granny’s house out in the wilds of Donegal, a house which is certainly habitable, was a lovely little house in its day, is neat and tidy and well looked-after, but was built in 1896 with an outdoor toilet and no fixed bath, has stone floors covered with oil cloth, a coal fire with no central heating, had an indoor toilet added in the 1950s, had a fixed bath added in the 1970s, has draughts ripping through it, and is quite probably haunted, then by all means let David McWilliams and Morgan Kelly do so. But, let them also try suggesting to a young couple getting married today that they should live in it, and see how well the suggestion goes down. I’m afraid that such a couple are far more likely to be seduced by the ‘property porn’ (as David and Morgan call it) associated with modern housing development and will prefer to live in a 21st century-built house with all mod cons.

Both David McWilliams and Morgan Kelly have claimed that hardly any houses built in the past few years have been sold because of the overhang of empties. This report shows otherwise. Of 101,445 houses built in the past few years, 78,195 are occupied and 23,250 empty. The percentages are 77.1% and 22.9%. The report also gives figures for the number of houses not yet completed and empty. I can’t actually see the point of those figures myself, unless you’re expecting people to move in before the roof has been added. Remember, these are recently-built houses we are talking about, not those built 10, 15, 20 years ago. The fact that such a high proportion of recently-built houses are now found to be occupied, the opposite of what we have been repeatedly told, makes it a very good day for the construction industry and a very bad day for NIRSA’s, David McWilliams’ and Morgan Kelly’s reputations.

From comments on the propertypin it would seem that this report is has a narrow reference set;

It only has reference to post-April 2007 housing estates where there is vacancy above 10%.

It therefore a) excludes estates where vacancy is below 10%, b) excludes unoccupied/unfinished one-off housing, c) it assumes that estates started prior to April 2007 are both finished and have occupancy levels above 90%.

So to echo hoganmahew it seems as if they got this figure for empties by only counting certain empties.

“All told, there were 244,590 units built between January 2006 and December 2009 (that were connected to the ESB electricity grid) (datasheet 3). This is despite the fact that in April 2006, the Census revealed that 266,322 housing units were unoccupied (216,533 vacant units and 49,789 holiday homes; 15% of stock; datasheet 11).”

NIRSA reckon there are about 300,000 empty dwellings of which about 120,000 are surplus (i.e. above normal vacancy and not obsolete).

The report is available here:
It is quite an easy read, so you should be able for it.

But then, accuracy has never really been your thing, has it @John, not when polemic will suffice.

@hoganmahew, this document states that it ‘includes all housing developments of two or more dwellings that had commenced by the time of the survey, (summer 2010)’.

They don’t have much sense of occasion. They could have held off on the report for 10 days.

Thanks, I saw that. I wasn’t sure what to make of it though, as it seemed overly broad.

The estate I lived in as a child had two or more dwellings in it. It certainly had commenced by the time of the survey. Indeed, it was completed in 1975. Clearly this wasn’t going to be surveyed. So there must have been other criteria. No?

This report is a complete whitewash. It is evident to me that it was driven by Finneran not Cuffe. It is equally evident that Cuffe had neither the backbone nor the foresight to do any different given his antics at the launch.

The report analyses estates with at least one home occupied. As many or most of developers are bankrupt the DoEHLG is on the hook to complete these things net of completion bonds if they exist.

Therefore the DoEHLG did an audit on THEIR potential liabilities and dressed it up as something else.

A one off or a completely empty estate is somebody elses problem. Nobody is pressurising them to call in bonds. These estates were not addressed at all because no DoEHLG contingent liability exists in the case of these units…bar some demolition post dereliction in future perhaps. Detroit comes to Drumsna sort of vibe.

The end result, correlating with the CIF position that ‘only’ 35k odd units overhang the market was the net result. This is crazy.

We have 300,000 empties today and 50,000 holiday homes.

Of the 300,000 the possibility exists that some 100-150k will be used as PPRs or Holiday homes or weekend PPRs or what have you. The other 150k will slide into neglect and deriliction as surely as Moyross and O’Malley Park in Limerick have.

It’s The Empties Stoopid !!!!

In the unlikely event that they only counted estates started after April 2007 the figures would be of little relevance since there are plenty of estates started in 2005 and 2006 that have both empty and uncompleted units.

From Prof Rob Kitchin in July 2010 “I would expect the DEHLG house count to come in well under the 120K rate (probably nearer the CIF rate). It only has reference to post-April 2007 housing estates where there is vacancy above 10%. It therefore a) excludes estates where vacancy is below 10%, b) excludes unoccupied/unfinished one-off housing, c) it assumes that estates started prior to April 2007 are both finished and have occupancy levels above 90%. There are lots of houses in pre-2007 estates that are empty; the same with one-offs. reports about 115K houses for sale in state and 20K for rent. In other words, it’s a partial survey and we’ll need to wait for census 2011 for a fuller picture.”
For the source of the above see 5th comment down at

As far as I could make out yesterday the Minister referred to estates built “in the past few years”. The DoEHLG press office advised me that the full report might be uploaded to the website next week. At that point hopefully you will the identities of the estates and can say “ah, but what about these other estates”.

A NAMA exec in Sept 10 told an audience of bankers in the UK that NAMA had some control over 6,000 flats within the M50 alone. According to the report yesterday there were some 6,816 dwellings (flats/houses) in all of Dublin that were built and vacant.

I personally feel the Minister was less than open yesterday in explaining the scope of the review (“past few years” – what’s that?) and as predicted by many, including I imagine the folk at NIRSA, the media has gone with the “vacancies far less than previously estimated”.

The numbers do not contradict the NIRSA report at all. See this post for an explanation –

The DEHLG still agree that vacancy including holiday homes is above 300K, that vacancy exc. hol homes is above 228K and oversupply is c.120K. The DEHLG report is about unfinished housing estates, not vacancy or oversupply. In other words it has certain parameters – to look at the most problematic part of the housing market and to quantify the issue. In every county the number of unfinished estates exceeds the NIRSA estimate for ghost estates, again not unsurprisingly because ghost estates are a subclass of unfinished estates.

The NIRSA report never disputed the CIF number of unsold brand new homes. These are just one part of the oversupply issue. They do dispute the notion, however that unsold houses are the extent of the issue. They are clearly not. The bottom line is we have 1.96m houses and 1.63m households. The number to focus on is still largely oversupply, not simply unfinished estates.

The assets are an opportunity for those who sold out before the peak. If sold, they will sell for less than the market rate as they are alongside possibly derelict housing. They may also not have tax reliefs etc due to lack of papers even if such reliefs continue.

BARGAINS! The banks, (if they exist at the time!) may discriminate on loans too, further depressing the price. The prospect of further sales at lower prices etc… Not having been lived in, will also affect the price as will the standard of roads and amenities, always the last to be finished!

These will need to be marketed aggressively and soon!

Thursday’s release appears to have been prompted by a leak to RTE.

This would explain why a Word doc. is provided.

Leaks are not always useful and a leak from an IBM plant tracking service of a report to the Irish Times has been copied this morning by other outlets including RTE.

One headline reads: ‘Ireland is top destination for inward investment – IBM’ which is not claimed.

The author of the IBM report is quoted by the IT as saying: “Ireland’s job creation effort last year was very successful”. – – which it wasn’t.

@ JohnTheOptimist

Goodbody economist Dermot O’Leary published an estimate of vacant housing last April. Link to report here.

@hoganmahew – Clearly there are differences in definition. I was not thinking about the 300k total vacancy rate but the oversupply. For the county that I am most familiar with, NIRSA estimated an oversupply or available stock that is far in excess of that implied by the DoE (vacant, unfinished and unstarted). The DoE numbers make sense to me for that county.

Anyone who started building a housing estate after April 2007 must have either had little choice or little idea about what was going on in the economy.

Work had already been halted on some estates in Dub by that point, although it was probably only spotted by those who were looking out for it.

@ Rob , your comment is entirely correct and sums up the problem we have had for many years now.

“The bottom line is we have 1.96m houses and 1.63m households. The number to focus on is still largely oversupply, not simply unfinished estates.”

That was ALSO the bottom line on thepropertypin and it never changed….and will not absent Bulldozing and 2 into 1 wallsmashing excursions in apartment blocks.

In April 2008 I explained this oversupply issue in the simplest possible terms.

( NB Oct 2010 Rob states 1.63m used out of 1.96m extant and I agree, I personally calculated 1.61m used out of 1.96m extant in April 2008)

From that link.

“The key points are.

1. Our average household size is indeed shrinking .

Household sizes at the last 3 censii were :

3.15 in 1996
2.95 in 2002
2.8 in 2006

A decline of about 0.04 a year with rounding up if in any doubt

2. We are building way to fast to absorb this shrink even with immigration factored in .

So thats 1.94m houses out there now (Feb 2008) and habitable.

So we take 1.61m used out of the 1.94m extant and we have 330k empty ( Feb 2008 mind )

Its 1.96m total now seeing as we built over 20k houses since I did those figures in February up to end 2007 … 1.94m then + 20k net adds even after retiring a few .

I assume these net adds are ALL empty so 330k empties in Feb 2008 + 20k empties since = 350k empties .”

You need to stop conflating oversupply and unfinished estates. Oversupply is principally made up of secondhand houses (plus the 44,000 in unfinished estates – those built and those in the pipeline). Oversupply then is principally made up of one-offs and individual houses/apartments in estates, rather than large clumps in unfinished estates. I think it is highly unrealistic to have expected vacancy rates to go down between 2006 (census) and 2009 when 215,000 additional houses were built and the number of households hardly went up. If you can provide some data as to oversupply in the area you are referring to rather than anecdote then I think we can have a proper debate. Anecdote was part of what led us into the problem.

What I thought was interesting though was the CIF issuing a statement that there is probably only 8/9 months housing supply in the market currently. The Department of the Environment, Heritage and Local Government in their own report with DKM Consultants issued just over a year ago state:
“The oversupply of units available in the market – over and above what is considered as the “normnal” level of vacant units – is estimated at 136,000 units on average, which is equivalent to around four years of current housing demand.”

I think the key thing about yesterday’s study is that is relates to housing developments where construction had commenced but not been completed. No account was taken of completed developments with empty housing units.

You should also factor the shrinkage in household size into it Rob. On trend the average Irish Houshold should now be 2.6 persons having shrunk at its normal 0.04 persons per annum times 4.5 years since the last census.

For various reasons I think the long term shrinkage rate has annenuated somewhat and would trend towards 2.7 persons per household today.

Taking the population estimate in April 2010 and rounding it up for a slight 6 monthly population increase in the ensuing 6 months and dividing that by 2.7 we get

4,470,700 rounded to 4.475m persons and that divided by 2.7 = 1.66m households right now.

The number of homes with electricity is 2m so that means 340,000 empties or holiday homes.

The number of near complete units but as yet ex an electricity supply is around 30,000 ( my guess) so that is 370,000 empty and habitable..or close to habitable. Many of the ‘near’ habitables are one offs.

A survey that only ‘managed’ to find 33,000 of them is simply risible in my opinion. A deliberate insult.

If emigration accelerates the number of empties will continue to rise, inexorably. I am astonished that this has not yet happened to be honest.

Emigration will tend to occur from newer and non obsolescent stock rather than older near obsolescent stock which shall the be retired upon the death of elderly people.

Either way I don’t think yesterday’s report gives us the full picture of potential overhang in the Irish market.

@Edgar Morgenroth – I think it would have been really useful to see a more detailed analysis -certainly from our point of view it would have been very helpful to allow us to eliminate estates straightaway that would not be suitable for social housing.

@ Aoife

There is no assumably ‘normal’ level of empties in Ireland and the rate has varied from 7% of the stock to 17% of the stock over the past 30 years alone…rising steadily all the while.

While we can always estimate the number of homes and the number of households from first principles I believe that any discussion of a ‘normal’ empty rate is entirely frivolous.

@Rob – My definition of oversupply comprises unoccupied houses (new or secondhand), which are either on the market or which the owner would like to sell. Those in the pipeline that have not been started are irrelevant here – they won’t be started and so will not be available! Whether part finished units are relevant is more difficult to assess but lets accept that they will be completed.

Maybe the county I am talking about is an outlier but as I see it given my definition, oversupply is almost equivalent to the number of available properties in recently finished or part finished estates (there are not large numbers of unoccupied second hand homes around, there are not many unoccupied new one-offs around, nor are there large numbers of unused holiday homes that might come on the market) – and the number from the DoE report is a fraction of your oversupply for that county. How do you explain the difference?


You are seemingly underestimating the number of one offs that have been completed to first fix at least. These are frequently built as ‘green hedges’ against changes in planning laws in rural areas. They also account for most of the construction activity in the past 3 years.

You may also be understimating the number of these built ‘on the side’ by construction industry personnel and while they may not wish to sell them they may have to sell them if their banks insist they do.

@2Pack – I don’t think they will amount to enough to change the picture. It may well be an outlier – it struck me when I first saw the (NIRSA) numbers and I am still puzzled. Of course there is no way to verify my concerns.

I agree overall with the NIRSA numbers because they are mor eor less the same as the numbers I and others posited and discussed on ThePropertyPin in 2007/2008.

I would think the one offs are significant because:

1. There are at least 30,000 of them, probably up to 50,000 but some are plausibly holiday homes rather than residences.
2. Many are in the Ghost Estate belts ….eg Leitrim and Longford for example…and statistically very significant in those particular areas and are generally built to a higher standard than the Ghost estates nearby.
3. Some are in the commuter counties around the main cities/employment centres and will inevitably influence proper policy making in the post Finneran years.

But virtually none of these one offs are in Dublin you will be glad to hear 🙂

I think you really need to go and play around with the data. It is all supplied in an excel spreadsheet attached to the NIRSA report and on the DEHLG website. Our model is factoring in household growth, holiday home demand (5/100) and obsolescence (6/1000). DEHLG/DKM and UCD have the same oversupply figures, they have just not broken it down by county (and we know the vacancy rate in 2006 and the number of units completed since then). We have no idea of the number of people who may want to sell but are locked out from doing so because of negative equity. We therefore cannot take the number of houses on the market (33,000 new +c.50,000 secondhand) as an indicator of the true potential housing market (and once you factor in one-offs and pipeline you’re not too far from 120K). The 10,000K houses in the pipeline are not those with planning permission that have not been built, they are under-construction (they have foundations/walls, etc but are not sealed). There are also undoubtedly investor buy to flip houses caught in the oversupply figure. I really think you have to do the analysis on the data rather than make assertions on anecdote. It’s like saying we’re going to have a soft landing because I’ve spoken to a few bankers and they tell me its okay. I’ve had a glance around at the landscape and it doesn’t seem so bad to me. Look, where that got us.

@Rob – “We have no idea of the number of people who may want to sell but are locked out from doing so because of negative equity.” unless they want to emigrate they still would have to live somewhere so this only matters for the spatial distribution of supply.

From years of experience it is useful to check the numbers that come out of anlaysis. If the latest DoE numbers for my county on completed vacant and started units (I exclude the ones where no construction has started) are correct then these amount to less than 25% of your oversupply – where are the other 75%? There would have to be an enormous amount of other vacant units around. This is not anecdote it is a simple reality check that should be part of normal academic rigor which you seem unused to given your defensive response.

Incidentally, I would have thought that citing other relevant literature is part of academic rigor in gerography too – John FitzGerald did some work on vacant houses, published in 2005 but strangely not referred to in your paper.

I can’t recall any warnings from you before the bust.

The big issue is the level of oversupply – clearly as the NIRSA paper put it “calculating the level of oversupply in the Irish housing market is not straightforward due to the paucity of data”. In that context it ought to be possible to question the numbers without getting accused (by Rob) of trying to downplay the problem, which I certainly do not want to do.

I don’t doubt the overall conclusion of the NIRSA paper – we have a lot of oversupply and particurly in the lower Shannon area where construction vastly exceeded demand – but I am puzzled by the one result I refer to.

@2Pack – “one offs are significant” – but are they vacant? Looking at commencement notices since April 2006 you find that about half were for one-offs (interestingly the number of commencement notices was subject to a strongly negative trend from January 2004). If half of these are now occupied (in my view a very conservative estimate) and you add the other half to the DoE numbers you get to about 50% of the NIRSA oversupply numbers. Of course if there had been a lot of excess supply in April 2006 one would have to add that, but the vacancy rate was below average (and had been in 2002 as well) and houses were selling fast for some time after April 2006.

What might account for a discreapnacy? Of course the DoE numbers could be an undercount – a possibility especially if, as was referred to above they only cover estates started after April 2007. The commencements show a bulge in 2006/07 – that would support my first comment that the numbers seemed to differ, but contradict my second point that the DoE numbers looked closer to the mark. Perhaps the demand for holidays homes was higher – I doubt it. Perhaps, household formation was faster – doubt that too. Perhaps the rate of obsolescence has increased – this is very likely but would not amount to that many units.

Perhaps I am being defensive, but I would prefer criticism to be based on evidence not anecdote. I have provided my evidence, you have asserted you don’t believe it, but provide nothing to back up your view other than you don’t believe it. Interestingly the DEHLG do not contest our figures – in fact their figures pretty much match ours as detailed above, and UCD came to the same conclusions as ourselves through another study. In the 2006 census there were 266,332 vacant units. Since then 215,000 additional units have been built. The number of households have gone up by about 170,000 (1.46m to 1.63m) according to the CSO. There are 1.96m housing units in the state and 1.63m households. Where do you think the vacant units have gone? In your county or any other? And yes, 75% of vacant units may well be outside of unfinished estates, which are a small part of the overall housing units (6.1%). They are vacant one-offs and secondhand units and investment property in estates.

As for citation – the paper has 88 references! I think that constitutes relating the findings to the literature. I missed one out, does that invalidate the findings? I’m afraid not. Because I missed one out, I don’t know what I’m talking about? Again, a bogus argument.

As for making a prediction before the crash, this was not an area I worked on back then. So what? Because I didn’t say anything before, I’m not allowed to now? If that were the case nearly all the economists writing on this blog would be invalidated from saying anything!

If you can explain away the CSO’s vacancy/housing units/household figures above, I’d be interested to hear it as no-one else has been able to.

@ All

The CSO figures cannot be explained away at all no matter how people try. There was a very strong correlation between empty houses in april 2006 and domestic electricity connections at end 2005 ( once completions for q1 2006 were added in) , the difference was in the low thousands ona sample of 1.8m premises/connections ….around 0.1% or 0.2%

I am very confident in the utility of meters as a proxy for habitable premises. On the Property Pin we decided after much discussion in 2007 and 2008 that there was, after all , a very simple way to prove the issue one way or another. Ghosts basically don’t use lekky 🙂

So we ask the energy regulator CER.IE for hard data on ANNUAL ELECTRICITY CONSUMPTION in the each of the

lowest decile, (average)
second lowest decile, (average)
third lowest decile, (average)

Of All Irish Homes.

There are slightly over 2m Domestic Meters installed in the field.

If, as I suspect, the vacancy rate is near enough 400,000 then there should be a sizeable jump between the usage figures for the second and third lowest deciles but not as great a jump between the first and second deciles.

If the vacancy rate is nearer ( but above 200,000) the serious jump in consumption should be betwee the first and second deciles. Simples.

Even an economist could not get it wrong.

I hereby nominate Edgar to _ASK_ the CER for this important data and when will you report your findings back to us in this discussion thread Edgar perchance ? 🙂 🙂

@Rob – I’ll go over this once again – once you accept the DoE numbers and a 50% occupancy rate among the one-offs built since 2006 then the numbers for oversupply you have are far in excess of that total. Neither of us know for certain what the level of occupancy in one-offs is but I believe in sampling and if the level was as low as it would need to be to make these and the DoE numbers add to yours for my county then I think I would have noticed – this would either imply that the DoE numbers are to low or for this county your estimate is to high or both to a smaller degree.

At the very least it is worth pointing out that your estiamtes are based on assumptions. You seem to treat your numbers as fact – if we knew the facts we would not be arguing. In so far as assumptions can be wrong (and they invariably are at least to some degree) and the fact that you applied some assumptions across all counties even though there is significant heterogeneity across counties (e.g. on holiday homes) there is scope for (significant) mis-estimation (I also gave some possible reasons for this above).

I think the FitzGerald paper is important because it identified the level of vacancy for a number of census years (at the county level 1991 – 2002 and nationally from 1971) and it also compared some key aspects across countries. It also contained a strong recommendation for property taxes on second or vacant dwellings. I mapped some of the data at the time, which showed the degree of spatial heterogeneity very well but these were not part of the publication.

“Because I didn’t say anything before, I’m not allowed to now?” – you seemed to insinuate that I did not either, which is not correct (I can supply you with referecnes and quotes if you like). I am the last person to stop someone from voicing opinions or concerns.

Edgar, are you getting those electricity usage figures for us from the CER or not?? If not then say so please and I will task another minion 🙂

The Fitzgerald paper from 2004 was the original work on the subject of empties in Ireland and is still the single most important work on the matter.

This topic was visited, interestingly, by The good Prof last year on Irisheconomy and some added comments appeared later on from a contributor to this blog named “John” all around this time last year.

I do wish that Prof John Fitzgerald himself would update his 2005 paper or at least produce a pre census ‘monogram’ even if only to shut that 2Pack fella up 🙂

@2Pack – it would indeed be interesting to look at that data – you are absolutely right about them being an excellent measure of whether a property is lived in permanently, a holiday home or vacant (another thing to add to the to-do-list).

What is your take on the latest DoE numbers? All I have done is a simple ‘reality check’ and I come up with something that puzzles me (I was surprised by the NIRSA numbers for my county but not nationally or indeed in most places).

At the end of the day though vacancies are not that interesting in themselves – what really matters is supply or oversupply (and of course demand). The meters don’t tell me anything about whether any of these units are on the market or will ever be on the market. The vacant units are only interesting if they end up on the market and not all will. For whatever reason there have always been significant numbers of vacant properties but clearly now there are lots more than ever before. If you look at the 91, 96 and 02 censuses (see FitzGerald 2005) you find vacancy rates of about 8%, 6% and 7% respectively (excluding holiday homes) – these would seem to be something of a ‘natural rate’. For 06 this was 12%.

I am certainly not putting forward our model as indisputable fact, but as a model underpinned by evidence gleaned from reputable sources – DEHLG and CSO. The 2006 census data and housing completion data I think we can treat as a pretty robust starting points. Naturally as a geographer, I agree that there is spatial heterogenity across the country which is why we want to compute projected county rates to reveal the differences between locations (rather than just have a national figure). This is also why the model uses projected local household growth based on the previous ten years (not including holiday homes), rather than a flat rate applied to all counties, and it uses the census 2006 county vacancy and holiday home rates as a starting point. It also has a very generous obsolescence rate (6/1000) given the distribution in age of housing stock in Ireland and a very generous base vacancy rate (6/100) (it should realistically be between 3-5%). The model therefore does shows very low oversupply in some counties, especially those around Dublin, and much higher rates elsewhere, notably the upper Shannon region, based on localised parameters.

Maybe your county is an outlier, it well could be, but based on our analysis of available data the national oversupply above a 6% base rate is c.120K. The DEHLG agree and UCD think its 50K more than that (using a 5% base rate)! Given this, do you really think that oversupply in your unnamed county, or any other county, is just slightly above the unfinished house data (which as I’ve detailed is just 6.1% of all units in the country)? Either way, given the difference between numbers of households and housing units there is clearly a problem!

I’m not insinuating that did not work on the area, but you were clearly suggesting that it was problematic to say anything since I hadn’t been working in the area (or why even raise the point?).

I doubt we’re going to agree on all of this, but the debate has been interesting and better than the abuse I’ve occasionally had hurled down at the phone from builders, developers and estate agents! Thankfully, the DEHLG’s survey has proved our point. 2800 unfinished estates is a massive problem to be addressed. The extent of oversupply is, however, I still think much greater than these estates (and I take oversupply to be different to overhang, ie unsold units). If you are talking about overhang, then we might be talking past each other to a certain extent.

‘The vacant units are only interesting if they end up on the market and not all will.’

Okay, your next comment popped up whilst I was writing my response. I think you are talking about overhang, I’m talking oversupply. Oversupply means in excess of household demand for occupying units, excluding holiday homes. Overhang is unsold units. I’ve had the same conversation with CIF. My suspicion is that the difference between ours and DKM/DEHLG oversupply rate (c.120K) and the unfinished homes (44K) is this: (a) unsold one-offs, (b) investor property – buy-to-flip, buy-to-let that is vacant [27% of BoI’s mortgage business in 2006 was investor mortgages, and similar rates probably exist for 2005 and early 2007], (c) excess vacant secondhand houses. My view is that just focusing on the overhang means ignoring the fact that a significant proportion of the oversupply will come onto the market once it starts to recover. It simply can’t be forgotten about because its sold. If that happens, the problem will be extended. We need supply and demand tightened right back up.


What Cuffe said he would do and what was actually done are two completely different things.

The DoE numbers are an honest attempt to begin to quantify the liability to the exchequer and to local authorities to complete these things and to take them in charge. They were spun as something else by Cuffe and especially by Finneran.

If nobody lives in them the estates were unenumerated as there is no liability to take them in charge ‘ex populo’ , same with one offs.

My personal theory on why we accumulated vast numbers of empties, particularly post 1996 is simple. One made more money per annum between 1996 and 2006 on an empty house ( capital appreciation on paper ) than one made renting and managing such an asset.

Why do anything when doing nothing is so profitable….except that too many genuises copped onto this utterly brilliant plan nationwide as we know 🙂

But there is very very little disagreement between the Pin analysis ( particularly in 2007 and 2008) on the total number of units involved and that of NIRSA.

On the Pin we also discussed the issue of 2 meters one premises (a granny flat for the free social electricity) and 1 meter 4 premises ….you can have a house in ‘up to’ 4 bedsits with only one ESB meter and also 1 meter 2 premises ( illegally built one offs pre the ‘retention’ in the main) and concluded they largely cancelled each other out in terms of census enumeration. That was in a private forum and is not searchable and as I recall, there was a right ding dong on the matter around October 2008.

I feel that slicing the data into deciles has more uses than that of enumerating the empties, I am sure that Richard Tol would think of one or two 🙂

We were discussing the ESB usage enumeration approach on the PIN in 2007. It is not inconceivable that a PQ could also elicit this data but I see no privacy issues or commercial issues if the data is released tranched into deciles by the CER.

@Rob – this discussion is certainly useful if perhaps a bit to heated at one point (sorry I got a bit narked). Maybe we should take some of this off-line allthough this has proably gone as far at it can right now. I suspect we are not that far apart as I have no reason to doubt your totals (and I do understand the methodology).

My approach to spatial data on Ireland is to check it gainst my own county (or even ED) which I know well. In most cases this quickly reveals any issues. Your numbers surprised me and the DoE numbers looked closer to the mark for my county (no point in going over that again).

Anyway, I agree with your definition of supply. To me it means what is on the market plus what would be on the market if things were a little different (latent supply). Oversupply is the number of units that are in excess of demand. What is a normal level of supply and demand under the current circumstances is an interesting question. If one thought of the medium-term then some of the current uncertainty will presumably be eliminated as will financing constraints. You incorporate one other important aspect of demand into your numbers – household numbers via extrapolation based on past trends – but clearly lots of other factors are also important (especially income). I would also not underestimate the ‘staying power’ of some owners (suppliers) to wait until things get better especially as they might try the rental market. Not all vacant units end up on the market – some will simply stay empty and obsolescence is likely to increase. Incidentally John FitzGerald estimated a depreciation/obsolescence rate of 0.6% (same as yours) for the period 1996 -2002.

@2pack – your personal theory has much to recommend it. Tax incentives come into it as well. John FitzGerald used the ESB connections in his paper in 2005. The topic has come up in discussions since then (as I say I will look into it and as it happens these numbers would come in handy for something completely unrelated). PQs are never as useful as good relationships with the datacollector!

Tax Incentives were important in incentivising oversupply in certain areas but are largely a historical rather than an economic factor nowadays.

In fact the original large scale incentives, the “seaside resort” scheme are going ex incentive at present and are therefore an ‘exciting opportunity’ as an estate agent would say for an academic to examine the market effect of the end of a tax incentive scheme …which shall be the fate of the more common rural renewal and urban renewal areas in future.

I could not personally recommend whether an academic who undertakes such a study should ideally be an economist or a geographer save to say that academic papers tend to look a tad ropey if any sections are prefaced with ‘2Pack on the Pin said that ………..’ meaning that I doubt if anybody will ask me about it anyway 🙂

From that exercise we can guess the ‘stickyness’ of these investors and from that again we can calculate whether these schemes shall be the decrepit slums of the near future or whether they will continue in use.

The last of these myriad Tax Relief schemes was the Town Renewal Scheme which encouraged rampant idiocy in many towns that needed no market supports of any sort when the scheme was launched.

That one ran from 2000 to end 2006 after which any site that was 15% complete could continue construction to mid 2008. Loughrea County Galway was a town that ‘benefitted’ from this latter scheme where Athenry Tuam and Ballinasloe did not …as a control.

These are the towns that were/are designated for the Town Renewal Scheme. Many feature elegaically on

Hacketstown, Tullow, Muinbheag, Tinnahinch/Graiguenamanagh

Cavan, Baileborough, Cootehill, Ballyjamesduff

Scarriff, Kilrush, Ennistymon, Sixmilebridge, Miltown Malbay

Cloyne, Charleville (Rathluirc), Kanturk, Fermoy, Skibbereen, Doneraile, Bantry

Moville, Ramelton, Ballybofey/Stranorlar, Ardara, Ballyshannon

Portumna, Loughrea, Ballygar, Headford, Clifden

Listowel, Killorglin, Castleisland, Caherciveen

Kilcullen, Rathangan, Monasterevin, Castledermot, Kilcock

Callan, Thomastown, Pilltown, Castlecomer, Urlingford

Mountrath, Portarlington, Rathdowney, Mountmellick

Abbeyfeale, Croom, Rathkeale, Castleconnell, Kilmallock

Carlingford, Dunleer, Ardee, Castlebellingham

Ballinrobe, Claremorris, Newport, Belmullet, Foxford

Oldcastle, Kells, Duleek, Trim

Clones, Ballybay, Castleblayney

Clara, Edenderry, Ferbane, Banagher


Rosses Point, Bellaghy/Charlestown

Nenagh, Borrisokane, Templemore, Littleton, Cashel, Cahir, Killenaule, Fethard

Cappoquin, Kilmacthomas, Portlaw, Tallow

Kilbeggan, Moate, Castlepollard

Ferns, Taghmon, Bunclody, Gorey

Dunlavin, Carnew, Tinahely, Rathdrum, Baltinglass

The question now is what do we do with the empty units? Obviously some as suitable for social and affordable housing, whilst many are not, so do we demolish or in keeping with other European countries start to refuse permission for unsustainable one off housing in the countryside, which is serving to increase our dispersed rural settlement pattern and in turn reduce our competitiveness in terms of waste collection, waste to energy, postal collection, transportation and thats not even considering the environmental cost in terms of the impact on our water system.

We do nothing Charlie. That is because there are huge numbers of one offs in the Upper Shannon too and we want to ensure that we take out that oversupply too. We have no made any attempt to enumerate these.

While you are correct in saying that demolition is inevitable the possibility exists that we may demolish one offs in lieu and move their denizens to the Ghost estates where they are, as you said, more serviceable.

We may then decide to cluster the population closer to the work, eg Longford and Cavan towns, and to demolish more enthusiastically in more rural areas.

Finally we will wish to ensure this farce does not happen again and will have to have planning laws in force to ensure it does not.

Sadly the population projections from the DoEHLG that underpin the most recent RPGs that will in turn influence every County Development Plan for the next 10 years are a complete fiction and guarantee a most unsatisfactory outcome.

5m euro to address the issues, amounting to 14k per estate…. didnt developers have to submit a bond when seeking planning permission from the local authorities? Where has this money gone and why isnt that ringfenced to deal with this issue?


Yo! Edgar. Did you ever get that ESB Metering data, even if only grouped by Decile to control against the Imminent Census Releases on The Empties.

Comments are closed.