Prompted by a colleague, I’ve been browsing the CSO Census report, The Roof over our Heads.  It is full of information from the Census 2011 on households and housing in Ireland.  I’ll probably blog about some of the other material at some point, but I thought it might be useful to point to some of their data on housing vacancy, a familiar topic on this blog.

In the report, the CSO produce an interesting map of all vacant residential address points in the country classified as vacant houses, vacant apartments and holiday homes.  There is little chance of identifying individual properties from this map as it is a scale of 1: 1 million, but by plotting the individual units as opposed to shading in areas we can get a sense of the scale of the issue (which in numeric terms is: 168,427 vacant houses; 61,629 vacant apartments; 59,395 holiday homes; out of total stock of 1,994,845 residential units).

Map of vacant properties in Ireland

Map of vacant properties in Ireland

There is clearly a patterns to holiday homes, concentrating on the coast, as well as the upper and lower Shannon.  Vacant apartments are mainly confined to large urban areas.  And whilst, there is much media talk at present concerning a shortage of family homes in Dublin, the data reveal there is no shortage of apartments.  In fact, there are 16,321 empty apartments in Dublin City, let alone the other Dublin local authorities.  As for vacant houses, they are everywhere.  The few blank spots are mountains or remote areas.

The CSO report also provide some data on towns with the highest levels of vacancy, both including and excluding holiday homes.  The table below lists the seven towns with the highest levels of vacancy excluding holiday homes.  In the case of Tulsk and Ballaghaderreen, two places I have some familiarity with, there is a strong correlation with the presence of unfinished estates.  However, as we have discussed elsewhere, unfinished estates are just one element of vacancy given that there are only 16,881 vacant properties on such estates, meaning there is a high degree of background vacancy in many locations beyond unfinished estates (see our AIRO VacantIreland interactive mapping tool that let’s you examine vacancy at Small Area level and individual unfinished estates).

most vacant towns

Rob Kitchin


The latest tranche of Census 2011 results for Northern Ireland were released yesterday. They provide information of demography, identity, health, housing, education, labour markets, and travel and migration at a variety of geographic scales: 18 Assembly Areas, 26 Local Government Areas, 582 electoral wards, 890 Super Output Areas, and 4,537 Small Areas. Data is available for download here and accompanying mapping boundaries here (Great work by NISRA and a good example of open data)

The rich diversity of data released, and its detailed geographic resolution, enables the general public, policy makers, government and business to better understand the people and places of Northern Ireland in 2011, and the trajectories of change over time, and provides a fresh evidence base for formulating new policy and business plans. Indeed, fresh evidence was needed as Census 2001 has been used as a core base for policy formulation right up to this new release, despite it being over a decade old. What the data makes clear is that whilst there is some continuity, there has also been much change with respect to Northern Irish society and economy over the past decade. By mapping the data and undertaking time-series analysis it will be possible to understand the processes shaping different facets of everyday life and to model future scenarios for planning purposes.

To get started on all of this we have developed an interactive mapping tool for the Northern Ireland Output Areas (OA) and selected some interesting variables for Day 1: Population, National Identity, Religion, Qualifications and Unemployment. Have a look at the new mapping tool here


Over the coming weeks we’ll add to this tool and will also start on our new INTERREG funded project (with colleagues at ICLRD) that will allow us to develop a very comprehensive All-Island Census Mapping Atlas that will look at change on the island from 2001 to 2011.

Justin Gleeson

Since the launch of our National Census Mapping Viewer we’ve have been doing some additional work on the Small Area (SA) datasets and are now in a position to add the new maps to the viewer. The availability of SA level data is a major step forward for socio-demographic mapping and evidence informed planning in Ireland and provides a completely new insight to the trends and patterns that are in place across the country.

If we take Maynooth as an example we can see that up to this point the best level of data we had was for the Maynooth ED as a whole. The introduction of the new SA geography now means that Maynooth can be broken down into 49 individual pieces of information for each census variable. As you can imagine this allows for a much greater level of analysis and understanding of what’s happening in the town when looking at variables such as unemployment, population cohorts (0-14, 65plus etc), health, disability, housing type etc.

This morning we’ve added SA maps to the ‘Population’, ‘Religion’ and ‘Nationality’ themes. Users now have a choice of viewing each variable at either the ED or SA spatial scale. Rather than keeping the legends the same for each variable (at ED and SA level) we have opted to let the data distribution define the legends by using ‘natural breaks’ for each variable. For example, this means that a yellow colour on the ED map may not be the exact same range as on the SA map. This is just something to be aware of. Some examples of the maps are below:

Population 65 plus in DLR at SA level

Polish population in Cork City at SA level

Religion (No Religion/Not Stated) mapping in Sligo at SA level

Over the next week we are going to add to the other themes and have organised a release schedule as follows:

  • Population, Religion and Nationality: Friday, 31th August
  • Education and Social Class: Monday, 3rd Sept
  • Principal Economic Status, Industry of Employment and Occupation: Wednesday, 5th Sept
  • Housing, Transport and Communication: Friday, 7th Sept
  • Health and Disability: Monday, 10th September

You can access the National Census Mapping Viewer here:

To view all our other census mapping tools click here: AIRO Census home page

AIRO team

This morning saw the release of the second summary publication presenting the results of the 2011 census relating to socio-economic topics. Some quick reflections on employment trends. The headline figures are shocking, but don’t come as a surprise. In April 2011 the labour force stood at 2.23 million, an increase of about 123,000 or 5.8 per cent since 2006. The average annual increase of 1.1 per cent reflects a serious slow down compared to the previous, inter-censal period when growth averaged 4 per cent. Numbers at work fell by about 123,000 over the five-year period while the number of unemployed workers increased by about 245,000, up 136.7 per cent.

 In April 2011, the unemployment rate as defined by the Census was 19 per cent (this is not the same as the official unemployment rate as defined by the ILO, which stood at 14.3 per cent). The 19 per cent figure arguably underestimates the scale of the problem if we take account of the fact that a substantial number of workers have moved out of the labour force for a range of reasons. Apart from the emigration vent, the report shows that the student population grew by 16.9 per cent since 2006. In particular, the male student population has experienced a strong increase with participation rates for 19 to 24 year of males increasing from 27.1 in 2006 to 38.9 per cent in 2011. Against the background of very high male youth unemployment rates (41.1% for 20-24 year old males) we may assume that quite a number of young males are finding refuge in the education sector. Overall, the total number of people outside the labour force was up by about 110,000.

Geographically, there are substantial differences in how the various local authorities fared. Limerick City, Donegal and Waterford City are topping the unemployment league with unemployment rates between 25 and 29 per cent. On the other end, below average unemployment rates are evident in Dublin City (18.5%) and its suburbs Dun Laoghaire-Rathdown (11.2%) and Fingal (16%), the commuter belt counties Kildare ( (17.9%) and Meath (18%), Cork County (14.8%), Limerick County (17.5%) and Galway County (18.1%). The greatest percentage point increases in the inter-censal period were experienced in Offaly, Wexford and Limerick City. Dun Laoghaire-Rathdown, Dublin City and Galway City experienced the smallest increase.

The general pattern is one of relatively low unemployment rates in the commuter belts of the main cities. The main cities are characterised by higher unemployment rates than their commuter zones but Dublin City (18.5%), Galway City (18.6%) and, to a lesser degree, Cork City are performing much better than Waterford City and Limerick City. In addition Dublin City, Galway City and Cork City have experienced amongst the lowest percentage point increases. Unemployment appears to be more a regional issue than a city-hinterland issue. These figures corroborate the findings of a recent analysis by Proinnsias Breathnach and me of the annual Forfás Employment Survey which monitors employment trends in firms which have received assistance from the government’s economic development agencies (see here). We showed how almost 80 per cent of jobs created by new foreign firms in the last decade were located in Dublin, Cork and Galway. The share of these three Gateways of all foreign employment rose markedly, from 49% to 58%, while Waterford and Limerick are losing share.

The Census analysis at Electoral Division level shows that substantial intra-city differences exist. Over half of the country’s unemployment 81 ‘blackspots’ (EDs where unemployment rates exceeded 35%) are found in Limerick City, Cork City, Dublin, City, Waterford City and South Dublin.

At an industrial level the greatest loss in employment occurred in the construction sector (-120,000) followed by manufacturing (-50,000). Most of the services sectors experienced employment growth. Employment in education grew by over 36,000, up 28.4%. Interestingly, employment levels in the financial services sector were up 9 per cent since 2006. I suggest that here the internationally traded segment is making up for the job-losses on the domestic front. Even ignoring the anomaly of the Irish construction industry, these figures reflect an ongoing international trend towards a tertiarisation of the economy. The services sector now accounts for 78 per cent of all employment in Ireland.

Whether this is a good or bad development clearly depends on the type of service jobs and the related earnings. The dynamics of persons at work by intermediate occupational group give us some insight – although the taxonomy is not unproblematic and lumps quite different occupations into the same categories (for a discussion see, Breathnacht, 2007). Most of the better paid informational economy occupations are experiencing employment growth. Managers and executives increased from about 123,000 to about 138,000, up 12.3 per cent. Other high-earning occupations experiencing growth include scientific and technical, health and related, teachers, central and local government, computer software and other professional. On the other end, we note gains in the relatively low earning personal services (up 7.8%) and sales (4.0%) occupations. Importantly, these two categories now account for over a quarter of all people at work. The falls in clerical and office occupations (-15.3%), engineering and allied trades (-27.3%), electrical trades (-30.2) and other manufacturing (-52.4%) may point to “a disappearing middle”. Although more detailed analysis is required the data seem to suggest a continued professionalization as well as polarisation of our society, a trend that started in the 1990s (Breathnach, 2007).

The spatial distribution of the various occupations has important implications for inter-regional spatial polarisation and balanced regional development. The data on socio-economic groups show clearly how the higher-income earning social groupings ‘employers and managers’ and ‘higher professionals’ are disproportionately concentrated in the East Region (See Rob Kitchin’s post today).

Yesterday the CSO released a new tranche of the Census 2011 results. The new tables relate to the population of towns and rural areas and population by area.  We’ve taken the data an visualized it on AIRO as interactive maps and graphs:

Map of population and population of towns (see map below)

Components of Population Change 2011 (Birth,Deaths & Migration)

Urban/Rural Population Split at County Level – Census 2011

Irish Towns categorised by Population, Area and Change (see image below)

With respect to towns, 62 per cent of the total population now live in urban areas.  Nearly all towns grew between 2006-2011, with some growing very strongly (10 by over 50% – Saggart, Courtown Harbour, Newcastle, Carrigtwohill, Ballymahon, Rathnew, Kinsealy-Drinan, Annacotty, Ballyjamesduff, Sixmilebridge).  Eight towns shrunk in size, with Templemore decreasing by 13.1%.


Justin Gleeson and Rob Kitchin

With the preliminary results of Census 2011 we can start to undertake some longitudinal analysis of the headline data.  Clearly a key component of the present crisis has been overdevelopment and the subsequent property crash. At a national level, whilst population increased by over a million people between 1991 and 2011 (1,055,550, 29.9%), and households increased by c.680,000[*1] (61.6%), the number of houses built in the same period was at least 869,949 (76.7% increase) (actually more given that some housing was replacement for stock that has subsequently become obsolete).  What this means is that whilst the population and households rose dramatically, house building managed to run ahead of household demand.  And if average household size had not been consistently falling between 1991 and 2011 from 3.46 to c. 2.7, then the level of oversupply would be significantly higher than it presently is.

One of the areas that has come under scrutiny with respect to the property crash has been the Upper Shannon Renewal Scheme area. This scheme, started in June 1998, provided tax relief or incentive allowances for development and covered Co. Cavan, west of the River Erne, all of counties Leitrim and Longford, north Co. Roscommon, east and south Co. Sligo.  Figures 1 provides data on vacancy including holiday homes, housing stock, % vacant stock for 1991, 1996, 2002, 2006 and 2011, and population and household change 1991-2011 for these five counties and the state.  Figures 2-4 graph these data and Figure 5 is house building per annum 1970-2009.  The data aren’t ideal, in that the vacancy rate includes holiday homes, but they still give a good impression of what has happened[*2].

What the data reveal is that development in these counties, like the state as a whole, ran way ahead of population/household growth and other forms of demand (holiday/second home), especially from 2002 onwards as the Upper Shannon scheme worked to accelerate construction activity.  This led to the strange phenomena of increasing population/households, increasing housing stock and increasing levels of vacancy (to over 20% for all 5 counties; base vacancy is usually expected to be between 3-6% in a normal housing market).  In Leitrim, for example, the county in the state with the highest level of overall vacancy (30.4%), housing stock grew faster (8,155; 81%) than population growth (6,477; 25.6%), and nearly twice that of household growth (4,432; 53.7%) and vacancy consequently grew by 3,733 (205%) from 1,820 to 5,553 units.  Roscommon and Sligo similarly built more houses than the population grew, let alone households.  All five counties show a marked increase in the housing vacancy level.  Even allowing for obsolescence and replacement, and demand for holiday homes, it is clear that housing was being built in excess of demand and in response to the tax incentives (as clearly illustrated by Figure 5).  The result is a significant oversupply of stock and a helping hand in the collapse of the banks (see Figure 6 for vacancy levels per ED).

Figure 1: Vacancy, housing stock, population and household change, Upper Shannon Renewal counties 1991-2011


Growth in housing stock, Upper Shannon Renewal Scheme, 1991-2011

Figure 3: Growth in vacancy, including holiday homes, Upper Shannon Renewal Counties, 1991-2011

Figure 4: % vacant, Upper Shannon Renewal area, 1991-2011

Figure 5: House completions in the Upper Shannon Renewal area, 1970-2009, ESB connections, Source DECLG

Figure 6: 2011 housing vacancy by ED

Rob Kitchin

*1 we don’t as yet know the total number of permanent households as temporary absent and visitor data has not been released.  What we do have is the total number of housing units (2,004,175) and the total amount of vacant units (294,202).  If we take one away from the other, there were 1,709,973 units that usually have someone resident, which is usually within a few thousand of the reported household figure for Census night.

*2 there is also a slight discrepancy in the stock and vacancy rate data as reported in the CSO data in that 91, 96, 02 all exclude temporary absent properties in the stock figures; 06, 11 include such properties in the stock figures, but not in the vacancy figures – the difference to the stock and vacancy rate including and excluding temp absent properties is very marginal, c. 0.1-0.3%.  In other words, the figures 2-4 would appear all but identical if the slight alterations were undertaken to make the data fully compatible.

To follow on from yesterday’s post concerning the initial Census 2011 results with respect to housing vacancy.  In that post we included a table of data detailing housing unit numbers, vacancy levels and change between 2006-2011 by local authority.  In the Census 2011 preliminary report the CSO has provide two graphs (vacancy at county level, Figure 1 – see below, and increase in housing stock, Figure 3) and a maps of housing vacancy at ED level (Figure 2).  What these figures show is a marked geography to vacancy, with the five Upper Shannon Renewal Scheme counties of Cavan, Leitrim, Longford, Roscommon and Sligo, along with Kerry, Mayo, Donegal, Clare and Wexford, showing significant levels of vacancy (all over 20%). All of these counties had increases their vacancy rates of over 10%, with Donegal, Cavan, Roscommon, Leitrim, Kerry and Clare having increases of over 20%.

Holiday homes are a significant contributory factor to vacancy in for four of these counties.   We don’t have the 2011 figures yet, but in 2006 holiday homes as a percentage of all vacancy was 52% in Wexford,  43.5% in Donegal, 37% in Clare and 36.5% in Kerry (as evidenced by the high rates of vacancy along their coastal fringes – Figure 2).  So roughly between a third and a half of vacancy in these counties is accounted for by holiday homes.  In the other six counties, however, in 2006, holiday homes accounted for less than a third of all vacant houses (Cavan 12.9%, Leitrim 26.7%, Longford 7.4%, Roscommon 15.9%, Sligo 23.1% and Mayo 29.6%).  This pattern is unlikely to have altered much in the last five years.  In other words, other factors are at play in all ten counties.

Seven of the ten counties  increased their housing stock by over 15% between 2006-2011, with the except of Kerry, Mayo and Clare that increased by over 12% (Figure 3).  Only in Cavan (13.9%), Wexford (10.3%) and Longford (13.3%) did population increase exceed 10%.  In other words, house building was exceeding population growth in all these counties.  As Figure 4 shows, many EDs in these counties had decreases in population between 2006 and 2011 and there is a strong match to those EDs with very high levels of vacancy (Figure 2).  These are also the EDs with low population densities (Figure 5).  Issues such as migration and natural fluctuations (ratio of births to deaths) also contribute.  As Figure 6 shows, several of these counties have low rates of natural increase.

Given the decrease in population in some parts of these counties, and the low rates of growth elsewhere in them, the data suggest that there is a significant oversupply of housing stock in them that may take many years to fill given present demographics.  We’ll be able to work out the exact levels of oversupply per county and likely length to fill once the full census results are out next year and we have a bit more information.

There has been some confusion across discussion boards since the Census results were released as to what constitutes a vacant house, and whether these figures include houses that were simply vacant on the night of the census and also houses in unfinished estates.  The CSO state the following: “‘In identifying vacant dwellings, enumerators were instructed to look for signs that the dwelling was not  occupied e.g. no furniture, no cars outside, junk mail accumulating, overgrown garden etc., and to find out from neighbours whether it was vacant or not. It was not sufficient to classify a dwelling as vacant after one or two visits. Similar precautions were also taken before classifying holiday homes.  Dwellings under construction and derelict properties are not included in the count of vacant dwellings. In order to be classified as under construction, the dwelling had to be unfit for habitation because the roof, doors, windows or walls had not yet been built or installed.”

In other words, enumerators visited homes several times and talked to neighbours to see if the house was a primary residence of occupation.  If the house was temporarily vacant on census night it was recorded as such and these figures are not included in the vacancy rate (as per previous censuses).  Houses in unfinished estates were included unless they were at a stage of construction that precluded habitation.  In other words, the 10,000 under-construction houses in the DECLG unfinished estates survey, and one assumes the vast majority of the 10,000 near-finished houses, will not have been included.


Figure 1: Vacant dwellings as a percentage of total housing stock by county, 2011

Figure 2: Percentage of dwellings vacant in each Electoral Division, 2011


Figure 3: Percentage increase in the number of dwellings by county, 2006- 2011


Figure 4: Percentage change in the population of Electoral Divisions, 2006 - 2011

Figure 5: Population density per square kilometer of Electoral Divisions, 2011

Figure 6: Natural increase by County, 2006-2011

Rob Kitchin

As regular readers of this blog will know, some of the contributors to Ireland After NAMA have been involved in debates over levels of housing vacancy in Ireland post the property crash and financial crisis.  We now no longer need to rely on models to predict the level of housing vacancy as the preliminary results of the Census 2011 provide a detailed breakdown for each local authority (see Table below – click on the image to enlarge – or Census 2011, Table 7).

The total stock of houses in the country grew by 234,562 (13.25%) between 2006 and 2011, rising from 1,769,613 units to 2,004,175 units.  The overall level of housing vacancy, including holiday homes, increased by 10.5% from 266,322 units in 2006 to 294,202 units in 2011. The overall level of vacant housing stock dropped slightly from 15% to 14.7%, but is effectively static, with growth in vacancy being matched proportionally by growth in housing units, which shrank markedly from 2008 onwards.

The much quoted figure in the media of 300,000 vacant units then has then been proven roughly right.  Of course, as we have pointed out several times, the real issue is oversupply not overall vacancy.  To calculate oversupply we need to subtract the number of holiday homes and also the base vacancy rate (calculated in Ireland as 6% overall stock; in 2011, 120,250) from overall vacancy.  The number of holiday homes have not yet been released, but it seems likely that oversupply will be 80-100,000 (it was estimated by the DECLG to be c.122-147,000 at the end of 2009) given growth in households between 2006-2011 and the tail off of construction in recent years.

Housing Vacancy in Ireland 2011, Source Census 2011

To look at the county level, there are ten local authorities where overall vacancy is over 20%.  These include Longford (21.8%), Wexford (20.9%), Clare (21.3%), Kerry (26.5%), Leitrim (30.4%), Mayo (24.8%), Roscommon (23%), Sligo (22.2%), Cavan (22.1%), Donegal (28.5%).  Only 10 LAs saw a reduction in the number of vacant dwellings, the largest of which was Fingal that saw a decrease by -425 housing units.  All other LAs saw a growth in vacancy, nine by over 20%: Carlow (32.2%), Clare (21.2%), Kerry (21.8%), North Tipp (24.9%), Leitrim (24.1%), Roscommon (24%), Cavan (24.7%), Donegal (26.4%), Monaghan (23.3%).

Once we have the holiday home figures we’ll be able to calculate a geography of oversupply, but it is clear from these figures that there is a pronounced pattern of vacancy.

Rob Kitchin