Regional Towns: Growth or decline? Can we tell?

Population change is an important issue and one of the key reasons that we conduct a census in Ireland.  However, as well as being one of the most important indicators of change, it is also an emotive issue.  Population growth in most cases is considered a good thing, an indicator of a vibrant economy and society, while population decline is taken to indicate stagnation and under development.  This is particularly the case in relation to well known, well defined areas such as counties or, in the case of this post, of towns.

Important strategic policies such as the National Planning Framework (NPF)  and the (RSES)  are currently being prepared, and these (along with policies such as Realising our Rural Potential- An Action Plan for Rural Development   have long time horizons and rely on population data as an important benchmark of development.  Therefore robust intercensal comparisons are critical.

Population data from Census 2016 for towns was made available with the publication of Profile 2- Population Distribution and Movements on 11 May.  In its initial release on StatBank[1], tables of town populations for 2011 and 2016 were provided.  In the background notes (Appendix 2) to Profile 2 the CSO noted that there had been boundary changes to 80 towns for which populations were given.  However, the towns were not named, listed or highlighted in the original data available on StatBank and data on town population was provided for all towns for 2011 and 2016.

It has now become apparent that the 2011 data that was originally provided related to the old boundaries and so the 2011 population was not directly comparable to that in 2016. This has been amended (Tables E2014 and E2016 amended on 9.06.17) and different tables are now provided in StatBank.  The 2011 data is no longer provided for the towns which have had boundary changes.  This will prevent inaccurate comparisons and also means that they can now be identified by users.

The change means that people downloading the data now will not make a direct comparison with 2011 for those towns, but for many of us who looked at the data just after its release the comparisons had already been made, commented on or published.

The number of boundary changes was very significant. In the table of 200 towns with population of more than 1,500, 71 had boundary changes.  Of the 41 towns in the Western Region with population of more than 1,500, 15 towns had had boundary changes making comparison with 2011 population data invalid.  The most notable of these is Ballina for which original published data showed a decline of 915 people (-8.25%).  This led to discussion and investigation by regional newspapers[2].

All of the five towns[3] in the Western Region with a population of more than 10,000 have been affected by boundary changes (each of these showed falls in population ranging from  -8.25% to -0.33%).  The extent to which the boundary change is responsible for the population change is unclear.  The CSO does note that in many cases the physical area of the town was reduced:

Census towns which previously combined legal towns and their environs have been newly defined using the standard census town criteria (with the 100 metres proximity rule). For some towns the impact of this has been to lose area and population, compared with previous computations.[4]

Among the seven towns[5] in the Western Region with population between 5,000 and 10,000 six had had boundary changes (the exception being Roscommon).  The population change in these towns, compared to the 2011 figure based on the old boundaries, varied from -0.79% (Buncrana) to +9.76% (Loughrea).  It is hard to assess the extent to which the population change between 2016 and 2011 is reflective of boundary changes or other factors.

So we are now in the situation where we know which towns have had boundaries changes (unlike the situation when Profile 2 was initially released), but we don’t know the extent of the boundary changes and how much they influenced the towns’ Census populations.

It would be very useful if the CSO could provide revised 2011 figures for those towns with boundary changes.  This would allow for direct comparison with 2011 and show clearly whether a town’s population grew or declined.  It would also provide clarity about the effect of the boundary change on the town population.

When the Small Area Statistics (SAPS) are published (20 July 2017) there will be greater detail on local population changes and it may be possible to be clear about where towns have grown and declined and the magnitude of the actual population changes (as compared to those population changes resulting from boundary changes).

Conclusion

It is important that where there are significant alterations to boundaries used or where methods change between Censuses they are very clearly highlighted in any tables published, especially when they relate to headline figures such as population change or population density.  This would mean that a user would not be led to assume that, because the data has been published alongside older data by the Central Statistics Office, it is comparable.

This might seem to be an issue only of concern to those who enjoy analysing data.  It is not.  Changes to town populations can have very significant implications for resource allocations both at a Local Authority level, regionally and nationally.

Would the NPF be more likely to focus on the development of a town that appears to be thriving and showing population growth or one which seems to have stagnated or declined?

Similarly those looking to invest in services and infrastructure, either public (e.g. broadband, education or healthcare) or private (e.g. cinemas, leisure), may think twice if a town seems to be in decline.

Indicators other than population change are used in decision making, but population is still one of the most important.  It is therefore essential that we have good reliable data, for which any changes in methods or boundaries have been very clearly highlighted[6]….

 

Helen McHenry

 

[1] StatBank is one of the CSO’s databases and its main data dissemination service.

[2] See Mayo News,  16 May 2017, http://www.mayonews.ie/news/30029-cso-confirm-ballina-s-population-and-increased-not-deceased and also Western People, 29 May 2016, ‘Misleading Census data’

[3] Ennis, Letterkenny, Sligo, Castlebar and Ballina.

[4] In addition 26 new census towns were created for the 2016 Census.

[5] Shannon, Tuam, Buncrana, Ballinasloe, Westport, Roscommon and Loughrea.

[6] …and not just in the small print or footnote which may fall off the bottom of a page…..