Abstract

Many countries do not maintain a Central Population Register (CPR) that can serve as a direct foundation for producing population statistics, and Ireland is among them. A necessary preliminary step in this process is the construction of a Statistical Population Spine (SPS) and annual Statistical Population Frames (SPF) derived from administrative records. The underlying principle of a Statistical Population Frame is that it can function as a substitute for a CPR by providing a count of individuals in the population at a specified reference point or over a defined reference period. An ideal SPF would contain one record for every statistical unit within the target population, with each unit uniquely identifiable through a stable identifier.
Although Ireland lacks a CPR, it does maintain a comprehensive system of unique personal identifiers, the Personal Public Service Number (PPSN), which is stored within a central records system (CRS). This CRS represents the ever-resident population. Ireland’s SPS is compiled from administrative records in several domains including children, education, social welfare, tax, health and property ownership and rental. Signs of life rules are used to determine the usually resident population each year to generate the SPF. CSO enriches these SPFs with address information including Ireland’s Eircode used as the linking variable for buildings/dwellings. The Eircode was first introduced in Ireland in 2015 and is state of the art. Over ten years later, it is used widely among the public, businesses and public sector.
CSO is now considering how these SPFs can be used as sampling frames for social surveys where the individual is surveyed. This will require consideration of how attributes can be added for post-stratification, weighting and calibration etc. Some of these attributes may have to come from the enumerated census and hence imputation strategies for persons not on the last census need to be considered.
An admin-based household spine is also under active consideration. This would allow timely indicators to be developed for pertinent issues such as fuel poverty and housing affordability. Many social survey frames are based on households and use the enumerated census as a basis. However, the census becomes out of date in the inter censal period.
Rules for assigning people to households are complex, as are common definitions for attributes such as household income. An agreed definition of household income derived from administrative sources is challenging, income as defined in the SILC survey may not match the definition of income used by the revenue collecting agency and those interested in developing an indicator around fuel poverty may have another view entirely. Ideally, we will need a set of principles that can be applied to administrative definitions to derive the required statistical concept.
While CSO envisages that the person and building spines will be compiled from administrative sources (with census attributes added) it is possible that the enumerated census is the best starting point for the household spine with administrative data used to update it in the intervening years.