Ireland’s Central Statistics Office (CSO) reported two different monthly views in its Live Register release for September 2026, published on 2 October. The seasonally adjusted total fell from 173,200 in August to 172,200, a monthly decrease of 1,000 people. The unadjusted total was 167,193. That raw figure was 1,950, or 1.2%, above September 2025’s 165,243. These comparisons do not conflict: the adjusted series removes recurring seasonal patterns and helps assess short-term direction, while the unadjusted series counts registrations in the reference week. Treating one as a substitute for the other would mix different measures. The month-to-month adjusted movement and the year-on-year unadjusted change answer different questions, and each needs its own period and basis stated alongside the number.

The Live Register is not a population survey that counts everyone who is unemployed, and it is not Ireland’s official unemployment rate. Under the CSO definition it includes people under 65 claiming Jobseeker’s Benefit (JB), qualifying applicants for Jobseeker’s Allowance (JA), people on the newer Jobseeker’s Pay Related Benefit (PRB), and certain other social-insurance contribution registrants. Coverage rules exclude some groups, including systematic short-time workers from JB and smallholders or some self-employed people from JA. Some registered people may not be receiving money: claims awaiting a decision, suspended awards, disallowed claims and applications for credited contributions can also be included. The total of 167,193 is therefore a count of people or claims registered under this administrative definition. It does not represent every person without a job, every person seeking work or only people receiving a payment.

Duration data show different directions within the annual increase. In September, 113,676 people had been registered for less than one year. They made up 68.0% of the total, and this group was 3,004 people, or 2.7%, larger than a year earlier. The number registered for one year or more fell to 53,517, down 1,054 or 1.9% year on year. Together the two duration groups total 167,193. Because shorter-duration registrations increased while longer-duration registrations decreased, the overall rise should not be read as uniform deterioration across all durations. The largest single age group was 25 to 34, with 40,845 people, or 24.4% of the total. Men made up 56.0% of those registered, and 72.1% were Irish nationals, according to the release.

Scheme comparisons also require accounting for a change in the administrative system. The number on JA was 3,306, or 3.0%, higher than in September 2025, while JB claims fell by 8,710, or 48.4%. There were 7,828 people on PRB. That scheme began on 31 March 2025 for people who became fully unemployed on or after that date. The CSO explicitly says the introduction of PRB contributed to the year-on-year fall in JB. Changes in the JB and PRB categories are therefore not a simple like-for-like labour-market comparison under an unchanged programme definition: the route and scheme available to some claimants changed. The overall Live Register still describes registrations within its scope, but administrative rules can affect the composition of its components over time.

The CSO compiles the figures from administrative records supplied by the Department of Social Protection; this is not a household survey. The reference month is the Monday-to-Sunday week containing the month’s last Thursday. The CSO distinguishes the Live Register from the Labour Force Survey (LFS), which provides the official unemployment rate. The two sources measure different concepts and populations, so their levels should not be compared as if interchangeable. Seasonal adjustment uses X-12-ARIMA through age and sex components, and earlier adjusted values can be revised when a new month is added. The CSO says monthly revisions are usually small, while an annual model review can bring larger changes. The adjusted figure of 172,200 should accordingly be read as a revisable statistical estimate, not a fixed census count.