evidoria

← Back to browse

Good practice Imported

PEX (Probability of Exit) — Ireland's Decade-Long Statistical Model for Long-Term Unemployment Risk

Ireland · Dublin · See the Ireland profile

Since 2012, Ireland's DSP has used a statistical model, PEX, to flag jobseekers at risk of long-term unemployment; a 2019 evaluation found 20% of claimants had no PEX score at all. The model was recalibrated in 2022 and remains in use today.

PEX (Probability of Exit) — Ireland's Decade-Long Statistical Model for Long-Term Unemployment Risk

Details

Promoter
Department of Social Protection (Intreo), with the Economic and Social Research Institute (ESRI)
Period
Model built on 2006 cohort data; operationalised as PEX from 2012; revised 2014; recalibrated 2022; in active use as of 2024
Keywords
public employment services, labour market activation, welfare caseload targeting, statistical profiling

Description

Ireland's Department of Social Protection has, since 2012, used a statistical profiling tool called PEX (Probability of Exit) inside its Intreo network of employment offices. Built on a probit regression trained on a 2006 cohort of jobseekers tracked for 18 months (O'Connell, McGuinness & Kelly, Economic and Social Review, 2012), PEX estimates each new claimant's probability of remaining unemployed after 12 months, using around eight characteristics — age, education, literacy, recent unemployment history, transport access and others. Claimants with a low PEX score (meaning a higher predicted risk of long-term unemployment) are meant to receive more intensive case-officer engagement.

A 2019 government-commissioned evaluation (ESRI Research Series No. 81, 'An Initial Evaluation of the Effectiveness of Intreo Activation Reforms') found that the model's rollout fell short of its design in practice: 20% of individuals in Intreo offices in 2013 had no PEX score at all, and others had missing answers to the profiling questionnaire that made their scores unusable for actually targeting activation support. This is a documented implementation gap acknowledged in an official evaluation, not a hypothetical risk.

Despite this, the Department has continued to invest in the model rather than abandon it: the World Bank's 2023 Public Employment Services Review of Ireland confirms PEX remains the country's core statistical profiling tool for caseload prioritisation, and the ESRI recalibrated it again in 2022 using updated administrative data, finding the predictive patterns from 2006 and a newer 2018 cohort were broadly similar 'despite a global recession in between.' PEX is one of the longer-running statistical profiling systems in European public employment services, illustrating both the durability of the approach and the persistence of implementation gaps between a model's technical design and its use on the ground.

Read the full analysis: https://www.esri.ie/publications/predicting-the-probability-of-long-term-unemployment-and-recalibrating-irelands

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful