Reading the Irish PMIs – and the Sentiment Counterpoint

The indices on the Irish dashboard

Three Purchasing Managers’ Index series cover the Irish private sector, all published by S&P Global on behalf of AIB and based on monthly questionnaires sent to purchasing executives. The Manufacturing PMI surveys around 250 industrial firms; the Services PMI around 450 private-sector services firms; the Construction PMI covers building activity across housing, commercial and civil engineering. Each is a diffusion index where 50 marks no change.

A separate quarterly product, the S&P Global Dublin PMI, is commissioned by the four Dublin Local Authorities and produced by Grant Thornton inside the Dublin Economic Monitor. It uses the same methodology but stratifies its panel to give Dublin and Rest of Ireland tracks – the only published regional decomposition of Irish business activity.

Alongside the PMIs, the Credit Union Consumer Sentiment Index has been published monthly since 1996 (originally KBC/ESRI; now Core Research with commentary from Austin Hughes). It uses University of Michigan methodology, runs on a 0–200 scale where 100 is neutral, and surveys 1,000 adults each month. It is the longest continuous Irish economic indicator alongside the AIB PMIs.

PMIs use a 50 = no change scale; the EU ESI is built around a long-run average (100) where the standard deviation is 10 points of 100. Italics denote comparator indices used in Section 5; primary analysis uses the first five rows.

Eight years of national PMIs

The Manufacturing PMI dataset starts in May 1998, making it almost twenty-eight years old – the longest continuous business-activity index for Ireland. Services and Construction begin in June 2000. The analysis below uses the post-January 2018 window (99 monthly observations, each verified from the published S&P Global / AIB press release). This is the largest window for which we hold every monthly value as a verified data point. Pre-2018 monthly PMI data is held by S&P Global on subscription and was not available to confirm for this article; we do not estimate or reconstruct values we cannot verify.

Eight years is still a substantial window. It covers four cleanly distinct periods – the late expansion of 2018–2019, the Covid shock and rebound of 2020–2021, the inflation cycle of 2022–2023, and the slow recovery of 2024–2026. That is long enough for the volatility numbers to mean something and short enough that the survey panel and seasonal adjustments have stayed broadly consistent.

Source: AIB / S&P Global monthly PMI press releases (Jan 2018 – Apr 2026).

The three indices behave very differently. Services has historically been the most consistently positive: 91% of months above 50 over the verified window, a mean of 54.1, and outside the Covid shock had not registered a contraction since February 2021 – until April 2026 ended that five-year streak. Manufacturing is the steadiest: 71% of months in expansion and the smallest typical month-on-month move of the three. Construction is by far the most volatile: 59% of months in expansion, a standard deviation of 8.1 over the verified window, and the deepest single trough in the dataset (a PMI of 4.5 in April 2020).

April data has now landed and tells a sharper sectoral story than Q1 alone. Manufacturing accelerated again to 54.9, the highest reading since May 2022 – though firms reported that a meaningful portion of new orders was driven by clients bringing forward purchases ahead of expected supply disruption and price rises, rather than underlying demand strength. Services fell to 49.7, its first contraction since February 2021, with Transport, Tourism & Leisure and Financial Services pulling the headline below 50. Construction dropped sharply to 47.1 from 53.2 in March – a contraction driven by housing (44.4), with civil engineering still falling (46.7, the twelfth consecutive month) and only commercial (50.3) just above the no-change line. So two of the three PMIs are now in contraction territory and only Manufacturing is expanding. The shared theme across all three reports is sharply rising input costs, with the war in the Middle East cited repeatedly as a driver of fuel, freight and raw materials inflation.

The regional split – Dublin vs Rest of Ireland

The headline AIB PMIs are national figures. The Dublin Economic Monitor commissions a separate S&P Global panel that splits the same survey methodology into two geographies: Dublin (the four local authorities) and Rest of Ireland (the remaining 26 counties). The result is the only public regional decomposition of Irish private-sector activity. Quarterly cadence and twenty-four observations from Q2 2020.

Source: Dublin Economic Monitor / S&P Global (quarterly releases, 2020–2026)

Two patterns dominate. The first is that both regions move broadly in step. The second is that Dublin’s line is consistently choppier – bigger spikes higher in the rebounds, deeper dips in the downturns. Same direction, more amplitude.

Source: Dublin Economic Monitor / S&P Global. Q1 2026 gap (+5.9 pts) is the widest in the dataset.

Across the 23 quarters from Q3 2020 onwards (excluding the Q2 2020 lockdown outlier), Dublin has come out ahead of the Rest of Ireland in roughly three quarters out of four, with an average gap of about 1.4 PMI points. The gap has narrowed and even reversed during specific episodes – most notably Q1 2021, when Dublin (42.0) trailed the Rest of Ireland (47.0) by five points during the Level 5 construction shutdown that ran from 8 January to 11 April 2021. Beyond that one episode, Dublin has generally been the firmer reading.

Q1 2026 is the standout quarter in the dataset. Dublin posted a headline reading of 55.6 (the strongest since Q2 2022) while the Rest of Ireland slipped to 49.7, marginally into contraction. The +5.9 point gap is the widest Dublin lead on record. Dublin Construction (60.7) and a Dublin Manufacturing rebound (59.9) drove the divergence. Outside the capital, Manufacturing actually contracted (48.0).

The consumer counterpoint – sentiment is telling a different story

The PMI story has turned sharper in April – Manufacturing accelerating, but both Services and Construction now in contraction. Running alongside all three, the Credit Union Consumer Sentiment Index has been heading lower for months and has now fallen for two consecutive months. The April reading of 53.3 is its lowest level since December 2022, down from 56.7 in March and 65.2 in February. That is over 30 points below the long-run survey average of 83.5 – a deep gap on a 0–200 scale, and one that pre-dates the April PMI weakening by months.

Source: Credit Union Consumer Sentiment Index (Core Research / Irish League of Credit Unions). The series began in 1996 (originally KBC Bank / ESRI; transferred to Core Research / ILCU in October 2022). The chart shows the verified Jan 2024 – Apr 2026 monthly window.

The shape tells the recent story. Sentiment held in the low-70s through 2024 – itself well below the long-run average of 83.5, but stable. Then in early 2025 it tipped over: A 16-point drop between February and April 2025, which coincided with the US tariff announcements (the press commentary at the time named those as the main driver, though other concerns were live too). From mid-2025 onward the index drifted in the 58–65 range. March 2026 broke lower with an 8.5-point drop, and April lost another 3.4 points to land at 53.3 – the lowest reading since December 2022. The Core Research commentary cited continuing Middle East volatility and rising retail fuel costs as the main drivers, with the largest single component fall being the outlook for jobs. How much of the move is really those named factors, versus accumulated background concerns about AI’s labour-market impact, interest rates, housing costs or general uncertainty, is harder to say. To put the current reading in cycle context against earlier periods (financial crisis, Covid), see the Core Research / ILCU monthly archives at the citation below.

European comparators – ESI and the OECD leading indicator

Two further indices belong on the dashboard for one specific use case each: Comparing Ireland to its European peers and looking 6–9 months out. Neither replaces the PMI, but each picks up something the PMI cannot.

The European Commission’s Economic Sentiment Indicator (ESI) is a single index built from five separate surveys covering industry, services, retail trade, construction and consumer confidence. Run by DG ECFIN since 1985 using the same methodology across all EU member states, it is normalised so that 100 equals the long-run average and the standard deviation is 10 points. A reading of 95 means the index is half a standard deviation below normal. Critically, the ESI weights consumer confidence into the same number as business confidence, where the PMI does not.

The OECD Composite Leading Indicator (CLI) is built to anticipate turning points in the economy, usually 6–9 months ahead. It combines several adjusted series (new orders, share prices, the gap between short- and long-term interest rates, consumer expectations) into a single index that sits around 100. The CLI is designed to look forward; the PMI describes what is happening this month.

For most Irish business decisions, the PMI is the right primary signal and the ESI/CLI sit beside it as comparators. The exception is companies whose customers are pan-European: For those, the ESI’s harmonisation makes it the more useful index, and the PMI becomes the comparator.

The combined view – all four indices on a shared timeline

Looking at each index in isolation tells you about that sector or that audience. Looking at all of them together, on the same time axis, tells you how the Irish economy actually moves. Four panels share an x-axis: Manufacturing, Services and Construction PMIs across the verified Jan 2018 – Apr 2026 window, with the Credit Union Consumer Sentiment Index on the bottom panel for the period it’s verified in this dataset (Jan 2024 – Apr 2026).

Sources: AIB / S&P Global monthly PMI press releases (Jan 2018 – Apr 2026); Credit Union / Core Research CUCSI (Jan 2024 – Apr 2026; the bottom panel is left blank prior to Jan 2024 rather than reconstructed).

The chart settings reward close inspection. Look at how Construction was the most volatile through the verified window, with multiple sharp drops outside the Covid period. Look at the 2022 inflation shock: Construction took the deepest hit through Q4 2022, Services held up best, Manufacturing crossed below 50 only briefly. You can see which indices move first and which follow when the panels share a time axis: Notice how the consumer-sentiment drop from February to April 2025 (−16 points) preceded the Services PMI weakening that built through late 2025 and Q1 2026 and finally tipped Services into contraction in April 2026 – the first negative Services reading in five years.

What to track over the rest of 2026

Three watch items emerge from a comprehensive read of the indices, each tied to a specific decision.

Two further structural observations matter for any business with ongoing exposure to the Irish economy.

• The PMI most often quoted in news commentary (the AIB Manufacturing PMI) is the steadiest signal, not the most informative one. For sectoral exposure, decompose into Manufacturing, Services and Construction; for regional exposure, use the Dublin PMI; for consumer exposure, use the CUCSI. The AIB Manufacturing headline is a useful shorthand but not a complete dashboard.
• The divergence between business activity and consumer sentiment is starting to resolve – toward sentiment, not toward PMIs. The two indices measure different things – PMI is about current order books, sentiment is about expectations of the year ahead – so they can move apart for stretches. Within the verified Jan 2018 – Apr 2026 window, the most prolonged divergence ran through 2022–2023 (sentiment fell hard during the Ukraine war and energy shock while PMIs held up), and that one resolved with sentiment partially recovering rather than PMIs collapsing. The 2025–2026 divergence opened ten months ago; the April Services contraction is the first material sign that this one may resolve the other way, with business activity catching down to sentiment. Whether that pattern extends through May and June is the most important question for 2026 planning.

What these indices can and cannot tell us

The PMIs and the CUCSI are validated, well-established indicators – but each has real limits worth flagging:

• PMIs tell you direction, not size. A reading of 55 means more firms reported expansion than contraction; it does not say by how much. A 53 in a deeply concentrated sector (Manufacturing, dominated by pharma and tech) and a 53 in a diffuse sector (Services) are not directly comparable as growth signals.
• The Dublin/RoI split is quarterly, not monthly. That lag matters when you need a short-term read; the Q1 2026 release was published in late April 2026.
• “Rest of Ireland” bundles very different places together. Cork and the South-West, the Border Region, the Midlands and the West behave differently. A 49.7 RoI reading averages those experiences; it does not tell us where the contraction sits geographically.
• Consumer sentiment is volatile around shock events. The CUCSI moves on tariff news, energy prices and geopolitical events as much as on underlying household conditions. Two consecutive months matter more than any single reading.
• Cause attribution is inherently fuzzy. When sentiment falls, press commentary will reach for the most-discussed shock at the time – Brexit, Covid, the Ukraine war, US tariffs, Middle East conflict, AI-related labour-market anxiety. Multiple stressors are usually present simultaneously, and the respondents themselves cannot disaggregate which is doing the work. Treat single-cause explanations in any press release (this one included) as illustrative, not definitive.
• None of these indices predict the future. They describe the present (PMI, sentiment) or the very recent past (Dublin PMI). For a view of what may come next, use the OECD CLI or ESRI commentary alongside them.

None of these limits invalidate the central findings. The headline trends – manufacturing in expansion, services in contraction, Dublin pulling away from Rest of Ireland, consumer sentiment at multi-year lows – are robust to reasonable changes in window or method. They simply mean the indices are best read as a set, not in isolation, and that explanations of *why* the indices are moving should be held more lightly than the readings themselves.

Conclusion

Five numbers describe the Irish economy at the end of April 2026, and they say different things. Manufacturing PMI at a near four-year high. Services PMI in contraction for the first time since February 2021. Construction PMI back in contraction, with the housing sub-index down to 44.4. Dublin’s headline PMI at a near four-year high while the Rest of Ireland slips into contraction. Consumer sentiment at its lowest level since December 2022, having fallen for two consecutive months. Each is a real signal; none is the whole picture.

Quoting a single PMI headline as the read on the Irish economy – usually the AIB Manufacturing reading, because it is the first published each month – underplays the sectoral split, the regional split and the consumer counterpoint. For any business decision that depends on Irish economic conditions, the work is in reading the indices as a set: PMIs for activity, the Dublin/RoI split for geography, the CUCSI for the consumer, and the OECD CLI for the medium horizon. Each adds something the others cannot. And the *explanations* of why any of them are moving should be held more lightly than the readings themselves – the press-release narrative is one interpretation, not the only one.

Validated data summary

Source: AIB / S&P Global monthly PMI releases (April 2026 reports for Manufacturing, Services and Construction, published 1, 6 and 12 May 2026). Full series (Jan 2018 onwards) used in volatility analysis.

All values computed from monthly press release figures. Pre-2018 monthly data is held by S&P Global on subscription and was not available for this article.

Source: Dublin Economic Monitor / S&P Global Market Intelligence quarterly releases. Full 24-quarter series used in analysis.

Sources