Quarterly Bulletin No. 1 2025

Box 1: The Effect of Economic Policy Uncertainty on the Forecast

By Gabriel Arce Alfaro and Stephen Byrne^[13]

Recent developments in U.S. trade policy have introduced heightened uncertainty to the global economy, given the volatile nature and potentially large impact of import tariffs. What are the possible implications for economic activity in Ireland of the recent rise in uncertainty? Periods of heightened policy uncertainty can have significant impacts on the economic outlook. Economic policy uncertainty influences the decisions of households and firms, leading to delayed investment, hiring and consumption.^[14] For example, households concerned about the potential future redundancies may delay purchases of big-ticket durable goods or more expensive leisure goods and services. Similarly, firms might delay or potentially cancel new investments in factories, machinery and technology.

Typically, policy uncertainty measures are text-based, tracking the frequency of uncertainty-related words appearing in the media. These measures are forward looking in the sense that they reflect real time uncertainty expressed by journalists, economists and commentators.^[15] Figure 1 plots two such measures of uncertainty: the index of Economic Policy Uncertainty (EPU) for Ireland, constructed from Irish newspapers by Rice (2023) and the World Uncertainty Index, which is the average of uncertainty indices for 143 countries derived from Economist Intelligence Unit country reports constructed by Ahir et al. (2022) (PDF 1.37MB). These measures are highly correlated, but the Irish EPU measure is preferable for examining uncertainty in Ireland as it tracks uncertainty about policy actions both domestically and globally that are relevant for Ireland. Both the Irish EPU and global uncertainty measures increased significantly in the last months of 2024 and first months of 2025. This increase was primarily due to uncertainty about the scope, timing and economic effects of possible changes to tariffs and other regulatory changes in the United States. The increase in the uncertainty measure for Ireland in tandem with the rise in global uncertainty reflects the importance of US multinationals and trade more generally to developments in the Irish economy.^[16]

Box 2: Value-Added Trade, Global Value Chains and Production Networks in Irish Exports.

By Michael O’Grady^[20]

Global trade faces substantial challenges in 2025, due to rising protectionism, economic fragmentation and trade disputes. Geopolitical tensions can reshape trade patterns, as countries prioritise resource security and align trade with strategic partners. Given the potential impacts of these factors on global trade flows, this Box presents an overview of Ireland’s positioning in Global Value Chains, its dependence on US imports for export-oriented production, and the main countries where Irish exports are finally absorbed.

Supply chains operate as systems of value addition, where each producer purchases inputs, adds value, and passes the goods or services to the next production stage. As part of these systems, Global Value Chains (GVCs) refer to the series of interconnected activities required to produce and deliver goods and services that span multiple countries. Each stage of the chain adds value to the final product, with individual stages often involving distinct locations, components and services. Since Covid-19 and the resulting supply chain issues that limited the availability of intermediate products, there has been renewed interest in decomposing the value added (VA) components of gross trade flows at both the country and country-sector levels.^[21]

The most basic decomposition splits the value-added content of exports into three segments: Domestic Value Added, Foreign Value Added, and Double Counting^[22]. Using the 2024 release of the OECD Inter-Country Input Output (ICIO) Tables, we estimate these values for Irish trade flows, and compare them to Ireland’s main trading partner economies (Figure 1).

Ireland is considerably more dependent on the use of foreign value-added content as an input into exports of goods and services than its key trading partners. This is due to the strong concentration of multinational enterprises (MNEs) in the Irish economy, and the international fragmentation of their production processes through global value chains.

GVC participation can be decomposed into two distinct components: backwards and forwards linkages.^[23] Backward GVC linkages represent the share of foreign value-added in total exports of a country. Similarly, forward GVC linkages represent the amount of domestic value-added embodied in exports that are further re-exported to other countries. We can generate a breakdown of GVC participation, using the decomposition method of Borin and Mancini (2023), as a share of total gross exports for Ireland and its main trading partners (Figure 2).

Backwards GVC linkages account for almost half of the total amount of value added in Irish exports. In contrast, no other trading partner’s backwards GVC share exceeds 30 per cent. In contrast, Ireland’s forward GVC linkage share (10%) is lower than all other trading partners, suggesting that Ireland exports a high volume of finished products, or intermediates that eventually form part of final demand in the importing country and not part of further re-exports.

The effects of trade disruptions and exogenous trade shocks can vary considerably across sectors given their different extent of trade integration and GVC participation. Using the ICIO tables, we again generate the basic export flow decomposition, this time for the 12 most export intensive sectors in the Irish economy (Figure 3).

The three largest exporting sectors (Computer Programming & Information services, Pharmaceutical Products and Financial & Insurance services) all rely heavily on imported intermediate goods and services, with a larger share of foreign VA content in their exports than domestic VA. Across these 12 sectors combined, foreign VA accounts for 47.5% of total gross exports.

We can decompose the components of Figure 3 using the approach of Borin and Mancini (2023) to obtain estimates of Ireland’s sectoral backwards and forwards linkages to global value chains (GVCs), presented as a share of gross sectoral exports (Figure 4). Again, Computer Programming & Information Services, Pharmaceutical Products, and Financial & Insurance Services (plus Air Transport) are the sectors most integrated in backwards GVCs, accounting for over 60% of total Irish backwards GVC participation. Administrative and Support services is the only sector that has a higher percentage share of export-value added linked to forward GVCs (24 per cent) than to backward GVCs (19 per cent).

The analysis of value-added trade flows is not limited to the global setting. At the country level, the US (31 per cent) represents the largest single-country source of imported value-added in gross exports.^[24] Given concerns over US trade relationships, and the risk of future barriers to trade, it is useful to identify the sectors most heavily exposed (in $ terms) to a reduction in the ability to import US goods and services (Figure 5). Consistent with Figures 3 and 4, the Computer Programming & Information services sector is heavily dependent on US imports for their export content, particularly services imports. As foreign content accounts for 71 per cent of total export value added in this sector, almost 23 percent of total export value added derives from US services imports alone.

Combined, the Computer Programming & Information services, Pharmaceutical Products, and Financial & Insurance services sectors account for over two-thirds of the US imports that are re-exported out of the Irish economy. However, the Petroleum Products (13.2%) and Computer, Electronic & Optical Products (3.9%) sectors have the largest proportion of US goods imports as a share of total sectoral imports. Thus, these sectors are more exposed to US trade realignment, or a general reduction in the volume of US-based goods available for import.

Finally, it is useful to consider the role of specific countries’ final demand in activating Irish production and trade flows. While bilateral trade statistics provide information on the initial importer of Irish goods and services, this may not necessarily be the country where final demand goods and services are consumed or intermediates are absorbed for domestic use. In the event of economic shocks or trade restrictions, Irish exporters could face increased costs or reduced demand for their goods and services. Thus, knowing these “final destination” export values provides a more complete indicator of the specific country-level risk exposures to the Irish economy.

Across our largest trading partners, the EU accounts for almost one-third of total Irish export absorption, while the US (20 per cent) is the largest single-country absorber of Irish export value added (Figure 6). Consistent with the increased importance of services in the Irish economy over the last two decades, services sector exports make up the majority (57 per cent) of aggregate export value added. Only the US imports a higher proportion of value added from goods-producing sectors than services sectors.

Overall, the various decompositions of the ICIO data shows three main sectors that are dependent on foreign value-added imports for their exports of goods and services: Computer Programming & Information Services, Pharmaceutical Products and Financial & Insurance Services.^[25] These sectors are highly dependent on backwards GVC-related trade, with the import of US services intermediates (and to a lesser extent, US goods intermediates) of considerable importance to their production processes. The decompositions also show the importance of the EU and the US as final destinations for Irish exports. Any introduction of tariffs between the USA and EU will not only impact the direct trade linkages between Ireland and the USA, but could have second-round effects due to the reduction in demand for Irish intermediate exports that are re-exported from the EU to the US, and from the US to the EU.

Box 3: Holding Steady? An Analysis of Low Unemployment in Ireland

By Enda Keenan^[26]

The monthly unemployment rate has been below 5 per cent in Ireland since February 2022, a value consistent with the economy being at or close to full employment.^[27] This is the longest, uninterrupted period of low unemployment observed in available data back to 1983 (Figure 1). Other measures of labour market slack such as the long-term unemployment rate (those unemployed for over one year) in the Labour Force Survey (LFS) reached an historical low of 1.1 per cent in Q4 2024. While analysis and commentary typically focusses on changes in the headline unemployment rate and the stock of unemployed workers, underpinning changes in the aggregate stock are flows into and out of unemployment from other labour market states. This Box examines transition rates between labour market states and assesses how this relates to changes in the composition of the stock of unemployed workers over time.

Within the LFS, a person aged over 15 years must be classified as either in employment (E), in unemployment (U), or outside of the labour force (N). The stock of unemployment in a given period is determined by the number of people in unemployment in period t-1 and the inflow/outflow to the other labour market states (E and N) in period t. ^[28] During economic expansions, the outflow from unemployment to employment (UE rate) typically increases, i.e. the job-finding rate rises, which contributes to a decrease in the stock of unemployed persons. On average, 33 per cent of unemployed persons have moved to being employed each quarter since 2022, which remains higher than pre-2019 levels (Figure 2).^[29] The outflow from unemployment to inactivity (UN) has broadly returned to its pre-pandemic average, while the share of persons remaining in unemployment (UU) has been steadily declining since 2012. On the inflow side, both EU and NU rates remain in line with long-run levels.

Additionally, movements between employment (E) and inactivity (N) may not affect the unemployment stock but can influence the headline unemployment rate. A greater flow from inactivity to employment (NE) can lead to higher employment levels that then contributes to higher labour force levels, reducing the unemployment rate through the denominator. The NE transition rate has averaged 6.4 per cent of the stock of inactive persons since 2022 compared to a long-run average of 4.8 per cent between 1999 and 2019.^[30] As a small open economy, migration also plays an important role in Irish labour market developments.^[31] Non-Irish citizens accounted for 73 per cent of the growth in the working age population since 2022, which highlights the contribution of net inward migration to the overall labour supply.

While the flows from unemployment to employment have improved since the pre-pandemic period, it is important to look at the composition of unemployed workers between comparative years of sub 5 per cent unemployment to see if changes in characteristics may have facilitated a higher job-finding rate. The share of persons aged 45 years and older, a cohort that can experience age bias and relatively greater difficulties in the job matching process, in unemployment has risen from 18 per cent in 2007 to 26 per cent in 2024 (Table 1). This 8pp increase is lower than the 12pp increase of this age cohort in the wider labour force as an ageing population now sees persons aged 45 years account for a larger share of the population.^[32] Characteristics negatively associated with the job-finding rate have declined such as lower educational attainment and longer periods in unemployment. Those with only upper secondary or lower education has fallen from 71 per cent of the unemployment stock in 2007 to 46 per cent in 2024, mirroring trends in the wider labour force of increased human capital development.^[33] During downturns, workers with relatively lower educational attainment are often the first to be laid off and the last to be rehired, leading to persistent unemployment. The share of long-term unemployed persons has decreased from 33 per cent to 23 per cent since 2019. While the share of those with previous work experience has remained relatively stable, the cross-section of persons with work experience and tertiary education has risen from 15 per cent in 2007 to 34 per cent in 2024. Higher skill-levels can make workers more adaptable to labour market demands, while higher educational attainment correlates with shorter unemployment spells.

In conclusion, the analysis above indicates that increases in the job-finding rate have contributed to low unemployment levels. This improvement has been backed by currently favourable macroeconomic conditions raising labour demand.^[34] Job postings and other measures of labour demand have slowed since 2022, with recent developments showing a reduction in firm unit profit margins (See Quarterly Bulletin Figure 16). This may affect firms’ ability to retain labour in anticipation of slower economic activity and contribute to increases in unemployment. ^[35] Inflow rates to unemployment have been relatively steady over time, while there has been a persistent decline in the share of persons remaining in unemployment for extended periods. Positive developments relating to lower long-term joblessness and higher educational attainment of unemployed persons are likely to have aided the job-finding rate to rise over time. These developments may have contributed to a structural change in the job-finding rate relative to previous decades that consequently may lead to lower structural unemployment in the coming years. Although unemployment could rise in the short to medium term due to cyclical pressures and Ireland’s exposure to global economic developments, the available pool of highly educated and skilled workers could facilitate a faster, subsequent economic recovery.

Box 4: New technologies and the Irish labour market

By Anil Yadav^[36]

Recent advances in new technologies – ranging from artificial intelligence (AI) to advanced robotics – are transforming labour markets and reigniting global debates about the impact of new technologies on employment. Historically, waves of innovation have been associated with increased productivity and improved standards of living, albeit often at the cost of disrupting employment for workers in some occupations. Similarly, as businesses adopt new innovations, they create opportunities while automating or transforming routine tasks. This dynamic reshapes skill demands and job structures, compelling policymakers and employers to reassess workforce development and welfare policies to ensure that the labour force remains both adaptable and resilient in the face of ongoing change.

This Box explores the exposure and complementarity of the Irish labour market to AI, setting the scene for further work planned in this area. Economists utilise various measures of exposure to assess the potential impact of new technologies on the labour market. We employ the complementarity-adjusted AI occupational exposure (C-AIOE) index, developed by Felten et al. (2021) and Pizzinelli et al. (2023), to create an exposure index specific to Ireland. The C-AIOE is a composite measure that captures both the susceptibility of particular sectors or occupations to AI technologies (exposure) and the extent to which these technologies can complement existing roles (complementarity). Importantly, C-AIOE is a relative measure, requiring a careful interpretation of the findings. A high exposure means an occupation is more exposed to AI compared to others, but it does not imply that the sectors or occupations will be entirely replaced by AI, nor does low exposure imply immunity from AI driven changes. Instead, it indicates their relative exposure compared to other sectors or occupations. A similar measure was constructed in a recent report by Department of Finance and Department of Enterprise, Trade and Employment, which analysed the potential impact of AI on Ireland’s labour market.^[37] This Box builds on this work by presenting AI exposure by sector, occupation and by considering firm size.

Figure 1 presents the exposure and complementarity of the workforce across all NACE sectors in Ireland to AI technologies, based on Irish Labour Force Survey (LFS) 2023 data. The sector level C-AIOE measure is computed as a weighted sum of occupational exposure, where the weights correspond to the share of each occupation within each sector. The figure underscores that some sectors are more exposed than others. Sectors such as ICT (Information and Communication), Finance and Insurance, Real Estate, and PST (Professional, Scientific, and Technical) exhibit a higher exposure (bottom-right quadrant), with over 90 percent of workers in these sectors being exposed to AI. Notably, approximately 75 percent of workers in ICT as well as Finance and Insurance, Real Estate sectors, are employed in low-complementarity occupations, meaning these roles are less likely to benefit from AI integration. In contrast, nearly 45 percent of workers in PST are engaged in high-complementarity occupations. Conversely, AFF (Agriculture, Forestry and Fishing), Construction, and Transportation and Storage sectors are relatively least exposed to AI technologies. Workers in these sectors largely work in low-exposed occupations, and even those in exposed occupations are typically in high-complementarity positions. Similarly, in the Education and Health sectors, the majority of workers are in high-exposure, high-complementarity roles. In Panel A of Figure 1, the bubble size presents relative employment level across sectors, with larger bubbles indicating sectors with higher employment, and in Panel B, the bubble size reflects relative Gross Value Added across sectors, where larger bubbles correspond to sectors with higher value added. Sectors in the top right quadrant (high exposure and high complementarity) in Panel B tend to have below-average Gross Value Added per worker, suggesting that further AI integration could lead to substantial productivity improvements and higher marginal gains in output per worker.

Figure 2 maps out individual occupations across the same two dimensions – AI exposure and complementarity – at the ISCO-08 2-digit level for Ireland, with each point representing the working population in that occupation. Occupations that combine high exposure and low complementarity (bottom-right quadrant) are more susceptible to potential AI disruption or significant task restructuring. Among those are occupations such as customer services clerks or general clerks. These perform certain administrative and routine-intensive tasks that may face greater challenges from AI as a proportion of their tasks can be automated with fewer opportunities for augmentation. Conversely, occupations in the high exposure, high complementarity (top-right) quadrant might experience evolving skill requirements but may also benefit from AI tools that enhance human decision-making and creativity. Examples include engineering professionals, production and service managers where AI can augment human capabilities. Meanwhile, occupations in low exposure roles (left-hand-side quadrants) tend to involve inherently human-centric tasks that are less likely to be automated yet by AI assistance. Examples of these occupations are Handicraft and printing workers, trade workers.

Larger firms are more likely to adopt AI technologies at a faster pace given their financial resources and ability to integrate new technologies (McElheran et al., 2024^[38]; Kren and Lawless, 2024^[39]). Figure 3 presents the difference in AI exposure and complementarity between larger firms (more than 500 employees) and smaller firms (less than 500 employees) across all NACE sectors. A positive value on the exposure axis indicates that larger firms are more likely to adopt AI technologies compared to smaller firms, while a positive value on the complementarity axis suggests that larger firms are better positioned to leverage AI adoption for productivity gains. The figure shows that in all sectors, larger firms – employing approximately 24 percent of the workforce – exhibit greater AI exposure compared to smaller firms, suggesting that they will be at the forefront of technology adoption. However, complementarity varies across sectors between large and small firms, being higher in sectors such as Agriculture and Health and lower in sectors such as ICT and Finance. These differences highlight how the impact of AI adoption may vary significantly across sectors.

In summary, understanding the way in which AI and the broader digital transformation is likely to have an impact on the Irish labour market will require a deepening understanding of both the demand side and supply side of the labour market. On the demand side, changing skills requirements, as reflected in online job postings, provides valuable insights into how employers in Ireland are integrating AI and other new technologies into their businesses, as well as the pace and scale of such adoption across sectors and occupations. Future analysis by Central Bank staff will focus on using online job postings data to identify emerging trends in skills demand and the occupational roles most affected by technological advancements. On the supply-side, data on educational attainment, skill levels, and workers’ ability to move between across sectors and occupations will allow us to understand how adaptable the Irish labour force is likely to be in the face of ever-increasing pace of change in the labour market. These elements are crucial for anticipating how the Irish workforce and economy can respond to ongoing technological advancements. Taken together, these demand- and supply-side perspectives will help build a fuller understanding of the interplay between technology and the labour market.

Box 5: Revisions to Government Expenditure Ceilings

By Rónán Hickey^[40]

The purpose of this Box is to compare actual government expenditure to initial budgetary ceilings over the past decade.^[41] This follows significant upward revisions to government spending allocations last year. At the beginning of 2024, gross voted expenditure of €96.3bn was planned for the year as a whole. With spending almost €3bn ahead of profile by September, however, increases to ceilings occurred in the October and November Fiscal Monitors (Figure 1), and spending ended the year at €103.7bn, or 7.7 per cent higher than initially planned. This reflected developments in both current and capital expenditure, whose initial allocations increased by €5.8bn and €1.6bn (6.9 and 12.5 per cent) respectively during the year. Overspends - relative to initial estimates - were also quite broad based across government vote groups, with expenditure in Health, Social Protection, Housing^[42] and Education all ending the year more than €1bn ahead of the allocations outlined at the beginning of the year.

Reflecting these developments, we estimate that net voted core Exchequer expenditure (which takes account of discretionary revenue measures) increased by 8.8 per cent in 2024, considerably higher than the benchmark of 5 per cent growth in the previous Government’s net expenditure rule.^[43] While Budget 2024 had anticipated that the threshold would be breached once again, a much lower growth rate of just over 6 per cent was anticipated.^[44] GNI* growth is now estimated to have been stronger in 2024 than was anticipated at the time of Budget 2024 (7.9 per cent growth in Budget 2025 compared to a previous Government projection of 4.9 per cent). The net expenditure rule was, however, designed to ensure sustainable public expenditure growth, with annual increases in spending anchored to the economy’s estimated trend growth (5 per cent) rather than being driven by cyclical conditions. This means that in the event of a positive output growth surprise, such as occurred in 2024, spending should remain anchored to the growth rate permitted under the spending rule rather than drifting upwards. This is to reduce the risk of procyclical dynamics taking hold.

There has been a notable increase in the divergence between initial spending ceilings and actual spending outturns in recent years. Overspends – relative to initial allocations – averaged 1.5 per cent per annum in the five years to 2019, with this figure increasing significantly in the subsequent five-year period (Figure 2). The largest such overspend occurred in 2020 - when gross voted expenditure ended the year almost €15bn above its initial ceiling - against the backdrop of the Covid-19 pandemic. This reflected the magnitude of this exceptional shock to the economy and the need to protect households and businesses. Subsequent measures could not have been anticipated when the Revised Estimates for Public Services was finalised in December 2019.^[45] Large upward revisions to budgetary ceilings have continued in subsequent years, however, with last year’s gap between the initial spending allocation and the actual expenditure outturn the highest in four years, both in nominal and percentage terms. In the post Covid-19 period, from 2022 to 2024, developments on the current expenditure side have driven 90 per cent of these overspends.

Stronger than planned expenditure has reflected developments across a number of vote groups in the post pandemic period (Figure 3). The Health vote group has been responsible for around one-quarter of the total overspend relative to initial allocations,^[46] with Social Protection driving a further 20 per cent as the Government has introduced significant cost of living supports. These were initially announced as temporary measures in 2022 but have been repeated for three consecutive years up to the most recent budget. Expenditure in the Education and Environment vote groups has also been above initial budgetary allocations. Cost of living measures were again a key factor driving the latter, with the introduction of energy credits explaining much of the deviation from the budgeted amount. Spending in the Department of Housing was responsible for almost 20 per cent of additional expenditure last year, and two-thirds of the total overspend on the capital side.

Expenditure overruns can arise for a number of reasons. The original resource allocation could prove insufficient to meet demand due to unexpected developments. The emergence of the Covid-19 pandemic in 2020 and the large increase in migrants and energy prices following Russia’s invasion of Ukraine in 2022, provide examples where such shocks have increased the demand and/or price of services. Alternatively, an overrun could arise because the initial budget allocation underestimated the cost of maintaining the existing level of services (ELS). If this estimate is too low, upward revisions to expenditure will occur. This appears to have been the case in 2024, given the estimated cost of ELS for the coming year increased from 2.4 per cent of the core current base in Budget 2024 to 4 per cent in Budget 2025 (almost doubling in nominal terms to €3.4bn). Finally, the higher outturn can also reflect poor expenditure control, particularly when ceilings are set just before the year in question begins. The expenditure overruns observed over recent years likely reflect a combination of these factors. It is also notable that the average increase in spending per capita has been significantly stronger than average growth in population over the past three years.

Reflecting the Exchequer developments outlined above, the broader measure of general government (GG) expenditure has increased sharply in nominal terms in Ireland over the past five years.^[47] The evolution of expenditure as a percentage of GNI* shows a different picture; it is estimated to have declined by 1.1 per cent in Ireland since 2019, continuing a broader downward trend (Figure 4). However, a sustainable fiscal policy must consider revenue developments as well as those on the expenditure side. Headline GG revenue has strengthened as a percentage of GNI* since 2019, driving the strong improvement in the GG surplus that has occurred over this period. But underlying revenue – which excludes the impact of the Apple state aid case and excess corporation tax receipts – is estimated to have fallen by 3.2 per cent. With public expenditure as a share of national income remaining broadly flat since 2022, the underlying GG budget deficit has increased over recent years. This highlights the potential impact of a loss of excess corporation tax receipts on the budget balance and the need to broaden out the tax base. The pattern of expenditure overruns that have been evident over recent years – the scale of which have increased substantially compared to the period before 2020 – emphasises the importance of establishing and adhering to an effective fiscal anchor.