The OECD average investment rate stood at 22.6% in 2024, down slightly from 23.0% in 2023. This level was below the pre-financial crisis average (2000-07), while remaining above the pre-pandemic average (2010-19).
By industry, real estate investment as a share of GDP declined widely across OECD countries in 2024, contrasting with a slight increase on average in information and communication. Relative to 2010-19, investment as a share of GDP increased in the majority of countries in both industries.
By asset type, dwellings and buildings were the main drag on investment rates in 2024, declining in most OECD countries. This weakness was partly offset by ICT assets, which – relative to 2010-19 – saw increases in more countries than any other asset type, rising in over three quarters of them.
ICT and R&D investment rates have risen broadly since 2020, but unevenly across OECD countries, leading to wider cross-country gaps.
OECD Compendium of Productivity Indicators 2026
Investment plays a central role in driving productivity growth by maintaining and upgrading the capital stock, accelerating the adoption of new technologies, and facilitating the diffusion of knowledge across firms and sectors. The composition of investment across industries and assets is constantly evolving. The concentration of investment in particular industries reflects and reinforces structural shifts in the economy, with implications for productivity dynamics at the aggregate level. Different asset types serve distinct economic functions, from expanding tangible productive capacity to building the intangible knowledge base that underpins innovation. As economies become more digitalised and knowledge-intensive, investment in ICT and R&D has taken on particular importance, a role further amplified by the rise of artificial intelligence and its growing demands on digital infrastructure (Gal et al., 2026[1]).
Recent investment dynamics have unfolded against a challenging macroeconomic backdrop. The recovery from COVID-19 pandemic coincided with a surge in inflation, driven by pent-up demand, supply bottlenecks, and rising energy prices. In response to elevated inflation, central banks raised policy interest rates between 2022 and 2023, in part normalising from historically low levels. As inflation gradually moderated through 2024, monetary conditions began to ease, while mild fiscal tightening was underway across many OECD countries. Changes in monetary and fiscal policy influence investment through their effects on financing conditions, expected demand and profits. Geopolitical tensions and broader global uncertainty have added further complexity to investment decisions throughout this period (OECD, 2024[2]; OECD, 2024[3]).
This chapter examines how investment has evolved across OECD and accession candidate countries in 2024, the latest year for which data are widely available. It begins with aggregate investment rates measured as Gross Fixed Capital Formation (GFCF) as a share of Gross Domestic Product (GDP) (Box 3.1), before turning to the composition of investment by industry and asset type.
Investment rates are calculated as the ratio of GFCF to GDP, based on national accounts data. GFCF is generally compiled using a combination of census data, surveys, and administrative and business data.
The 2008 System of National Accounts (SNA, 2009[4]) recommends a minimum level of asset breakdown (Table 3.1). However, the SNA 2008 asset classification captures only a selection of all relevant asset types, particularly in the area of intangible assets.
The 2025 System of National Accounts (SNA, 2025[5]) expands the asset boundary to include additional intangible assets, most notably data and other components of intellectual property products (IPP). While there is not yet a standardised definition for investment in Artificial Intelligence (AI) as a separate asset category, AI investment is closely related to several existing asset types such as ICT equipment, computer software and databases, and research and development (R&D) (Fonteneau et al., 2025[6]). These changes aim to improve the measurement of investment in the context of an increasingly digital economy.
Note: SNA 2025 based data with sufficient cross-country coverage are not expected to be available before around 2030. The SNA 2025 grouping shown here reflects a preliminary OECD elaboration and may be revised.
Source: OECD elaboration based on the 2008 and 2025 Systems of National Accounts.
The OECD average investment rate declined in 2024, falling back below its pre-financial crisis average (2000-07) while remaining above its pre-pandemic average (2010-19)
Investment rates declined in 2024 compared with 2023 in 35 out of 42 OECD and accession candidate countries for which data are available. In many cases the decline was modest, with investment rates falling by less than one percentage point in 21 countries. Consistent with this broad but moderate slowdown, the OECD average investment rate, calculated as a weighted average using current-year GDP PPP weights, decreased from 23.0% in 2023 to 22.6% in 2024. Despite this decline, investment rates in 2024 remained above their pre-pandemic average (2010-19) in 26 countries.
Investment rates varied significantly across regions and countries. In 2024, investment rates ranged from 31.3% of GDP in Türkiye to 15.4% in Luxembourg. Relatively high investment rates were sustained across the Asia-Pacific region, with Korea, Indonesia, Japan, Australia, and New Zealand all exceeding the OECD average of 22.6%. In Latin America, Mexico and Chile had marginally above-OECD average rates of around 23%, while Colombia and Costa Rica recorded lower rates of around 16%. In Europe, Iceland, Czechia, and Switzerland recorded the highest rates at around 26%, whereas Luxembourg, Greece, and Poland posted rates around ten percentage points lower.
Ireland stands out as a significant outlier, recording a sharp decline of 7.5 percentage points from 2023 to 2024. Assessing Irish investment data is complicated by the presence of multinational firms (see Chapter 2) whose large transactions – including volatile swings in onshored intellectual property assets – can heavily impact national accounts without reflecting underlying domestic investment conditions (Andersson et al., 2024[7]). The recorded decline should therefore be interpreted with caution (Figure 3.1).
GFCF as a percentage of GDP, total economy
Note: GFCF and GDP are measured at current prices. OECD figures are calculated as weighted averages using current-year GDP PPP weights.
Source: Authors’ calculations based on OECD National Accounts.
Over a longer horizon, comparison with the low benchmark of the 2010s makes the 2024 investment picture appear more favourable than it might otherwise seem. The investment rate peaked prior to the global financial crisis (GFC) and did not fully recover during the 2010s. Dlugosch et al. (2025[8]) suggest that weak investment in the 2010s reflected subdued aggregate demand, elevated uncertainty and structural factors, including the shift towards intangible assets, more cautious corporate financial behaviour, and declining business dynamism. The post-pandemic rebound in 2022-23 represented a catch-up towards the pre-GFC level. From this longer-term perspective, the decline in 2024 signals a partial unwinding of that rebound, with the investment rate still exceeding its pre-pandemic average but falling back below its pre-GFC level (Figure 3.2).
GFCF as a percentage of GDP, total economy
Note: GFCF and GDP are measured at current prices. Weighted averages are calculated using current-year GDP PPP weights. The pre-GFC average line shows the mean of annual OECD weighted averages over 2000-07; the pre-COVID average line shows the mean of annual OECD weighted averages over 2010-19.
Source: Authors’ calculations based on OECD National Accounts.
Real estate investment as a share of GDP declined widely across OECD countries in 2024, contrasting with a slight average increase in information and communication
Changes in the aggregate investment rate can be decomposed into industry contributions by examining movements in investment as a share of GDP across industries. Investment in real estate activities, relative to GDP, declined in 19 of the 23 countries in 2024 compared with 2023, with the largest falls observed in Estonia, Israel and Hungary. In other industries, changes were more limited and varied across countries, with no single industry driving a pronounced move in either direction. As a share of GDP, investment in information and communication increased slightly, on average, with the largest rises observed in Iceland, Finland and Belgium, and only modest declines in Luxembourg and the Slovak Republic. Investment in information and communication industry covers a range of assets, including but not limited to ICT assets (Figure 3.3, Panel A).
Compared with 2010-19, information and communication saw the broadest increases in investment as a share of GDP, rising in 27 of the 32 countries. Growth in investment relative to GDP was also seen in the majority of countries in professional, administrative, financial and insurance activities. In Austria, Germany and Luxembourg, public administration, education and human health recorded the largest increase relative to 2010-19. However, weaker contributions from other industries meant that total investment rates increased only modestly or remained below their 2010-19 averages (Figure 3.3, Panel B).
Percentage points
Note: GFCF and GDP are measured at current prices. Wholesale, retail, transportation and accommodation combines ISIC Rev.4 industries G-I; Public administration, education and human health combines O-Q; Professional, administrative, financial and insurance activities combines K, M and N; and Others combines A, B, D-F and R-U. Full industry names follow ISIC Rev.4. Countries for which data refers to 2023: Bulgaria, Italy, Latvia, Netherlands, Norway, Poland, Portugal, Spain and United States. Ireland is excluded, as GFCF in 2024 was significantly affected by substantial exports of intellectual property associated with multinational corporate restructuring (Central Statistics Office, 2024[9]).
Source: Authors’ calculations based on OECD Productivity Database.
Dwellings and buildings drove down investment rates across OECD countries in 2024, partly offset by an increase in ICT assets
The decline in overall investment rates across OECD countries between 2023 and 2024 was primarily driven by a contraction in dwellings and buildings, which dragged down total investment rates in 23 out of 30 countries. This is consistent with the industry-level results discussed above, given the large share of dwellings in real estate investment. Machinery (excl. ICT equipment) also pushed investment rates lower in 21 countries, though its drag was generally smaller and more uneven. A notable exception was the Slovak Republic, where a large increase in machinery investment was recorded. The overall decline was partly offset by ICT investment, which increased relative to GDP in 15 out of 30 countries, including 11 where total investment rates fell. As an asset category, ICT investment can occur across industries and is not limited to investment by the information and communication service industry (Figure 3.4, Panel A).
Compared with 2010-19, investment rates for total economy increased in more than half of the countries in 2024. In many of these countries, ICT assets were the main driver of the increase, with ICT investment rates rising in 27 of the 35 countries. The largest increases in ICT investment rates were observed in Sweden, Switzerland and Japan. Investment rates for dwellings and buildings, as well as IPP (excl. computer software and databases), also increased in the majority of countries, although generally by less than ICT assets (Figure 3.4, Panel B).
Percentage points
Note: GFCF and GDP are measured at current prices. Countries for which data refers to 2023: Colombia, Latvia, New Zealand and Portugal. Country for which data refers to 2022: Norway. Ireland is excluded, as GFCF in 2024 was significantly affected by substantial exports of intellectual property associated with multinational corporate restructuring (Central Statistics Office, 2024[9]).
Source: Authors’ calculations based on OECD Productivity Database.
ICT investment rates have risen broadly since the pandemic, but unevenly across OECD countries, leading to wider cross-country gaps
The broad increase in ICT investment partly reflects the 2020 rise across OECD economies, associated with the expansion of digital infrastructure and the shift to remote working at the onset of the COVID-19 pandemic. Thereafter, trajectories diverged markedly. Japan and the United States maintained strong investment momentum, with ICT investment rates continuing to increase beyond their 2020 levels. By contrast, ICT investment in the euro area and the rest of the OECD rose in 2020 but then largely plateaued, remaining at the 2020 level (Figure 3.5).
Within the euro area, ICT investment remained particularly low in Luxembourg, the Slovak Republic, Germany, Croatia and Ireland – all below 2% of GDP in 2024 (Figure 3.6). The relatively subdued ICT investment of the euro area is consistent with concerns about a “middle technology trap”, whereby Europe’s industrial structure remains concentrated in more mature sectors, limiting investment in emerging digital technologies (Draghi, 2024[10]).
This divergence was also reflected in a widening dispersion of ICT investment rates across OECD countries. Between 2010-19 and 2024, the average gap between countries in the top and bottom quartiles widened by around 60%, from 2.1 to 3.3 percentage points.
GFCF as a percentage of GDP, total economy
Note: GFCF and GDP are measured at current prices. Other OECD figures are computed as weighted averages of OECD economies excluding the United States, Japan and the euro area, using current-year GDP PPP weights. Japan’s 2020 benchmark year revision re-estimated software investment using the 2020 Input-Output tables and expanded coverage in the 2021 Economic Census, raising the 2020 estimate from JPY 11.1 trillion to JPY 19.1 trillion and thereby revising up measured ICT investment (Cabinet Office, 2025[11]).
Source: Authors’ calculations based on OECD Productivity Database and OECD National Accounts.
This wider dispersion was evident in 2024 ICT investment rates. Switzerland recorded the highest ICT investment rate, at 6.5% of GDP, followed by Sweden, Japan and Czechia at 5.5%, 5.3% and 4.7%, respectively. At the other end of the distribution, Poland remained below 1% of GDP, while Bulgaria and Mexico each recorded 1.3%.
Across ICT asset types, software and databases dominated in most countries, underlining the growing importance of intangible digital assets. Their share exceeded 80% of ICT investment in 2024 in France, Switzerland, Israel and Bulgaria. Greece, Croatia and Mexico were exceptions, where IT equipment accounted for a larger share (Figure 3.6).
GFCF as a percentage of GDP, total economy, 2024
Note: GFCF and GDP are measured at current prices.
Source: Authors’ calculations based on OECD Productivity Database.
R&D investment remained strong across OECD countries, with signs of growing disparities
Research and development (R&D) is one of the largest components of intangible investment, alongside software and databases. Although R&D is not classified as an ICT asset, it remains relevant for understanding AI-related investment and productivity developments (Rubinton and Patro, 2026[12]). More broadly, as digital transformation shifts business investment towards intangible assets, R&D investment provides an important lens through which to assess evolving investment patterns (André and Gal, 2024[13]; Dlugosch et al., 2025[8]).
While this section focusses on R&D investment as recorded in the SNA, it is useful to distinguish it from Gross Domestic Expenditure on R&D (GERD) under the Frascati Manual (OECD, 2015[14]). In the SNA, R&D is capitalised as an IPP asset, while GERD treats R&D as an activity and measures expenditure dedicated to R&D activities. The two measures can differ because of the treatment of capital inputs used in R&D, the recording of R&D undertaken in the course of software development, and cross-border transactions in R&D assets. In particular, R&D investment in the SNA includes domestically produced R&D, plus imports and minus exports of R&D assets, whereas GERD measures R&D performed domestically (Ker and Galindo-Rueda, 2017[15]).
In 2024, Switzerland, Ireland, Korea and Denmark recorded some of the highest R&D investment rates, each above 4% of GDP. Compared with 2023, rates increased in 11 out of 34 countries. Among the remaining countries, declines were mostly modest, with decreases of less than 0.1 percentage points in 21 out of 23 countries. Ireland was a notable exception: although its R&D investment rate remained high compared with other countries, it fell sharply from 11.4% of GDP in 2023 to 5.1% in 2024, reflecting substantial exports of intellectual property associated with multinational corporate restructuring (Central Statistics Office, 2024[9]). By contrast, GERD in Ireland remained stable at 1.5% of GDP in 2023 and 1.4% in 2024, illustrating how R&D investment can diverge from expenditure on R&D performed domestically in countries with large cross-border transactions in R&D assets.
Despite the modest declines between 2023 and 2024, R&D investment rates remained above their 2010-19 averages in the majority of countries. The largest increases were recorded in Denmark, Croatia, Korea and Iceland, where rates in 2024 exceeded their 2010-19 averages by 1.4, 1.3, 0.9 and 0.8 percentage points, respectively (Figure 3.7).
As with ICT, compared with 2010-19, cross-country differences in R&D investment rates widened. Excluding Ireland, the average gap between the top and bottom quartiles rose by around one-quarter, from 2.6 percentage points in 2010-19 to 3.2 percentage points in 2024. Together, the widening gaps in ICT and R&D investment point to an increasingly uneven shift towards digital and knowledge-based capital across OECD countries.
GFCF as a percentage of GDP, total economy
Note: GFCF and GDP are measured at current prices.
Source: Authors’ calculations based on OECD Productivity Database and OECD National Accounts.
References
[7] Andersson, M. et al. (2024), “Intangible assets of multinational enterprises in Ireland and their impact on euro area activity”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4857698.
[13] André, C. and P. Gal (2024), “Reviving productivity growth: A review of policies”, OECD Economics Department Working Papers, No. 1822, OECD Publishing, Paris, https://doi.org/10.1787/61244acd-en.
[11] Cabinet Office (2025), Towards the 2020 Benchmark Year Revision of the National Accounts, Economic and Social Research Institute, Tokyo (in Japanese), http://www.esri.cao.go.jp/jp/sna/seibi/kouhou/pdf/2020kijun/20251119announce.pdf.
[9] Central Statistics Office (2024), Press Statement Quarterly National Accounts and International Accounts Q2 2024, https://www.cso.ie/en/csolatestnews/pressreleases/2024pressreleases/pressstatementquarterlynationalaccountsandinternationalaccountsq22024/ (accessed on 7 May 2026).
[8] Dlugosch, D. et al. (2025), “Understanding the weakness in business investment: A cross-country analysis”, OECD Economics Department Working Papers, No. 1836, OECD Publishing, Paris, https://doi.org/10.1787/89bd437d-en.
[10] Draghi, M. (2024), The future of European Competitiveness, https://doi.org/10.2872/9356120.
[6] Fonteneau, F. et al. (2025), “Advancing the measurement of investments in artificial intelligence”, OECD Artificial Intelligence Papers, No. 47, OECD Publishing, Paris, https://doi.org/10.1787/13e0da2f-en.
[1] Gal, P. et al. (2026), “Business investment in the face of the digital transformation: Initial evidence”, OECD Economics Department Working Papers, No. 1859, OECD Publishing, Paris, https://doi.org/10.1787/4f89aa3e-en.
[15] Ker, D. and F. Galindo-Rueda (2017), “Frascati Manual R&D and the System of National Accounts”, OECD Science, Technology and Industry Working Papers, No. 2017/06, OECD Publishing, Paris, https://doi.org/10.1787/edb6e020-en.
[2] OECD (2024), OECD Economic Outlook, Volume 2024 Issue 1: An Unfolding Recovery, OECD Publishing, Paris, https://doi.org/10.1787/69a0c310-en.
[3] OECD (2024), OECD Economic Outlook, Volume 2024 Issue 2: Resilience in Uncertain Times, OECD Publishing, Paris, https://doi.org/10.1787/d8814e8b-en.
[14] OECD (2015), Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development, OECD Publishing, Paris, https://doi.org/10.1787/9789264239012-en.
[12] Rubinton, H. and B. Patro (2026), “Tracking AI’s contribution to GDP growth”, St. Louis Fed On the Economy, https://www.stlouisfed.org/on-the-economy/2026/jan/tracking-ai-contribution-gdp-growth.
[5] SNA (2025), System of National Accounts 2025, United Nations, https://doi.org/10.18356/9789211073225.
[4] SNA (2009), System of National Accounts 2008, United Nations, https://unstats.un.org/unsd/nationalaccount/docs/SNA2008.pdf.