Finland: Selected Issues


Finland: Selected Issues

Structural Shocks, Productivity, and Growth1

Finland has gone from being a top performing advanced economy to a growth laggard since 2007 as it has suffered a unique combination of structural and cyclical shocks. The rapid decline of the (previously) high productivity ICT sector in recent years has weighed on overall growth and productivity, compounding the effects of the longer-run decline of the wood and paper industry. An analysis of industry level data indicates that shifts in the sectoral distribution of labor and capital towards lower productivity sectors is also contributing to slower aggregate productivity growth. Firm level analysis suggests that the aggregate TFP impact of reallocating resources within sectors is limited, though there is more scope to reallocate resources between sectors. Policy options to raise productivity, output, and employment are examined.

A. Before the Crisis

1. In the decade prior to the global financial crisis Finland grew faster than comparator countries. Over 1997–2007, Finland’s average real GDP growth was 3.9 percent per year. This significantly exceeded the average Euro area growth rate of 2.4 percent over the same period. Finland’s Nordic neighbors also grew more slowly, with average growth rates between 0.5 and 1.8 percentage points lower than Finland’s.


Average Real GDP Growth Rates, 1997-2007


Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: IMF World Economic Outlook and Fund

2. Strong total factor productivity (TFP) growth was the most important driver of real GDP and labor productivity growth. Estimates of TFP derived from an aggregate production function approach suggest that, on average, TFP growth accounted for about 60 percent of annual real GDP growth. This is confirmed by decompositions of the growth of labor productivity (real GDP per hour worked) into contributions from TFP growth and various types of capital deepening, following Dabla-Norris and others (2015).2 Figure 1 illustrates the contributions to labor productivity growth from information and communications technology (ICT) capital deepening, non-ICT capital deepening, human capital, and TFP growth.3 In this decomposition TFP growth accounts for nearly 70 percent of labor productivity growth on average in Finland over 1997–2007. In comparison, in Sweden (Core Europe) TFP growth’s contribution is slightly more (less) than ½ of the total labor productivity growth, on average, while in the United States it is around ⅓.

Figure 1.
Figure 1.

Contributions to Labor Productivity Growth, 1997–2007

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: Conference Board, EU KLEMS, Dabla-Norris and others (2015), and Fund staff calculations.Notes: Labor productivity is measured as real GDP per hour worked. Core Europe includes Austria, Belgium, Denmark, France, Germany, Luxembourg, and the Netherlands.

3. Rapid TFP growth was partly due to a compositional shift of industries towards higher productivity sectors, especially the ICT sector. Industry level data from the EU KLEMS database provides a more detailed picture of different industries contributions to value added and productivity growth, as well as changing allocations of factors across industries. Looking at the contributions of different industries’ TFP growth to total TFP growth over 1996–2007, we find that during this period the ICT sector’s contribution accounts for slightly more than half of the average total industry TFP growth.4 Due in part to high TFP growth, which averaged 9.4 percent per year over 1995–2007, the ICT sector’s share of total industry value added increased from 6.5 percent in 1995 to 10.7 percent in 2007. Three-fourths of this increase was driven by the growth of the electrical and optical equipment manufacturing industry. This was the industry classification applied to Nokia, which by 2008 was the world’s largest mobile phone handset maker, accounting for nearly 40 percent of all mobile phones sold.


Finland: Value Added Shares and Average TFP Growth Rate, 1995–2007


Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations.

4. The expansion of the ICT sector also helped offset the secular decline in non-ICT manufacturing, particularly that of the traditional wood and paper industry. In many advanced economies, manufacturing’s share of total value added has been gradually declining over the past few decades. However, Finland managed to maintain a relatively stable manufacturing value added share of 23–25 percent up until 2007. This is despite the decline starting in the mid-1990s of one of its most important and productive manufacturing industries, wood and paper products, which suffered from declining global demand for its products and increasing competition from emerging markets. The wood and paper industry declined from 7.2 to 4.0 percent of total value added between 1995 and 2007. The growth of electrical and optical equipment industry, from 3 to 6 percent of total value added over the same period, offset the decline of the wood and paper industry.


Manufacturing Share of Total Value Added


Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations.

B. Structural Shocks and Sectoral Shifts

5. Since 2007, Finland’s economic performance has deteriorated sharply. Finland’s GDP fell by more than 8 percent in 2009. After a partial rebound in 2010–11, the economy sank into recession again in 2012–14. As a result, Finland has gone from being one of the top performing advanced economies before to crisis, to being one of the worst performing ones over 2008–14. It has underperformed its Nordic neighbors, including Denmark, which suffered a housing bust during this period.


Average Real GDP Growth Rates, 2008-14


Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: IMF World Economic Outlook and Fund

6. The 2008–09 crisis coincided with technological and competitive changes in the global mobile phone industry that contributed to the collapse of Nokia’s handset business. In 2007 Apple released the iPhone, which caused massive changes in the mobile phone industry over the next few years. Nokia’s made strategic mistakes in how it initially responded to competition in the smart phone segment of the market, causing it to quickly lose market share to Apple. At the same time, increasing competition from Asian firms such as Samsung began to undermine Nokia’s position at the lower end of the global mobile phone market, especially in developing and emerging markets. The difficulties caused by these technological and competitive pressures were compounded by the negative effects of the global financial crisis and subsequent euro area crisis on demand in Nokia’s core European market. As a result, even as total industry value added was declining, the electrical and optical equipment industry’s share of total value added collapsed. It halved between 2007 and 2009, falling from 6 to 3 percent, then continued to decline to 1.3 percent in 2012. The electrical and optical equipment industry alone was responsible for about ¼ of the decline in total industry real value added in 2009. It has also resulted in a substantial deterioration in labor productivity, both in the electrical equipment industry and in the economy overall.


Manufacturing Industries’ Shares of Total Value Added


Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations.

Contributions to Total Real Value Added Growth

(Percentage points and percent)

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations.

Real Value Added per Hour, 2007 and 2012

(Euros per hour, 2005 euros)

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations.

7. The fall in demand during the crisis also accelerated the secular decline in the wood and paper industry. After a gradual 3 percentage point decline from 1995 to 2007, the wood and paper industry’s share of value added dropped another 1½ percentage points between 2007 and 2009. It then rebounded slightly in 2010 and has largely stabilized (in relative terms) since then. However, this is in the context of an economy that has been in recession for three years, so in absolute terms the wood and paper industry is still shrinking.

8. Aggregate and sectoral TFP growth rates are substantially lower than before the crisis. Even excluding 2008–09, when TFP growth was very negative, the contribution of TFP to real GDP has dropped dramatically. On average over 2010–14, TFP has contributed 0.3 percentage points to annual real GDP growth compared to the pre-crisis average of more than 2 percentage points. Looking at 30 industries in the EU KLEMS data set, 24 had negative average TFP growth over 2008–09. For 2010–12, nearly half of the industries still had negative average TFP growth, with the electrical equipment industry performing the worst (-15 percent).

9. The structural shocks to key manufacturing industries and continued weakness of both domestic and external demand have also depressed investment. Private investment as a share of GDP has declined by 5 percentage points between 2007 and 2014. This is partly driven by the response of relatively capital intensive manufacturing industries to the structural and demand shocks they have experienced since 2007. For industries where lower investment is due to weak demand, we can expect investment to pick-up when demand eventually recovers. However, industries facing permanent structural shocks will need to actively disinvest. This can be seen in both the electrical and optical equipment and wood and paper industries where the real fixed capital stock declined roughly 12 percent between 2007 and 2012.

10. Along with slower TFP growth, weaker investment is a drag on labor productivity growth. Investment is necessary to replace depreciated capital and increase the capital stock with more advanced equipment and software, which provide capital services in the production process. Before the crisis, capital services (ICT and non-ICT combined) were contributing as much as 1 percentage point to the annual growth rate of real value added. Since the crisis, capital services contribution to real value added growth has fallen to less than half the pre-crisis average, which further contributes to the slowdown in labor productivity growth.

11. The sectoral composition of employment has gradually shifted as industries’ fortunes have changed. To understand how labor inputs (measured by number of people employed) have been shifting across industries, we decompose the contributions from each industry over different sub-periods. We can see that all industries were increasing employment in the period 1997–2001, but the pattern began to shift during 2002–2008. Despite being a significant driver of value added growth, employment in the ICT sector barely changed over 2002–2008, as Nokia off-shored much of its production. Meanwhile, employment in the non-ICT manufacturing sector was shrinking even before the crisis due to the decline of the wood and paper industry. During 2009–2012 ICT sector employment began to shrink and the non-ICT manufacturing sector shed even more workers, but with the decline more broadly based across manufacturing industries. The only sectors with an increase in employment during 2009–2012 are the public sector and other services (e.g., finance, personal, and professional services), which, with the exception of finance, are all relatively low productivity sectors.


Change in Employment and Average Productivity

(Thousands of employees and thousands of euros per employee)

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations

12. Changes in the capital stock exhibit patterns similar to that of labor. In particular, the ICT and non-ICT manufacturing sectors’ contributions to total capital stock growth were generally negative after 2003. In contrast, the public sector, other services, and other production (e.g., primary industries, construction, and utilities) all experienced positive growth in their capital stocks. However, even in those sectors the growth of the capital stock has slowed since the crisis, contributing to the ½ percentage point lower average growth rate of the total capital stock since 2009.


Industry Contributions to Capital Stock Growth

(Percent and percentage points)

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: EU KLEMS and Fund staff calculations.

13. The relative shift in the allocation of capital and labor to lower productivity sectors reduces Finland’s potential growth rate. Aggregate TFP can be considered as the weighte average of the TFP of different sectors of the economy. As capital and labor are increasingly employed in sectors with lower productivity and slower TFP growth, aggregate TFP growth will decline. Currently, staff forecasts for aggregate TF growth suggest it will be half of its pre-crisis average over the medium-term.


TFP Growth

(Percent, annual average)

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: Statistics Finland and Fund staff calculations.

14. Shifting resources towards more productive sectors would boost aggregate labor productivity and TFP growth. For example, shifting 3 percent of labor inputs (measured as hours worked) from lower productivity sectors (i.e., trade, other production, and the public sector) to higher productivity sectors (i.e., ICT and non-ICT manufacturing) would increase labor productivity by nearly 2½ percent immediately. It would also raise the TFP growth rate, as the sectors where resources are being reallocated to have typically experienced higher TFP growth.

C. Firm Level Analysis

15. Firm level data allows us to investigate the within-industry distributions of firm size, productivity, labor, and capital. Using firm level data from Orbis, a worldwide database of primarily private company information, we can analyze the structure of industries in terms of the distribution of firms by size and how factors are allocated within each industry. This firm level data is also used for counterfactual analysis later. In particular, we focus on the post crisis period due to the more limited coverage of the firm level data before 2009.

16. The Orbis data for Finland provides over 90 thousand firm-year observations over 1994–2014, though coverage before 2009 is limited. The value added of Finnish firms in the Orbis database accounts for less than 20 percent of total value added (less than 26 percent of value added excluding the public sector) from the EU KLEMS data before 2009. For 2009–12 the coverage improves, with the firms in the Orbis database accounting for 66 percent of the total value added (88 percent of the value added excluding the public sector) from the EU KLEMS database. In fact, Nokia is one of the firms included in the Orbis database only from 2009. We exclude agriculture and mining sectors from the analysis and group firms into six industries: Wood and Paper, ICT Goods and Services, Other Manufacturing, Construction, Trade, and Other Services.

17. Small and medium size firms make up the greatest share of the number of firms, but very large firms dominate in terms of value added. The distribution of labor (share of employees in the sector) and capital (share of fixed assets) across firm sizes is highly skewed towards very large firms, and these firms tend to be relatively more capital intensive (Figure 2). Labor productivity is highest in very large manufacturing firms (see Appendix I for details).

Figure 2.
Figure 2.

Stylized Facts for Key Industries

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

18. Firm level data can be used to estimate TFP for different firm sizes and sectors. Average TFP can be estimated by regressing real output measured by real gross value added on labor and capital inputs (all in logarithms). The coefficients of labor and capital input can be interpreted as the share of labor and capital in a Cobb-Douglas production function, while the constant term represents the average TFP level across firms and time (see Table A2 in Appendix II). Fixed effect dummies are added to the regressions to estimate the effect of firm sector and size characteristics on TFP and a time fixed effect to control for the impact of overall macroeconomic conditions.5 Specifically, we estimate:


where LGVAi,t, LNi,t and LKi,t are the logarithms of real added value (in euro), the number of employees, and the real fixed assets (in euro) for firm i at year t, respectively. This model includes fixed effects for time, YRt, firm size, SZi, and sector, SECTi.

19. The results show substantial TFP differences across sectors and firm sizes. The firm size fixed effects are significant and suggest that firm size is positively correlated with TFP, with small firms 13 percent less productive than very large firms on average. Amongst the size cohorts, the gap is largest between small and medium sized firms, with the latter being 6 percent more productive on average (Figure 3). The sectoral differences, though smaller in magnitude, are significant as well. Wood and Paper, Trade. And Other Services are the least productive sectors, while ICT Goods and Services and Utilities are the most productive.

Figure 3.
Figure 3.

TFP differences by Firm Size and Sector

Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

20. This difference in TFP due to firm size can be used to assess the impact of structural changes on the aggregate level of TFP. A counterfactual scenario assumes that firms in each size cohort increase their productivity to match the productivity of the next larger cohort. The productivity of the V. Large firms—the cohort with the highest TFP—is assumed to remain the same. Similar scenarios are also constructed at the sectoral level – firms in a particular sector increase their TFP to match the TFP next larger cohort of the given sector.


Counterfactual Analysis


Citation: IMF Staff Country Reports 2015, 312; 10.5089/9781513517162.002.A001

Sources: Orbis and IMF Staff Calculations

21. Productivity gains from adjustment within a sector appear limited. The baseline counterfactual analysis yield productivity gains of less than 1 percentage point because over 75 percent of value added (and employment) is concentrated in very large firms that are already more productive. Thus, an increase in the productivity of smaller firms (or the reallocation of resources from small to large firms) has a marginal impact on aggregate productivity. To significantly increase aggregate productivity, the very large firms need to become more productive. Focusing on individual sectors, the gains are somewhat larger for the services sectors, where the very large firms are less dominant, but even in these sectors the aggregate productivity gain is less than 1.5 percent.

D. Potential Policy Responses

22. A number of potential policy options exist to facilitate the reallocation of resources across sectors and raise productivity, employment, and output in key sectors. Both product and labor market reforms can have significant effects on the allocation between sectors. They can also have productivity enhancing effects on certain industries, especially trade and manufacturing. In addition, efforts to increase the use of high-skilled labor, invest in R&D, ICT capital, and infrastructure can increase employment and output.

23. In particular, we look at reform impact estimates from Dabla-Norris and others (2015) and assess their applicability to Finland. Dabla-Norris and others (2015) estimate the cumulative short-term (3 years) and medium-term (5 years) effects of different “reform shocks” on TFP, employment, and output using industry level data for a panel of countries. The “reform shocks” they examine include:

  • Reducing product market regulations (PMR). This is based on the weighted average OECD PMR Indicator of Regulation Impact. In Finland’s case, this would include measures such as liberalizing shop opening hours and reducing restrictions on large retail stores in cities.

  • Easing employment protection legislation (EPL). The OECD’s EPL indicator is used as a proxy for overall rigidities in the labor market. See second chapter of the Selected Issues Paper for a detailed discussion of Finnish labor market issues and related reforms.

  • Reducing the labor tax wedge. The labor tax wedge measure comes from the OECD Taxing Wages database. This is relevant for Finland where the labor tax wedge is relatively high.

  • Increasing the intensity of use of high-skilled labor. The measure is the share of hours worked by employees with tertiary education in each sector from the EU KLEMS database. Broad support for higher education obviously matters here and is an area in which Finland is relatively strong. However, active labor market programs (ALMP) are another route to help unemployed workers improve their skills or retrain in new industries, which can facilitate labor reallocation between sectors and increase labor productivity.

  • Boosting R&D expenditure. The measure is R&D expenditure by industry as a share of value added, taken from the OECD. R&D spending is another area where Finland has been very strong, with significant public support for R&D and with the ICT sector being a particularly R&D intensive sector. However, as the sectoral composition of the economy changes, it is worth examining how well the current system is suited to support R&D and innovative activities by firms and whether new modes of encouraging R&D (e.g., R&D tax credits) would be appropriate.

  • Investing in ICT capital. From the EU KLEMS database, the share of ICT capital services in total capital services by industry is used to measure the intensity of ICT capital usage. Though it is an advanced economy, Finland has scope to increase ICT capital intensity in certain sectors, especially the public sector where the ICT capital intensity is less than ⅓ of the aggregate economy’s ICT capital intensity.

  • Investing in infrastructure. This measure includes roads, phone lines, and electricity generation capacity. While Finland’s infrastructure is well developed, there is undoubtedly scope to improve transportation and other infrastructure, especially in the Helsinki region.

24. Most reforms have positive effects on TFP over the medium-term, though the extent of the effect will depend on the country’s initial conditions. With the exception of easing EPL and increasing the intensity of high-skilled labor usage, Dabla-Norris and others (2015) estimate the reforms they consider could raise aggregate TFP in the average country by a few percentage points over 3–5 years. Investing more in R&D and ICT capital is estimated to have even larger effects over the medium-term, with the potential to raise aggregate TFP by around 10 percent. TFP in the manufacturing and ICT sectors especially benefit from increasing R&D spending and ICT intensity (by as much as 5–10 percent) in the medium-term. However, the effects of some reforms may be on the lower side of the estimates as Finland is already fairly close to the frontier and the scope for significant increases in R&D spending or ICT capital intensity may be more limited.

25. Reforms also typically have a positive impact on output in different sectors. The cross country regression results suggest that decreasing product market regulations can raise total output by up to 4 percent in the medium-term. The impact would probably be less than 4 percent in Finland, where the PMR indicator is slightly below the OECD average, but still positive. As with TFP, increasing R&D and ICT capital have the largest medium-term effects on output, boosting it more than 4 percent. Most of the increase stems from the impact of these measures on the manufacturing, finance and professional services, and ICT sectors. Infrastructure investment is also estimated to raise total output by between 2 and 4 percent in the medium-term, primarily through its effects on the other production, trade, and finance and professional services sectors.

26. The effects of “reform shocks” on employment are more varied, as in some cases policy measures can decrease employment in certain sectors. In aggregate, reducing product market regulations and investing in infrastructure has the greatest short-term and medium-term effects on employment. Increased investment in ICT capital can have the most deleterious effect on employment, possibly because new ICT capital can serve as a substitute for labor (e.g., through greater automation of production). The effects of other measures vary across industries. For example, while boosting R&D is good for the finance and professional services sectors, it tends to decrease employment in the trade sector (potentially by more than 4 percent in the medium-term).

E. Conclusions

27. Finland faces the challenge of raising productivity after the substantial decline of key high productivity sectors just as demographic pressures are mounting. Finland’s rapidly aging population is driving a slowdown in labor force growth, which needs to be addressed by labor market policies.6 The slower growth in the labor supply and rising number of retirees relative to workers makes the need to raise productivity even more pressing in order to maintain the country’s social welfare system.

28. Large firms already account for most resources and value added, so the scope for within industry reallocation is probably limited. In some euro area countries, small and medium-sized firms account for more substantial shares of employment and value added. However, in Finland larger firms account for the bulk of employment and output. Hence, measures to shift within industry allocations of capital and labor may have small effects in aggregate. Policies that instead promote the growth of small and medium-sized firms (e.g., support for R&D, assistance with becoming exporters), so they become larger and more productive firms, may prove more fruitful.

29. However, there is scope for policy measures to boost productivity and output, including through facilitating the reallocation of resources between sectors. For Finland, measures that improve TFP in low productivity sectors include product market reforms that could raise TFP in the Trade sector. Additionally, increased infrastructure investment can boost output in the short-run in relatively high productivity sectors such as manufacturing, ICT, and finance, as well as in lower productivity sectors such as trade and other production (e.g., agriculture). Increasing R&D spending could not only raise output in the medium-term, it could also generate a shift in the composition of employment away from the lower productivity trade sector (e.g., due to R&D resulting in more efficient logistics) and towards the financial and professional services sector. Measures to increase the availability and intensity of use of high-skilled labor can also raise employment in the higher productivity manufacturing and ICT sectors, which is where increasing ALMP could have a positive impact.


  • Budina, Nina, Mustafa Saiyid, and Xingwei Hu, 2015, “Obstacles to Firm Growth in Spain,” in Spain—Selected Issues Paper, second chapter, IMF Country Report 15/233 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Dabla-Norris, Era, Si Guo, Vikram Haskar, Minsuk Kim, Kalpana Kochhar, Kevin Wiseman, and Aleksandra Zdzienicka, 2015, “The New Normal: A Sector-Level Perspective on Productivity Trends in Advanced Economies,” Staff Discussion Note 15/03 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • O’Mahony, M., and M.P. Timmer. 2009, “Output, Input and Productivity Measures at the Industry Level: The EU KLEMS Database,” The Economic Journal, Vol. 119 (538): F374F403.

    • Search Google Scholar
    • Export Citation

Appendix I. Stylized Facts for Key Industries

  • Small and medium size firms make up the greatest share of the total number of firms, but very large firms dominate in terms of value added. We utilize the Orbis database classification of firm sizes: Small, Medium, Large and Very Large (see Appendix II for definitions). The shares of small and medium firms are about equal in the manufacturing industries and the trade sector (including distribution, retail, and wholesale trade), with approximately 40 percent of firms in each size category. The shares of large and very large firms are relatively higher - around 25 percent of total firms in manufacturing industries and 15 percent in the trade sector. In the Construction and Other Services sectors, small firms account for about 60 percent of the total and medium firms another 30–35 percent, with large and very large firms constituting less than 10 percent of total firms. In contrast, very large firms account for over 75 percent of total value added in the aggregate, with the share of such firms ranging from just over 60 percent of value added for the Other Services sector to close to 90 percent for the ICT sector. Small and medium firms account for less than 5 percent of value added in the manufacturing sectors.

  • The distributions of labor and capital are skewed. The distribution of labor (share of employees in the sector) and capital (share of fixed assets) across firm sizes is highly skewed in all industries, though more so in the manufacturing and Trade sectors. In all of the manufacturing sectors, about 80 percent of employees are engaged at very large firms and these firms own more than 90 percent of the fixed assets. In contrast, small and medium firms combined employ less than 7 percent of the people and owe less than 2 ½ percent of the capital in these manufacturing sectors, despite accounting for roughly 80 percent of the firms. In the trade sector, very large firms employee 75 percent of the people in the sector and own just less than 90 percent of the fixed assets. The distribution is less skewed in Construction, where 52 percent of employees work for very large firms, with small and medium firms employing nearly 30 percent of the workers. However, even in the construction sector, about 70 percent of the fixed assets are owned by very large firms while small and medium sized firms own about 17 percent of the fixed assets. Other Services is also less skewed, with 60 percent of the people and 70 percent of fixed assets employed in very large firms.

  • Large and very large firms are more capital intensive. Given the skewness of the distributions of both labor and capital across firm sizes, it is not immediately obvious how much capital intensity (fixed assets per employee) varies across firm sizes. Notably the capital intensity of small firms is broadly similar across industries (between 43 and 54 thousand euros per employee) and the capital intensity of medium sized firms is similar to that of small firms. In the manufacturing sectors, capital intensity of large and very large firms is substantially higher than for small and medium firms, especially in the Wood and Paper sector, where capital intensity of very large firms is more than six times higher than in small firms. Very large firms in the Other Manufacturing, ICT, and Trade sectors all have capital intensities about three times greater than small firms. The distribution of capital intensity is most compressed in Construction and Other Services sectors, with very large firms’ capital intensity about twice that of small firms.

  • Labor productivity is highest in large manufacturing firms. The distribution of labor productivity (value added per employee) also exhibits skewness across firm sizes, though less than that of capital intensity. Again, the labor productivity of small firms across industries is relatively similar (between 53 and 66 thousand euros per employee). The productivity of very large firms differs substantially though, with the productivity of very large Wood and Paper producers nearly twice that of very large firms in the Trade sector. Interestingly, while very large firms have the highest labor productivity in most sectors, this is not the case for the Trade and Other Services sectors, where the labor productivity of large firms is higher. Within industries, the productivity of very large firms in the manufacturing sectors is between 50 and 120 percent higher than that of small firms. In the Trade and Other Services sectors the productivity of very large firms is only about 20 percent greater than that of small firms. Notably, though they are not typically more capital intensive, medium size firms are significantly more productive than small firms, except in the Other Services sector. Excluding the Other Services sector, medium size firms are between 20 and 50 percent more productive than small firms.

Appendix II. Firm Size Definitions and Regression Results

Table A1.

Firm Size Categories

article image
Sources: Orbis and Fund staff calculations.
Table A2.

Panel Regression for Log of Real Value Added

article image
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1Sources: Orbis and Fund staff calculations.

Prepared by Nathaniel Arnold and Pragyan Deb.


We would like to thank Era Dabla-Norris, Kevin Wiseman, Minsuk Kim, and Jovana Sljivancanin for providing us with data and programs used in Dabla-Norris and others (2015).


TFP is derived as a residual after accounting for the contributions of the various capital types to labor productivity. Hence, measurement errors in the ICT, non-ICT, and human capital series will be captured by TFP.


In the EU KLEMS data, TFP is calculated as the residual factor in real value added after accounting for real capital and labor inputs. See O’Mahony and Timmer (2009) for the details of the EU KLEMS data construction.


See Budina and others (2015) for a similar analysis of Spanish firm level data.


See second chapter of Selected Issues Paper, for more on this.

Finland: Selected Issues
Author: International Monetary Fund. European Dept.