Jan 22, 2025 Establishment Size Enterprise Note No. 44 Establishment Size Distribution in the European Union* M. Nazım Tamkoç T his Brief studies the establishment size distribution in the European Union countries. A well-established literature has established that an abundance of small establishments among lower-income countries is evidence of the misallocation of inputs. To investigate this phenomenon, this Brief analyzes both the mean size (in terms of employment) and the employment share of the top 10 percent of establishments within an economy, across European Union countries and regions at the NUTS1 and NUT2 levels. Results show that higher-income countries have larger establishments on average and a higher concentration of employment in the top 10 percent of establishments than lower-income countries. ese ndings hold when looking at both the NUTS1 and NUTS2 regional levels. Moreover, establishment size increases with the age of the establishment and the level of foreign ownership and exports. Finally, a comparison of overall establishment size distribution reveals that lower-income countries have a higher prevalence of smaller establishments and fewer larger establishments than higher-income countries, which con rms the predictions of the misallocation literature. The relationship between smaller firm size and unproductive establishments. As a result, the mean and misallocation of resources establishment size in less developed countries is smaller compared to that in richer countries. ere is consensus in the literature that di erences e logical question is then, what distorts allocative in total factor productivity (TFP) are the main e ciency in lower-income countries that results in determinant of di erences across countries in income smaller establishments? In other words, what are the per capita (Klenow and Rodriguez-Clare, 1997; factors behind that are of behind misallocation. One of Prescott, 1998; Hall and Jones, 1999; Caselli, 2005; the most studied policies driving these distortions is and Hsieh and Klenow, 2010). One of main size-dependent (or size-correlated) distortions (see explanations for TFP di erences has been misallocation Restuccia and Rogerson (2008); Güner, Ventura, and of resources (Hsieh and Klenow, 2009; Restuccia and Xu (2008); Hsieh and Klenow (2009); Bartelsman, Rogerson, 2013)—situations where inputs are not Haltiwanger, and Scarpetta (2013); Garicano, Lelarge, allocated to their most e cient uses within a country. and Van Reenen (2014); García-Santana and Mas erefore, a major aim of this literature is to quantify (2014); Fattal-Jaef (2022); López and Torres (2020); the productivity and output gains from reallocation of and Tamkoç (2024), among many). Size-dependent resource toward their optimal usage. distortions refer to policies that promote small Recent studies have established that the mean size businesses and/or put heavier taxes or regulations on of establishments in wealthier countries is, on average, larger establishments. Examples are abundant around larger than in less developed countries (Bento and the world. Establishments with more than 50 workers Restuccia, 2017, 2020; Poeschke, 2018; Tamkoç, face heavier policies in France; larger establishments 2023a). is result follows as it is an indication that in face stricter labor market regulations in Germany and such economies, resources (such as labor and capital) Italy; interest rates are subsidized for small and medium are able to nd their way to bigger establishments enterprises in Romania; and small businesses are levied Hence, the misallocation of resources implies higher with lower tax rates in Poland—to name just a few. concentration of employment and output in smaller Richer economies may also su er from distortions that *A liations: World Bank, Development Economics, Enterprise Analysis. For correspondence: mtamkoc@worldbank.org. Acknowledgments: is Brief is a part of a series focusing on issues of regional disparities and growth opportunities in the EU-27 area. e series is a product of the World Bank’s Enterprise Analysis team (DECEA) and has bene tted from generous support from the EU DG REGIO directorate. e team would also like to thank Norman V. Loayza and Jorge Rodriguez Meza for comments and guiding the publication process. Nancy Morrison provided excellent editorial assistance. Objective and disclaimer: e ndings in this series of Briefs do not necessarily represent the views of the World Bank Group, its Executive Directors, or the governments they represent. All Briefs in the series can be accessed via: https://www.worldbank.org/en/research/brief/global-indicators-briefs-series. Enterprise Note No. 44 hinder the growth of establishments (Hsieh and e second nding emerging from the data is that Klenow, 2014). at is, the misallocation literature the employment share of the top 10 percent of does not imply the absence of size-dependent policies in establishments (in terms of employment) is higher in higher-income countries; rather, it suggests that such countries and regions with higher GDP per inhabitant. policies may distort the allocation of resources that e employment share of the top 10 percent of could result in further growth. ere are other potential establishments at a given location is the ratio of the explanations for the relative abundance of smaller number of workers at the top 10 percent of establishments, such as nancial frictions (Midrigan establishments (in terms of establishment size) in that and Xu, 2014; Moll, 2014); informality (Loayza, 1996; location to the total number of workers in that same Leal, 2014; Ulyssea, 2018; Sarıkaya, Tamkoç, and location. is higher concentration of employment in Torres, 2023; Tamkoç, 2023b); and time tax (Tamkoç larger establishments in higher-income countries is and Ventura, 2023). As a reference, Hopenhayn (2014) robust if the larger size is more narrowly de ned as the and Restuccia and Rogerson (2017) survey the top one percent of establishments or establishments misallocation literature. with more than 250 workers. e literature also notes that eliminating In turn, the third nding emerging from the data is distortions—factors that cause misallocation—from that the right tail of the overall establishment size the economy does not result in an equilibrium where all distribution of higher-income countries is thicker than production is concentrated in a single establishment the right tail of lower-income countries. e right tail employing all the workers. Since heterogeneous of the establishment size distribution of each country is establishment sizes primarily arise from decreasing calculated as the slope of the regression line where the returns in production, the removal of distortions leads logarithm of the fraction of establishments with more to a bigger mean size while production continues to than a number of workers is regressed on the logarithm occur in establishments of various sizes. of the number of workers. Overall, all these results is Brief explores whether the establishment size indicate that lower-income economies in the EU have distribution across 27 European Union (EU) countries many small establishments and fewer larger (the EU27) exhibits patterns that imply an excessive establishments than wealthier economies in the region. abundance of smaller establishments. e main data source of the study is the World Bank Enterprise Mean establishment size differences across Surveys (WBES), which were conducted between the EU 2018 and 2021 and contain rich information about establishment characteristics, nance, labor, To rst see whether the expected relationship infrastructure, and government-establishment relationships between overall income level and average establishment among many others, from the non-extractive, size holds in Europe, panel a of gure 1 plots the (log) non-agricultural private sector of each country (see the of the mean establishment size for EU countries on the rst Brief in this series). Other data, such as gross y-axis and the (log) of GDP per inhabitant on the domestic product (GDP) per inhabitant at purchasing x-axis. Each dot represents a country, and the solid line power standards (PPS) and employment at disaggregated represents a linear t from a regression of the log of locations, are taken from Eurostat. mean size regressed on the log of GDP per inhabitant. e rst nding that emerges from the data is that is regression is weighted by the overall employment the mean establishment size is higher in economies with of each location, so that more populous countries carry higher GDP per inhabitant. Establishment size is more weight in the regression. e regression de ned as the number of permanent and full-time coe cient is displayed at the top of the panel, with *** workers, including all employees and managers. e indicating statistical signi cance at the 1 percent level. positive relationship between mean establishment size Panels b and c repeat the same exercise for NUTS1 and and the level of income persists even after controlling NUTS2 regions, respectively.1 for establishment characteristics such as age of the Indeed, panel a implies that higher-income establishment, gender and experience of the top countries have larger establishments, on average, than manager, foreign ownership level, percentage of sales lower-income countries. In fact, the average coming from exports, being part of a establishment size among the 27 EU countries is 30.2 multi-establishment rm, and the sector of activity. e workers. is average encompasses comparatively regression results show that the age of the lower-income countries such as Poland and Greece, establishment, foreign ownership, export levels, and which have a mean size of 13.9 workers and 17.9 being a part of a multi-establishment rm have a workers, respectively, as well as higher-income positive correlation with establishment size. On the countries such as Denmark and Luxembourg, with other hand, establishments with female top managers averages of 59.2 workers and 43.4 workers, respectively. tend to be smaller than those with male top managers. e relationship between mean establishment size and 2 Enterprise Note No. 44 Figure 1 Higher-income EU countries and regions have larger establishments a. Across countries b. Across NUTS1 regions c. Across NUTS2 regions Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES) and Eurostat. Note: Each dot represent a country (in panel a), a NUTS1 region (in panel b), or a NUT2 region (in panel c). In each panel, the y-axis is the average establishment size (in terms of number of workers) and x-axis is the GDP per inhabitant at PPS in the EU27 from 2020. Both axises are on a log scale. Solid lines in the panels are the regression line where the log of average establishment size is regressed on the log of GDP per inhabitant. The slope of the regression line is denoted as coef. EU27 = 27 countries of the European Union (EU); NUTS = Nomenclature of Territorial Units for Statistics; PPS = purchasing power standards. a. NUTS1 refers to subnational regions with populations ranging from 3 million to 7 million. NUTS2 refers to subnational regions with between about 800,000 to 3 million inhabitants. *** p<0.01 GDP per inhabitant is positive and statistically across countries, in terms of the mean size and GDP per signi cant at the 1 percent level. e magnitude of the inhabitant. For instance, the highest and lowest regression coe cient implies that roughly doubling NUTS1 regions in terms of mean establishment size are GDP per inhabitant is associated with a 79.7 percent both in France: there are 104.1 workers per increase in establishment mean size, on average. is establishment on average in the Hauts-de-France positive relationship between the mean size and GDP region (FRE), whereas the mean size of Corsica (FRM) per inhabitant con rms the prediction of is 12.3 workers. e former is located on the English macro-development models. Channel on the border with Belgium and hosts a major Similarly, panel b shows that the positive automobile industry, while the latter region is famous relationship between mean establishment size and GDP for its tourism activity. Across 90 NUTS1 regions of per inhabitant holds at the NUTS1 level. ere is the EU, the average size of the workforce is 27.8 signi cant variation across NUTS1 regions, within and workers. e right tail of the distribution appears to be 3 Enterprise Note No. 44 dominated by NUTS1 regions in Germany. ere are at the NUTS1 level and 14.6 at the NUTS2 level. 29 NUTS1 regions whose mean establishment size is Figure 2 shows that the positive relationship higher than the EU NUTS1 average and more than between mean establishment size and GDP per one-third of them (11) are in Germany. Despite large inhabitant is not primarily driven by the sectoral variation across NUTS1 regions, the signi cant composition of regions. e share of manufacturing is coe cient from the regression of establishment size on de ned as the percentage of establishments operating in GDP per inhabitant implies that a 10 percent increase the manufacturing sector in a region. Panel a presents in GDP per inhabitant is associated with a roughly 3.7 the relationship between the establishment mean size percent increase in establishment mean size, on average. and the share of manufacturing, while panel b Panel c repeats the same exercise for 186 NUTS2 illustrates the relationship between the share of regions in the EU. As regions are disaggregated further, manufacturing and GDP per inhabitant at the NUTS1 the slope of the regression line decreases, but it is still level. At the 10% level, neither regression slope is positive and statistically signi cant. e average size of signi cant. all NUT2 regions is 28.1 workers. Of these, 16.9 percent of the NUTS2 regions whose mean Establishment size is larger for foreign- establishment size exceeds the overall average (71) are owned and older establishments and smaller in Germany and 8.5 percent are in the Netherlands. among establishments with a female top e Hovedstaden region of Denmark, the capital manager region of Denmark, and Nord-Pas-de-Calais region of France, which is within the FRE NUTS1 region How does the relationship between establishment discussed earlier, have the highest mean size in EU, size and level of development relate to establishment whereas Zachodniopomorskie and Podkarpackie characteristics? ree important establishment NUTS2 regions of Poland have the smallest mean characteristics are considered: foreign ownership, age, establishment size. e magnitude of the relationship and gender of top managers. Only the results at the between average establishment size and GDP per NUTS1 level are presented, but results hold at the inhabitant declines in more disaggregated samples. country- and NUTS2 levels, as well. Incidentally, the variation in average establishment size Foreign ownership could relate to larger is highest at the most disaggregated (that is, the NUTS establishment size if foreign rms may only nd it 2 region) level. While the standard deviation of the worthwhile to build or acquire larger establishments, mean size is 9.4 workers at the country level, it is 13.9 particularly if they must manage the operations of those The sectoral composition of regions is not driving the positive relationship between the mean Figure 2 establishment size and GDP per inhabitant Share of manufacturing across NUTS1 regions a. Size and manufacturing share b. Manufacturing share and GDP Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES) and Eurostat. Note: Each dot represents a NUTS1 region with populations rangind from 3 million to 7 million. In panel a, the y-axis is the average establishment size on a log scale, and the x-axis is the share of establishments in the manufactuing sector. In panel b, y-axis is the share of manufacturing and x-axis is the GDP per inhabitant at PPS in EU-27 from 2020 in log-scale. Solid lines in each panel are the regression line where the y-axis is reggresed on the x-axis values. The slope of the regression line is denoted as coef. EU-27 = 27 countries of the European Union; NUTS = Nomenclature of Territorial Units for Statistics; PPS = purchasing power standards. 4 Enterprise Note No. 44 establishments at an arm’s length from abroad (Chari, increase in establishment mean size, as indicated by the Chen, and Dominguez 2009). Likewise, if surviving signi cant and positive regression line. e gure establishments are able to grow reliably over their life excludes Luxembourg because the average foreign cycle, older establishments will also be larger (Hsieh ownership in Luxembourg is 43 percent, which is an and Klenow 2014). But female-managed outlier. establishments may face additional constraints, such as Figure 4 shows that NUTS1 regions with older barriers to accessing nance, that could result in lower establishments tend to have larger establishments, on observed size; additionally, the presence of average. e average age of establishments in NUTS1 gender-based discrimination could imply that only regions is 28.1 years (in contrast, the average is 30.2 smaller-sized establishments operate with a female percent in high-income regions). ere are 14 NUTS1 manager (Chiplunkar and Goldberg 2023; Ranasinghe regions whose average age is below 20 years. e 2024). average establishment began operating after 1991 (after Figure 3 shows that NUTS1 regions with higher the collapse of the Soviet Union) in 14 NUTS1 rates of foreign ownership tend to have larger regions. With the exception of two NUTS1 regions in establishments, on average. Foreign ownership is Italy, regions with average age below 20 are de ned as the percentage of establishments owned by concentrated in Eastern Europe and the Balkans. private foreign individuals, companies, or Moreover, four out of the ve oldest NUTS1 regions organizations. e average foreign ownership across all are located in Germany where the average NUTS1 regions is 5.4 percent, compared to an average establishment age exceeds 40 years. of 8.7 percent in high-income NUTS1 regions. ere If female managers face higher distortions or are nine NUTS1 regions with less than 0.5 percent barriers to operations due to discrimination, they foreign ownership: three in Italy, ve in Poland, and would employ fewer workers compared to their male one in Portugal. e average mean establishment size in counterpart. Figure 5 shows that NUTS1 regions with these regions is 15 workers, which is 53 percent of the a higher share of female top managers tend to have NUTS1 average. An increase in foreign ownership of 1 smaller establishments, on average. e average share of percentage point is associated with a 3.5 percent female top managers in NUTS1 regions of the EU is Figure 3 Regions with higher rates of foreign ownership tend to have larger establishments Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES). Note: Each dot represents a NUTS1 region with populations ranging from 3 million to 7 million. The y-axis is the average establishment size on a log scale, and the x-axis is the average percentage owned by foreigners. The solid line is the regression line where the log-average establishment size is regressed on the percentage owned by foreigners. The slope of the regression line is denoted as coef. NUTS = Nomenclature of Territorial Units for Statistics. *** p<0.01 5 Enterprise Note No. 44 Figure 4 Regions with older establishments tend to have larger establishments Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES). Note: Each dot represents a NUTS1 region with populations ranging from 3 million to 7 million. The y-axis is the average establishment size on a log scale, and the x-axis is the average age of establishments in a region. The solid line is the regression line where the log-average establishment size is regressed on the log-average of establishments in a region. The slope of the regression line is denoted as coef. *** p<0.01 Figure 5 Regions with a higher share of female top managers tend to have smaller establishments Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES). Note: Each dot represents a NUTS1 region with populations ranging from 3 million to 7 million. The y-axis is the average establishment size on a log scale, and the x-axis is the average age of establishments in a region. The solid line is the regression line where the log-average establishment size is reggresed on the share of establishments with female top manager in a region. The slope of the regression line is denoted as coef. *** p<0.01 6 Enterprise Note No. 44 17.9 percent. A signi cant and negative regression e results indicate that the level of development is coe cient of the average establishment size on the share a signi cant determinant of establishment size even of female management at the NUTS1 level indicates after accounting for all establishment-level that one percentage point increase in the share of characteristics. While the coe cient of log-GDP per female top managers decreases the mean establishment inhabitant is relatively smaller than that of reported in size by about 1.8 percent. In NUTS1 regions, the mean panel b of gure 1, it is still positive and statistically size of an establishment with less than 10 percent signi cant. After controlling for establishment-level female top managers on average is 32.2 workers, characteristics, the estimate of the elasticity of whereas the establishments with more than 25 percent establishment size with respect to GDP per inhabitant female top managers employ 20.7 workers, on average. is 0.11. In addition, the age of an establishment, Table 1 reports the result of an establishment-level regression of the natural logarithm of the number of exporting status, being part of a multi-establishment permanent full-time workers on several establishment establishment, and foreign ownership signi cantly characteristics. In addition to GDP per inhabitant (at increase the size of an establishment even after the NUTS1 level), the regressions include the controlling for the level of development (that is, GDP experience of the top manager in the sector (measured per inhabitant). In contrast, and consistent with the in years), the percentage of sales coming from exports, analysis at the NUTS1 level, the establishments run by and whether the establishment is part of a female top managers are 14 percent smaller than multi-establishment rm. establishments run by their male counterparts. The level of development, the age of an establishment, exporting status, being part of a Table 1 multi-establishment establishment, and foreign ownership significantly increase the size of an establishment Dependent Var: Log # of (1) (2) (3) (4) (5) (6) (7) Workers Log of GDP per inhabitant 0.30*** 0.28*** 0.25*** 0.20*** 0.20*** 0.13** 0.11** (at NUTS1 level) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) Experience of the Top 0.003** -0.003* -0.004** -0.004** -0.005*** -0.005*** Manager (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) Log of age 0.26*** 0.23*** 0.23*** 0.22*** 0.22*** (0.03) (0.03) (0.03) (0.03) (0.03) Percentage of Sales Coming 0.76*** 0.75*** 0.65*** 0.56*** from Exports (0.09) (0.09) (0.09) (0.09) Female Top Manager -0.15*** -0.14*** -0.14*** (0.05) (0.05) (0.05) Multi-establishment 0.67*** 0.65*** (dummy) (0.06) (0.06) Percentage Owned by 0.43*** Foreigners (0.09) Constant -0.75 -0.56 -0.92* -0.38 -0.28 0.36 0.61 (0.53) (0.53) (0.53) (0.54) (0.54) (0.54) (0.54) Sector (dummy) Yes Yes Yes Yes Yes Yes Yes # of Observation 18841 18513 18427 18315 18303 18303 18175 Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES) and Eurostat. Note: Huber-White robust standard errors are reported in parenthesis. Results are at the NUTS1 level, for regions with populations ranging from 3 million to 7 million. *** p<0.01, ** p<0.05, * p<0.1 7 Enterprise Note No. 44 The employment share of larger establishments average employment share for the top 10 percent of establishments is 58.3 percent across EU 27 countries. As discussed, the data show that mean Denmark has the highest employment share of large establishment size across countries and regions varies establishments, at 76.9 percent, while Greece has the greatly. Consequently, it could be misleading to talk lowest, at 44.5 percent. e statistically signi cant about the largest establishments across regions based on regression coe cient of employment share on GDP per the number of workers. For example, the top 10th percentile establishments by size in the inhabitant at the country level implies that a 10 percent Hauts-de-France NUTS1 region of France corresponds increase in GDP per inhabitant is associated with a 1.7 to 140 workers, whereas the top 10th percentile by percentage point increase in the employment share of establishment size in the Corsica NUTS1 region of large establishments. France is 30 workers. Panel b presents a similar pattern with the Figure 6 shows that the employment share of the employment share of establishments with at least 250 top 10 percent of establishments is larger in workers. e regression coe cient in panel b is bigger higher-income countries and regions (panel a). e than the coe cient in panel a, indicating stronger The employment share of the top 10 percent of establishments, and of establishments with Figure 6 at least 250 workers, is larger in higher-income countries and regions a. Across countries b. Across countries with 250+ threshold c. Across NUTS1 regions d. Across NUTS2 regions Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES) and Eurostat. Note: Each dot represent a country (panel a), a NUTS1 region populations ranging from 3 million to 7 million (panel b), and a NUT2 region with populations ranging from about 800,000 to 3 million (panel c). In each panel, the y-axis is the employment share of top 10 percent of establishments, and the x-axis is the GDP per inhabitant at PPS in EU27 from 2020. Both axises are on a log scale.The solid lines in both panels are the regression lines, where the employment share of top 10 percent of establishments is regressed on the log of GDP per inhabitant. The slope of the regression line is denoted as coef. EU-27 = 27 countries of the European Union; NUTS = Nomenclature of Territorial Units for Statistics; PPS = purchasing power standards. *** p<0.01 8 Enterprise Note No. 44 relationship between the employment share of lowest employment share of large establishments, with establishments with at least 250 workers and level of about 30 percent. development. e stronger relationship seems to be Panel d presents the results at the NUTS2 level. Even at mainly driven by very low employment shares in this level, higher-income regions have higher relatively lower-income countries. For example, the employment shares. On average, the employment share employment shares of the top 10 percent of of the top 10 percent of establishments is 56.4 percent. establishments and of establishments with at least 250 Some countries seem to have high variation in the workers are 76.9 percent and 62.3 percent, respectively, employment share of top 10 percent of establishments in Denmark. In contrast, the employment shares of the at the NUTS2 level. For example, more than 80 top 10 percent of establishments and of establishments percent of employment is concentrated at top 10 with at least 250 workers are 55.7 percent and 20.7 percent of establishments in Opolskie (PL52) and percent in Romania, respectively. e remainder of the Småland and the islands (SE21) NUTS2 regions of Brief focuses on the employment share of top 10 Poland and Sweden, respectively, whereas the percent of establishments because there are NUTS1 employment share of large establishments in and NUTS2 regions that do not have any Podkarpackie (PL82) (Poland) and Stockholm (SE11) establishments with 250 or more workers. (Sweden) NUTS2 regions is about 35 percent. Despite Panel c documents the positive relationship the large variation in employment shares among between large establishments’ employment share and NUTS2 regions within and across countries, the GDP per inhabitant at the NUTS1 level. e average regression coe cient of 0.06 is signi cant and implies employment share of the top 10 percent of that higher-income regions have higher concentration establishments is 55.9 percent among all NUTS1 of employment at large establishments, on average. regions. FRE has the highest employment share: large establishments employ nearly 88 percent of all workers Establishment size distribution in the region, very close to the mean establishment size presented in the previous section. On the other hand, Figure 7 plots the Log-Log establishment size the Insular Italy NUTS1 region (ITAG) of Italy has the distribution in four selected EU countries (Denmark, Figure 7 Romania and Poland have many smaller plants than Denmark and the Netherlands Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES). Note: Each line represents a country’s Log-Log establishments size distribution line. y-axis is the logarithm (base 10) of the fraction of establishments with at least X workers and x-axis is the logaratihm (base 10) of the number of workers (X). All lines are smoothed using locally weighted regression. 9 Enterprise Note No. 44 Figure 8 Comparing slopes of log-log establishment size distribution plots across EU countries Source: Author’s own calculation based on World Bank Enterprise Surveys (WBES). Note: Each line represents a country. y-axis is the slope of regression line where the logarithm (base 10) of the fraction of establishments with at least X workers is regressed on the logaratihm (base 10) of the number of workers (X). x-axis shows GDP per inhabitant. Solid regression line is obtained by regressing the slope of Log-Log establishment size distribution on the log-GDP per inhabitant at the country level. Coef denotes the slope of the refression line. EU27 = the 27 countries of the European Union (EU); PPS = purchasing power standards. *** p<0.01 Netherlands, Poland, Romania). e horizontal axis However, the right tail is thicker in Denmark than in denotes the logarithm (base 10) of the number of Netherlands. e establishment size distribution in workers in an economy. e vertical axis is the Romania and Poland looks very di erent from those of logarithm (base 10) of the fraction of establishments Denmark and Netherlands. Romania and Poland have that have the same number or more than the many smaller plants and, for example, the fraction of corresponding number of workers. e data cover only establishments with more than 100 or 1000 workers is establishments with more than 5 workers. For example, much smaller in these countries than in Denmark. the fraction of establishments with more than 5 workers Figure 8 shows that the slope of Log-Log in each economy is 1 (or 100 percent) where log10 of 1 establishment size distribution plot is larger (or smaller equal zero. e establishment size distributions in in absolute terms) in higher-income countries. e gure 7 are smoothed with locally weighted regressions. slope of the establishment size distribution is calculated Because Log-Log plots of establishment size distributions have a value zero at 5 workers, the slope of by regressing the logarithm of the share of the line is indicative of the establishment size establishments (y-axis of gure 7) on the logarithm of distribution across countries. Establishment size the number of workers (x-axis of gure 7) for each distribution lines with a larger slope (or a smaller slope country. While the establishment size distribution slope in absolute terms) indicate the existence of many large is -0.95 in Denmark, it is about -1.42 in Hungary, establishments, while lines with a smaller slope (or a Romania, and Bulgaria. Con rming the results larger slope in absolute terms) suggest that smaller presented above, the relationship between the slope of establishments are abundant. the establishment size distribution and the level of e fraction of establishments with fewer than 100 development is positive and statistically signi cant at workers is very similar in Denmark and Netherlands. the 1% level. 10 Enterprise Note No. 44 Concluding remarks Hall, R. E., and C. I. Jones. 1999. “Why Do Some Countries Produce So Much More Output per Worker than Others?” Quarterly is Brief tests the implication of misallocation Journal of Economics 114 (1): 83–116. Hopenhayn, H. A. 2014. “Firms, Misallocation, and Aggregate models using the WBES establishment-level data across Productivity: A Review.” Annual Review of Economics, 2014, vol. 6, 27 EU countries, 90 NUTS1 regions, and 186 NUTS2 issue 1, 735–70. regions. Results con rm the ndings of the literature Hsieh, C. T., and P. J. Klenow. 2009. “Misallocation and Manufacturing TFP in China and India.” Quarterly Journal of that misallocation result in overabundance of small Economics 124 (4): 1403–48. establishments and the concentration of employment at Hsieh, C. T., and P. J. Klenow. 2010. “Development Accounting.” small establishments in lower-income countries. American Economic Journal: Macroeconomics 2 (1): 207–23. Hsieh, C. T., and P. J. Klenow. 2014. “ e Life Cycle of Plants in India and Mexico.” Quarterly Journal of Economics 129 (3, August): 1035–84. Klenow, P. J., and A. Rodriguez-Clare. 1997. “ e Neoclassical Revival in Growth Economics: Has It Gone Too Far?” In NBER Notes Macroeconomics Annual 1997, Vol. 12, edited by B. S. Bernanke 1 NUTS refer to the EU geographical classi cation system, the and J. J. Rotemberg, 73–103. MIT Press. Nomenclature of Territorial Units for Statistics. NUTS1, the Leal, J. C. 2014. “Tax Collection, the Informal Sector, and largest unit, has populations ranging from 3 million to 7 million. Productivity.” Review of Economic Dynamics 17 (2): 262–86 NUTS2 refers to subnational regions with between about 800,000 Loayza, N. V. 1996. “ e Economics of the Informal Sector: A Simple to 3 million inhabitants. Model and Some Empirical Evidence from Latin America.” Carnegie-Rochester Conference Series on Public Policy 45 (December): 129–62. López, J., and J. Torres. 2020. “Size-Dependent Policies, Talent Misallocation, and the Return to Skills.” Review of Economic Dynamics 38 (October): 59–93. References Midrigan, Virgiliu, and Daniel Yi Xu. 2014. "Finance and Bartelsman, E. J., J. Haltiwanger, and S. Scarpetta. 2013. Misallocation: Evidence from Plant-Level Data." American “Cross-Country Di erences in Productivity: e Role of Allocation Economic Review, 104 (2): 422–58. and Selection.” American Economic Review 103 (1): 305–34. Moll, Benjamin. 2014. "Productivity Losses from Financial Frictions: Bento, P., and D. Restuccia. 2017. “Misallocation, Establishment Size, and Can Self-Financing Undo Capital Misallocation?" American Productivity.” American Economic Journal: Macroeconomics 97: 71–87. Economic Review, 104 (10): 3186–3221. Bento, P., and D. Restuccia. 2021. “On Average Establishment Size Poeschke, J. 2018. “ e Firm Size Distribution across Countries and across Sectors and Countries.” Journal of Monetary Economics 117 Skill-Biased Change in Entrepreneurial Technology.” American (C): 220–42. Economic Journal: Macroeconomics 10 (3): 1–41. Caselli, F. 2005. “Accounting for Cross-Country Income Di erences. Prescott, E. C. 1998. “Needed: A eory of Total Factor In Handbook of Economic Growth, Vol. 1A, edited by P. Aghion and Productivity.” International Economic Review 39 (3): 525–51. S. N. Durlauf, 679–741. Elsevier. Ranasinghe, A. 2024. “Misallocation across Establishment Gender.” Chari, A., W. Chen, and K. M. E. Dominguez. 2009. “Foreign Journal of Comparative Economics 52 (1, March): 183–206. Ownership and Firm Performance: Emerging-Market Acquisitions Restuccia, D., and R. Rogerson. 2008. “Policy Distortions and in the United States.” NBER Working Paper No. 14786, National Aggregate Productivity with Heterogeneous Establishments.” Bureau of Economic Research, Cambridge, MA. Review of Economic Dynamics 11 (4): 707–20. Chiplunkar, G., and P. K. Goldberg. 2023. “Aggregate Implications of Restuccia, D., and R. Rogerson. 2013. “Misallocation and Barriers to Female Entrepreneurship.” NBER Working Paper No. Productivity.” Review of Economic Dynamics 16 (1): 1–10. 28486, National Bureau of Economic Research, Cambridge, MA. Restuccia, D., and R. Rogerson. 2017. “ e Causes and Costs of Fattal-Jaef, R. 2022. “Entry Barriers, Idiosyncratic Distortions, and the Misallocation.” Journal of Economic Perspectives 31 (3): 151–74. Firm Size Distribution.” American Economic Journal: Sarıkaya F., M. N. Tamkoç, and J. Torres. 2023. “Informality as a Macroeconomics 14 (2): 416–68. Francis, D., and F. Jolevski. 2023. “Assessing the Business Barrier to Plant Growth.” Working Paper. Environment in the EU-27: An Overview.” DECIG Brief, World Tamkoç, M. N. 2024. “Bribery, Plant Size and Size Dependent Bank, Washington, DC. Distortions.” Journal of Development Economics, 171 García-Santana, M., and M. Mas. 2014. “ e Reservation Laws in Tamkoç, M. N. 2023a. “Managers, Talent Misallocation and India and the Misallocation of Production Factors.” Journal of Productivity.” Working Paper. Monetary Economics 66 (C): 193–209. Tamkoç, M. N. 2023b. “Fading Away Informality by Development.” Garicano, L., C. Lelarge, and J. Van Reenen. 2014. “Firm Size Working Paper. Distortions and the Productivity Distribution: Evidence from Tamkoç, M. N., and G. Ventura. 2023. “Rules and Regulations, France.” American Economic Review 104 (11): 3557–96. Managerial Time and Economic Development.” Working Paper. Güner, N., G. Ventura, and Y. Xu. 2008. “Macroeconomic Ulyssea, G. 2018. “Firms, Informality, and Development: eory and Implications of Size-Dependent Policies.” Journal of Monetary Evidence from Brazil.” American Economic Review 108 (8): Economics 55 (6): 1038–51. 2015–47. e Enterprise Note Series presents short research reports to encourage the exchange of ideas on business environment issues. e notes present evidence on the relationship between government policies and the ability of businesses to create wealth. e notes carry the names of the authors and should be cited accordingly. e ndings, interpretations, and conclusions expressed in this note are entirely those of the authors. ey do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its a liated organizations, or those of the Executive Directors of the World Bank or the governments they represent.