WPS6493 Policy Research Working Paper 6493 Regulation, Trade and Productivity in Romania An Empirical Assessment Donato De Rosa Mariana Iootty Ana Florina Pirlea The World Bank Europe and Central Asia Region Finance and Private Sector Development June 2013 Policy Research Working Paper 6493 Abstract Inappropriate regulation can influence productivity in particular, would have led to an annual increase performance by affecting incentives to invest and in productivity of 7 percent. Realizing the benefits adopt new technologies, as well as by directly curbing from trade integration depends to some extent on competitive pressures. Results of a labor productivity regulation. In this regard, the effects of regulation on growth model for European countries suggest that productivity growth are found to be positive, regardless improving the regulatory environment—proxied by of the indicator used to measure regulation, and both the Worldwide Governance Indicators regulatory through direct and indirect channels (by increasing the quality indicator—and boosting effective exposure to speed at which a country catches up with productivity competition through increasing trade integration— leaders). Simulation results also show how countries with expressed as the ratio of exports plus imports to gross different levels of regulatory quality would benefit from domestic product—have positive effects on productivity a regulatory improvement: had Romania improved its growth. In Romania a 10 percent increase in openness regulatory environment to the same level as Denmark in to global trade over 1995–2010 would have boosted 2010, its annual productivity growth would have been 14 productivity growth by 9.7 percent per year. A 10 percent higher over 1995–2010. percent increase in openness to European Union trade, This paper is a product of the Finance and Private Sector Development, Europe and Central Asia Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at dderosa@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Regulation, Trade and Productivity in Romania: An Empirical Assessment 1 Donato De Rosa 2 (World Bank, dderosa@worldbank.org) Mariana Iootty 3 (World Bank, miootty@worldbank.org) Ana Florina Pirlea 4 (World Bank, apirlea@worldbank.org) Key words: cross country output convergence, aggregate productivity and regulation. JEL: O12, O47 Sector Board: Finance and Private Sector 1 The authors would like to thank Mamta Murthi, Peter Harrold, Francois Rantrua, Ismail Radwan, Pedro L. Rodriguez, Arabela Aprahamian and Catalin Pauna for useful comments. Views expressed in this paper are those of the authors and should not be held to represent those of the World Bank Group or its Executive Directors. 2 Senior Economist (FPDOK), corresponding author. 3 Economist (ECSSF2) 4 ETC (ECSF2) 1 1. INTRODUCTION Several theoretical and empirical studies are devoted to examining the question of whether less productive economies are catching up with their more productive counterparts. A particular strand of the empirical literature has tried to highlight the channels through which institutional settings can affect productivity differences across countries. The regulatory dimension is identified as one such channel, based on the premise that inappropriate regulation can shape the incentives to invest and adopt new technologies, as well as effectively curb competitive pressures, thus limiting the effects that the spur of competition can have on productivity enhancements. Based on a set of OECD countries and using industry-level data, Nicoletti and Scarpetta (2003) 5 measure the extent to which multi-factor productivity (MFP) growth is affected by measures of product market regulation (reflecting the intensity of competition). They find that restrictive regulation in manufacturing tends to reduce MFP growth mainly via the process of technological catch-up. Using a similar approach, Conway et al. (2006) 6 and Arnold, Nicoletti and Scarpetta (2008) 7 find evidence that anti-competitive regulation is harmful particularly for technology-driven productivity improvements in ICT-intensive sectors. 8 Using a different approach, Anos Casero and Udomsaph (2009) 9 analyze micro- level data from countries in Europe and Central Asia to estimate total-factor productivity (TFP) and examine how changes in business environment and productivity growth are correlated in the 2001-2004 period. Successful efforts to improve the business environment are found to have a strong impact on firm productivity. Openness to trade can also expected to positively affect productivity growth. Traditional models of the innovation-focused growth literature (see Grossman and Helpman (1991) 10 and Romer (1990) 11 for seminal interpretations) highlight three main channels through which trade could raise productivity: through diffusion of intermediate goods (or, implicitly, technologies); through an expansion of the market for the output of innovation; and through diffusion of general knowledge. The effective impact of increasing openness to trade on income and productivity growth is likely to depend on the country’s 5 Nicoletti, G. and S. Scarpetta (2003) “Regulation, productivity and growth: OECD evidence�, Economic Policy 18 (36), 9–72 6 Conway, P., D. De Rosa, G. Nicoletti and F. Steiner (2006), "Product Market Regulation and Productivity Convergence", OECD Economic Studies No. 43: 39-76 7 Arnold, J., G. Nicoletti and S. Scarpetta (2008),"Regulation, Allocative Efficiency and Productivity in OECD Countries: Industry and Firm-Level Evidence", OECD Economics Department Working Papers, No. 616 8 All these studies use the framework proposed by Aghion and Griffith (2005) in which productivity growth in a country/sector depends on the pace with which the country/sector grows in relation to the country/sector leader. In this sense, productivity growth of a lagging country depends on how fast the leader grows and the speed at which the productivity gap is being closed. This, in turn, depends on the policy environment in the follower country/sector, especially with reference to policies that promote firm rivalry and market entry. Apart from this common framework, these three studies have some differences. Arnold, Nicoletti and Scarpetta (2008) use micro-level data and focus on MFP growth. Nicoletti and Scarpetta (2003) also measure MFP but using industry level information, while Conway et al (2006) also rely on industry level data but measuring labor productivity growth. Finally, all these studies rely on the OECD sample of countries. 9 Anos Casero, P. and C. Udomsaph, (2009) “What Drives Firm Productivity Growth?� World Bank Policy Research Working Paper Series N.4841,. 10 Grossman, G.M., and E. Helpman. (1991). Innovation and growth in the global economy. Cambridge, MA: MIT Press. 11 Romer, P (1990) “Endogenous Technological Change� The Journal of Political Economy, Vol. 98, No. 5, Part 2 2 institutional framework. 12 As highlighted by Aghion, Howitt and Mayor-Foulkes (2005) 13 and Acemoglu, Johnson and Robinson (2005) 14, the positive long-run effects of trade on growth would only arise when combined with an institutional framework of a certain quality, which would include different types of laws, regulations, enforcement of property rights, social arrangements, etc. This study attempts to shed light on the effect of trade and regulation on productivity convergence while focusing on the specific case of Romania. The main results of the analysis can be summarized as follows: • Openness to trade (measured both in relation to global trade as well as to trade with EU partners) is shown to have a positive impact on productivity convergence when controlling for the level of regulatory quality. The potential gains from increasing trade intensity are particularly high in Romania: a ten percentage point increase in openness to global trade over the 1995-2010 period would have boosted productivity growth by 9.7 percent per year. A 10% increase in openness to EU trade, in particular, would have led to an annual increase in productivity of 7%. • As the realization of the benefits from trade integration depends to some extent on regulation, the effects of regulation on productivity growth are shown to be positive, regardless of the indicator used to measure regulation, and both through direct and indirect channels. Simulation results also show how countries with different levels of regulatory quality would benefit from a regulatory improvement: had Romania improved its regulatory environment to the same level as Denmark in 2010, its annual productivity growth would have been 14 percent higher over the 1995-2010 period. • Productivity growth in the leader country has a positive and significant impact on productivity growth in the follower countries. • The more distant a country is from the technological frontier, the greater is the scope for productivity improvements arising from the catching-up process. • The coefficient of human capital (measured as education achievement but not necessarily representing the skills of the labor force) does not appear to be statistically significant, suggesting that the effect of quality of human capital -- at least in the measure used here -- is fully captured by the productivity gap term. 2. COUNTRY-LEVEL ANALYSIS To test the effect of regulation on catch-up to best practice, a model of labor productivity based on Conway et al. (2006) is adapted and estimated at the country level. In this framework, the productivity growth of a country depends on its ability to keep pace with the leader by either innovating or taking advantage of technology transfers. This potential is influenced by the policy environment in follower 12 In fact, there might be also negative consequences of openness to trade, particularly when a greater international trade leads a country to specialize in sectors with relatively growth potential (Grossman and Helpman (1991)) 13 Aghion, P., P. Howitt, and D. Mayor-Foulkes. (2005) “The effect of financial development on convergence: Theory and evidence� Quarterly Journal of Economics 120: 173–222 14 Acemoglu, D., Johnson, S. and J. A. Robinson (2005) “Institutions as the Fundamental Cause of Long-Run Growth.� In Philippe Aghion and Steve Durlauf, eds., Handbook of Economic Growth. 3 countries that can affect, positively or negatively, the incentives to increase productivity. The following specification is used: 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑡 = 𝛿�∆𝑙𝑛𝐿𝑃𝑖𝑡 � + 𝜎(�𝑟𝑜𝑑𝑔𝑎�𝑖𝑡−1 ) + 𝛾𝑅𝑒𝑔𝑖𝑡 + 𝛼(𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑡−1 + 𝛽ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡−1 + �𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 + ∑𝑖 𝛾 Country𝑖 + ∑𝑡 𝜃 time𝑡 +∈𝑖𝑡 Eq.(1) where i and t denote countries and years, respectively. 𝐿𝑃 is labor productivity, defined as value added per employee, where both value added and number of employees are based on Eurostat data. 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑡 reflects the productivity shocks in the leader country; productivity growth in the frontier country widens the production possibility set leading to faster productivity growth in follower countries. 15 �𝑟𝑜𝑑𝑔𝑎�𝑖𝑡−1 is the productivity gap and is defined as the log ratio of the level of productivity in each country relative to that of the productivity leader. Assuming that a country's distance from the technological leader measures the scope for technology transfer, the 𝜎 coefficient captures the effect of differences in productivity levels between each country and the productivity leader on productivity growth. As �𝑟𝑜𝑑𝑔𝑎�𝑖𝑡 varies in the interval [−∞, 0] , a negative coefficient would suggest that the higher the productivity gap the greater is the scope for productivity improvements that might arise from the technological catch-up process. 𝑅𝑒𝑔𝑖𝑡 is the indicator of economy-wide regulation and proxies the policy environment of each country. In this particular exercise, it is measured by the regulatory quality indicator from the Worldwide Governance Indicators (WGI) database and is available from 1995 to 2010. The indicator captures the ability of a government to formulate and implement sound policies and regulations that permit and promote private sector development. The indicator is standardized between �2.5 and 2.5 with high scores corresponding to better outcomes. 16 The 𝛾 coefficient measures the direct influence of regulation on productivity growth. In order to reflect the possible effect of regulation on the speed with which each country catches up to the productivity leader, the model includes an interaction between regulatory quality and the productivity gap. Thus, the 𝛼 coefficient measures the indirect effect of regulation on productivity growth. As �𝑟𝑜𝑑𝑔𝑎�𝑖𝑡 is always negative, a favorable regulatory environment (captured by higher – and positive – values of 𝑅𝑒𝑔𝑖𝑡 ) would imply a negative value for the interaction term. As such, a negative and significant value of the 𝛼 coefficient would suggest that a better regulatory environment accelerates the catch up process. Figure 1 presents the evolution of the regulatory quality indicator for Romania and some selected EU countries (while Table II, Annex, presents the indicator values for all countries). 15 The productivity leader is allowed to change over time. 16 Details of how this indicator is constructed are presented in Kaufman, K., A. Kraay, and M. Mastruzzi. (2009). “Governance Matters VIII: Aggregate and Individual Governance Indicators, 1996-2008�, World Bank Policy Research Working Paper No. 4978. 4 Figure 1: Regulatory Quality indicator:1996-2010 2.5 Bulgaria Czech Republic 2 Denmark Estonia France 1.5 Germany Hungary Italy Latvia 1 Lithuania Netherlands Poland 0.5 Romania Slovakia Slovenia 0 Spain Sweden United Kingdom -0.5 Source: WGI dataset available on http://info.worldbank.org/governance/wgi/index.asp ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡−1 is included to control for a possible (direct) influence (not reflected in the productivity gap term) of the quality of human capital on the ability of the economy to respond to shocks in the productivity frontier. It is defined as the proportion of the workforce with secondary education, with data drawn from the World Development Indicators (WDI) dataset. 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 measures trade openness accounting for the effect of competition from global trade on the country’s ability to converge to the productivity frontier. Two different measures are tested: i) the country’s openness to global trade, defined as the sum of country i’s global exports and imports divided by country i’s GDP; and ii) the openness to trade within the EU region, defined as the sum of country i’s exports to and imports from the EU region divided by country i’s GDP. Both trade and GDP data are from UNCTAD. Country𝑖 is a vector of country dummies included to account for unobserved time-invariant factors that might affect productivity growth in a particular country, while time𝑡 is a vector of time dummies included to control for possible productivity shocks that might occur in a given year. The model is estimated over the 1995-2010 period for a set of 30 countries: EU countries, including Romania, plus Iceland , Norway and Switzerland) 17. Tables 1 and 2 below provide summary statistics for the variables listed in Eq(1) while the estimation results are presented in Table 3. 17 See Table I, annex, for the set of countries covered by the current analysis. 5 Table 1: Summary Statistics of Variables (all countries, 1995-2010 period) Std. 25th 75th Variables N Mean Dev pctile pctile ∆𝑙𝑛𝐿𝑃 450 0.05 0.07 0.02 0.08 ∆𝑙𝑛𝐿𝑃𝑙𝑒𝑎𝑑𝑒𝑟 450 0.03 0.04 0.01 0.06 �𝑟𝑜𝑑𝑔𝑎� 480 -0.96 0.80 -1.59 -0.38 𝑅𝑒𝑔 480 1.23 0.43 0.95 1.60 ℎ𝑢𝑚𝑎𝑛𝐾 480 47.74 16.85 38.60 60.25 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒 480 101.95 48.16 66.00 133.00 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒 _𝐸𝑈 480 52.96 26.09 33.00 70.00 Table 2: Summary Statistics of Variables (Romania, 1995-2010 period) Std. 25th 75th Variables N Mean Dev pctile pctile ∆𝑙𝑛𝐿𝑃 15 0.08 0.16 0.01 0.17 ∆𝑙𝑛𝐿𝑃𝑙𝑒𝑎𝑑𝑒𝑟 15 0.03 0.04 0.01 0.06 �𝑟𝑜𝑑𝑔𝑎� 16 -2.38 0.35 -2.65 -2.12 𝑅𝑒𝑔 16 0.23 0.27 0.03 0.50 ℎ𝑢𝑚𝑎𝑛𝐾 16 57.26 4.72 52.20 61.13 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒 16 69.75 8.27 62.50 76.00 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒 _𝐸𝑈 16 42.63 7.41 35.00 48.00 Productivity growth in the leader country is found to have a positive and significant impact on productivity growth in the follower countries. The coefficient of the productivity gap term is negative and significant, suggesting that the more distant a country is from the technological frontier the greater is the scope for productivity improvements arising from the catching-up process. The coefficient of human capital (measured as education achievement but not necessarily representing the skills of the labor force) does not appear to be statistically significant, suggesting that the effect of quality of human capital -- at least in the measure used here -- is fully captured by the productivity gap term. 18 18 In principle any effect of quality of human capital (or of any other input) on productivity growth by country should be captured by the productivity gap term. Human capital is included as a distinct variable in the model to check whether it has a direct effect on the ability of the country to catch up with the productivity frontier. 6 Table 3: Productivity Convergence and Regulation: country-level analysis for 1995-2010 period (robust std errors in parenthesis) Dependent variable: growth in labor productivity 1 2 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑡 0.4869*** 0.5013*** (0.103) (0.103) �𝑟𝑜𝑑𝑔𝑎�𝑖𝑡−1 -0.1058*** -0.1199*** (0.021) (0.021) 𝑅𝑒𝑔𝑖𝑡 0.0401* 0.0411* (0.022) (0.022) (𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑡−1 -0.0256* -0.0226* (0.013) (0.013) ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡 0.0006 0.0008 (0.001) (0.001) 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 0.0006** (0.000) 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝐸𝑈𝑖𝑡−1 0.0010*** (0.000) Constant -0.2778*** -0.2203*** (0.059) (0.061) Country dummies Yes Yes Time dummies Yes Yes R-squared 0.451 0.453 N. obs 450 450 Note: * significant at 10%; ** significant at 5%; *** significant at 1% Openness to global trade is shown to have a positive impact on productivity convergence. On average, ten extra percentage points of global trade intensity over the 1995-2010 period would lead to an increase in average annual productivity growth of 0.6 percent per year. 19 An increase in trade openness within EU region by 10 percentage points would lead to an average annual increase in productivity growth of 1.0 percent. When measuring the marginal effect of trade intensity by country, the potential benefits of trade integration are particularly high for Romania (see Figure 2) 20. All else equal, an increase of 10 percentage 19 As the dependent variable is in logarithmic format, the exact percentage change in the predicted productivity growth associated with a change in the regressor X is computed as �exp�𝛽 ̂ ∆𝑋� − 1�𝑥100where 𝛽 ̂ is the estimated coefficient. 20 The marginal effect of trade openness by country was computed in two steps. First, through the estimation of the following equation: 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑡 = 𝛿�∆𝑙𝑛𝐿𝑃𝑖𝑡 � + 𝜎(�𝑟𝑜𝑑𝑔𝑎�𝑖𝑡−1 ) + 𝛾𝑅𝑒𝑔𝑖𝑡 + 𝛼(𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑡−1 + 𝛽ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡−1 + �𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 + � 𝜗Country𝑖 ∗ 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 + � 𝛾 Country𝑖 + � 𝜃 time𝑡 +∈𝑖𝑡 𝑖 𝑖 𝑡 and then through the computation of the following term 𝜕(∆𝑙𝑛𝑃𝑟𝑜𝑑 )�𝜕 (𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒 ) for each country. The exact percentage change in the predicted productivity growth associated with a change in openness to trade is computed as 𝜕(∆𝑙𝑛𝑃𝑟𝑜𝑑 ) �𝑒𝑥� �� � ∗ ∆𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒� − 1� ∗ 100. 𝜕(𝑂�𝑒𝑛 _𝑡𝑟𝑎𝑑𝑒) 7 points in Romania’s openness to global trade over the 1995-2010 period would boost the country’s productivity growth by 9.7 percent per year, the highest value among the countries for which the average marginal effect is statistically significant, where the effect in the country with the second highest predicted effect -- Latvia – is only 4.9 percent, 50% less than for Romania. The same exercise for trade intensity within the EU region shows similar results. Romania, again, is the country that could potentially benefit most from increasing trade intensity: ten extra percentage points of openness to trade within the EU region would increase productivity growth by almost 7% per year. Figure 2: Increase in average annual productivity growth over the 1995-2010 period given an increase of ten percentage points of openness to trade Openness to Global Trade Openness to EU Trade 12.00 8.00 6.99 9.72 7.00 10.00 6.00 8.00 5.00 4.49 4.61 percent percent 4.00 6.00 4.89 3.00 2.43 2.54 4.00 2.00 2.39 2.57 2.00 1.00 0.00 0.00 Slovakia Czech Bulgaria Latvia Romania Slovakia Czech Republic Latvia Romania Republic Note: this figure shows the results of those countries for which the average marginal effect of openness to trade was statistically significant. Table III, Annex, presents the estimated marginal effect of openness to trade for all countries. Reaping the benefits from trade integration depends to some extent on regulation. In this regard, the direct effect of regulation on productivity growth is found to be positive (and significant), indicating that a less burdensome regulatory environment is conducive to productivity growth. Based on the model, regulatory improvement would also have indirect consequences for productivity growth by impacting the speed of the catch up process. As argued by Conway et al (2006), this could be due to the fact that less burdensome regulation would increase the incentives and reduce the costs of incorporating new technologies into the production process. The coefficient of (Reg ∗ prodgap)it−1 is negative and statistically significant at 10% level suggesting that well-functioning regulation is positively associated to a faster catch-up process. The total effect of the regulatory environment on productivity growth (combining the direct and indirect channels) can be assessed by estimating the average marginal effect of the 𝑅𝑒𝑔𝑖𝑡 variable in Eq(1). Results show that, all else equal, improving the regulatory indicator by one standard deviation (0.43) over the 1995-2010 period is associated with an increase of 2.86 percent in average annual productivity growth. 21 21 The average marginal effect of regulatory quality was 0.0648; the 2.86 percent results from the following computation [exp(0.0648*0.43)-1]*100. 8 To show how countries with different levels of regulatory quality would benefit from an improved regulatory improvement, we estimate the average marginal effect of the 𝑅𝑒𝑔𝑖𝑡 variable by country and then simulate the gains in productivity growth across countries if they all engage in a sustained reform effort aimed at achieving the best practice of the countries in the sample. Following a reasonable conjecture, this hypothetical improvement would not have the same magnitude in all countries, as it would be conditioned on the last observed value for regulatory conditions in 2010. Specifically, the reform would imply, for each country, augmenting the 𝑅𝑒𝑔𝑖𝑡 variable from its observed value in 2010 to the value of the country with the highest observed value in the same year, which is Denmark (with a value of 1.9). In other words, the benchmark for regulatory reform used in this exercise is the implementation of the regulatory settings related with the country with the best environment in 2010. The closer a country is to this 2010 benchmark the less ambitious would be the reform it would implement. It is worth acknowledging that the simulated values produced by this kind of exercise should be only considered as indicative parameters. Regulatory reform is not assumed to change the estimated average parameters of the model, and for this reason, it would not change the estimated average relationship between the outcome variable and the other covariates. Nevertheless, this kind of exercise can be a useful instrument to have a sense of the magnitude of the impact on the economy of a possible policy change. Figure 3 presents the results of a simulation that considers the impact on productivity growth due to a regulatory reform initiative that improves the country’s regulatory environment to the level of Denmark in 2010. The average increase in annual productivity growth ranges from 0.5 percent for Finland to 15.5 percent for Bulgaria. Romania stands out with the second highest productivity dividend (14.1 percent). Figure 3: Increase in average annual productivity growth (over 1995-2010) given a move to a regulatory environment equivalent to Denmark in 2010 18.00 16.00 14.00 12.00 percent 10.00 8.00 6.00 4.00 2.00 0.00 Greece Austria Ireland Netherlands Cyprus Slovenia United Kingdom Belgium Malta Estonia Sweden Romania Latvia France Iceland Italy Lithuania Bulgaria Slovakia Poland Portugal Hungary Spain Finland Germany Czech Republic Norway Note: The regulatory environment is measured by the regulatory quality indicator (WGI data). Results for Switzerland and Luxembourg are not presented as the corresponding marginal effects of regulatory quality were not statistically significant. Table IV, Annex, presents the average marginal effect of regulatory quality for all countries. The figure is based on the estimation of Eq(1) with 𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆 variable. Results are robust to the use of 𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆𝑬𝑼 variable. 9 3. SECTOR-LEVEL ANALYSIS In order to shed light on the sectoral dimension of productivity growth, the following equation is estimated: 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑗𝑡 = 𝛿�∆𝑙𝑛𝐿𝑃𝑖𝑗𝑡 � + 𝜎��𝑟𝑜𝑑𝑔𝑎�𝑖𝑗𝑡−1 � + 𝛾𝑅𝑒𝑔𝑖𝑡 + 𝛼(𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 + 𝛽ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡−1 𝑗 + � 𝜃 time𝑡 +∈𝑖𝑗𝑡 + �𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 + � 𝛾 Country𝑖 + � 𝜋 𝑠𝑒�𝑡𝑜𝑟 𝑖 𝑗 𝑡 Eq.(2) where: i, j and t denote countries, sectors (Nace 1.1) 22 and years, respectively. LP stands for labor productivity and is defined as sectoral value added per employee, based on Eurostat data. Reflecting data availability, the model is estimated at sector level over the period of 2003-08 for the same countries of Eq(1), except for Switzerland and Malta, given the lack of data on sectoral productivity for these countries. All the variables are defined as in Eq(1), except 𝑅𝑒𝑔 for which an additional proxy for the quality of the regulatory environment is used. 23 In addition to the regulatory quality indicator - from the WGI database - a set of alternative indicators constructed through a principal component analysis (PCA) based on Doing Business (DB) data is used. The first (All DB) is a comprehensive index including all DB indicators in all topics. The second (Business Start-up_DB) measures the complexity of a number of procedural aspects related to the entry and exit process; it encompasses the indicators for starting a business, registering property and closing a business. The third (Business Operations_DB) captures the burden of regulations faced by an enterprise in managing regular operations; it covers the topics of dealing with construction permits, paying taxes, trading across borders and employing workers. The fourth and last index (Institutional Environment_DB) measures the quality of the legal and institutional framework and aggregates the indicators for protecting investors, getting credit and enforcing contracts. All PCA indices based on DB data are coded such that higher values indicate simpler regulation. 24 It is worth mentioning that none of the regulatory proxies are available at the sector level, only at the country level. The same happens with 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒 and ℎ𝑢𝑚𝑎𝑛𝐾. Table VI (Annex) provides summary statistics for all proxies of regulatory environment used for the estimation at sector level while Figure 4 shows the evolution of the comprehensive PCA indicator on DB data (Table VII, Annex, shows the values for all PCA indicators for all comparator countries). 22 See Table IV in the annex. 23 The productivity leader is allowed to change over time and across sectors. 24 In this sense, the PCA indices – based on DB data - are consistent with the regulatory quality indicator (based on WGI data); even though the PCA indices are scaled differently, they are also defined in such a way that higher values mean better regulatory environment. 10 Box 1 – Constructing alternative proxies of regulatory environment: Principal Component Analysis using Doing Business data The PCA indices are linear combinations of Doing Business sub-indicators, where each sub-indicator is multiplied by an optimal weight that best accounts for the variance of the indicators. For example, the principal components analysis index for business entry is calculated through the following equation: Business Entry index = w0*(Starting a Business) + w1*(Registering Property)+ w2*(Closing a Business), where the weights w0, w1, and w2 are the weights that lead to the greatest variance. All principal components analysis indices are coded so higher numbers indicate less complex regulation. Figure 4: PCA indicator (All_DB):2003-2008 100 Bulgaria Czech Republic 90 Denmark Estonia France 80 Germany Hungary 70 Italy Latvia Lithuania 60 Netherlands Poland Romania 50 Slovakia Slovenia 40 Spain 2003 2004 2005 2006 2007 2008 Note: Authors’ elaboration based on DB data The results of estimating Eq(2) are given in Table 4. Consistent with the previous set of results at the country level, labor productivity growth in less productive sectors is positively influenced by the speed at which productivity grows in the same sectors in the best performing countries. In addition, the coefficient of productivity gap is, again, always negative. 11 Table 4: Productivity Convergence and Regulation: sector level analysis for 2003-08 period (robust standard errors in parenthesis) 1 2 3 4 5 6 7 8 9 10 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑡 0.0257*** 0.0334*** 0.0276*** 0.0309*** 0.0302*** 0.0246*** 0.0326*** 0.0278*** 0.0300*** 0.0298*** (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) �𝑟𝑜𝑑𝑔𝑎�𝑖𝑗𝑡−1 -0.0548*** -0.0568*** -0.0946*** -0.0483*** -0.0217 -0.0556*** -0.0576*** -0.0947*** -0.0489*** -0.0217 (0.007) (0.014) (0.022) (0.013) (0.017) (0.007) (0.014) (0.022) (0.013) (0.017) ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡−1 0.0017 0.001 0.0004 0.0011 0.0015 0.0015 0.001 0.0003 0.0011 0.0013 (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑡−1 0.0005 0.0004 -0.0003 0.0005 0.000 (0.001) (0.001) (0.001) (0.001) (0.000) 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝐸𝑈𝑖𝑡−1 0.0003 0.0005 -0.0007 0.0005 -0.0003 (0.001) (0.001) (0.001) (0.001) (0.001) 𝑅𝑒𝑔𝑖𝑡 0.0515* 0.0439* (0.0340) (0.025) (𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑡−1 -0.0110** -0.0104** (0.004) (0.004) 𝐴𝑙𝑙_𝐷𝐵𝑖𝑡 0.0047*** 0.0047*** (0.002) (0.002) (𝐴𝑙𝑙_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 -0.0002** -0.0001** (0.000) (0.000) 𝑆𝑡𝑎𝑟𝑡 − 𝑢�_𝐷𝐵𝑖𝑡 0.0081*** 0.0081*** (0.002) (0.002) (𝑆𝑡𝑎𝑟𝑡 − 𝑢�_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 0.0003 0.0003 (0.000) (0.000) 𝐵𝑢𝑠. 𝑂�𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠_𝐷𝐵𝑖𝑡 0.0025** 0.0024* (0.001) (0.001) (𝐵𝑢𝑠𝑂�𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 -0.0003 -0.0002 (0.000) (0.000) 𝐼𝑛𝑠𝑡 − 𝐸𝑛𝑣_𝐷𝐵𝑖𝑡 0.002* 0.0019* (0.001) (0.001) (𝐼𝑛𝑠𝑡 − 𝐸𝑛𝑣_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 -0.0007*** -0.0007*** (0.000) (0.000) _cons -0.5913*** -0.7969*** -0.9979*** -0.6575*** -0.5876** -0.5314*** -0.7765*** 0.000 -0.6335*** -0.5541** (0.157) (0.179) (0.221) (0.171) (0.229) (0.154) (0.184) (0.000) (0.175) (0.234) R-squared (within) 0.1348 0.1312 0.1247 0.1334 0.1379 0.136 0.1322 0.1239 0.1347 0.1379 N. obs 5129 5129 5129 5129 5129 5129 5129 5129 5129 5129 Note: Country, sector and time dummies were included but not reported. * significant at 10%; ** significant at 5%; *** significant at 1% 12 The direct effect of regulation on productivity convergence is found to be positive regardless of the proxy used to measure the quality of the regulatory environment. Regarding the indirect effect of regulation, the coefficient of the (𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 variable is shown to be negative and statistically significant for three out of the five proxies of regulatory environment, suggesting that, at least for these cases, the effect of regulation on the speed of the catch-up process is positive. The total effect of regulation on productivity growth – when assuming both direct and indirect effects – is also shown to be positive: estimations of the marginal effect of the quality of the regulatory environment on productivity growth points to a positive (and statistically significant) effect for all proxies used (see Table VI in the Annex for the results). Figure 5 below displays the estimated effect on productivity growth due to a one standard variation in each regulatory proxy over the 2003-08 period. Results suggest that productivity dividends are higher when the regulatory framework is measured by the indicator Institutional Environment_DB, which captures the quality of the legal and institutional framework for protecting investors, getting credit and enforcing contracts. Figure 5: Increase in average annual productivity growth (over 2003-08 period) given a one standard deviation increase in each regulatory proxy 8.0 6.9 7.0 6.1 6.0 4.8 4.9 5.0 percent 3.8 4.0 3.0 2.0 1.0 0.0 Note: Reg is the regulatory quality indicator, according to WGI data. One standard deviation of each regulatory proxy is computed over the whole sample of countries for the 2003-08 period. See Table VI, Annex, for the respective values considered. To identify the sectors in which the impact of regulation on productivity growth is higher, the average marginal effect of the regulatory variable was computed. 25 Results show that the marginal impact of the regulatory environment on productivity growth is positive and statistically significant for all sectors, regardless of the proxy used. Table 5 displays the effects on productivity growth associated with a one standard deviation (0.38) increasing of regulatory quality indicator (Reg) over the 2003-08 period. 26 Sectoral 25 See Table VIII in the Annex for the estimated marginal effects of all regulatory proxies used. 26 The effects on productivity growth due to variation of other proxies of regulatory environment were also computed. Results are available upon request. 13 productivity dividends resulting from regulatory improvement range from 3.9 to 5.2 percent annually. The low variance of these estimations suggests that potential gains of regulatory upgrading seem to be equally distributed across sectors. However, among the sectors listed, the non-manufacturing ones – like Electricity, gas and water supply; Water transport; Air transport, Post and telecommunications, Retail trade and Other business activities – are particularly relevant, as these are normally the most regulated in the economy and inappropriate regulation of these activities tends to have “trickle down� effects on the rest of the economy, as all industries tend to be heavy users of non-manufacturing inputs. Table 5: Increase in average annual productivity growth by sector given a one standard deviation augment of regulatory environment over 2003-08 period Estimated Sector effect (%) Mining of coal and lignite; extraction of peat 5.2 Manufacture of wearing apparel; dressing; dyeing of fur 5.2 Manufacture of leather and leather products 5.2 Manufacture of textiles 5.1 Hotels and restaurants 5.1 Manufacture of furniture; manufacturing n.e.c. 5.1 Retail trade, except of motor vehicles and motorcycles; repair of personal and household goods 5.1 Manufacture of wood and wood products 5.0 Manufacture of other transport equipment 5.0 Manufacture of food products and beverages 5.0 Manufacture of fabricated metal products, except machinery and equipment 5.0 Other business activities 5.0 Manufacture of machinery and equipment n.e.c. 5.0 Manufacture of rubber and plastic products 5.0 Manufacture of electrical machinery and apparatus n.e.c. 5.0 Forestry, logging and related service activities 5.0 Sale, maintenance and repair of motor vehicles and motorcycles; retail sale of automotive fuel 4.9 Fishing 4.9 Land transport; transport via pipelines 4.9 Construction 4.9 Manufacture of radio, television and communication equipment and apparatus 4.9 Manufacture of other non-metallic mineral products 4.9 Manufacture of medical, precision and optical instruments, watches and clocks 4.9 Manufacture of motor vehicles, trailers and semi-trailers 4.9 Publishing, printing and reproduction of recorded media 4.9 Manufacture of pulp, paper and paper products 4.9 Recycling 4.9 Agriculture, hunting and related services 4.9 Manufacture of basic metals 4.8 Wholesale trade and commission trade, except of motor vehicles and motorcycles 4.8 Manufacture of office machinery and computers 4.8 Other mining and quarrying 4.8 14 Computer and related activities 4.8 Supporting and auxiliary transport activities; activities of travel agencies 4.8 Air transport 4.7 Manufacture of chemicals, chemical products and man-made fibres 4.7 Post and telecommunications 4.7 Water transport 4.6 Research and development 4.6 Manufacture of tobacco products 4.5 Electricity, gas and water supply 4.5 Renting of machinery and equipment without operator and of personal and household goods 4.4 Extraction of crude petroleum and natural gas; service activities incidental to oil and gas extraction, excluding surveying 4.3 Manufacture of coke, refined petroleum products and nuclear fuel 4.3 Real estate activities 3.9 Note: one standard deviation of regulatory quality indicator (0.38) is computed over the whole sample of countries for the 2003-08 period (see Table VI, Annex). Results from Table 4 show that openness to trade – measured at the country level – does not present a statistically significant effect on productivity convergence at sector level. When testing alternative measures for openness to (overall) trade at sector level - covering agriculture, fishing, mining and manufacturing sectors at 2 digit level of Nace 1.1 - estimation results in Table 6 suggest that trade openness has in fact a positive (and statistically significant) effect on productivity growth. 27 27 Openness to (global) trade at sector level is defined as the sum of sector j’s global exports and imports divided by sector j’s production value. Data for global imports and exports by sector is from Comtrade, while production value information is from Eurostat. 15 Table 6: Productivity Convergence and Regulation: sector level analysis for 2003-08 period with openness to trade data at sectoral level (robust standard errors in parenthesis) 1 2 3 4 5 𝑙𝑒𝑎𝑑𝑒𝑟 ∆𝑙𝑛𝐿𝑃𝑖𝑡 0.0278*** 0.0336*** 0.0296*** 0.0315*** 0.0311*** (0.009) (0.009) (0.009) (0.009) (0.009) �𝑟𝑜𝑑𝑔𝑎�𝑖𝑗𝑡−1 -0.0735*** -0.0762*** -0.1073*** -0.0693*** -0.034 (0.009) (0.018) (0.028) (0.016) (0.024) ℎ𝑢𝑚𝑎𝑛𝐾𝑖𝑡−1 0.0005 -0.0002 -0.0001 -0.0001 0.0006 (0.002) (0.002) (0.002) (0.002) (0.002) 𝑂�𝑒𝑛_𝑡𝑟𝑎𝑑𝑒𝑖𝑗𝑡−1 0.0001** 0.0001** 0.0001** 0.0001** 0.0001** (0.000) (0.000) (0.000) (0.000) (0.000) 𝑅𝑒𝑔𝑖𝑡 0.1184** (0.048) (𝑅𝑒𝑔 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑡−1 -0.0044** (0.002) 𝐴𝑙𝑙_𝐷𝐵𝑖𝑡 0.0049** (0.002) (𝐴𝑙𝑙_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 -0.002** (0.000) 𝑆𝑡𝑎𝑟𝑡 − 𝑢�_𝐷𝐵𝑖𝑡 0.0069** (0.003) (𝑆𝑡𝑎𝑟𝑡 − 𝑢�_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 0.0003 (0.000) 𝐵𝑢𝑠. 𝑂�𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠_𝐷𝐵𝑖𝑡 0.0030** (0.001) (𝐵𝑢𝑠𝑂�𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 -0.0001 (0.000) 𝐼𝑛𝑠𝑡 − 𝐸𝑛𝑣_𝐷𝐵𝑖𝑡 0.0019* (0.001) (𝐼𝑛𝑠𝑡 − 𝐸𝑛𝑣_𝐷𝐵 ∗ �𝑟𝑜𝑑𝑔𝑎�)𝑖𝑗𝑡−1 -0.0006* (0.000) _cons -0.6108*** -0.7523*** -0.9707*** -0.6276*** -0.5760** (0.146) (0.186) (0.273) (0.168) (0.263) R-squared (within) 0.1259 0.1262 0.1249 0.1257 0.1252 N. obs 3155 3155 3155 3155 3155 Note: Country, sector and time dummies were included but not reported .* significant at 10%; ** significant at 5%; *** significant at 1% The computation of the marginal effect of trade openness on productivity convergence by sector allows identifying the industries for which productivity growth is most responsive to a variation in trade intensity. Figure 6 displays the effects on productivity growth due to a one standard deviation increasing of trade openness by sector over the 2003-08 period. 28 Only the sectors for which the estimated effect was statistically significant are presented. Manufacture of basic metals is the activity with the highest productivity dividend: a one standard deviation increase in the sector’s openness to trade over the 2003-08 period would lead to a productivity growth almost 34 percent higher (yearly). The average value of trade openness for this sector in Romania is largely inferior to the sector average of all countries considered in the analysis; this 28 The presented effects of trade openness on productivity growth refer to the model that includes Reg as the regulatory proxy used. The trade effects considering alternative regulatory proxies were also computed. Results are available upon request. 16 suggests that there is space for Romania to increase its trade integration which can largely benefit productivity convergence. The same seems to happen to the other selected sectors. Figure 6 - Increase in average annual productivity Figure 7- Average openness to trade by sector: 2003-08 growth by sector given a one standard deviation period augment of sectoral openness to trade over 2003-08 period 18.0 17.1 40.0 15.9 33.7 16.0 35.0 14.0 30.0 12.0 percent per year 25.0 9.9 10.0 percent 20.0 8.0 15.0 6.0 10.0 7.7 6.2 4.0 5.0 2.0 0.0 0.0 Manufacture of basic Manufacture of motor Mining of coal and Manufacture of basic Manufacture of motor Mining of coal and metals vehicles, trailers and lignite; extraction of metals vehicles, trailers and lignite; extraction of peat semi-trailers semi-trailers peat All countries Romania Note: one standard deviation of trade openness for the manufacture of basic metals, manufacture of motor vehicles, trailers and semi-trailers, and mining of coal and lignite (63.7, 15.5 and 42.8, respectively) are computed over the whole sample of countries for the 2003-08 period. 4. CONCLUDING REMARKS This paper attempts to shed light on the effect of trade and regulation on productivity convergence while focusing on the specific case of Romania. A model of labor productivity based on Conway et al. (2006) was adapted and estimated, first, at country the level over the 1995-2010 period for a set of 30 countries: EU countries, including Romania, plus Iceland, Norway and Switzerland. Proxying the regulatory environment by the regulatory quality indicator from the WGI dataset, the country level model has provided results which can be summarized as follows: • Openness to global trade (measured both in relation to global trade as well as to trade with EU partners) is shown to have a positive impact on productivity convergence when controlling for the level of regulatory quality. The potential gains for increasing trade intensity in Romania are particularly high: a 10% increase in openness to global trade over the 1995-2010 period would have boosted productivity growth by 9.7 percent per year. A 10% increase in openness to EU trade, in particular, would have led to an annual increase in productivity of 7%. • As the realization of the benefits from trade integration depends to some extent on regulation, the effects of regulation on productivity growth are shown to be positive, regardless of the indicator used to measure regulation, and both through direct and indirect channels. Simulation results also show how countries with different levels of regulatory quality would benefit from a regulatory improvement: had Romania improved its regulatory environment to the same level as Denmark in 2010, its annual productivity growth would have been 14 percent higher over the 1995-2010 period. 17 • Productivity growth in the leader country has a positive and significant impact of productivity growth in the follower countries. • The more distant a country is from the technological frontier, the greater is the scope for productivity improvements arising from the catching-up process. • The coefficient of human capital (measured as education achievement but not necessarily representing the skills of the labor force) does not appear to be statistically significant, suggesting that the effect of quality of human capital -- at least in the measure used here -- is fully captured by the productivity gap term. In order to shed light on the sectoral dimension of productivity growth, the same model was estimated at sector level over the period of 2003-08 for the same countries, except for Switzerland and Malta. In this sector level model, besides the regulatory quality indicator - from WGI database - a set of alternative indicators constructed through a principal component analysis (PCA) based on Doing Business (DB) data was used. The sector level results can be summarized as follows. • Trade openness at sector level has a positive (and statistically significant) effect on productivity growth. The computation of the marginal effect of trade openness on productivity convergence by sector has pointed manufacture of basic metals as the activity to show the highest productivity dividend. As the average value of trade openness for this sector in Romania is largely inferior to the sector average of all countries considered in the analysis, this suggests that there is space for Romania to increase its trade integration to the benefit of productivity convergence. • The total effect of regulation on productivity growth – when assuming both direct and indirect effects – was also shown to be positive: estimations of the marginal effect of the quality of the regulatory environment on productivity growth points to a positive (and statistically significant) effect for all proxies used. In particular, results suggest that productivity dividends are higher when the regulatory framework is measured by the indicator Institutional Environment_DB, which captures the quality of the legal and institutional framework regarding protecting investors, getting credit and enforcing contracts. In order to identify the sectors in which the impact of regulation on productivity growth is higher, the average marginal effect of the regulatory environment variable was computed and results have shown that the marginal impact of the regulatory environment on productivity growth is positive and statistically significant for all sectors, regardless of the proxy used. Sectoral productivity dividends resulting from regulatory improvement range from 3.9 to 5.2 percent yearly. • Consistent with the previous set of results at country level, labor productivity growth in less productive sectors is positively influenced by the speed at which productivity grows in the same sectors in the best performing countries. In addition, the coefficient for the productivity gap was, again, negative. 18 ANNEX Table I –List of Countries 1 Austria 2 Belgium 3 Bulgaria 4 Cyprus 5 Czech Republic 6 Denmark 7 Estonia 8 Finland 9 France 10 Germany 11 Greece 12 Hungary 13 Iceland 14 Ireland 15 Italy 16 Latvia 17 Lithuania 18 Luxembourg 19 Malta 20 Netherlands 21 Norway 22 Poland 23 Portugal 24 Romania 25 Slovakia 26 Slovenia 27 Spain 28 Sweden 29 Switzerland 30 United Kingdom 19 Table II – Evolution of Regulatory Quality indicator* Country 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Denmark 1.70 1.70 1.70 1.70 1.70 1.70 1.70 1.70 1.80 1.80 1.70 1.80 1.90 1.90 1.90 1.90 Finland 1.35 1.50 1.65 1.80 1.85 1.90 1.90 1.90 1.90 1.80 1.70 1.60 1.50 1.60 1.80 1.80 Netherlands 1.75 1.80 1.85 1.90 2.00 2.10 2.00 1.90 1.70 1.80 1.70 1.70 1.80 1.80 1.80 1.80 Ireland 1.55 1.60 1.65 1.70 1.75 1.80 1.75 1.70 1.60 1.60 1.50 1.80 1.80 1.90 1.70 1.70 Luxembourg 1.65 1.60 1.55 1.50 1.75 2.00 1.95 1.90 1.80 1.80 1.60 1.70 1.70 1.70 1.70 1.70 Sweden 1.45 1.40 1.35 1.30 1.40 1.50 1.60 1.70 1.60 1.70 1.50 1.50 1.60 1.70 1.70 1.70 Switzerland 1.30 1.40 1.50 1.60 1.70 1.80 1.75 1.70 1.70 1.70 1.50 1.50 1.60 1.60 1.60 1.70 United Kingdom 2.00 2.00 2.00 2.00 1.85 1.70 1.70 1.70 1.70 1.80 1.60 1.80 1.90 1.80 1.60 1.70 Germany 1.60 1.50 1.40 1.30 1.45 1.60 1.60 1.60 1.60 1.50 1.50 1.60 1.60 1.50 1.50 1.60 Austria 1.70 1.60 1.50 1.40 1.50 1.60 1.60 1.60 1.60 1.50 1.60 1.70 1.70 1.60 1.50 1.50 Liechtenstein 1.50 1.50 1.50 1.50 1.50 1.50 1.60 1.70 1.60 1.50 1.60 1.30 1.40 1.40 1.50 1.50 Norway 1.55 1.50 1.45 1.40 1.20 1.00 1.15 1.30 1.30 1.50 1.50 1.30 1.40 1.40 1.40 1.50 Cyprus 1.35 1.30 1.25 1.20 1.15 1.10 1.15 1.20 1.20 1.20 1.30 1.30 1.30 1.40 1.30 1.40 Estonia 1.45 1.40 1.35 1.30 1.30 1.30 1.35 1.40 1.30 1.30 1.30 1.30 1.40 1.50 1.50 1.40 Malta 0.85 0.90 0.95 1.00 1.05 1.10 1.10 1.10 1.20 1.10 1.00 1.10 1.10 1.20 1.40 1.40 Belgium 1.30 1.20 1.10 1.00 1.10 1.20 1.25 1.30 1.30 1.40 1.20 1.30 1.40 1.40 1.30 1.30 France 1.05 1.00 0.95 0.90 0.95 1.00 1.00 1.00 1.20 1.20 1.20 1.20 1.30 1.30 1.20 1.30 Czech Republic 1.05 1.00 0.95 0.90 0.85 0.80 1.00 1.20 1.20 1.10 1.10 1.10 1.00 1.20 1.30 1.20 Spain 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.30 1.30 1.30 1.20 1.20 1.20 1.20 1.20 Hungary 0.70 0.80 0.90 1.00 1.00 1.00 1.15 1.30 1.10 1.20 1.10 1.20 1.20 1.20 1.10 1.10 Slovakia 0.55 0.50 0.45 0.40 0.45 0.50 0.70 0.90 1.00 1.20 1.20 1.10 1.00 1.10 1.10 1.10 Latvia 1.05 1.00 0.95 0.90 0.80 0.70 0.80 0.90 1.00 1.00 0.90 1.00 1.00 1.00 1.00 1.00 Lithuania 1.35 1.20 1.05 0.90 0.85 0.80 0.95 1.10 1.00 1.10 1.00 1.00 1.10 1.10 1.00 1.00 Poland 0.75 0.70 0.65 0.60 0.60 0.60 0.65 0.70 0.70 0.80 0.80 0.70 0.80 0.80 0.90 1.00 Iceland 0.90 1.00 1.10 1.20 1.35 1.50 1.50 1.50 1.60 1.60 1.60 1.60 1.50 1.10 1.00 0.90 Italy 0.80 0.80 0.80 0.80 0.85 0.90 0.90 0.90 1.00 1.10 1.00 0.90 0.90 0.90 0.90 0.80 Portugal 1.35 1.30 1.25 1.20 1.10 1.00 1.15 1.30 1.20 1.20 1.30 1.10 1.10 1.10 1.10 0.80 Slovenia 1.00 1.00 1.00 1.00 0.85 0.70 0.75 0.80 0.80 0.90 0.80 0.80 0.80 0.80 0.90 0.80 Greece 0.55 0.60 0.65 0.70 0.75 0.80 0.90 1.00 1.00 0.80 0.90 0.80 0.90 0.90 0.80 0.70 Romania 0.05 0.10 0.15 0.20 0.05 -0.10 -0.05 0.00 -0.10 0.20 0.20 0.50 0.50 0.60 0.60 0.70 Bulgaria -0.20 -0.10 0.00 0.10 0.15 0.20 0.40 0.60 0.60 0.70 0.60 0.60 0.60 0.70 0.60 0.60 *countries are ordered by 2010 values Note: values for 1995, 1997, 1999 and 2001 were imputed 20 Table III – Average marginal effect of Openness to Trade by Country (country level analysis, 1995-2010 period) Average marginal Average marginal effect effect Country �(∆��𝑷𝒓�𝒅) �(∆��𝑷𝒓�𝒅) [ ] [ ] �(𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆) �(𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆𝑬𝑼) Austria 0.000 0.000 Belgium 0.000 0.000 Bulgaria 0.001 0.004** Cyprus 0.002 -0.003 Czech Republic 0.003** 0.003** Denmark 0.000 -0.001 Estonia -0.001 -0.001 Finland 0.000 -0.001 France 0.000 -0.001 Germany -0.001 -0.002 Greece 0.000 0.000 Hungary 0.000 0.000 Iceland -0.009 -0.020 Ireland 0.001 0.001 Italy -0.003 -0.009 Latvia 0.005*** 0.005** Lithuania 0.000 0.000 Luxembourg 0.000 0.000 Malta 0.001 0.001 Netherlands 0.000 0.000 Norway -0.005 -0.013 Poland 0.000 0.000 Portugal 0.002 0.002 Romania 0.009*** 0.007*** Slovakia 0.002*** 0.002** Slovenia 0.001 0.001 Spain 0.001 -0.003 Sweden -0.005 -0.009 Switzerland 0.000 -0.001 United Kingdom 0.004 -0.001 Note: * significant at 10%; ** significant at 5%; *** significant at 1% 21 Table IV – Average marginal effect of Regulatory Quality by Country (country level analysis, 1995-2010 period) Average marginal Average marginal Effect$ Effect# Country �(∆��𝑷𝒓�𝒅) �(∆��𝑷𝒓�𝒅) [ ] [ ] �(𝑹𝒆𝒈) �(𝑹𝒆𝒈) Austria 0.05** 0.05** Belgium 0.047* 0.047* Bulgaria 0.111*** 0.104*** Cyprus 0.062*** 0.06** Czech Republic 0.082*** 0.078*** Denmark 0.05** 0.05** Estonia 0.092*** 0.087*** Finland 0.051** 0.051** France 0.051** 0.051** Germany 0.053** 0.052** Greece 0.055** 0.054** Hungary 0.087*** 0.082*** Iceland 0.051** 0.05** Ireland 0.046* 0.046* Italy 0.049** 0.049** Latvia 0.097*** 0.091*** Lithuania 0.094*** 0.089*** Luxembourg 0.042 0.043 Malta 0.07*** 0.067*** Netherlands 0.052** 0.052** Norway 0.044* 0.045* Poland 0.082*** 0.078*** Portugal 0.07*** 0.068*** Romania 0.101*** 0.095*** Slovakia 0.088*** 0.083*** Slovenia 0.07*** 0.068*** Spain 0.058** 0.057** Sweden 0.052** 0.051** Switzerland 0.041 0.042 United Kingdom 0.049** 0.049** Note: * significant at 10%; ** significant at 5%; *** significant at 1% $ Results based on Eq(1) with 𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆 variable; # Results based on Eq(1) with 𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆𝑬𝑼 variable. 22 Table V-List of NACE 1.1 Sectors Agriculture, hunting and related services Air transport Computer and related activities Construction Electricity, gas and water supply Extraction of crude petroleum and natural gas; service activities incidental to oil and gas extraction, excluding surveying Fishing Forestry, logging and related service activities Hotels and restaurants Land transport; transport via pipelines Manufacture of food products and beverages Manufacture of basic metals Manufacture of chemicals, chemical products and man-made fibres Manufacture of electrical machinery and apparatus n.e.c. Manufacture of fabricated metal products, except machinery and equipment Manufacture of radio, television and communication equipment and apparatus Manufacture of furniture; manufacturing n.e.c. Manufacture of leather and leather products Manufacture of machinery and equipment n.e.c. Manufacture of medical, precision and optical instruments, watches and clocks Manufacture of motor vehicles, trailers and semi-trailers Manufacture of office machinery and computers Manufacture of coke, refined petroleum products and nuclear fuel Manufacture of other transport equipment Manufacture of other non-metallic mineral products Manufacture of pulp, paper and paper products Manufacture of rubber and plastic products Manufacture of tobacco products Manufacture of textiles Manufacture of wearing apparel; dressing; dyeing of fur Manufacture of wood and wood products Mining of coal and lignite; extraction of peat Other mining and quarrying Other business activities Post and telecommunications Publishing, printing and reproduction of recorded media Research and development Real estate activities Recycling Renting of machinery and equipment without operator and of personal and household goods 23 Retail trade, except of motor vehicles and motorcycles; repair of personal and household goods Sale, maintenance and repair of motor vehicles and motorcycles; retail sale of automotive fuel Supporting and auxiliary transport activities; activities of travel agencies Water transport Wholesale trade and commission trade, except of motor vehicles and motorcycles Table VI – Summary Statistics of Regulatory Variables (all countries, 2003-2008 period). stats Mean sd p25 p75 𝑅𝑒𝑔 1.27 0.38 1.00 1.60 𝐴𝑙𝑙 _𝐷𝐵 71.03 10.16 63.98 77.96 𝑆𝑡𝑎𝑟𝑡 − 𝑢�_𝐷𝐵 82.57 8.06 76.51 87.54 𝐵𝑢𝑠 . 𝑂�𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠 _𝐷𝐵 78.29 8.87 73.37 83.77 𝐼𝑛𝑠𝑡 − 𝐸𝑛𝑣 _𝐷𝐵 69.52 10.27 64.52 75.80 24 Table VII – Values of PCA - DB indicators# 𝑨��_𝑫𝑩* 2003 2004 2005 2006 2007 2008 Denmark 86.31 86.16 87.48 88.54 90.73 90.28 Ireland 89.10 88.98 89.16 89.25 89.27 89.32 United Kingdom 89.14 89.16 89.07 89.07 89.02 87.65 Norway 84.43 83.90 84.33 83.82 83.39 82.40 Sweden 80.62 79.95 81.98 83.46 83.39 82.10 Finland 81.29 81.62 81.70 81.75 81.68 81.23 Belgium 77.91 79.48 79.49 79.62 80.35 80.67 Austria 77.16 77.12 77.18 77.48 78.01 77.03 Switzerland 76.23 76.29 76.45 76.57 77.63 76.66 Netherlands 74.44 74.48 75.12 75.62 76.30 76.03 Estonia 73.09 73.29 74.47 74.49 75.32 74.93 Germany 74.62 74.62 74.28 75.11 75.10 74.07 France 55.10 56.83 56.98 68.12 68.89 73.11 Lithuania 67.09 69.29 71.57 72.11 71.85 71.67 Latvia 67.83 67.85 69.33 71.61 71.86 71.52 Cyprus 68.46 68.46 68.46 68.46 68.46 67.81 Portugal 63.64 63.64 64.72 67.27 67.76 67.70 Hungary 65.44 65.45 65.50 65.83 66.70 67.61 Bulgaria 59.11 59.09 59.68 61.80 66.89 67.43 Slovakia 60.99 62.38 64.33 65.53 65.72 65.71 Luxembourg 65.40 65.40 65.40 65.40 65.41 64.85 Czech Republic 61.74 61.85 61.72 62.76 66.18 64.82 Poland 63.88 63.96 64.00 65.05 64.93 64.75 Spain 63.82 63.80 64.92 64.93 64.81 64.19 Italy 61.45 63.87 65.09 63.68 63.01 63.81 Romania 47.05 47.48 49.72 62.06 63.66 63.74 Slovenia 57.44 57.51 57.88 57.95 56.60 59.43 Greece 51.85 51.87 52.01 53.69 53.59 55.09 𝑺𝒕𝒂𝒓𝒕 − 𝒖𝒑_𝑫𝑩* 2003 2004 2005 2006 2007 2008 Norway 99.1 98.0 99.0 99.8 98.9 98.7 Finland 96.0 96.7 96.8 96.9 96.7 96.6 Sweden 95.2 93.8 94.2 94.3 94.2 94.3 Denmark 87.6 87.6 89.1 89.9 94.1 94.1 Netherlands 90.6 90.7 90.7 91.7 91.4 90.8 Ireland 90.2 90.2 90.3 90.5 90.4 90.6 United Kingdom 89.9 89.9 89.9 89.9 89.8 89.8 Belgium 81.4 84.8 84.9 85.2 86.9 87.0 Lithuania 81.5 86.4 86.5 87.4 87.2 87.0 25 Austria 86.8 86.7 86.8 86.9 86.7 86.5 Switzerland 85.7 85.9 86.3 86.4 86.4 86.4 Portugal 78.6 78.6 79.8 84.5 85.9 86.1 Estonia 81.6 82.1 83.7 83.8 85.9 85.6 Slovakia 72.4 75.5 79.7 81.8 81.4 84.3 Cyprus 83.5 83.5 83.5 83.5 83.5 83.5 Germany 83.1 83.1 82.4 82.5 82.5 82.3 Spain 79.5 79.5 82.6 82.7 82.6 81.7 Italy 79.2 84.3 84.4 81.8 80.0 81.7 Hungary 76.5 76.5 76.8 77.4 78.5 81.7 Latvia 77.1 77.2 76.9 78.9 78.9 79.0 Bulgaria 74.0 73.9 74.0 75.7 75.4 78.7 Romania 71.7 72.6 75.2 75.6 78.3 78.6 France 73.5 77.5 77.8 77.8 78.8 78.5 Slovenia 71.6 71.7 72.0 72.1 72.4 77.4 Luxembourg 76.1 76.1 76.1 76.1 76.1 76.1 Poland 72.2 72.3 72.4 72.7 73.2 73.3 Czech Republic 66.6 66.8 67.4 68.3 71.2 71.9 Greece 65.1 65.1 65.4 69.9 69.7 70.6 𝑩𝒖𝒔. 𝑶𝒑𝒆𝒓𝒂𝒕𝒊��𝒔_𝑫𝑩* 2003 2004 2005 2006 2007 2008 Denmark 96.84 96.84 97.62 97.68 97.66 96.88 Ireland 90.77 90.60 90.60 90.60 90.69 90.68 United Kingdom 91.07 91.07 91.07 91.07 91.08 88.76 Sweden 89.47 89.47 90.35 90.35 90.36 88.27 Switzerland 86.14 86.14 86.14 86.24 88.02 86.51 Norway 86.24 86.24 86.24 84.90 84.90 83.33 Austria 83.84 83.84 83.84 84.00 84.90 83.33 Estonia 82.97 82.97 83.85 83.76 83.78 83.29 Belgium 83.27 83.27 83.27 83.28 83.33 83.26 Finland 83.42 83.42 83.42 83.47 83.47 82.78 France 60.22 60.22 60.22 75.15 75.15 81.74 Germany 80.94 80.94 80.94 82.11 82.08 81.44 Netherlands 78.76 78.76 79.64 79.73 80.87 80.87 Cyprus 78.63 78.63 78.63 78.63 78.63 78.63 Czech Republic 77.73 77.73 76.07 77.14 79.65 77.79 Lithuania 76.38 76.38 77.17 77.27 77.03 76.88 Latvia 74.06 74.06 74.84 76.44 77.04 76.67 Italy 73.26 73.26 74.97 75.01 75.01 74.97 Luxembourg 75.91 75.91 75.91 75.91 75.92 74.87 Portugal 73.28 73.28 74.17 75.07 75.08 74.61 Poland 74.87 74.87 74.87 75.75 74.86 74.47 Hungary 73.45 73.45 73.45 73.47 74.28 73.75 26 Bulgaria 67.14 67.14 67.14 68.12 75.02 72.63 Spain 71.89 71.89 71.89 71.89 71.81 71.37 Slovakia 71.73 71.73 72.61 72.72 72.52 69.96 Greece 68.54 68.54 68.54 68.66 68.68 69.87 Romania 48.23 48.23 49.02 67.35 66.77 66.23 Slovenia 67.14 67.14 67.14 67.14 65.81 65.26 𝑰�𝒔𝒕 − 𝑬�𝒗_𝑫𝑩* 2003 2004 2005 2006 2007 2008 Ireland 91.50 91.50 92.22 92.22 92.22 92.22 United Kingdom 90.75 90.75 90.57 90.57 90.57 90.57 Latvia 77.02 77.02 81.37 83.43 83.43 83.43 Belgium 77.95 77.95 77.95 77.95 77.95 79.40 Romania 72.17 72.17 74.23 75.22 77.26 78.28 Norway 75.89 75.89 75.89 75.89 75.89 75.89 Finland 75.35 75.85 75.85 75.85 75.85 75.85 Germany 77.69 77.69 77.80 77.80 77.80 75.75 Austria 74.80 74.80 74.80 75.52 75.52 75.61 Denmark 74.01 73.29 73.45 75.49 75.49 75.49 France 67.61 67.61 67.61 71.70 73.74 73.74 Lithuania 66.59 66.59 72.76 72.76 72.76 72.76 Bulgaria 64.52 64.52 66.58 68.64 70.69 71.42 Netherlands 70.92 70.92 70.92 70.92 70.92 70.92 Estonia 70.63 70.63 70.63 70.63 70.63 70.63 Sweden 61.85 61.85 65.94 70.55 70.55 70.55 Slovakia 66.75 67.83 67.83 67.83 69.88 69.88 Czech Republic 67.57 67.57 70.53 70.53 70.53 68.49 Hungary 68.40 68.40 68.40 68.40 68.40 68.40 Poland 63.96 63.96 64.20 65.19 66.98 66.98 Switzerland 66.61 66.61 66.61 66.61 66.61 66.27 Luxembourg 65.54 65.54 65.54 65.54 65.54 65.54 Spain 62.63 62.63 62.63 62.63 62.63 62.63 Portugal 60.98 60.98 60.98 60.98 60.98 61.70 Slovenia 55.18 55.18 56.39 56.39 54.33 57.72 Cyprus 55.73 55.73 55.73 55.73 55.73 53.76 Greece 47.32 47.32 47.32 47.32 47.32 48.31 Italy 48.15 48.15 48.15 48.25 48.25 48.25 #countries are ordered by 2008 values *The All DB PCA index is a comprehensive index including all DB indicators in all topics. The PCA index of Business Start- up_DB encompasses the DB indicators for starting a business, registering property and closing a business. The PCA index of Business Operations_DB covers the topics of dealing with construction permits, paying taxes, trading across borders and employing workers. The PCA index of Institutional Environment_DB aggregates the indicators for protecting investors, getting credit and enforcing contracts Note: The PCA indicators were originally computed for all countries around the world (for which the DB data is available). The table displays only the values for the set of countries covered by the current analysis (EU countries plus Iceland, Norway and 27 Switzerland). For this reason, the 100 value is not always presented, which means that the country with the least complex regulation in the world is not among those included here. Table VIII– Average marginal effect of different regulatory measures (country/sector level analysis, 2003-08 period) Results based on Eq(2) Results based on Eq(2) with 𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆 variable with 𝑶𝒑𝒆�_𝒕𝒓𝒂𝒅𝒆𝑬𝑼 variable Variables coef coef �(∆��𝑷𝒓�𝒅) [ ] 0.125*** 0.113** �(𝑹𝒆𝒈) �(∆��𝑷𝒓�𝒅) [ ] 0.006*** 0.006*** �(𝑨��_𝑫𝑩) �(∆��𝑷𝒓�𝒅) [ ] 0.006** 0.006** �(𝑺𝒕𝒂𝒓𝒕−𝒖𝒑_𝑫𝑩) �(∆��𝑷𝒓�𝒅) [ ] 0.004*** 0.004*** �(𝑩𝒖𝒔.𝑶𝒑𝒆𝒓𝒂𝒕𝒊��𝒔_𝑫𝑩) �(∆��𝑷𝒓�𝒅) [ ] 0.006** 0.006** �(𝑰�𝒔𝒕−𝑬�𝒗_𝑫𝑩) Note: * significant at 10%; ** significant at 5%; *** significant at 1% 28