The determinants of the distance to TFP frontier among European firms
M. Dolores Añón Higón
Juan A. Máñez
María E. Rochina-Barrachina
Amparo Sanchis*
Juan A. Sanchis
Universitat de València and ERI-CES
(Preliminary version, please do not quote)
Abstract
In this paper we characterize TFP frontier firms at the industry level within the
European Union during the period 2003-2014, and explore the determinants of
the firms’ distance to the frontier. We find that larger, more capital-intensive, and
more labour skilled firms are closer to the productivity frontier. In contrast, older
firms are further away from the frontier. In addition, we obtain that a number of
economic and institutional factors, such as tertiary education, trade openness,
easiness in getting credit and governance quality, all positively affect the catching
up of laggards towards the productivity frontier. We also examine the moderating
effect of the Great Recession on these determinants and find differentiated
effects when analysing separately manufacturing, non-financial market services
and other production.
* Corresponding author: Facultat d’ Economia, Universitat de València, Avinguda
Tarongers, s/n, 46022 València; e-mail: [email protected]
Financial support under Framework contract ENTR/300/PP/2013/FC-WIFO is
gratefully acknowledged. The content of this paper does not reflect the official
opinion of the European Commission. Responsibility for the information and
views expressed in this paper lies entirely with the authors.
1
1. Introduction.
During the last decade, especially after the financial crisis starting in 2008, there
has been a renewed interest in understanding the factors driving productivity
differences across countries, industries and firms. A number of papers have shed
light on how cross-country differences in economic outcomes relate to differences
in the within-industry productivity dispersion across firms (Bartlesman et al., 2013;
Restuccia and Rogerson, 2008; Hsieh and Klenow, 2009). Some studies have
also addressed the factors explaining these productivity differentials, as well as
the determinants of their persistence (Syverson, 2011). However, little is known
about the drivers behind convergence of laggards to the best performing firms in
their industry (technological frontier firms). There is also little evidence on the role
of the Great Recession as a moderating factor affecting how firm and country
heterogeneity explains the productivity gap between frontier and non-frontier
firms and, thus, the within-industry productivity catching-up process.
The aim of this paper is fill this gap by providing fresh evidence on the factors
making firms to converge to the better performing firms in their industry at the
European level. In particular, we analyse the factors determining the distance-tofrontier of laggard firms. With this aim, we test how firms’ characteristics and
country’ economic and institutional features may influence a firm’s TFP relative
position to the frontier. We also take into account the role of the cycle as a
moderating factor in the impact of these factors and distinguish among firms in
manufacturing, non-financial market services and other production. Finally, we
aim also at providing guidance on how policy makers may further stimulate the
productivity growth of firms lagging behind the productivity frontier. Fostering the
productivity level of laggard firms within industries is an important policy issue,
especially in countries with more skewed industrial structures where more firms
linger far from technological excellence. In this context, it is crucial to understand
the factors determining the distance-to-frontier of laggard firms.
Using European firm-level data from AMADEUS for the period 2003 to 2014, we
identify the technological frontier firms by industry within the European Union, and
explore which firm and country-level characteristics determine the distance of
other firms in the same industry to the EU frontier firms. To analyse the
determinants of the within-industry distance to the TFP frontier we first provide a
2
description of the characteristics of the most productive firms, that is, the frontier
firms, in comparison with the non-frontier firms. We then analyse the
determinants of firms’ TFP gaps. To take into account the moderating effect of
the cycle, we analyse two different sub periods, namely, the period prior to the
last financial crisis, which we called pre-crisis years, 2003-2007, and the postcrisis years, 2008-2014.
This paper contributes to the scarce existing literature on the characteristics of
firms operating at the productivity frontier and on the determinants of the distance
to the frontier. Andrews et al. (2016) analyse productivity divergence from frontier
firms using a micro-level dataset of firms in OECD countries. Other related
papers using firm-level dada to analyse the catching up to the technological
frontier are Bartelsman et al. (2008) for a group of five European countries and
the US, Iacovone and Crespi (2010) for México, and Van der Wiel et al. (2008) for
the Netherlands. However, we are not aware of any existing study attempting to
capture the determinants of the catching up to the EU technological frontier, and
the moderating role of the Great Recession.
Our main findings may be summarized as follows. First, from a descriptive point
of view, our results indicate that relative to non-frontier firms, frontier firms are
larger, more capital intensive, have higher value added, earnings and profits, and
pay higher wages per employee. Regarding persistence of firms in the frontier
status, we observe that, generally, persistence is higher among frontier firms in
manufacturing as compared to frontier firms in non-financial market services.
Third, the regression analysis of the determinants of the TFP gap between
frontier and non-frontier firms within industries suggests that, in general, larger,
more capital-intensive and more labour skilled firms are closer to the frontier. In
contrast, older firms are further away from the frontier. At the country level, we
obtain that a number of economic and institutional factors, such as tertiary
education, trade openness, easiness in obtaining credit and governance quality,
all play a significant role in reducing the distance of laggards to the frontier. As
regards the role of the cycle in moderating the effect of these factors, we obtain
different patterns when analysing separately manufacturing, non-financial market
services and other production. In particular, we obtain that in manufacturing and
other production, the effect of these factors is more relevant during the post-crisis
3
period, whereas for firms in non-financial market services the effect of these
determinants seems to be stronger during the pre-crisis period.
The rest of the paper is organized as follows. Section 2 presents the data we use,
based on AMADEUS, and describes the measurement of firms’ productivity.
Section 3 provides a careful characterization of the best performing or EU frontier
firms versus non-frontier firms. Section 4 presents a regression analysis of the
determinants of TFP gaps, looking at firm specific and country level factors.
Finally, section 5 summarizes the main results and concludes.
2. Data and measurement of productivity.
The data we analyse are drawn from AMADEUS, a database providing firms’
balance sheet information, such as value added, assets, age, and number of
employees, for all EU countries. We use this information to estimate TFP at the
firm level in a cross-country setting.
Following Arnold et al. (2008), Gal (2013), and Andrews et al. (2015), we focus
on the set of firms with more than 20 employees as this helps to obtain a better
coverage and a more balanced sample. Arnold et al. (2008) point out that
AMADEUS does not have satisfactory coverage of firms with less than 20
employees; therefore, the resampling procedure targets the size-sector-country
distribution of the true population of firms with 20 employees and more. Andrews
et al. (2015) for ORBIS highlight a similar problem, and also exclude firms with
less than 20 employees.1
Further, to ensure that the computation of TFP is feasible for the largest possible
number of firms we follow Gal (2013), who suggests an imputation methodology,
improving the original coverage of AMADEUS. Our working sample is largely
coherent with that in Gal (2013) for the overlapping countries. Also following Gal
(2013), currency conversion based on PPPs and deflation are applied to ensure
comparability of productivity measures across countries and over time.
1
A key drawback of ORBIS and AMADEUS is that it is a selected sample of larger and more
productive firms, which tends to result in smaller and younger firms being under-represented in
some economies.
4
We construct a firm-level TFP using a Cobb-Douglas production function that is
estimated with the information contained in the AMADEUS database. 2 This
production function for firm i in sector p, year t, and country c may be defined as:
β
β
Lp
Y = f (Liptc , K iptc ,Wiptc ) = Liptc
K iptcKp exp(ω iptc )
! iptc
(1)
where Y is value added, L is labour, K is capital and ω is the TFP. We suppose
that capital is a dynamic input and evolves according to a given law of motion that
is not correlated with contemporary productivity shocks (i.e., is a state variable),
while labour and materials are inputs that can be freely adjusted when the
company is affected by a shock of productivity (i.e., both are variables inputs
without adjustment costs). We consider the following log version of the production
function (1):
y = β0p + βLp liptc + βKpk iptc + ω iptc + ηiptc
! iptc
(2)
where yiptc is the natural log of value added, liptc is the natural log of labour and
kiptc is the natural log of capital. As for the unobservables, ω iptc is the firm
productivity (not observed by the econometrician but observable or predictable by
firms) and ηiptc is a standard i.i.d. error term that is neither observed nor
predictable by the firm.
Under the assumption that capital is a state variable, whereas labour and
intermediate materials are inputs that can be adjusted whenever the firm faces a
productivity shock, Olley and Pakes (1996, hereafter OP) show how to obtain
consistent estimates of the production function coefficients using a semiparametric procedure (see also Levinsohn and Petrin, 2003, hereafter LP, for a
closely related estimation strategy).
However, here we follow Wooldridge (2009), who argues that both OP and LP’s
estimation methods can be reconsidered as consisting of two equations which
can be jointly estimated by GMM: the first equation tackles the problem of
2
From the original sample we perform a cleaning related to the variables we use to calculate TFP.
In particular, we drop observations with a negative value in capital, value added and labour. After
estimating TFP we drop observations with a TFP growth rate above and below the 99 and 1% of
the TFP growth distribution, respectively. Further, we also drop observations with a TFP in levels
above and below the 99.75% and 0.25%, respectively.
5
endogeneity of the non-dynamic inputs (that is, the variable factors); and, the
second equation deals with the issue of the law of motion of productivity.
We start considering first the problem of endogeneity of the non-dynamic inputs.
Correlation between labour and productivity complicates the estimation of
equation (2), because it makes the OLS estimator biased and the fixed-effects
and instrumental variables methods generally unreliable (Ackerberg et al., 2007).
Both OP and LP’s methods use a control function approach to solve this problem,
by using investment in capital and materials, respectively, to proxy for
“unobserved” firm productivity.
In particular, the OP method assumes that the demand for investment in capital,
i = i(k iptc ,ω iptc ) , is a function of firms’ capital and productivity. To circumvent the
iptc
problem of firms with zero investment in capital, the LP’s method uses the
demand for materials (intermediate inputs), miptc = m(k iptc ,ω iptc ) , instead, as a proxy
variable to recover ‘unobserved’ firm’s productivity. We follow this last approach.
As the demand of intermediate materials is assumed to be monotonic in
productivity, it can be inverted to generate the inverse demand function for
materials and, therefore, the productivity.
Wooldridge (2009) proposes to estimate jointly a system of two equations by
GMM using the appropriate instruments and moment conditions for each
equation. Ackerberg et al. (2006) showed that there exists an identification
problem in the first step estimation of the coefficients of variable inputs (affecting
the labour input) in OP and LP methods that rely on a two-step estimation
procedure, and derived a mixture of OP and LP’s approaches to solve the
problem. However, theirs is still a two-step estimation procedure. More recently,
Wooldridge (2009) has argued that both OP and LP’s estimation methods can be
reconsidered as consisting of two equations which can be jointly estimated by
GMM in a one-step procedure. This joint estimation strategy of equations has the
following advantages: i) it increases efficiency relatively to two-step procedures;
ii) it makes unnecessary bootstrapping for the calculus of standard errors; and, iii)
it solves the aforementioned identification problem.
After estimating the production function we obtain estimates of firms’ TFP using
the following expression:
6
ω = y iptc − β̂Lp liptc − β̂Kpk iptc
! iptc
(3)
where ω iptc is the estimate of the log TFP for firm i belonging to industry p in
period t and country c. We use time constant and sector-specific coefficients for
labour and capital that are common across all countries to facilitate both
international and over time comparability of the resulting productivity levels.
The data we analyse correspond to the period 2003 to 2014. The coverage of EU
member states is subject to data availability in AMADEUS. Thus, in the case of
Cyprus and Lithuania, there is no information in the sample. Luxembourg, Malta,
and Latvia are excluded from our sample due to insufficient observations. Thus,
our data cover 23 EU Member States. The industry classification is based on
NACE Rev. 2, limited to the market sectors and excluding agricultural, mining,
real estate and financial sectors.3 These sectors are more prone to issues in
accurately measuring output and they are more affected by lack of information.
As for the financial sector, since AMADEUS does not generally include banks, the
coverage would not be representative of this sector at the European level.
3. Descriptive analysis of EU frontier firms.
There is a renew interest in understanding the factors behind productivity growth
and in particular the drivers of productivity convergence among EU countries.
There are two main perspectives in the study of catching-up mechanisms to the
frontier: the macroeconomic and the microeconomic approach. In the former, the
unit of analysis is either the country or the region (Barro and Sala-i-Martín, 1992)
and the objective is to identify the frontier at the country or region level, and to
test whether productivity growth in other territories is related to the existing gap to
the frontier (see Sala-i-Martí, 1996; Griffith et al., 2004a; Acemoglu et al., 2006;
Kneller and Stevens, 2006; Aghion et al., 2008; Amable et al., 2010, among
others). Under this approach, it is assumed that all firms in a given region catchup and converge towards the frontier, disregarding heterogeneity across firms
and ignoring the fact that growth may be influenced by mechanisms of selection
and reallocation, as well as uncertainty (Jovanovic, 1982; Melitz, 2003).
3
See Table A.1 in the Appendix for a classification of the industries used.
7
The microeconomics approach, instead, aims at overcoming some of these
limitations, and in particular the heterogeneity across firms (Griffith et al., 2004b;
Acemoglu et al., 2007; Alvarez and Crespi, 2007; Bartelsman et al., 2008;
Bartelsman et al., 2015, Andrews et al., 2015; Ding et al., 2016). The unit of
analysis is the firm, and it is based on the identification of a best-practice frontier
firm (or group of firms) reflecting the most advanced technology within an industry
and country (or groups of countries). Although some micro studies based on a
single country are able to address the issue of firm-level heterogeneity, and they
are potentially helpful in understanding whether catching up differ across types of
firms, they may not be able to correctly identify the ‘true’ frontier, which in many
industries may depend upon firms operating in neighbouring countries.
In light of these considerations, micro data for all (potentially relevant) countries
within a given geographical area or development level is useful in identifying the
“global” frontier, which may differ from the national frontier firms. The scarce
literature using firm-level data in a multi-country analysis uses a small number of
countries to identify the global frontier firms. For example, Bartelsman et al.
(2008) focus on the United States and five European economies. Iacovone and
Crespi (2010) and Van der Wiel et al. (2008) analyse Mexico and The
Netherlands, respectively, and use the data from Bartelsman et al. (2008) to
identify the global frontier. However, a small sample of countries casts doubts
about the correct identification of the global frontier. An exception to these studies
is the work of Andrews et al. (2015), who analyse the global frontier considering a
sample of 23 OECD countries.
In this paper we rely on the microeconomics approach in order to derive the “EU
frontier firms” considering the EU as a well-defined area to be analysed.
Following Andrews et al. (2015), for measuring the productivity frontier we use an
absolute number of firms within each industry, calculated as the 5% of the
median number of firms (across years).4 Frontier firms are identified using the top
We calculate the median number of firms across years for each industry, and the 5% of this
median is the number of firms that we keep constant across years for each industry, and use this
number to calculate the frontier with the most productive firms.
4
8
5% globally most productive firms at the EU level (within each industry and
year).5
Once the EU frontier firms have been identified, we first characterize the EU
frontier firms, analyse its composition by country of origin and then, examine
firms’ persistence in the EU frontier.
3.1. A characterization of the EU frontier firms.
To analyse the EU frontier firms we describe cross-sectional differences in key
characteristics between EU frontier and non-frontier firms along a number of
measurable dimensions. This part of the analysis is purely descriptive. The main
dimensions considered include: TFP (measured as described in section 2), age,
employment, value added, capital intensity, earnings, 6 profits and wages per
employee. A test of difference in means over these dimensions determines the
extent to which EU frontier firms differ significantly from non-frontier firms.
Table 1 reports differences in average characteristics at aggregate level for the
EU frontier firms, as defined above, relative to non-frontier firms along a number
of firm characteristics over the period 2003 to 2014. In addition, this table
provides cross-sectional differences in average characteristics for the years 2006
(pre-crisis) and 2013 (post-crisis), respectively. Table 1 also shows these
differences over the total period by the three aggregate groups of industries:
manufacturing, non-financial market services and other production industries.7 In
all cases the differences in means between frontier and non-frontier firms are
based on the classification according to the TFP productivity measure.
[Insert Table 1 around here]
5
An alternative approach to define the productivity frontier is to take the top 100 most productive
firms within each industry and year. However, the tendency for the number of firms in AMADEUS
to expand over time suggests the adoption of a definition of frontiers based on a fixed number of
firms. As a robustness check we have also used the top 100 most productive firms and the results
are qualitatively similar.
6
Earnings are measured by EBITDA (earnings before interest, taxes, depreciation and
amortization).
7
The industry classification is based on NACE Rev. 2, limited to the market non-farm/agricultural
non-financial sectors (that is, excluding industries A, B, and O to P due to the lack of information).
See the Appendix for a classification of industries.
9
Over the period 2003 to 2014, firms at the EU frontier are on average four times
more productive than non-frontier firms, indicating that the TFP gap between
frontier and non-frontier-firms is considerable and therefore suggesting that the
potential for catching up growth in TFP remains huge.8 In addition, EU frontier
firms are on average older,9 larger, more capital intensive, have higher value
added, earnings and profits, and pay higher wages per employee than nonfrontier firms (this characterisations is also obtained by Andrews et al. 2016 for
the global frontier firms using a sample of 24 OECD countries). This features also
hold when we look ant manufacturing, non-financial services and other production
industries.
Comparing the reported statistics for the year 2006 (pre-crisis) and the year 2013
(post-crisis), we observe that there is a reduction in the average TFP of both
frontier and non-frontier firms. Also from 2006 to 2013 both frontier and nongexperience a reduction in value added, earnings and profits. Finally, frontier
firms increased in size, as measured by number of employees, and decreased its
wages per employee.
3.2. The EU frontier composition by country of origin.
Table 2 shows the percentage of firms in the EU frontier corresponding to each
country in the European Union in the two selected years (2006, 2013). The
information is provided separately for manufacturing, non-financial market
services and other production, but also for the total of these sectors.
By looking at the country composition of the EU frontier for all sectors, we
observe that those countries with higher participation are Germany, France, the
UK, Italy, Spain, Sweden and Belgium. We also observe that from 2006 to
8
Note that TFP is measured in logs, therefore, relative to non-frontier firms, frontier firms are on
average exp 1,4=4 times more productive.
9
Although the significant difference in age is lost over time.
10
2013,10 countries that increase their presence in the EU frontier are France, Italy,
Sweden and Belgium. The other three countries decrease their presence.
By sectors, the same countries appear as leaders in the frontier in manufacturing
and non-financial market services. The main differences are that the participation
of Germany and Belgium in non-financial market services, although relevant,
decreases over time. In other production, Germany, France, UK, Italy, Spain and
Sweden have a notable presence in the frontier, and from 2006 to 1013 Germany
and Sweden increase their presence and Belgium decreases its presence in the
frontier.
[Insert Table 2 around here]
3.3. Persistence of EU frontier and non-frontier firms.
To analyse firms’ persistence in the EU frontier, we report in Tables 3 to 6 the
percentage of firms remaining at the frontier, and the percentage lagging behind
the frontier, after one, two and five years, respectively. We report the results for
all sectors and for manufacturing, non-financial market services and other
production. The information in these tables corresponds to three different
samples. Sample 1 follows firms (frontier and non-frontier) from 2003 onwards.
Sample 2, follows firms (frontier and non-frontier) from 2006 onwards and
therefore minimizes the coverage problems of Sample 1 regarding several
countries before the year 2006.11 Finally, Sample 3 follows firms (frontier and
non-frontier) from 2006 onwards but restricting the sample to a balanced panel of
firms that remain in AMADEUS from 2006 to 2013. As compared to Samples 1
and 2, Sample 3 eliminates the effect of firms’ exit from AMADEUS in the
computation of persistence. The purpose of examining these three samples is to
provide a robustness check as regards the analysis of firms’ persistence.
[Insert Table 3 around here]
10
For several countries, there is a relevant drop in observations and coverage in AMADEUS from
the year 2013 to 2014. For this reason, and following Andrews et al. (2015), we present
information for 2013, although data corresponding to 2014 are included in the estimations.
11
In particular, Austria, Denmanrk and Ireland show a low number of observations in the earlier
years of the period analysed.
11
Table 3 reports firms’ persistence in the EU frontier. We observe that the
percentage of firms that manage to remain at the EU frontier after one year
ranges from 40.35% (in Sample 1) to 50.12% (in Sample 2), to 66.14% (in
Sample 3). In Sample 3, results indicate that 41.36% of firms remain at the EU
frontier after five years. In general, persistence is higher among frontier firms in
manufacturing than among firms in non-financial market services (this result is
similar in the three samples and it is also obtained by Andrews et al., 2015, using
the OECD-ORBIS productivity database from Gal, 2013; these authors analyse
manufacturing and non-financial market services, but do not consider other
production).12
[Insert Table 4 around here]
Table 4 reports persistence in the EU non-frontier. The percentage of firms that
remain at the EU non-frontier after one year is 97.07% (according to Sample 3),
and after five years is 80.94%. By sectors and regarding non-frontier firms, in
general, the intermediate level of persistence is for firms in other production
(independently of the sample). For the balanced sample (Sample 3) the higher
persistence in this status is for firms in non-financial services. However, for
Samples 1 and 2 it is for firms in manufacturing.
In Tables 5 and 6 we analyse persistence in the frontier and in the non-frontier,
respectively, at a more disaggregated industry level and using Sample 3. Results
in Table 5 indicate that the sectors that show high persistence in the frontier
status after five years in manufacturing industries are manufacturing of textiles,
wearing
apparel,
leather;
manufacturing
of
wood,
paper
and
printing;
manufacturing of coke and refined petroleum products; manufacturing of
chemicals and chemical products; manufacturing of pharmaceutical products; and
manufacturing of electrical equipment. In the case of services, a high rate of
persistence is obtained for transport and storage; postal and courier activities;
publishing, motion picture, video and television; and telecommunications. As
12
Andrews et al. (2015), using a Solow-residual based total factor productivity measure, find that
around half of the firms manage to remain at the global frontier from one year to the next, and
after five years, less than 20% of firms are still there. These results are quite similar to the ones
we obtain for persistence at the EU in our Samples 1 and 2. In Sample 3, persistence is higher
due to the fact that balancing the sample we avoid the effects of attrition in the AMADEUS
database when calculating the percentage of firms staying in the frontier after several years.
12
regards other production, higher persistence is observed for utilities and lower
persistence for construction (very low persistence after five years).
[Insert Table 5 around here]
[Insert Table 6 around here]
Results in Table 6 regarding persistence in the non-frontier status, indicate that,
in manufacturing, the industries with high persistence are: manufacturing of food,
beverages and tobacco; manufacturing of chemicals and chemical products;
manufacturing of rubber, plastic and non-metallic products; manufacturing of
computer, electronic and optical; manufacturing of electrical equipment;
manufacturing of machinery and equipment; and manufacturing of furniture,
jewellery, and musical products. In non financial market services, higher
persistence in the non-frontier status is found for wholesale trade, except of motor
vehicles; computer programming, consultancy; and professional, scientific and
technical activities. Finally, in other production this applies for electricity, gas,
steam and air conditioning supply.
Another proxy for persistence in the frontier (non-frontier) status is the average
number of years a firm remains at the frontier (away from the frontier). We have
calculated these mean values for frontier (non-frontier) firms during the period
2003-2014, using only information on firms belonging to the balanced sample
from 2006 to 2013. We find that those firms at the frontier (for at least one year)
remain there for around 30% of the years analysed, and the mean number of
years they remain there is around 3. Furthermore, we find that non-frontier firms
(for at least one year) remain at the non-frontier status around 98% of the years
analysed, and the mean number of years they remain there is around 10.5.
4. Determinants of the distance to the frontier.
We now analyse the factors determining the distance-to-frontier of laggard firms.
With this aim, we test how firms’ characteristics and country’ economic and
institutional features may influence a firm’s TFP relative position to the frontier. As
stated earlier, frontier firms are identified using the top 5% globally most
13
productive firms at the EU level (within each industry and year).13 We estimate a
model with the distance to the frontier (or technology gap) on the left hand side
and firms’ and country’ characteristics and other controls (such as industry,
country and year dummies) on the right hand side, as follows:
ln
TFPicpt
TFPptF
= α + β 'Z icpt + δ ' X ct + γ p + γ c + γ t + ε icpt
(4)
where subscript i denotes firm, subscript c country, subscript p a particular
industry and t the time period. The distance to the frontier of a particular laggard
firm is calculated as the log of the ratio of its own TFP over the average TFP of
the frontier firms belonging to the same industry and for each particular year.
Firms’ characteristics are captured by Zicjt (including size –measured by the
number of employees–, age, capital and wage per employee –as a measure for
skilled labour).14 Xct is a vector of variables capturing economic and institutional
characteristics at the country-year level. Among these variables we include a
measure of human capital at the country level, which is the percentage of 15-64
years old population with Tertiary Education, drawn from Eurostat. We expect this
variable to be related to the ability to absorb ideas and knowledge from the
technological frontier and, then, to have a positive impact on the technological
catching up of laggards. We also include a measure of Trade Openness
(calculated as the ratio of the sum of imports plus exports over GDP) from the
World Development Indicators. According to Melitz (2003), exposure to trade is
an important driver contributing to a better resource allocation, thereby affecting
within-firm productivity dispersion. It has been documented that more exposure to
trade implies selection of the most productive firms into the export markets and
market share reallocation by the exit of less efficient firms (Amiti and Konings,
2007, Fernandes, 2007).
13
Alternatively, we also used the definition of frontiers as the 100 top most productive firms for EU
industry-year frontiers, which provides very similar results. These results are not presented in the
paper but can be obtained from the authors upon request.
14
We use wages as a proxy for skills at the firm level. Although we acknowledge this is far from
perfect, this is the best available measure in the AMADEUS database for this purpose. We are
also aware of papers that use information on wages to control for differences in the quality of the
workforce (Gopinath et al., 2015).
14
In addition, from the World Bank's dataset Doing Business we use an indicator
capturing the distance to frontier (DTF) between a particular country’s
performance and the world’ best practice regarding the ease in getting credit
(DTF getting credit). This indicator tries to capture two types of related issues: the
strength of credit reporting systems and the effectiveness of collateral and
bankruptcy laws in facilitating lending.15 By construction, higher values of this
indicator implies greater ease in conducting business, thereby we expect this
variable to positively affect TFP growth.
Finally, from the Worldwide Governance Indicators we use different indicators in
order to obtain a summary measure of the Governance Quality: Government
Effectiveness, Regulatory Quality, Rule of Law, and Control of Corruption.
Government effectiveness captures perceptions of the quality of public services,
the quality of the civil service and the degree of its independence from political
pressures, the quality of policy formulation and implementation, and the credibility
of the government's commitment to such policies. Regulatory Quality reflects the
perceptions of the ability of the government to formulate and implement sound
policies and regulations that allow and promote private sector development. The
indicator Rule of Law captures the perceptions of the extent to which agents have
confidence in and abide by the rules of society, and in particular the quality of
contract enforcement, property rights, the police, and the courts, as well as the
likelihood of crime and violence. Finally, Control of Corruption captures
perceptions of the extent to which public power is exercised for private gain,
including both minor and grand forms of corruption, as well as "capture" of the
state by elites and private interests. Given the high correlation among these four
variables, we apply factor analysis to obtain a synthetic measure, Governance
Quality, which is a weighted sum of the original indicators.
The estimation results obtained by linear regression are presented in Table 7. All
standard errors are calculated as robust standard errors clustered by firm. All
regressions are run only with non-frontier firm distances. Results were similar
when also including frontier firms in the regressions, in which case the
15
In particular, this indicator measures two types of institutions and systems that may facilitate
access to finance and improve its allocation: (i) Legal rights of borrowers and lenders in secured
transactions and bankruptcy laws. (ii) Coverage, scope and quality of credit information available
through credit registries and bureaus. 15
corresponding log value for the dependent variable was set to zero. Coefficient
estimates for the regressors at the firm-level have the interpretation of elasticities
since both the dependent variable and firm-level regressors are in log form.
[Insert Table 7 around here]
In Column (1) of Table 7 we estimate the regression in expression (4) above.
Regarding firms’ characteristics, we find that larger and more skilled-labour firms
decrease the distance to the frontier. For instance, a one per cent increase in size
(as measured by the number of employees) reduces the distance to the frontier
by 0.18 per cent, and a one per cent increase in capital intensity reduces the
distance to the frontier by 0.026 per cent. In addition, older firms increase the
distance to the frontier.16 As regards to country-level economic and institutional
factors, we find that they play an important role as determinants of firms’ distance
to the TFP frontier. In particular, we obtain that higher tertiary education, trade
openness, easiness in getting credit, and quality of governance significantly affect
the technological catching up of laggards to the frontier. These results are in line
with existing literature. For instance, Iacovone and Crespi (2010) provide
evidence on the role of trade openness in the catching up of laggards towards the
frontier. Arnold et al. (2001) document how product market regulations that
restrain competitive pressure negatively affect firms’ productivity, implying that a
weak market and institutional framework may impose a burden on productivity
growth of firms.
In Column (2) we include the same regressors and controls than in Column (1),
but we also interact the regressors with a dummy variable taking value 1 for the
recession years (2008-2014). Hence, the estimates for non-interacted regressors
in this column correspond to the reference category, pre-crisis years (2003-2007).
In general, we obtain the same patterns than in Column (1) for pre-crisis years.
For post-crisis years, results indicate that larger, more capital intensive and more
labor skilled firms decrease the distance to the frontier (but in a weaker
magnitude to that in the pre-crisis period). Also older firms increase the distance
to the frontier (but in a weaker magnitude to that in the pre-crisis period). In
16
This result is in line with Conway et al. (2015) for New Zealand, who estimate a logistic
regression of frontier versus laggard firms within industries and find that frontier firms are more
likely to be younger.
16
addition, regarding country-level economic and institutional factors, we find that
they also play an important role in reducing firms’ distance to the TFP frontier
during the post-crisis period, although their effect is weaker for tertiary education
trade openness and easiness in getting credit, and stronger for governance
quality.
Finally, in Column (3), we consider the importance of distinguishing among
sectors, taking also into account the role of the cycle as a moderating factor of
the impact of the different determinants. 17 We interact all the regressors of
Column (1) both with dummies of pre-crisis and post-crisis years, and also with
dummies corresponding to three industrial sectors: manufacturing, non-financial
market services, and other production. Hence, in this final column, coefficients
have a direct interpretation and do not need to be interpreted with regards to
reference categories. Starting with firms’ characteristics, we find that size is an
important factor in reducing firms’ distance to the frontier in all sectors, and the
role of firms’ size in both periods is significantly different, being more important
during the recession years. As regards to firms’ age, we obtain that older firms
increase their distance to the frontier in all sectors, and the role of firms’ age in
both periods is statistically different, being more important during the recession
years for manufacturing and other production, and more important for the precrisis period in services. In addition, we find that capital intensity is an important
factor in reducing firms’ distance to the frontier in all sectors, and its impact in
both periods is significantly different, being more important during the pre-crisis
years in all sectors. Finally, our results show that firms with higher wages per
employee are significantly closer to the frontier. The different effect in pre-crisis
and post-crisis years is also statistically significant for the three sector groups,
being more relevant during the pre-crisis years in all sectors.
Regarding economic and institutional characteristics at the country level, our
findings suggest that they play a relevant role in reducing firms’ productivity gaps.
Tertiary education reduces firms’ distance to the frontier in all sectors and in both
periods, being more relevant during the pre-crisis period. Trade openness is also
17 To
properly assess the moderating role of the cycle, we perform tests on the equality of
coefficients for each variable in the two dummy periods, pre-crisis and post-crisis, rejecting the
null that they are equal at the one per cent level of significance. 17
an important factor in the productivity catching-up in all sectors and in its
relevance is similar in both periods. The easiness in getting credit has also a
positive effect in reducing the distance of laggards to the frontier, especially in
non-financial services during the recession years. Finally, we obtain mixed results
for the measure capturing the quality of governance, which increases firms’
distance to the frontier in manufacturing and non-financial services during the
pre-crisis years, but decreases firms’ distance to the frontier in all sectors during
the recession years, being the effect significantly different in both periods.
In summary, our findings suggest that firms’ size, capital intensity and human
capital are important drivers in explaining the catching up process towards the
frontier, whereas age seem to be negatively related to the convergence to the
frontier. At the country level, trade openness, tertiary education, the easiness of
getting credit and governance quality all play a positive role in reducing firms’
distance to frontier. In addition, our results indicate that it is important to consider
the heterogeneity among industries when analysing the determinants of the
distance to the TFP frontier, and also the moderating role of the business cycle.
In the case of manufacturing firms and firms in other production, we find that, in
general, firms’ determinants of the distance to the frontier are particularly relevant
during the post-crisis period (except for capital intensity, that seems to be more
important during the pre-crisis years). Regarding firms in non-financial market
services, the role of the cycle is more mixed: age and capital are more relevant
during the pre-crisis, but age and wages per employee seem to be stronger
during the recession period. Finally, regarding country-level factors and the role
of the cycle by sectors, the results on their impact are more mixed: both in
manufacturing and in non-financial services, tertiary education is more relevant in
pre-crisis and governance quality during recession; as for trade openness, their
relevance is similar in both periods in manufacturing and non-financial services,
and, finally, the easiness of getting credit seems to be more relevant for nonfinancial services during the recession.
5. Summary of results and concluding remarks
18
In this paper we analyse the determinants of within-industry distance to the TFP
frontier across EU economies. We first provide a description of the characteristics
of the best performing firms in terms of productivity, that is, the frontier firms, in
comparison with the non-frontier firms. We further analyse the determinants of
firms’ TFP gap between frontier and non-frontier firms, including factors at the
firm and country levels.
The analysis of the determinants of the TFP gap between frontier and non-frontier
firms within industries shows that, in general, larger, more capital intensive and
more labour skilled firms are closer to the frontier. In contrast, older firms are
further away from the frontier. At the country level, trade openness, tertiary
education, the easiness of getting credit and governance quality all play a positive
role in reducing firms’ distance to frontier. In addition, our results indicate that it is
important to consider the heterogeneity among industries when analysing the
determinants of the distance to the TFP frontier, and also the moderating role of
the business cycle.
Although is very difficult to provide a well-targeted policy advice, our findings
suggest a number of implications. First, given the characteristics of firms at the
technological frontier, it seems advisable to pursue active policies promoting size,
labour quality and capital intensity at the firm level. Macroeconomic policies
towards more educated workforce and trade openness, by helping at absorbing
advanced technologies and a better reallocation of resources, may also serve to
enhance the technological catching up of laggards towards the frontier. Our result
on the importance of the easiness in getting finance also points towards the need
to implement policies that stimulate a more competitive banking system, with a
reduction in informational asymmetries between lenders and borrowers
(eliminating borrowing constraints). Finally, our findings also suggest that
improving the quality of governance seems to be an important institutional factor
that policy makers should consider.
References
19
Acemoglu, D., P. Aghion and F. Zilibotti (2006), “Distance to Frontier,
Selection and Economic Growth”, Journal of the European Economic Association,
4(1), 37–74.
Acemoglu, D., P. Aghion, C. Lelarge, J. Van Reenen and F. Zilibotti
(2007), “Technology, Information and the Decentralization of the Firm”, Quarterly
Journal of Economics, 122(4), 1759–1799.
Ackerberg, D., Benkard, C. L., Berry, S., & Pakes, A. (2007). Econometric
tools for analyzing market outcomes. Handbook of econometrics, 6, 4171-4276.
Aghion, P., R. Burgess, S. Redding and F. Zilibotti, (2008), “The Unequal
effects of Liberalization: Evidence from Dismantling the License Raj in India”,
American Economic Review, 98(4), 1397–1412.
Alvarez, R. and G. Crespi, (2007), “Multinational Firms and Productivity
Catching-up: the Case of Chilean Manufacturing”, International Journal of
Technological Learning, Innovation and Development, 1(2), 136-152.
Altomonte, C., and G.I.P. Ottaviano, (2011), “The Role of International
Production Sharing in EU Productivity and Competitiveness”, EIB Papers,
Volume 16, No. 1.
Amable, B., L. Demmou and I. Ledezma, (2010), “Product Market
Regulation, Innovation, and Distance to Frontier”, Industrial and Corporate
Change, 19(1), 117–159.
Andrews, D., C. Criscuolo and P.N. Gal (2015), “Frontier Firms,
Technology Diffusion and Public Policy: Micro Evidence from OECD Countries”,
OECD Productivity Working Papers 2, OECD Publishing.
Andrews, D., C. Criscuolo and P.N. Gal (2016), “The global productivity
slowdown, technology divergence and public policy: a firm-level perspective”,
Hutchins Center Working Papers 24, Brookings.
Arnold, J.M., G. Nicoletti and S. Scarpetta (2008), “Regulation, Allocative
Efficiency and Productivity in OECD Countries: Industry and Firm-Level
Evidence,” OECD Economics Department Working Papers 616, OECD
Publishing.
Barro, R. J. and X. Sala-i-Marti (1992), “Convergence”, Journal of Political
Economy, 100(2), 223–251.
Bartelsman, E., S. Dobbelaere and B. Peters (2015), “Allocation of Human
Capital and Innovation at the Frontier: Firm-level Evidence on Germany and the
Netherlands”, Industrial and Corporate Change, 24(5), 875-949.
Bartelsman, E., J. Haskel and R. Martin (2008), “Distance to Which
Frontier? Evidence on Productivity Convergence from International Firm-level
Data”, CEPR Discussion Papers, No.7032.
Bartelsman, E.J., Haltiwanger, J. and S.Scarpetta (2013a), “Cross Country
Differences in Productivity: The Role of Allocation and Selection”, American
Economic Review, Vol. 103, No. 1, Feb 2013, pp. 305-34.
Conway, P., Meehan, L. Zheng, G. (2015), “Do New Zealand Firms Catch
up to the Domestic Productivity Frontier?”, New Zealand Productivity Commission
Working Paper 2015/3.
Ding, S., Sun, P. and W. Jiang (2016), “The Effect of Import Competition
on Firm Productivity and Innovation: Does the Distance to Technology Frontier
Matter?”, Oxford Bulletin of Economics and Statistics, 78, 197–227.
Gal, P.N. (2013), “Measuring Total Factor Productivity at the Firm Level
using OECD-ORBIS, Economics Department Working Papers Nº 1049, OECD.
20
Gopinath, G., Kalemli-Ozcan, S., Karabarbounis, L., and C. VillegasSanchez (2015), “Capital Allocation and Productivity in Southern Europe”, NBER
Working Paper No. 21453.
Griffith, R., Redding, S., and Van Reenen, J. (2004a), “Mapping the Two
Faces of R&D: Productivity Growth in a Panel of OECD Industries”, Review of
Economics and Statistics, 86(4), 883-895.
Griffith, R., Redding, S. and H. Simpson (2004b), “Foreign Ownership and
Productivity: New Evidence from the Service Sector and the R&D lab”, Oxford
Review of Economic Policy, 20(3), 533-564.
Hsieh, C. and P. Klenow, (2009), “Misallocation and Manufacturing TFP in
China and India", Quarterly Journal of Economics, 124(4): 1403-1448, November.
Iacovone, L. and G. Crespi (2010), “Catching-Up with the Technological
Frontier: Micro-level Evidence on Growth and Convergence”, Industrial and
Corporate Change, 19(6), 2073-2096.
Jovanovic, B. (1982), “Selection and the Evolution of Industry”,
Econometrica, 50(3), 649-670.
Kneller, R. and P. A. Stevens, (2006), “Frontier Technology and Absorptive
Capacity: Evidence from OECD Manufacturing Industries”, Oxford Bulletin of
Economics and Statistics, 68(1),1–21.
Levinsohn, J. and A. Petrin (2003),”Estimating Production Functions Using
Inputs to Control for Unobservables”, Review of Economic Studies 70(2), 317–
342.
Melitz, M.J. (2003), “The Impact of Trade on Intra-Industry Reallocations
and Aggregate Industry Productivity”, Econometrica, 71(6), 2003, pp.1695-1725.
Olley, S. and A. Pakes (1996), “The Dynamics of Productivity in the
Telecommunications Equipment Industry”, Econometrica, 64, 1263-1298.
Restuccia, D. and R. Rogerson (2008), “Policy Distortions and Aggregate
Productivity with Heterogeneous Establishments”, Review of Economic
Dynamics, 11(4), 707-720.
Sala-i-Marti, X. (1996), “Regional Cohesion: Evidence and Theories of
Regional Growth and Convergence”, European Economic Review, 40, 13251352.
Syverson, C. (2011), “What Determines Productivity?”, Journal of
Economic Literature, 49:2, 326-365.
Van der Wiel, H., H. Creusen, G. van Leeuwen and E. van der Pijll (2008),
“Cross your Border and Look Around”, DEGIT Conference Papers c013_005,
DEGIT, Dynamics, Economic Growth, and International Trade.
Wooldridge, J. M. (2009), “On Estimating Firm-level Production Functions
Using Proxy Variables to Control for Unobservables,” Economics Letters, Vol.
104, No. 3, pp. 112–114.
21
Appendix.
Table A.1. Industry classification
Manufacturing
Manuf. of food, beverages and tobacco
Manuf. of textiles, wearing apparel, leather
Manuf. of wood, paper, printing
Manuf. of coke and refined petroleum prod.
Manuf. of chemicals and chemical prod.
Manuf. of pharmaceutical products
Manuf. of rubber, plastic and non-metallic
Manuf. of basic and fabricated metal prod.
Manuf. of computer, electronic and optical
Manuf. of electrical equipment
Manuf. of machinery and equipment n.e.c.
Manuf. of motor vehicles, trailers
Manuf. of furniture; jewellery, musical prod.
Non financial market services
Wholesale and retail trade of motor vehicles
Wholesale trade, except of motor vehicles
Retail trade, except of motor vehicles
Transport and storage
Postal and courier activities
Accommodation and food service activities
Publishing, motion picture, video, television
Telecommunications
Computer programming, consultancy
Professional, scientific and technic. activities
Other production
Electricity, gas, steam and air cond. supply
Water suppl.; sewerage, waste management
Construction
22
Table 1. Means and differences in means of firm characteristics: EU frontier firms vs. non-frontier firms.
TFP
NF
F
Age
NF
F
Employment
NF
F
Value Added
NF
F
Capital
intensity
NF
F
Earnings
NF
F
Profits
NF
F
Wage p.e.
NF
F
Average 2003-2014
4.4 5.8***
24.1 31.9 ***
107.5 857.3 ***
4422
70157 ***
43.7 122.6 ***
1230
29023 ***
338
14149 ***
30.9 57.6 ***
Pre-crisis: 2006
4.4 5.8***
23.4 30.5 ***
109.7 789.4 ***
4558
66895 ***
39.4 120.8 ***
1265
28629 ***
381
14366 ***
31.1 65.2 ***
Post-crisis: 2013
4.3 5.7***
26.6 34.9 ***
105.2 855.8 ***
4383
67638 ***
49.4 130.7 ***
1141
27258 ***
307
13431 ***
31.8 59.7 ***
Manufacturing
4.3 5.6***
25.7 36.2 ***
107.0 662.6 ***
4639
66638 ***
43.6 82.8 ***
1402
27160 ***
376
13232 ***
30.2 54.5 ***
Non financial market services
4.5 5.5***
22.9 28.5 ***
89.9 499.2 ***
4463
66950 ***
62.8 253.8 ***
1490
32965 ***
420
15992 ***
33.5 58.1 ***
Other production industries
4.5 6.0***
22.9 28.9 ***
113.1 1134.4 ***
4210
74267 ***
38.5 123.1 ***
997
29652 ***
279
14482 ***
30.7 60.5 ***
By sectors:
Average 2003-2014
Note: EU frontier corresponds to the average of the top 5% EU most productive firms in each year and industry. F is the average of frontier firms; NF is the
average of non-frontier (all other) firms. TFP is measured in logs, employment by the number of employees. Value added, capital, earnings, profits and wages
measured in thousands of euros. Earnings are measured by EBITDA (earnings before interest, taxes, depreciation and amortization). Test in mean
differences with *** significant at 1%, ** significant at 5%, and * significant at 10%.
Table 2. Percentage of firms in the EU frontier by country origin in selected years.
Total
Country
Manufacturing
Nonfinancial
Market
Services
Other
Production
2006
2013
2006
2013
2006
2013
2006
2013
Austria
0.0
0.5
0.0
0.6
0.0
0.4
0.0
0.0
Belgium
5.6
6.5
5.5
8.5
6.7
5.3
2.0
1.3
Bulgaria
0.2
0.5
0.2
0.2
0.3
0.7
0.0
1.0
Croatia
0.2
0.3
0.2
0.2
0.1
0.4
0.3
0.3
C Republic
2.4
1.4
1.4
1.2
2.5
1.4
6.3
2.3
Denmark
0.0
2.7
0.0
1.6
0.0
4.6
0.0
1.3
Estonia
0.4
0.4
0.1
0.2
0.8
0.8
0.3
0.0
Finland
1.2
1.2
0.9
1.2
1.9
1.4
0.3
0.3
France
16.2
17.0
13.7
13.8
19.3
21.9
16.3
14.3
Germany
18.6
12.7
25.0
17.3
12.3
6.5
11.7
13.3
Greece
0.9
0.6
1.0
0.7
0.9
0.6
0.3
0.0
Hungary
0.3
1.2
0.6
1.1
0.1
1.1
0.0
1.7
Ireland
0.2
0.2
0.0
0.1
0.4
0.3
0.0
0.3
Italy
9.8
13.0
14.7
19.4
4.0
5.8
8.0
9.3
Netherlands
1.6
1.4
2.0
2.0
1.3
0.7
1.0
1.0
Poland
1.6
0.5
0.5
0.2
2.4
0.3
3.7
2.0
Portugal
2.0
1.7
1.2
1.7
2.8
1.7
3.0
1.7
Romania
0.8
1.4
0.8
0.8
0.5
1.8
2.0
2.7
Slovakia
3.3
1.0
2.6
0.6
4.9
1.5
0.7
1.0
Slovenia
0.0
0.1
0.0
0.1
0.0
0.2
0.0
0.0
Spain
13.5
12.0
11.7
10.8
12.6
11.3
24.0
19.0
Sweden
5.3
9.0
3.8
5.8
7.5
12.7
4.3
10.3
UK
16.0
15.0
14.1
11.9
18.7
18.6
15.7
16.7
Notes: EU frontier firms are the top 5% most productive firms in each year and industry
at the EU level. Denmark has no coverage in AMADEUS for the year 2006 in
the sample used for TFP measurement. Cyprus and Lithuania are not covered
by AMADEUS for the period analysed. Latvia, Luxembourg and Malta have
not been considered because of insufficient observations in the database.
Table 3. Persistence in the EU frontier.
Industry
Total
Manufacturing
Services
Other Production
Percentage of firms staying in the frontier after several years
Sample 1
Sample 2
Sample 3
Number of years
Number of years
Number of years
1
2
5
1
2
5
1
2
5
40.35%
42.31%
36.40%
45.00%
23.12%
24.00%
20.70%
27.33%
9.31%
10.46%
7.60%
10.00%
50.12%
52.77%
46.70%
50.00%
34.42%
37.31%
30.90%
33.67%
17.38%
20.23%
13.60%
17.67%
66.14%
65.70%
66.10%
68.89%
54.32%
55.23%
50.00%
60.00%
41.36%
41.88%
35.59%
53.33%
Notes: EU frontier firms are the top 5% most productive firms in each year and industry at the EU level. Sample 1 follows frontier
firms from 2003 onwards. Sample 2 follows frontier firms from 2006 onwards. Finally, Sample 3 follows frontier firms from 2006
onwards but restricting the sample to a balanced panel of firms that stay in AMADEUS database from 2006 to 2013.
Table 4. Persistence in the non-EU frontier.
Industry
Total
Manufacturing
Services
Other Production
Percentage of firms staying away from the frontier after several years
Sample 1
Sample 2
Sample 3
Number of years
Number of years
Number of years
1
2
5
1
2
5
1
2
5
83.53%
84.17%
82.61%
85.07%
64.76%
65.86%
64.02%
64.59%
49.04%
51.90%
47.08%
48.76%
88.97%
90.95%
87.48%
88.89%
76.65%
78.62%
75.08%
76.84%
60.12%
63.45%
58.08%
58.40%
97.07%
97.14%
96.93%
97.35%
89.95%
88.98%
90.70%
90.09%
80.94%
79.38%
82.39%
80.33%
Notes: Non-EU Frontier are all firms below the top 5% most productive firms in each year and industry at the EU level. Sample 1
follows non-frontier firms from 2003 onwards. Sample 2 follows non-frontier firms from 2006 onwards. Finally, Sample 3 follows nonfrontier firms from 2006 onwards but restricting the sample to a balanced panel of firms that stay in AMADEUS database from 2006
to 2013.
25
Table 5. Persistence in the EU Frontier (balanced sample).
Percentage of firms staying in the frontier after several years
Industry
Total
Manufacturing
Manuf. of food, beverages and tobacco
Manuf. of textiles, wearing apparel, leather
Manuf. of wood, paper, printing
Manuf. of coke and refined petroleum prod.
Manuf. of chemicals and chemical prod.
Manuf. of pharmaceutical products
Manuf. of rubber, plastic and non-metallic
Manuf. of basic and fabricated metal prod.
Manuf. of computer, electronic and optical
Manuf. of electrical equipment
Manuf. of machinery and equipment n.e.c.
Manuf. of motor vehicles, trailers
Manuf. of furniture; jewellery, musical prod.
Non financial market services
Wholesale and retail trade of motor vehicles
Wholesale trade, except of motor vehicles
Retail trade, except of motor vehicles
Transport and storage
Postal and courier activities
Accommodation and food service activities
Publishing, motion picture, video, television
Telecommunications
Computer programming, consultancy
Professional, scientific and technic. activities
Other production
Electricity, gas, steam and air cond. supply
Water suppl.; sewerage, waste management
Construction
Number of years
1
2
5
66.14% 54.32% 41.36%
68.42%
57.89%
52.94%
92.86%
71.43%
72.41%
50.00%
48.00%
80.00%
78.57%
50.00%
35.71%
61.11%
52.63%
47.37%
52.94%
89.29%
64.29%
62.07%
55.00%
36.00%
45.00%
64.29%
50.00%
21.43%
44.44%
31.58%
47.37%
64.71%
60.71%
46.43%
62.07%
20.00%
36.00%
25.00%
42.86%
25.00%
14.29%
38.89%
93.33%
50.00%
43.75%
78.57%
81.25%
61.54%
58.33%
71.43%
42.86%
33.33%
60.00%
37.50%
25.00%
50.00%
75.00%
53.85%
50.00%
57.14%
42.86%
0.00%
20.00%
25.00%
25.00%
50.00%
43.75%
30.77%
58.33%
50.00%
14.29%
0.00%
72.41%
72.73%
40.00%
68.97%
63.64%
0.00%
62.07%
54.55%
0.00%
Notes: EU frontier firms are the top 5% most productive firms in each year and industry at the EU level.
Balanced sample corresponds to Sample 3, which follows frontier firms from 2006 onwards but restricting
the sample to a balanced panel of firms staying in AMADEUS database from 2006 to 2013.
26
Table 6: Persistence in the non-EU Frontier (balanced sample).
Percentage of firms staying away from the frontier after several years
Industry
Total
Manufacturing
Manuf. of food, beverages and tobacco
Manuf. of textiles, wearing apparel, leather
Manuf. of wood, paper, printing
Manuf. of coke and refined petroleum prod.
Manuf. of chemicals and chemical prod.
Manuf. of pharmaceutical products
Manuf. of rubber, plastic and non-metallic
Manuf. of basic and fabricated metal prod.
Manuf. of computer, electronic and optical
Manuf. of electrical equipment
Manuf. of machinery and equipment n.e.c.
Manuf. of motor vehicles, trailers
Manuf. of furniture; jewelry, musical prod.
Non financial market services
Wholesale and retail trade of motor vehicles
Wholesale trade, except of motor vehicles
Retail trade, except of motor vehicles
Transport and storage
Postal and courier activities
Accommodation and food service activities
Publishing, motion picture, video, television
Telecommunications
Computer programming, consultancy
Professional, scientific and tech. activities
Other production
Electricity, gas, steam and air cond. supply
Water suppl.; sewerage, waste management
Construction
Number of years
1
2
5
97.07% 89.95% 80.94%
96.78%
96.66%
97.01%
85.71%
95.68%
95.74%
97.32%
97.36%
97.48%
96.93%
96.58%
97.44%
98.67%
89.82%
87.80%
88.31%
57.14%
89.46%
87.23%
89.71%
87.93%
89.92%
87.29%
86.97%
89.74%
92.49%
80.89%
73.43%
77.43%
0.00%
82.97%
75.00%
79.97%
77.50%
80.81%
80.45%
79.42%
78.92%
84.09%
96.63%
97.51%
97.19%
96.22%
72.22%
96.44%
96.28%
93.33%
96.83%
97.15%
90.89%
91.82%
90.69%
89.57%
50.00%
90.03%
87.59%
85.00%
91.11%
90.68%
82.46%
84.67%
82.46%
79.43%
50.00%
80.29%
75.93%
70.83%
82.86%
83.22%
97.46%
95.39%
97.62%
91.88%
85.12%
90.64%
87.06%
75.60%
80.44%
Notes: Non-EU frontier firms are all firms below the top 5% most productive firms in each year and industry
at the EU level. Balanced sample corresponds to Sample 3, which follows non-frontier firms from 2006
onwards but restricting the sample to a balanced panel of firms staying in AMADEUS database from 2006 to
2013.
27
Table 7. Distance to the EU Frontier
Dep. Variable: ln
TFPicjt
TFPjtF
Size
Size 2008-2014
Size Manuf. 2003-07
Size Manuf. 2008-14
Size Serv. 2003-07
Size Serv. 2008-14
Size Other 2003-07
Size Other 2008-14
Age
Age 2008-2014
Age Manuf. 2003-07
Age Manuf. 2008-14
Age Serv. 2003-07
Age Serv. 2008-14
Age Other 2003-07
Age Other 2008-14
Capital
Capital 2008-2014
Capital Manuf. 2003-07
Capital Manuf. 2008-14
Capital Serv. 2003-07
Capital Serv. 2008-14
Capital Other 2003-07
Capital Other 2008-14
Wage per employee
Wage per employee 2008-2014
Wage per empl. Manuf. 2003-07
Wage per empl. Manuf. 2008-14
Wage per empl. Serv. 2003-07
Wage per empl. Serv. 2008-14
Wage per empl. Other 2003-07
Wage per empl. Other 2008-14
Tertiary education
Tertiary education 2008-2014
Tertiary educ. Manuf. 2003-07
Tertiary educ. Manuf. 2008-14
Tertiary educ. Serv. 2003-07
Tertiary educ. Serv. 2008-14
Tertiary educ. Other 2003-07
Tertiary educ. Other 2008-14
Trade openness
Trade openness 2008-2014
Trade open. Manuf. 2003-07
Trade open. Manuf. 2008-14
Trade open. Serv. 2003-07
Trade open. Serv. 2008-14
Trade open. Other 2003-07
Trade open. Other 2008-14
DTF Getting credit
DTF Getting credit 2008-2014
DTF Get. Cre. Manuf. 2003-07
DTF Get. Cre. Manuf. 2008-14
DTF Get. Cre. Serv. 2003-07
DTF Get. Cre. Serv. 2008-14
DTF Get. Cre. Other 2003-07
(1)
0.182***
-0.009***
0.026***
0.848***
0.013***
0.001***
0.002***
-
(2)
(0.001)
(0.001)
(0.001)
(0.003)
(0.001)
(0.000)
(0.000)
0.152***
0.044***
-0.007***
-0.003***
0.041***
-0.022***
0.826***
-0.036***
0.016***
-0.005***
0.001***
0.000***
0.001***
0.000***
-
(0.001)
(0.001)
(0.001)
(0.001)
(0.001)
(0.001)
(0.003)
(0.003)
(0.001)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(3)
0.160***
0.206***
0.156***
0.197***
0.141***
0.180***
-0.012***
-0.016***
-0.005***
-0.001***
-0.016***
-0.020***
0.030***
0.010***
0.048***
0.030***
0.017***
0.003***
0.862***
0.878***
0.824***
0.855***
0.808***
0.836***
0.014***
0.009***
0.017***
0.011***
0.012***
0.012***
0.001***
0.001***
0.001***
0.001***
0.000***
0.001***
0.001***
0.001***
0.001***
0.002***
0.000***
(0.002)
(0.002)
(0.002)
(0.001)
(0.003)
(0.003)
(0.002)
(0.001)
(0.002)
(0.001)
(0.002)
(0.003)
(0.001)
(0.001)
(0.001)
(0.001)
(0.002)
(0.002)
(0.003)
(0.004)
(0.004)
(0.003)
(0.010)
(0.012)
(0.001)
(0.001)
(0.001)
(0.001)
(0.001)
(0.001)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
(0.000)
28
DTF Get. Cre. Other 2008-14
-0.000***
(0.000)
Governance quality
0.049*** (0.004) -0.002***
(0.004)
Governance q. 2008-2014
0.047***
(0.003)
Govern. qual. Manuf. 2003-07
-0.002***
(0.005)
Govern. qual. Manuf. 2008-14
0.037***
(0.004)
Govern. qual. Serv. 2003-07
-0.034***
(0.005)
Govern. qual. Serv. 2008-14
0.031***
(0.004)
Govern. qual. Other 2003-07
0.050***
(0.008)
Govern. qual. Other 2008-14
0.050***
(0.006)
Constant
-5.609*** (0.017) -5.603***
(0.019)
-5.562***
(0.022)
Observations
1,040,504
1,040,504
1,040,504
R-squared
0.771
0.773
0.775
Notes: EU frontier firms are the top 5% most productive firms in each year and industry at the EU
level. Size refers to the number of employees. All specifications include industry, country and year
dummies. Robust standard errors clustered by firm in parentheses. *** p<0.01, ** p<0.05, * p<0.1
29