World Bank, Kosovo Country Economic Memorandum, November 2021 (2021)
Summary
World Bank Group Kosovo Country Economic Memorandum, Raising Firm Productivity, dated November 2021. The memorandum covers firm characteristics and dynamics, firm productivity and access to credit, the impact of COVID-19 on firms, and conclusions and tentative policy implications, followed by references, a technical appendix and additional tables and figures. Its figures draw on Kosovo tax and company registry data for 2010–18, including active firms by size, entry and exit rates, firm survival, job creation and destruction, total factor productivity revenue (TFPR) by sector, and credit to the private sector. A regression table on drivers of access to credit in Kosovo covers 1,513 firms and 3,536 observations, and a figure compares digitalization and innovation in response to the COVID-19 outbreak.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Cite this document
“World Bank, Kosovo Country Economic Memorandum, November 2021 (2021).”
Original at openknowledge.worldbank.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/world-bank-kosovo-country-economic-memorandum-november-2021-2021-b9da06cd60/
, SHA-256 c7b1afae2f4e26760a749a65a062e2b75e530fe6cdb20e3032a94eb51f57889a.Full text
Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized
November 2021
WORLD BANK GROUP
Raising
Firm Productivity
Kosovo Country
1
Economic Memorandum
Kosovo Country Economic Memorandum
Raising
Firm
Productivity
Table of Contents
Executive summary 7
Firm Characteristics and Dynamics 11
1.1. Firm Characteristics 12
1.2. Firm Dynamics 22
Firm Productivity and Access to Credit 35
1.3. TFPR and Labor Productivity 36
1.4. Access to Credit and Productivity 43
1.5. Barriers to Productivity 47
The Impact of COVID-19 on Firms 48
Conclusions and Tentative Policy Implications 51
References 55
Technical Appendix 57
Additional Tables and Figures 61
2
Figures
Fig 1. Active Firms by Size and Total, Already Active, 2012–18, Fig 31. Labor, Capital, and
Average, 2010–18 p.12 Percent p.24 Intermediate Inputs, 2012–17,
Fig 2. Average Firm Size and Density, Fig 17. Drivers of Entry and Exit Rates, Index 100=2012 p.37
2012–18 p.13 Percentage Points p.24 Fig 32. TFPR and value added per
Fig 3. Turnover Comparisons by Firm Fig 18. Firm Size at Entry p.25 worker by Sector and Average
Size, Relative to EU Averages, p.38
Fig 19. Firm age and Employment
2018. Index EU=100 p.14 Growth p.26 Fig 33. Decomposition of TFPR
Fig 4. Active Enterprises by Main Growth, 2014–17, Percent
Fig 20. Employees by Firm Age and
Activity Compared to the EU, Change p.39
Size over Five Years, Average
Percent of Total p.14 2012–18, Index, Entry Year = 1 Fig 34. Characteristics of Entrants
Fig 5. Employment Share by Sector, p.27 Relative to the Median Firm.
Percent of Total p.15 Percent p.40
Fig 21. Firm Survival in Kosovo,
Fig 6. Firms, Employment, and Sales Average 2012–18, Percent p.27 Fig 35. TFPR, Employment, Revenue,
by Municipality, Percent p.15 and Value Added per Worker,
Fig 22. Thousands of Employees in
Logarithms, 2012–18 p.41
Fig 7. Exporters and Importers p.17 Active Firms, 2010–18 p.28
Fig 36. Total Factor Productivity
Fig 8. Exports by Sector p.18 Fig 23. Firms Reporting Employees,
Revenue (TFPR) by Sector,
Fig 9. Firms with Foreign Percent of Total Firms, 2010–
2013–17 p.42
Shareholders, percent of firms, 2018 p.29
Fig 37. Domestic Credit to the Private
2018 p.19 Fig 24. Contribution to Employment
Sector, Percent of GDP p.43
Fig 10. Gender of Firm Decisionmakers Growth by Sector, 2010–14 and
2015–18, Percent p.30 Fig 38. Credit Disbursed and
by Sector, Percent p.20
Employment by Sector,
Fig 11. Digital Firms, 2010–10, Percent Fig 25. Employment Growth, Firms
2010–18, Percent of Total p.44
of Total Firms p.21 by Sector Surviving 5 Years,
Average Number of Employees, Fig 39. Credit to the Private Sector by
Fig 12. Concentration in the Kosovar Type, Average 2010–18, Euros
2017–18 p.30
Economy p.21 and Percent of Total p.45
Fig 26. Net Job Creation, 2011–18 p.31
Fig 13. Net Firm Turnover, Entry. and Fig 40. Loans to the Private Sector by
Exit Rates, Percent of Firms, Fig 27. Job Creation, Destruction, and
Term (Years), Average 2015–18,
Average 2012–18 p.22 Employment by Firm Age and
Percent of Total p.45
Size, Average 2016–18 p.32
Fig 14. New Business Density and GDP Fig 41. Productivity and Access to
per Capita in Constant US$ PPP, Fig 28. Employment by Firm Age,
Credit p.47
2018 p.23 Percent of Total, Average
2016–18 p.32 Fig 42. Barriers to Investment and
Fig 15. Economic and Administrative Exports. Percent of firms p.47
Firm Entry and Exit, Number of Fig 29. TFPR, Sales, and Value Added
per Worker, 2013–17 p.36 Fig 43. Digitalization and Innovation
Firms, 2012–18 p.23
in Response to the COVID-19
Fig 16. Size of Firms Entering a Market Fig 30. Labor Productivity Relative to
Outbreak Compared, Percent
Relative to the Average Firm the EU, 2018, Percent p.37
of Firms p.49
Tables
Table 1. Selected Characteristics of Table 4. TFPR in Micro-Enterprises Table 8. Main Drivers of Lower TFPR
Firms in Kosovo, 2017, Percent p.19 Relative to Other Firm Classes p.41 Growth and Recommended Policy
Table 2. Five-Year Transition Matrix Table 5. Credit in Kosovo, Summary Changes p.53
by Firm Class, 2012 and 2013 Firm Statistics, 2018 p.44 Boxes
Cohorts, Average p.26 Table 6. Access to Credit by Firm Class,
Table 3. Ratio of Productivity in Average, 2015–18 p.46 Box 1. Measuring Firm Dynamics and
“Laggard” (90th percentile) to “Frontier” Productivity in Kosovo p.8
Table 7. Potential Correlates with
(10th percentile) Firms by Sector and Productivity in Kosovo p.48 Box 2. Decomposing the Sources of
Size, Average 2012-18 p.40 TFPR Growth p.39
© 2021 International Bank for Reconstruction and Development / The World Bank
1818 H Street NW
Washington DC 20433
Telephone: 202-473-1000
Internet: www.worldbank.org
This work is a product of the staff of The World Bank with external contributions. The findings,
interpretations, and conclusions expressed in this work do not necessarily reflect the views of The
World Bank, its Board of Executive Directors, or the governments they represent.
The World Bank does not guarantee the accuracy of the data included in this work. The boundaries,
colors, denominations, and other information shown on any map in this work do not imply
any judgment on the part of The World Bank concerning the legal status of any territory or the
endorsement or acceptance of such boundaries.
Rights and permissions
The material in this work is subject to copyright. Because The World Bank encourages dissemination
of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as
long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary
rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW,
Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org.
4
Acknowledgments
This note was prepared by Lucio Castro (Senior Economist), Leonardo Iacovone (Lead Economist),
and Aslı Şenkal (Senior Economist) and with contributions from Blerta Qerimi (Senior Private
Sector Specialist), Besart Myderrizi (Research Analyst), and Temel Taskin (Economist). The note
benefitted from the comments of Bledi Çeliku (Senior Economist), Enrique Blanco Armas (Practice
Manager), and Donato De Rosa (Lead Economist).
This note has been prepared as part of the Kosovo Country Economic Memorandum led by Aslı
Şenkal (Senior Economist, Task Team Leader). The work was overseen by Linda Van Gelder
(Country Director, Western Balkans), Massimiliano Paolucci (Country Manager, Kosovo and North
Macedonia), Jasmin Chakeri (Practice Manager, Macro, Trade, and Investment Global Practice), and
Denis Mehmedev (Practice Manager, Finance, Competitiveness, and Innovation Global Practice).
The team is grateful to Kosovo Agency of Statistics and Central Bank of Kosovo for their support in
providing data.
5
List of Acronyms
BPS World Bank business pulse survey
CEM Country Economic Memorandum
EBRD European Bank for Reconstruction and Development
ES Enterprise survey
EU European Union
FDI Foreign direct investment
GDP Gross domestic product
HHI Herfindahl index
KAS Kosovo Agency of Statistics
KCR Kosovo credit registry dataset
KPS Kosovo productivity survey
KTR Kosovo tax registry dataset
LP Labor productivity
ICT Information and communications technology
MSME Micro, small, and medium sized enterprises
NIS National innovation system
OECD Organisation for Economic Co-operation and Development
PPP Purchasing power parity
R&D Research and development
SME Small and medium sized enterprises
TFP Total factor productivity
TFPR Total revenue factor productivity
VA Value added
WB6 Western Balkans
WDI World Development Indicators
W&R Wholesale and retail trade
6
Executive summary
To boost economic growth and foster sustained formal job creation in Kosovo, igniting firm
productivity is crucial. Both demographically and historically, Kosovo is a small, young transition
economy. Although in the decade leading up to the COVID-19 crisis, the country’s economic growth
outperformed the other Western Balkans (WB6) peers,1 it was not able to provide enough jobs, especially
for women and the young, or to reduce high unemployment.2 Sustained job creation can only be driven
by private companies that are productive enough to grow and expand. That is why understanding firm
dynamics, particularly what drives productivity, the engine of economic development, is crucial to
identify barriers to growth and job creation, and to formulate policy reforms that will accelerate firm
growth and formal job creation.
Based on detailed micro-data, this note examines the characteristics and recent evolution of firms
in Kosovo, with particular attention to firm productivity. It benefits from a dataset compiled from
anonymized tax records complemented by data from the Kosovo Productivity Survey (KPS). To explore
the links between productivity and credit, it also draws on a unique dataset on access to credit from the
Kosovo Credit Registry (KCR). Box 1 describes the data and methodology used. Annex 1 provides further
technical details. The note also provides a snapshot of the impact of COVID-19 on firms, based on the
Business Pulse Survey (BPS)3 and administrative data.
For the last decade, the landscape of firms in Kosovo has been dominated by microenterprises with
low productivity, employment and survival rates. According to the administrative data, in 2010–18
micro companies (firms averaging fewer than 10 employees) and small enterprises (10–49 employees)
accounted for more than 98 percent of firms. Micro enterprises are less productive and have higher net
firm turnover,4 lower survival rates,5 and much lower employment growth. Moreover, the average annual
net rate of firm creation in Kosovo—the difference between firm entry and exit rates—was close to zero,
an indication of a lack of dynamism in the last decade, especially for a young economy.
Firm creation and growth, firm density, average size, and the likelihood of survival are all low, which
implies that there are severe constraints on private sector development. Firm density, the number
of firms per million inhabitants, is considerably lower than in the European Union (EU) and also than
what would be predicted by Kosovo’s income per capita. Partly reflecting the small size of the economy,
Kosovar firms are also much smaller than their peers in the EU and in the aspirational economies.6 Firm
survival is also much lower than in comparator countries, especially among micro enterprises.
Unlike firms in aspirational peers, Kosovo seems not to have many middle and large companies.
Only about 2 percent of the micro enterprises became small firms, 5.5 percent of small enterprises
become medium, and 4.6 percent of medium firms became large. Medium companies report about 37
percent less in sales than the EU mean. Large firms in Kosovo report the lowest proportion of sales, just
15 percent of the EU average, than large firms in aspirational peers.
Kosovo’s firms are only tenuously linked to global markets and the country is lagging in the share
of female-run companies. The share of exporter enterprises (0.3 percent per the Kosovo Statistics
Authority (KAS) and 3.4 percent per the Kosovo Productivity Survey [KPS]) is far below the averages
for the EU (7.4 percent); aspirational peers (e.g., 7.4 percent in Croatia); and the Western Balkans; only 6
percent have any foreign shareholders; and just 9 percent are female-run.
Impediments to firm growth and exit are numerous. It appears that constraints on firms are sizable
from the beginning and become more pervasive the larger the firm; companies increase the number of
their employees until they are mid-sized, then seem to practically stop growing. Administrative costs to
exit the market are also high in terms of liquidating assets and legally closing a company.
Positive and rising net job creation in 2015–18 was driven by higher formalization of jobs and the
increasing size of incumbent firms, especially young SMEs. The number of net jobs created annually—
the difference between gross job creation and gross job destruction—practically doubled in 2015–18
compared to 2011–14. Net job creation was driven by the increasing size of incumbent firms and higher
job formalization: the share of firms with employees increased from 40 percent in 2010 to 70 percent in
2018.
1 In addition to Kosovo, the WB6 countries are Albania, North Macedonia, Serbia, Montenegro and Bosnia and
Herzegovina.
2 World Bank 2018.
3 See Section 3 for further details on the BPS.
4 Firm turnover is the difference between firm entry and firm exit.
5 Share of active firms that continue reporting sales per year in total firm population.
6 In this note, aspirational peers are Croatia, Czechia, Latvia, Lithuania, Estonia, and Slovakia—all EU members.
7
In Kosovo labor productivity (LP) is far below LP in the EU and aspirational countries; annual
average LP growth was negative and average yearly total revenue factor productivity, a revenue-
based measure of productivity growth (TFPR),7 was low in 2012–17, the period for which data are
available. Clearly, there is room for further economic convergence with the more developed European
countries: LP in Kosovo is only a third of the EU average and is much lower than in the aspirational
economies. Moreover, LP, proxied by sales per worker, fell on average by –0.25 percent annually between
2012 and 2017, while TFPR increased by just 0.5 percent.
Declining productivity of incumbent firms explains Kosovo’s disappointing TFPR growth rate; firm
dynamism plays only a minor, and somewhat negative, role. Kosovo’s low TFP growth is explained by
the declining market share and productivity of incumbent firms, only partly compensated for by rising
average firm productivity. The latter component accounted for just a third of TFPR growth in 2012–17, far
below the average of developing countries for which there are comparable data. Exit of less productive
firms was only a minor factor in driving aggregate TFPR growth, and entry of new firms contributed
negatively to productivity growth.
Linked to Kosovo’s meager TFPR performance is the fact that, relative to the EU and other peers,
private investment in research and development (R&D) and management techniques is also low. In
spite of the strong empirical evidence of the positive impact of investment in R&D on firm productivity
(Cirera and Maloney 2017), only 1 percent of businesses in Kosovo report investing in R&D; the average
for the EU is 14 percent and among some aspirational economies like Croatia it is 7.4 percent.8 Private
investment in R&D), according to the KPS, is even lower among micro enterprises (0.1 percent) and small
companies (0.2 percent. Similarly, only about 6 percent of the plants report investing in management
training, despite the evidence of its beneficial effects on firm productivity (Bloom et al. 2010).
More productive firms create more jobs and sell more; exporters and foreign-controlled companies
tend to be larger, more productive, and more capital-intensive than firms oriented to the domestic
market and pay higher wages. Higher firm TFPR is associated with higher average revenues and more
workers per company, which underscores the importance of fostering firm productivity for economic
growth and job creation. Furthermore, the track record of exporters and foreign-owned firms highlights
the relevance of integration into global markets, especially for a small economy like Kosovo.
Access to credit is low and biased toward short-term financing; and financing is not allocated to the
most productive firms. Only one-third of firms have access to formal credit. About 70 percent of loans
granted have terms no longer than three years, suggesting that the local financial system is not providing
enough capital for long-term investment. More productive and larger firms in more concentrated sectors
are more likely to get access to credit; firms based in non-majority municipalities9 have less access. Micro
and small enterprises and startups are particularly deterred by the lack of access to formal financing.
The COVID-19 shock affected MSMEs the most, especially in sectors under lockdown. While most
of the firms adjusted by granting leaves and adjusting wages and worker hours, MSMEs and female-
run companies experienced the largest net job losses. Despite layoffs were largest in sectors under
lockdown, 2020 employment did pick up in most sectors, including hospitality. However, more than half
of companies reported they expected to fall into arrears end 2020; MSMEs in particular were financially
stressed.
Kosovo needs a multidimensional policy strategy to foster growth in firm productivity. Based on the
study findings and the results of other notes prepared for Kosovo’s Country Economic Memorandum
(CEM), this note proposes a policy strategy that targets the three main sources of firm productivity
growth: (1) firm productivity (the “within” component); (2) market reallocation (the “between”
component); and (c) firm dynamics (entry and exit).
In what follows, Section 1 examines the characteristics and recent dynamics of Kosovar firms. Section 2
analyzes the drivers and evolution of productivity, with emphasis on the links between productivity and
access to credit. It also assesses the main barriers to productivity growth. Section 3 sheds light on how
COVID-19 has affected Kosovar firms. Section 4 concludes by discussing tentative policy implications of
the analysis.
7 In the absence of firm level prices, this note uses a revenue-based measure of TFP (see Box 1). Please see
Cusolito and Maloney (2016) for a more detailed description of advantages and disadvantages of using TFPR
measures.
8 See World Bank (2018).
9 See Box 1 for the definition of non-majority municipalities.
8
Box 1. Measuring Firm Dynamics and Productivity in Kosovo
Total factor productivity (TFP) and labor productivity (LP) are two ways of computing productivity, the measure
of firm efficiency at combining or converting inputs (capital, labor, et al.) to produce goods or services (output).
More precisely, productivity can be defined as the variation in output that cannot be explained by changes in
factors of production.
While LP measures output per worker, and therefore looks only at employment, TFP incorporates not only
adjustments in other production factors (e.g. capital stock and intermediate inputs) but also, and more
importantly, technology, innovation, management practices, and other ways of augmenting firm output with
the same input endowment (Syverson 2011).
As no data on firm-level prices are available for Kosovo, this note calculates revenue TFP (TFPR) instead of
physical TFP. TFPR reflects not only firm physical efficiency at production but also prices that reflect product
quality and markups in addition to input costs (Cusolito and Maloney 2018). This note is unfortunately unable
to capture directly those price-related dimensions.
This chapter relies on two datasets containing detailed information on variables required to analyze firm
dynamics and estimate TFPR and LP. Administrative tax records from the Kosovo Statistics Authority (KAS)
is the main source of data. The KTR dataset contains records for about 30,000 formal establishments with
annual information on company characteristics (location by municipality, activity at four digits of NACE , year
of starting and closing operations), employment, revenues, exports, imports, inputs, costs, and investment in
durable goods.
A panel dataset for 3,600 establishments reporting the information required to estimate our preferred measure
of productivity for TFPR in 2012–17: investment, revenues, intermediate consumption, wages and employment
was built. (Annex 2 provides summary statistics.)
The KTR panel data set was complemented with information from the 2017 World Bank Kosovo Productivity Survey
(KPS). The KPS covered a representative sample of 3,083 firms. In addition to the variables contained in the KTR,
the KPS dataset provides among other variables detailed information on input purchases, worker skills, firm location
(region), and shareholding (foreign and domestic). We can match 3,314 firms from the KPS with the KTR dataset. (See
Annex 2 for summary statistics on the KPS dataset.)
Capital stock is estimated using investment data by establishment in durable goods from the KTR dataset
and combining it, where possible, with the estimates of firm total assets provided by the KPS dataset. LP is
calculated as the ratio of sales to the number of employees (except when we state that it refers to value-added
per employee); TFP is estimated by applying semi-parametric methods following Levinsohn and Petrin (2003),
given the small sample of firms reporting investment in the KTA dataset. (See Annex 1.A for further details.)
We also control for the potential effects of firm entry and exit in TFP growth by applying the decomposition
proposed by Meliz and Polanec (2015).
As firms reporting the data available to estimate TFPR are mainly medium and large firms, the panel dataset is
reweighted using information on firm employment shares by NACE 2-digit sector from the entire KAS dataset,
to give more weight to smaller companies to increase representativeness of the dataset, where small and micro
productive units are more dominant.
To assess the links between access to credit and productivity, a panel dataset with information by firm from
the Kosovo Credit Registry (KCR) with information on loan amount disbursed, loan term, lending by financial
institutions (banking and micro-finance), nominal interest rate per financial instrument, and loan payments,
among other variables, for the universe of firms for 2010–18 is used. The KCR and KTR datasets are matched by
firm, producing a panel with complete information for about 6,000 firms. (See Annex 2 for summary statistics
on the KCR panel dataset.)
Sales, wages, exports, imports, and other variables measured in euros were deflated by the Producer Price Index
(PPI). Investment was deflated using the corresponding investment price deflators. All monetary variables
were converted to 2018 euros. Observations reporting negative turnover, sales, or investment were identified as
missing. Finally, using KAS data, “non-majority municipalities” were identified as those subnational units with
a share of population of Serbian origin of 40 percent or more of the of total population. The other municipalities
are predominantly Albanian/Turkish majority.
9
01
Firm Characteristics and
Dynamics
This section presents an overview of the main characteristics of firms in Kosovo.
It describes the number and evolution of active firms, average firm size, number of
firms per capita, and firm and employment distribution by sector. It also assesses
the relative importance and features of firms engaged in international trade and
benefiting from foreign direct investment (FDI). Moreover, it examines the gender
composition of firm decisionmakers. Finally, it analyzes the economic concentration
of the Kosovar economy. (Annex 1 provides technical details on how firm
characteristics and dynamics in Kosovo were analyzed.)
1.1. Firm Characteristics
More than 98 percent of Kosovar firms are micro and small enterprises10 and the population
of active firms has been relatively stable for the last decade. Between 2010 and 2018, every year
an average of about 30,000 enterprises reported sales for at least two years (Figure 1A). While the
proportion of micro enterprises is marginally lower than in some aspirational peers, the country has
more small firms, except for Croatia. The share of medium and large enterprises is similar to Croatia
and Lithuania (Figure 1B).
Figure 1.
Active Firms by Size and Total, Average, 2010–18
a) Number of active firms b) Active firms by size, percent of total
34,000 0 50 100
Micro 90.3
32,000 30,866 (0-9) 90.9
30,299 30,250 93.4
29,441 96.0
30,000 29,041 28,709 97.1
28,556
27,918 28,123 Small 8.3
28,000 (10-49) 7.6
5.4
3.1
26,000 2.3
Medium 1.2
24,000 (50-249) 1.2
1.0
0.7
22,000 0.5 Kosovo
Large 0.2 Croatia
20,000 (250+) 0.3 Lithuania
2010 2011 2012 2013 2014 2015 2016 2017 2018 0.2
0.2 Czechia
0.1 Slovakia
Source: KTR, World Bank staff calculations.
Notes: Active units are those with positive turnover. Activities included: B-N (excluding K64.2) according to NACE Rev. 2. Micro firms
average 0–9 employees, small firms 10–49, medium firms 50–249, and large firms 250 and above.
In the last decade, the share of medium and large companies was practically unchanged,
possibly implying barriers to firm growth. There were no major changes in the share of medium
enterprises (50–249 employees) and large firms (250 and more).11 The share of micro enterprises
fell from 95 percent in 2010 to 90 percent in 2018, as the share of small firms rose from 4 to 8.3
percent. These figures suggest that there are barriers to growth beyond small size in Kosovo.
Average firm size is smaller than in many of the aspirational peers though similar to the EU
average, but firm density is below even what can be expected by the country´s income per
capita. The average firm in Kosovo employs 5 workers, far below most aspirational peers except
Czechia (4.5) and Slovakia (3.5), and close to the EU mean (5.1). Firm density reaches only 16,300
firms per million inhabitants, just over a third of the EU average (54,600), and lower than in any
of the aspirational economies (Figure 2B). Moreover, Kosovo has less firm density than its income
per capita would suggest, in contrast with aspirational peers and Western Balkans comparator
countries, except for Serbia (Figure 2C).
10 Micro firms average 0–9 employees and small companies 10–49.
11 KTR, World Bank staff calculations.
12 Firm Characteristics and Dynamics
Figure 2.
Average Firm Size and Density, 2012–18
a) Size: average number of workers b) Density: active firms, thousand per million inhabitants
8 Czechia 96.7
7.0
7 Slovakia 82.3
6.1
6 Lithuania
5.5 5.4 63.3
5.1 5.1
5 Estonia 62.5
4 3.8
3.5 EU (2020) 54.6
3
Latvia 53.7
2
Croatia 35.3
1
Kosovo 16.3
0
Croatia Latvia Lithuania Estonia EU(2020) Kosovo Czechia Slovakia 0 20 40 60 80 100 120
c) Active firms per million inhabitants. and GDP per capita (log, current US$, PPP), 2018
12.0 Firms per million
CZE Kosovo
people (log) SVK
R² = 0.3152 Comparators
11.5 MLT
LVA PRT SVN SWE
Aspirational peers
GRC LTU ITA Other EU countries
11.0 MNE ESP BELNLD
HRV HUNEST LUX
FIN AUT IRL
BGR
ALB POL EU FRA DNK
10.5 ROU CYP DEU
MKD
10.0
KOS
SRB
9.5
9.0
9.0 9.5 10.0 10.5 11.0 11.5 12.0 GDP per capita (PPP, in logs)
Source: KAS and Eurostat data, World Bank staff calculations.
Notes: Active firms are those with positive employment or turnover on each country adjusted by population size. Activities: B-N (excluding
K64.2) according to NACE Rev. 2. Firm size as total persons employed in the population of active enterprises / total active enterprises.
Employed persons are those aged 15 and over who perform part-time or full-time work for pay, profit, or family gain, according to
Eurostat.
Measured by sales, firms in Kosovo are also smaller than companies in aspirational peers and
the EU, especially the medium and large enterprises. Micro and small companies report sales that
are half the revenues reported by average EU peers. Sales reported by small Kosovar companies
are about 44 percent of the EU mean and large company sales are 15 percent smaller. While these
differences are not so large for Kosovo’s MSMEs, except relative to Slovenian peers, its large firms,
companies with more than 250 employees, report the lowest proportion of sales relative to the EU
average compared to aspirational peers (Figure 3).
Firm Characteristics and Dynamics 13
Figure 3.
Turnover Comparisons by Firm Size, Relative to EU Averages, 2018. Index EU=100
120 Kosovo
Latvia
98 Lithuania
89 86 Czechia
78 Slovenia
70 74
64
54 52 55 56 54
49 44 45
37
32
25
15
Micro Small Medium Large
(0-9 employees) (10-49 employees) (50-249 employees) (250+ employees)
Source: KTR and Eurostat data, World Bank staff calculations.
Notes: Total turnover in 2018 in euros of active businesses (except finance and insurance) divided by total active enterprises. Baseline is the
EU average (indexed to 100) for countries that report aggregate turnover by firm size. Activities: B-N (excluding K64.2) according to NACE
Rev. 2. Micro are firms with an average of 0–9 employees, small 10–49 employees, medium 50–249 employees, and large 250 and more.
A distinctive feature of Kosovo’s economy is the relative importance of wholesale and retail
(W&R) firms. More than 50 percent of firms are in W&R; 14 percent are in manufacturing, 11.2
percent in accommodation and food services, about 7 percent in construction, and the rest in other
services activities. W&R is much more important in Kosovo than in aspirational peers like Croatia
and Slovakia, where it only accounts for about 30 percent of firms (Figure 4).
Figure 4.
Active Enterprises by Main Activity Compared to the EU, Percent of Total
a) Kosovo
Wholesale & Retail trade Manufacturing Construction
48.4 14.7 8.5
b) Croatia
Transportation & Storage Administrative & Manufacturing Wholesale & Retail
25.3 support services 13.7 trade
16.2 12.6
c) Slovakia
Accommodation & Food Manufacturing Information &
services 16.2 Communication
25.3 12.6
Wholesale & Retail trade
13.1
Utilities
Source: KTR and Eurostat data, 2020, World Bank staff calculations.
Notes: Active units are those with positive turnover. Activities included: B-N (excluding K –Financial and insurance activities) according to
NACE Rev. 2. See Annex 1 for more detail on sector classification.
14 Firm Characteristics and Dynamics
W&R is also one of the largest employers, along with manufacturing and other service activities. It
employs about 33 percent of workers in Kosovo, followed by manufacturing, 15 percent; construction, 10.7
percent; accommodation and food services, 7.3 percent; and financial services, about 7 percent (Figure 5).
Figure 5.
Employment Share by Sector, Percent of Total
Wholesale and retail trade 33.8
Manufacturing 15.1
Construction 10.7
Accommodation and food services 7.3
Financial services 6.9
Information and communication 5.3
Administrative and support services 4.7
Electricity, gas, steam and air-conditioning supply 4.0
Transportation and storage 4.0
Professional, scientific and technical act. 3.3
Water supply; sewerage, waste mgmt. and remediation act. 2.7
Mining and quarrying 2.0
Real Estate 0.2
Source: KTR and Eurostat data, World Bank staff calculations.
Notes: Share of employed persons in active firms. Activities: B-N (excluding K64.2) of NACE Rev. 2. Micro firms have an average of 0–9
employees, small firms 10–49, medium firms 50–249, and large firms 250 and more.
Employment and sales are heavily concentrated in Pristina. Although only 24.1 percent of firms
are in Pristina, the municipality accounts for about 43 percent of Kosovar employment and about 44
percent of sales, similar to comparator countries like Moldova (Cojocaru 2016). Ferizaj is second in
importance with 6.6 of employment and about 7 percent of sales (Figure 6).
Figure 6.
Firms, Employment, and Sales by Municipality, Percent
a) Firms
Leposaviq
0.2
Mitrovicë
e Veriut
Zveçan 0.3
0.2
Zubin Podujevë
Potok
3.0
Istog 0.1 Mitrovicë Prishtinë
1.4 3.6 24.1
Pejë Skënderaj Vushtrri Novobërdë
5.5 1.3 2.5 0.1 Kamenicë
Klinë Drenas Obiliq
1.4 2.1 0.8 1.4
Deçan Malishevë Fushë Gjilan
Kosovë
1.2 1.6 6.4
Suharekë 2.9 Graçanicë
Junik Ranillug
2.8 0.0
Rahovec Lipjan Kllokot
2.5 2.4
Gjakovë Shtime Ferizaj
Partesh
5.6 1.1 8.5
Shtërpcë Viti
Mamushë
0.4 1.9
Prizren Kaçanik
10.5 Hani i 1.6
Elezit
0.2
Dragash
1.2
Firm Characteristics and Dynamics 15
Figure 6.
Firms, Employment, and Sales by Municipality, Percent
b) Employment
Leposaviq
0.1
Mitrovicë
e Veriut
Zveçan 0.1
0.0
Zubin Podujevë
Potok
1.5
Istog 0.0 Mitrovicë Prishtinë
1.1 3.7 42.7
Pejë Skënderaj Vushtrri Novobërdë
5.6 0.9 1.7 0.0 Kamenicë
Klinë Drenas Obiliq
0.9 2.0 0.9 0.6
Deçan Malishevë Fushë Gjilan
Kosovë
0.7 1.0 4.5
Suharekë 3.7 Graçanicë
Junik Ranillug
2.2 0.0
Rahovec Lipjan Kllokot
1.3 1.4
Gjakovë Shtime Ferizaj
Partesh
3.4 0.6 6.6
Shtërpcë Viti
Mamushë
0.1 1.1
Prizren Kaçanik
6.2 Hani i 0.8
Elezit
0.4
Dragash
0.4
c) Sales
Leposaviq
0.1
Mitrovicë
e Veriut
Zveçan 0.1
0.1
Zubin Podujevë
Potok
1.5
Istog 0.0 Mitrovicë Prishtinë
0.9 1.9 43.9
Pejë Skënderaj Vushtrri Novobërdë
5.9 0.9 1.6 0.0 Kamenicë
Klinë Drenas Obiliq
0.9 1.7 1.0 0.3
Deçan Malishevë Fushë Gjilan
Kosovë
0.7 1.1 2.9
Suharekë 4.8 Graçanicë
Junik Ranillug
2.0 0.1
Rahovec Lipjan Kllokot
1.3 1.7
Gjakovë Shtime Ferizaj
Partesh
2.1 0.7 7.3
Shtërpcë Viti
Mamushë
0.1 1.8
Prizren Kaçanik
5.3 Hani i 0.7
Elezit
0.5
Dragash
0.2
Source: KTR data, World Bank staff calculations.
Notes: Activities: B-N (excluding K64.2) according to NACE Rev. 2. Firms not reporting location were excluded.
16 Firm Characteristics and Dynamics
Kosovo has the lowest share of exporters among Western Balkans countries and a much lower
proportion of manufacturing exporters than aspirational peers except Czechia. Less than 4
percent of firms export and import; about 0.3 percent are solely exporters, and about 25 percent
are exclusively importers (Figure 7A). Self-reported KPS data indicates that 1.4 percent of firms
identified themselves as exporters and 34 percent as importers.12 Exporters and importers tend
to be medium and large firms (Figure 7B). Focusing on manufacturing, Figure 7C reveals that the
fraction of exporters is lower (6.4 percent) than in any aspirational country except Czechia (4.3
percent), and less than a fourth of the EU average (21.7 percent). According to the 2020 Business
Environment and Enterprise Performance Survey (ES), Kosovo also has the lowest proportion of
exporters among Western Balkans economies.13
Figure 7.
Exporters and Importers
a) Percent of total firms b) Percent of total firms by sector and size
Non-exporter & Importer and Only Only
Non-importer exporter Exporter Importer Exporters 87
83
79.4% 2.0% 0.5% 18.1% Importers
76
31
28
16 15
1
Micro Small Medium Large
(0-9) (10-49) (50-250) (+250)
c) Manufacturing exporters, percent of manufacturing firms
Estonia 46.0
Latvia 27.5
EU (2020) 21.7
Lithuania 17.8
Croatia 14.2
Slovakia 9.5
Kosovo 6.4 % of manufacturing firms
Czechia 4.3
Source: KTR and Eurostat data, 2020, World Bank staff calculations.
Note: Exporters or importers are those companies that report €10,000 or more in exports or imports to the KAS in the relevant year.
“Importer and exporter” classification refers to a company that both imports and exports in the same period. “Only exporters” or
importers) are firms that export or import but do not import or export. Figure 7 C depicts the number of manufacturing exporters relative
to the total manufacturing active firms as classified in Eurostat. Figures are average values for 2012–18. Activities: B-N (excluding K64.2)
according to NACE Rev. 2.
12 See Table 2 in Annex 2.
13 See https://www.beeps-ebrd.com/indicators/
Firm Characteristics and Dynamics 17
Manufacturing accounts for about 45 percent of exports. Mining, manufacturing, and some
service activities have the largest export-to-sales ratios. W&R is the second largest exporter at
24.6 percent, followed by ICT at 7.4 percent (Figure 8A). According to KPS, furniture is the main
manufactured export, accounting for about 38 percent. Exports account for about 40 percent of
sales in mining and about 14 percent of sales in manufacturing (Figure 8B). Transport, administrative
services, and ICT report export-to-sales ratios of about 10 percent. Traditionally non-tradable
activities, such as financial services, construction, and real estate, have the lowest export-to-sales
ratios. Because service exports in accommodation and food services are not captured by the tax
records, they are understated.
Figure 8.
Exports by Sector
a) Percent of total exports
Manufacturing Wholesale and retail trade Information Mining and Transport
45.1 24.1 and communication quarrying and
7.4 Storage
4.7
Utilities
6.5 Construction &
Real estate 2.7
Administrative
and support act.
b) Share of exports in total sales by sector
Mining and quarrying 36
Utilities 27
Manufacturing 22
Transport and storage 13
Information and communication 11
Administrative and support act. 11
Average 6
Professional, scientific and technical 5
Wholesale and retail trade 2
Construction and Real Estate 2
Financial services 2
Accommodation and food service 1
Source: KTR and KPS data, World Bank staff calculations.
Note: Exporters/importers are those companies that report €10,000 or more in exports/imports in the year of reference. “Importer and
exporter” reflects the company reports both imports and exports within the same period. Only exporters (importers) are firms that
export (import) but do not import (export). The share of exports in total sales corresponds to total exports divided by total sales. Figures
report the average values for 2012–18. Activities: B-N (excluding K64.2) according to NACE Rev. 2. See Annex 1 for more detail on sector
classification.
Only 6 percent of Kosovan firms have foreign shareholders (Figure 9A) and the share of firms
with at least 10 percent of foreign shareholding is far lower than in the other Western Balkans
countries. Foreigners hold more than 10 percent of the assets in less than 6 percent of firms. About
10 percent of shareholders in construction and transport firms are foreign, as are 8 percent in
utilities and 5 percent in manufacturing (Figure 9B). According to the Enterprise Survey data, only
1.3 percent of firms have at least 10 percent foreign ownership, compared to 11.8 percent of firms
elsewhere in the region.14
14 World Bank Enterprise Survey (2019). The figure is the average for Albania, Bosnia and Herzegovina,
Montenegro, North Macedonia, and Serbia.
18 Firm Characteristics and Dynamics
Figure 9.
Firms with Foreign Shareholders, percent of firms, 2018
a) Firms with foreign shareholding b) By sector
Mining and quarrying
With no foreign shareholders 93.8 Manufacturing 5.29
Utilities 8.27
With foreign shareholders 6.2 Construction 10.17
Wholesale and retail trade 6.78
With foreign shareholders Transport and storage 9.60
5.7
(>10% assets) 3.86
Accommodation and food services
With foreign shareholders Information and communication 4.61
(>50% assets) 0.8
Real Estate
Professional, scientific 2
and technical act. 2.36
Administrative and 2 support act. 2.17
Source: KPS data, World Bank staff calculations.
Notes: Fraction of firms that report having foreign shareholders (> 0%). Figures are weighted by firm weights.
Only 3.4 percent of firms invest in R&D, far less than in the EU and peer economies; investment
in training and management is also low. Table 1 shows that only 3.4 percent of firms report
investing in R&D against an EU average of 14 percent and, typical of comparators, Croatia with 7.4
percent (World Bank 2018). The proportion is especially low among micro (3.7 percent), and small
(12 percent), and medium enterprises (12.6 percent). Across all firm size classes, investment in R&D
is also lower than in the EU, Western Balkans comparators, and aspirational peers (World Bank
2018). Only about 6 percent of enterprises report investing in employee training and 8 percent in
management training; the percentage increases, however, with firm size.
Table 1. Selected Characteristics of Firms in Kosovo, 2017, Percent
Class Category Invest in Receive Have Tertiary Have Unpaid Share of plants Share of firms
R&D Government Educated Employees that provide that Train
Subsidies Workers Employee Managers and
Training Supervisors
Size Micro (0-9) 3.7 1.6 31.6 14.7 5.4 7.9
Small (10-49) 12 2.7 26.5 2.7 15.6 13.1
Medium (50-249) 12.6 4.0 23.6 1.7 29.9 23.7
Large (250+) 18.2 14.0 36.5 0.2 27.6 27.2
Shareholding Nonforeign 4.3 1.8 31.0 14.1 6.1 8.6
shareholding
Foreign 3.8 0.9 33.8 8.8 9.3 6.1
shareholding
Average All firms 3.4 1.7 31.2 13.8 6.3 8.4
Source: KPS data, World Bank staff calculations.
Note: figures were calculated using firm sample weights. On average micro have 0–9 employees, small firms 10–49 employees,
medium firms 50–249 employees and large firms 250 or more employees.
Only about 33 percent of workers have tertiary education and about 14 percent are unpaid—a
proxy for labor informality. Large firms received a larger proportion of government subsidies
than smaller firms. There are no significant differences between firms by size in workers’ skills,
although the share of tertiary workers is higher in larger firms (36.5 percent). The share of unpaid
workers reaches 15 percent among micro firms, although it is only 1.7 percent for medium and 1.7
percent for large ones. Although only 1.7 percent of all firms reported receiving subsidies, among
these more beneficiaries were large firms.
Firm Characteristics and Dynamics 19
Firms with foreign shareholders tend to have more skilled workers and fewer unpaid workers
but the share of companies investing in R&D is similar to domestic-owned companies—close to
4 percent (Table 1). About 9 percent of companies with foreign shareholders invest in employee
training compared to just 6 percent in solely domestic firms. In contrast, investment in management
and supervision is higher in enterprises with a foreign shareholder. Those companies also have a
slightly larger proportion of skilled workers (34 vs. 31 percent of domestic-owned firms), and far
fewer unpaid workers (8.8 vs. 14 percent).
In only 9 percent of the firms is the decisionmaker female15 (Figure 10A). Sectors with the
highest proportion of female decisionmakers (Figure 10B) are administrative services 15 percent;
professional services 10 percent; and W&R 10 percent. ICT, accommodation and food services. and
administrative and support services employ the largest shares of female workers, about 30 percent
each, followed by W&R 15 percent and professional services 10 percent. Mining, electricity and gas,
and construction have the fewest female workers (Figure 10C).
Figure 10.
Gender of Firm Decisionmakers by Sector, Percent
a) Gender of firm decisionmaker Percent of total firms
91% men women 9%
b) Firms with female decisionmaker by sector c) Female employees in 2017
Administrative and support act. 14.7% Information and communication 33.1%
Professional, scientific and technical 10.4% Accommodation and food services 31.8%
Wholesale and retail trade 10.3% Administrative and support services 30.3%
Real estate activities 27.1%
Real Estate 9.3%
Wholesale and retail trade 26.2%
Manufacturing 8.7% Professional, scientific
and technical activities 23.7%
Accommodation and food service 7.7%
Manufacturing 18.7%
Information and communication 6.9% Transportation and storage 17.9%
Transport and storage 5.6% Water supply; sewerage, waste 10.9%
mgmt. and remediation activities
Construction 3.8% Construction 6.9%
Electricity, gas, steam and
Utilities 1.3% air-conditioning supply 6.7%
Mining and quarrying 4.0%
Mining and quarrying 0%
Source: KPS data, World Bank staff calculations.
Note: Figures are weighted using firm sample weights. See Annex 1 for more detail on sector classification.
Female-run companies have a higher proportion of unpaid workers and received proportionately
less in government subsidies. According to the KPS data, firms with female decision-making
have more skilled workers (43 vs. 30 percent in male-run enterprises). In female-run firms about
19 percent of workers are unpaid vs. 13 percent in male-run companies, but the latter receive 1.8
percent of government subsidies and the former 1.4 percent.
Between 2010 and 2018 the proportion of firms in digital-intensive sectors consistently averaged
8.5 percent of the firm population despite a surge in 2017 (Figure 11).16
15 The decision maker is defined in response to the following question in the KPS: “Who makes the main decisions
in the company?” Multiple responses were allowed, including founder, family member of the founder, manager,
members of executive board, and gender.
16 In 2013 and 2017, changes in the number of firms in professional services, information and communication, and
administrative services activities accounted for more than 90 percent of the variation in the share of digital-
intensive enterprises.
20 Firm Characteristics and Dynamics
Figure 11.
Digital Firms, 2010–10, Percent of Total Firms
10 9.70
9.41
9.07
8.60
8.30
7.89
8 7.61
7.15 7.26
6
4
2
0
2010 2011 2012 2013 2014 2015 2016 2017 2018
Source: KAS data and Calvino et al. 2018; World Bank staff calculations.
Notes: Digital-intensive sectors are those with a high level of digitalization.
Kosovo’s economy has relatively high and persistent levels of concentration. For the economy as a
whole and proxied by the normalized Herfindahl Index (HHI) of sales, since 2013 concentration has
been declining only marginally (Figure 12A). By sector, economic concentration is highest in electricity
and gas (close to 1), followed by information and communication, and mining (each 0.53) and financial
services (0.43), and is lowest in accommodation and food services, construction, and W&R (Figure 12B).
Figure 12.
Concentration in the Kosovar Economy
a) Herfindahl Index, 2010–18 b) Herfindahl Index by sector, 2018
0.35 Electricity, gas, and related act. 0.92
Information and communication 0.44
0.30 0.38
0.280 Mining and quarrying
0.268 0.274 0.262 0.263
0.251 0.253 0.245 Financial services 0.27
0.25 0.246
Real estate activities 0.25
0.20 Administrative and support services 0.24
Water supply; sewerage, waste 0.15
manag. and remediation activities
0.15 Manufacturing 0.13
Wholesale and retail trade 0.11
0.10 Professional, scientific and 0.09
technical activities
Transportation and storage 0.05
0.05
Construction 0.04
0.00 Accommodation and food services 0.01
2010 2011 2012 2013 2014 2015 2016 2017 2018 0 0.20 0.40 0.60 0.80 1.00
Source: KTR; World Bank staff calculations.
Note: The normalized HHIS is 1 minus the squared sum of market shares of firm sales at four-digit level NACE Rev. 2. Shares are weighted by total
sector sales. The index ranges from 0 to 1, where 1 is the highest level of economic concentration. See Annex 1 for more detail on sector classification.
Firm Characteristics and Dynamics 21
1.2. Firm Dynamics
Firm dynamics have a crucial bearing on productivity growth. According to economic theory and the
available empirical evidence (Maloney and Cusolito, 2017), entry of new more productive firms, exit of
companies with lower productivity or the increase in the average productivity of incumbent firms are
key drivers for productivity growth in any economy. Thus, this section delves into the recent dynamics
of firms in Kosovo with the aim of setting the background for the analysis of firm productivity presented
below in Section 1.3.
For 2012–18 Kosovo’s net rate of firm creation was close to zero because firm entry and exit rates
have been similar. Although the firm entry rates are also similar to some of the most dynamic aspirational
economies (e.g. Lithuania and Latvia), it has much higher firm exit rates than the EU average and tracks
some of the most developed aspirational peers, resulting in an average net annual firm turnover rate of
just 0.3 percent (Figure 13A). During this period, only five sectors posted positive net firm turnover rates:
manufacturing 3.12 percent, mining 3.10 percent, utilities 3 percent, accommodation and food services
2.5 percent, and construction 2.3 percent (Figure 13B). As expected, entry and exit rates are much higher
for micro and small firms than for larger companies (Figure 13C).
Figure 13.
Net Firm Turnover, Entry. and Exit Rates, Percent of Firms, Average 2012–18
a) Entry, Exit, and Net Firm Turnover in Kosovo and Aspirational Peers
25 Entry rate
% of firms
21.2 Exit rate
20 Net turnover rate
17.3
15 15.0
13.5 13.2 12.8
10.6 10.0 11.1
10 8.3 9.2 8.0 8.9 8.8 8.8 8.3
5 3.9 4.4
2.8 2.8
0.3 1.2 0.0 0.5
0
Lithuania Latvia Kosovo Slovakia Estonia EU (2020) Croatia Czechia
b) Net Firm Turnover Rate by Sector
Manufacturing 3.1
Mining and quarrying 3.1
Utilities 3.0
Accommodation and food services 2.5
Construction 2.3
Average 0.3
-0.3 Real estate activities
-0.7 Information and communication
-0.7 Wholesale and retail trade
-1.4 Transportation and storage
-1.8 Administrative and support services
-1.9 Professional, scientific and technical activities
-3.0 Financial services
-4 -3 -2 -1 0% 1 2 3 4
c) Entry and Exit Rates by Firm Size, Percent of Total Active Firms
16 Entry Exit
14 14 13
12
10
8
6 5
4 4
2 2 2 2 2
0%
0-9 10-49 50-249 250+
Source: KTR and Eurostat data; World Bank staff calculations.
Note: Entry and exit rates are total active units that enter and exit the market divided by total active enterprises. A unit enters the market in t if it
is active in t and it was not active in t-1 and t-2; and exits the market i if it is active in t and it is not active in t+1 and t+2. EU entry and exit rates are
based on the same definitions and criteria. Activities: B-N (excluding K 64.2) of NACE Rev. 2. See Annex 1 for more detail on sector classification.
Micro are firms with an average of 0–9 employees, small 10–49 employees, medium 50–249 employees, and large 250 and more.
22 Firm Characteristics and Dynamics
Kosovo’s new business density is slightly above what its income per capita would predict but
lower than in some comparators and aspirational peers. Figure 14 plots new business density, a
proxy for firm entry used because international data are available and comparable, against income
per inhabitant adjusted for purchasing power parity (PPP). It shows that the number of new firms
per thousand people is far higher than Kosovo’s GDP per capita would predict. Nevertheless, some
Western Balkans and aspirational peers have higher new business density indicators. Unfortunately,
there are no internationally comparable data on firm exit rates.
Figure 14.
New Business Density and GDP per Capita in Constant US$ PPP, 2018
30 New registrations per 1,000 people aged 15-64
Kosovo
25
Comparators
20 Aspirational peers
Upper-middle income countries
15 High income countries
10
R² = 0.1546
5
0
8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0
GDP per capita
(PPP, in logs)
Source: Entrepreneurship Survey and database, http://www.doingbusiness.org/data/exploretopics/entrepreneurship
Administrative barriers to exit are high. Figure 15A shows the difference between the number
of firms that terminate their legal registration and the number that stop reporting sales, the latter
being the standard economic definition of firm exit.17 Between 2010 and 2018, 3,400 companies
“exited” annually on average, whereas fewer than 200 firms formally deregistered. The discrepancy
between economic entry and administrative registration rates is far less pronounced, suggesting
fewer major regulatory obstacles to creating new companies (Figure 14B).
Figure 15.
Economic and Administrative Firm Entry and Exit, Number of Firms, 2012–18
a) Entry b) Exit
Economic Economic
Administrative Administrative
6,000
5,426
6,000
5,000
4,410
5,000
4,320
3,982 3,907 4,009
3,601 3,739
3,449
4,000
3,422
4,000
3,419 3,368 3,285
3,056 3,238
2,809 2,913 2,784 2,971
2,774
2,508
3,000 3,000
2,000 2,000
1,000 1,000
200 388 323
141 158 192 191
0 0
2012 2013 2014 2015 2016 2017 2018 2010 2011 2012 2013 2014 2015 2016
Source: KTR; World Bank staff calculations.
Note: Activities: B-N (excluding K 64.2) of NACE Rev. 2. Administrative defines entry and exit using the firm’s registration year in the Tax
Agency registry; Economic defines entry and exit according to firm sales. See Annex 1.A for more detail.
17 See Annex 1 for further technical details on the economic definition of firm entry and exit rates.
Firm Characteristics and Dynamics 23
Entrants are much larger than firms that exit but smaller than incumbent firms. New firms are
only about one-third of the size of incumbent companies and about twice as large as firms exiting.
The relatively large size of firms at entry compared to companies already in the market (Figure 16)
could be related to high barriers to entering the market, perhaps connected to informal competition
(see World Bank 2018).
Figure 16.
Size of Firms Entering a Market Relative to the Average Firm Already Active, 2012–18, Percent
Entry
39
32
Exit
21
22
Incumbent
121
129
Source: KTR; World Bank staff calculations.
Notes: The average number of employees and turnover (in real 2018 euros) are calculated as the total number of employees and turnover
of each category divided by the number of firms. Activities: B-N (excluding K64.2) of NACE Rev. 2. See Annex 1 A for definitions of entry,
exit, and incumbent.
The change in entry rates was driven by the increasing participation of sectors with higher
rates of firm creation; the change in exit rates was driven by a rise in the average exit rate for all
sectors. More than 90 percent of the change in the entry rate between 2012 and 2018 was explained
by an increasing share of sectors with higher firm entry rates (“between-sectors”). In contrast, most
of the change in the aggregate exit rate resulted from a generalized increase in exit rates across all
sectors (“within-sectors”). Moreover, the change of about 1 percent in the entry rate was far lower
than the 0.72 percent change in the exit rate (Figure 17).
Figure 17.
Drivers of Entry and Exit Rates, Percentage Points
0 Percentage point 1
Within-sectors
Entry rate change Between-sectors
(2012-2018) Cross-change
0.09 0.64 +0.72
Change
Exit rate change
(2010-2016)
-0.04 1.01 +0.97
Source: KTR data; World Bank staff calculations.
Notes: The figure depicts changes in entry rates due to variation “within-sectors”, changes in the participation of industries with different
level of dynamism (“between-sectors”), and the covariance between changes in sector shares and level of firm dynamism (“cross-
change”). Activities: B-N (excluding K64.2) of NACE Rev. 2.
24 Firm Characteristics and Dynamics
Firm size at entry is lower than would be predicted by Kosovo’s income per capita and in
aspirational peers but similar to Western Balkans comparators. New firms experience most of
their employment growth in the first year, with growth extremely low thereafter. Firms enter with
an average of 3.3 employees, jump to 4.6 workers in their first year, then add workers very slowly
(Figure 18A). Moreover, entrants have a smaller median size than would be predicted by income
per capita and the aspirational economies, although similar to most Western Balkans comparators
(Figure 18B).
Figure 18.
Firm Size at Entry
a) Median number of employees after entry by firm age*
Employees
7
6 5.8 6.0
5.5
5.0
5
4.6
4
3.3
3
2
1
0
0 1 2 3 4 5 Age
b) Median number of employees at entry and GDP per capita in US$ PPP. 2018
25 Median size (employees) Kosovo
Comparators
20 Aspirational peers
Rest of the countries
15 R² = 0.3907
10
ALB
KOS MKD
5
SRB MNE
BIH
0
6 7 8 9 10 11 12
GDP per capita (PPP constant %, log)
Source: KTR and Enterprise Surveys data; World Bank staff calculations.
Notes: Average entry size and average size of active firms are the total number of employed persons divided by the number of total active
unit of each category. Activities: B-N (excluding K64.2) of NACE Rev. 2.
Over a five-year period, 33 percent of firms are born as and remain micro enterprises, 44.5
percent of the micro enterprises continue to be micro, and other firm classes similarly lack
dynamism. Focusing on the cohorts of firms in 2012 and 2013, Table 2 shows that about 58 percent
of small, 64 percent of medium, and 74.4 percent of large enterprises stay in the same class over a
five-year period. Only about 2 percent of micro enterprises became small firms, 5.5 percent of small
become medium, and 4.6 percent of medium firms become large.
Firm Characteristics and Dynamics 25
Table 2. Five-Year Transition Matrix by Firm Class, 2012 and 2013 Firm Cohorts, Average
Size class five years after the year of reference
Micro Small Medium Large Economic Exit
Size class at reference year (0-9) (10-49) (50-249) (250+)
Births 33.7 2.6 0.4 0.1 63.2
Micro (0-9) 44.5 2.3 0.0 0.0 53.2
Small (10-49) 17.9 57.7 5.5 0.0 18.9
Medium (50-249) 5.0 12.9 63.6 4.6 13.9
Large (250+) 3.5 0.0 8.1 74.4 14.0
Source: KTR data; World Bank staff calculations.
Notes: This table shows the share of firms by size class in the year of reference and five years later; we use data for the 2012 and 2013
firm cohorts to identify entrants and incumbents and track changes in the number of employees over a five-year period. To control for
potential idiosyncratic shocks in a particular year, calculations are made by taking the average between the two firm cohorts. Year of
reference refers to the first year the firm enter the panel, either 2012 or 2013. Economic exit are classified as those units that report
positive turnover in year of reference but do not report turnover in the next two years. Size class is defined according to the number of
employees of the firm in the year of reference, as noted in parentheses.
Firm mortality is higher for smaller companies and a notable proportion of small and medium
enterprises become smaller over time. In line with the international evidence (Hsieh and Klenow
2014; Cusolito and Maloney 2019) closure rates are highest among new firms (63 percent) and micro
enterprises (53 percent) and lowest among medium and large firms (about 14 percent). Moreover,
about 18 percent of small companies become micro-sized and about 13 percent of medium become
small (Table 2).
Micro firms have the lowest employment growth rates; small, and especially medium, firms
show higher and more stable growth. In line with the international evidence (Hsieh and Klenow
2014), younger firms have higher employment growth than older firms, especially in their first two
years after birth (Figure 19A). Over time micro firms consistently have lower employment growth
than small and medium enterprises, and their growth rates may turn negative or close to zero in 10
years (Figure 19b).
Figure 19.
Firm age and Employment Growth
b) Firm Age and Employment Growth by Number of Employees at Birth
a) Firm Age and Employment Growth (Percent)
(Size Class)
10% 35% Micro (0-9)
Employment growth (%) Employment growth (%)
Small (10-49)
30%
Medium (50-249)
25%
5% 20%
15%
10%
0% 5%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Firm age Firm age
Source: KTR data; World Bank staff calculations.
Notes: Annual growth rate is calculated as the number of employed persons in the period of reference relative to the number employed
in the year before. To increase sample size and analyze the relationship between firm age and employment growth more robustly, we
calculate age as the difference of a given year relative to the year the firm was registered. Figure 19B is based on data only available for
five years. Both figures include the entire universe of firms (e.g. incumbents, entrants and exiters).
26 Firm Characteristics and Dynamics
Conditional on surviving five years after their creation, companies born micro (0–9 employees)
and small (10–49) grow faster and more sustainably than those born medium (50–249), which
suggests that there are barriers to growth for larger firms. Figure 20 shows that micro enterprises
are able to increase the number of employees 2.5 times over a five-year period, but small companies
only 2 times. Medium companies increase employment by 1.7 times up to the third year but then
become practically stagnate. Although this declining rate of growth relative to firm size has been
observed in many countries (Syverson 2011; Cusolito and Maloney 2019), in Kosovo it suggests the
presence of restrictions on firm growth beyond a certain threshold of workers.
Figure 20.
Employees by Firm Age and Size over Five Years, Average 2012–18, Index, Entry Year = 1
2.5 Micro
(0-9)
2.0 Small
(10-49)
Medium
(50-249)
1.5
1.0
0 1 2 3 4 5
Source: KTR data; World Bank staff calculations.
Notes: The figure depicts the fraction of persons employed in active units relative to entry, by size class at birth solely including firms
that survived after their first year. Since there are many entry cohorts, we average employment figures across cohorts. Figures in
parenthesis in the x axis labels refer to number of employees.
Only about 74 percent of firms survive the year after they enter the market, 10 percentage
points less than such firms in aspirational peers and the EU. Considering all firms irrespective
whether they reported or not sales in previous years, this difference is also fairly constant over the
next four years of firm operations (Figure 21A). Conditional on surviving in the previous year, firm
survival in the first year after entry, the “valley of death”, is also markedly lower than in the EU and
in the aspirational peers, although it gradually converges in the subsequent years (Figure 21.b).
Figure 21.
Firm Survival in Kosovo, Average 2012–18, Percent
a) Firms that survive by age after entry Percent of firms b) Firms that survive for another year (Conditional on surviving)
100 Kosovo
EU (2020)
84.0 Czechia
Croatia 89.4
84.0 89.0
Slovenia 86.0
70.6
83.6
60.5
73.8
54.7 85.0 84.5
83.6
48.5 83.7
61.8
51.6
73.8
43.9
37.1
0 1 2 3 4 5 Age 0 1 2 3 4 5 Age
Firm Characteristics and Dynamics 27
Figure 21.
Firm Survival in Kosovo, Average 2012–18, Percent
c) Survival rates by firm size (number of employees)
100.0
95.0
89.6
86.1 83.3
95.5 81.9
87.5
80.9
73.2 72.2
61.1 63.2
58.0
Micro (0-9) 50.9
Small (10-49) 43.2
36.5
Medium (50-249) 30.8
0 1 2 3 4 5 6
Source: KTR and OECD Business Dynamic Indicators data; World Bank staff calculations.
Notes: the figure depicts the percent of active units that survive at different ages after entry. Since there are many entry cohorts, we
average the cohort figures. For example, the survival rate at year 1 is the number of units in the reference period (t) newly born in t-1
having survived to t divided by the number of enterprise births in t-1. For selected EU countries, we use 2010-2017 averages. Activities:
B-N (excluding K64. 2) of NACE Rev. 2. While Figure 20 a reports the annual survival rate for all firms, irrespective whether they report
sales in the previous years, Figure 20 b calculates the rate of survival per year conditional on the firm reporting sales in the panel in the
previous year
Only about 70 percent of micro firms survive the first year, and in subsequent years their
survival rates are much lower than those of small and medium enterprises. In the first year,
95.5 percent of small and 100 percent of medium enterprises also survive. In years two to five, the
survival differential between micro and medium companies is on average 30 percentage points (pp)
and between micro and large firms 40 pp (Figure 20C).
According to the administrative data employed in this note, between 2010 and 2018 Kosovo’s
economy created about 80,000 new jobs. Between 2011 and 2014 jobs created averaged just under
5,000, and from 2015 up to 2018, jobs created annually averaged 10,000 (Figure 22). Employment
creation did accelerate starting in 2015 and by 2018 had reached 180,600.18
Figure 22.
Thousands of Employees in Active Firms, 2010–18
Thousands of YoY change
employed persons
YoY change
200 40,000
181
169 Total employed
160 145 155 32,000
133 persons
125 130
120 118 24,000
100
18,278
80 13,764 16,000
11,801 9,988 11,827
40 6,436 5,461 8,000
3,058
0 0 0
2010 2011 2012 2013 2014 2015 2016 2017 2018
Notes: Employed persons in active firms (based on sales). Activities B-N (excluding K64.2) of NACE Rev. 2.
Source: World Bank’s own calculations based on KTR.
18 The administrative data broadly coincides with KTR information; according to KSA national accounts data, gross
employment, measured by number of employees, went up by about 50 percent between 2009 and 2019 and 20
percent between 2015 and 2019 (https://askdata.rks-gov.net/PXWeb/pxweb/en/ askdata/).
28 Firm Characteristics and Dynamics
Job creation was partly related to workers being registered by firms that had not reported
employees before. The proportion of active firms reporting employees went from 63 percent in
2014 to 71 percent in 2018 (Figure 23), and new registrations explained about 22 percent of net job
creation. Similarly, Table A1.1 in Annex 1 shows that the average number of employees per firm
practically doubled (from 3 to 6) between 2010 and 2018 but the median number basically did
not vary. Finally, Table A2.4 in Annex 2 also indicates that firms that started reporting employees
experienced faster employment growth in the years thereafter.
Figure 23.
Firms Reporting Employees, Percent of Total Firms, 2010–2018
Employers
Non-employers
49 50 54
60 63
71 69 69 71
51 50 46
40 37
29 31 31 29
2010 2011 2012 2013 2014 2015 2016 2017 2018
Source: KTR; World Bank staff calculations.
Notes: Activities B-N (excluding K64.2) of NACE Rev. 2.
Younger firms with higher productivity and access to credit had higher net job creation rates.
Firms with higher revenues per capita tend to experience higher net job growth rates (Table A2.4.
The magnitude of the coefficient is even larger when controlling for industry (NACE 4 digits),
geographic (municipality), and time (year) fixed effects. The negative sign of the age coefficient
suggests that younger firms experience higher net job creation. Access to credit is also positively
correlated with more rapid employment growth.
By sector, employment growth was mainly driven by W&R and other services and to a lesser
extent by manufacturing and construction. Other services and W&R accounted for about 25
pp out of nearly 35 percent of total job growth in both 2010–14 and 2015–18. Manufacturing and
construction explained between 4.3–4.9 and 2.8–4.8 percent of total job growth, depending on the
period; mining and utilities (electricity, gas, water and sewerage services) accounted for only a
minor share (Figure 24).
Firm Characteristics and Dynamics 29
Figure 24.
Contribution to Employment Growth by Sector, 2010–14 and 2015–18, Percent
2010- 2014 1.35
10.2 13.9 4.3 2.8 33.2
Other services Wholesale and retail Manufacturing Construction and real estate Mining Utilities
2015- 2018
-0.22 13.1 12.3 4.9 4.8 35.6
Source: KTR; World Bank staff calculations.
Note: Activities B-N (excluding K64.2) of NACE Rev. 2.
ICT and professional services are the most dynamic job creators. The data on job creation in the
five years after firm entry reveal that the most dynamic activities are information and communication,
which averages 6.7 employees, and professional services, 4. In industries with large numbers of
workers and contribution to total job creation, such as W&R, construction, and manufacturing, new
firms managed to create only half as many jobs as the more dynamic industries (Figure 25).
Figure 25.
Employment Growth, Firms by Sector Surviving 5 Years, Average Number of Employees, 2017–18
Manufacturing Gap between entry size
2.8
and size at age 5
Construction 2.4
Wholesale and retail trade 2.4
Transportation and storage 3.2
Accommodation and food services 1.4
Information and communication 6.7
Professional, scientific and technical activities 4.0
Employees 0 2 4 6 8 10 12 14
Source: KTR; World Bank staff calculations.
Notes: Difference between t the number of employees and the number of employees five year later conditional on the firm “surviving”
or reporting sales in the previous year. We report activities in which more than 10 companies enter the market. See Annex 1 for more
detailed sector classification.
From 2015 on, annual net job creation doubled, driven by rising job creation and falling job
destruction, almost all by incumbent firms. As a result, the economy has been creating an average
of about 10,000 net new jobs annually, about twice the net jobs created annually between 2011 and
2014 (Figure 26A). Rising net job creation among incumbent firms, and to a lesser extent, entry
of new firms, along with relatively stable exit rates, largely explain the rise in aggregate net job
creation (Figure 26B).
30 Firm Characteristics and Dynamics
Figure 26.
Net Job Creation, 2011–18
a) Net change in number of jobs
Employed persons
35,000 32,803
31,393
Job Creation
30,000 28,964
27,725 Net Job Change
25,949
25,000 22,606 23,772 Job Destruction
20,679
20,000
19,566
15,000 18,311 17,621 17,737
16,170 15,200
14,525 14,148
10,000
5,000
0
2011 2012 2013 2014 2015 2016 2017 2018
b) Incumbents entering and exiting
30,000
Incumbent (creation)
20,000 Entry
Net by incumbents
8,639 8,691 9,745
8,066 Exit
10,000 6,705
Incumbent (destruction)
3,090 2,563
18
0
-10,000
-20,000
2011 2012 2013 2014 2015 2016 2017 2018
Source: KTR data; World Bank staff calculations.
Note: Net job change equals gross job creation minus gross job destruction. Activities: B-N (excluding K64.2) of NACE Rev. 2. See Annex
1.A for detailed definitions of incumbent, entry, and exit.
Young SMEs are driving job creation. Although they contribute just 25 percent of total employment,
young SMEs accounted for nearly 40 percent of the jobs created annually between 2016 and 2018. As
firms mature, they lose some dynamism. Even among small and medium companies, mature and old
firms explain 50 percent of total employment but only one-third of gross job creation. Furthermore,
large, mature firms contribute less than proportionately to gross job creation (Figure 27).
Firm Characteristics and Dynamics 31
Figure 27.
Job Creation, Destruction, and Employment by Firm Age and Size, Average 2016–18
Contribution to job creation
52 51 Contribution to job destruction
Contribution to employment
41
36
30
24
22 21
16
2 2 3
Young (1-5) Mature and Old (>5) Young (1-5) Mature and Old (>5)
SMEs (1-249) Large (250+)
Source: KTR; World Bank staff calculations.
Notes: Using the same criteria as Criscuolo, Gal, and Menon 2014 we exclude entrants (0 years). Young are companies 0–5 years of
age; mature and old firms have lasted more than five years. Activities: B-N (excluding K64.2) of NACE Rev. 2. Micro firms average 0–9
employees, small 10–49, medium 50–249, and large 250 and more.
In 2010–18 young firms and startups, including entrants, accounted for about 45 percent of total
employment. On average, startups accounted for about 33 percent of total employment, young firms
for about 16 percent, and mature firms for the rest. However, there is considerable heterogeneity by
sector. Setting aside entrants, the proportion of jobs in younger firms is largest in accommodation
and food services, 74 percent, and professional and administrative services, about 60 percent each.
In contrast, about 95 percent of workers in financial services and water supply and 90 percent in
mining are employed by mature firms (Figure 28).
Figure 28.
Employment by Firm Age, Percent of Total, Average 2016–18
Accommodation and food services Entrants (0 years)
Start-ups (1-2)
Administrative and support services
Young (3-5)
Construction Mature and old (6+)
Electricity, gas and related act.
Financial and insurance services
Information and Communication
Manufacturing
Mining
Professional, sci. and tech. act.
Real Estate
Transportation and Storage
Water supply, sewerage and waste mgmt.
Wholesale and Retail trade
0% 20% 40% 60% 80% 100%
Source: KTR; World Bank staff calculations.
Notes: Entrant firms have less than a year in the market, startups up to 2 years, young up to 5 years, and maturing more than 5 years.
Activities: B-N (excluding K64.2) of NACE Rev. 2. Calculations for this figure do not match exactly the estimates in Figure 26 because
they incorporate entrants and startups and do not consider the contribution of age by firm size class. See Annex 1 for more detailed
sector classification.
32 Firm Characteristics and Dynamics
Firm Characteristics and Dynamics 33
02
Firm Productivity and
Access to Credit
This section first presents the main results of our estimates of TFPR and LP for
2013–17.19 It then examines the drivers of TFPR growth and compares TPFR in
domestic firms, firms with foreign shareholders, and exporters. It also assesses
productivity differences by the gender of the firm decisionmaker. Finally, it delves
into the links between access to credit and productivity and the main barriers to
productivity growth.
19 We exclude 2018 from the productivity analysis because reporting of key variables, such as
investment and purchases of inputs is significantly lower in that year than in previous years.
1.3. TFPR and Labor Productivity
Between 2013 and 2017, TFPR grew by 0.5 percent annually, but LP (value added per worker) fell
by –0.25 percent. Although TFPR growth in Kosovo was negative in 2013 (–0.5 percent), between
2014 and 2017 it grew at an annual average of about 0.7 percent. While LP growth, proxied by value
added per worker, was negative in 2013–15, it was marginally positive in 2016 at 0.17 percent and
2017 at 0.26 percent. During this period proxies for LP, sales per worker, and average wages per
worker, exhibited similar dynamics (Figure 29A). This timid recovery in productivity growth rates
coincided with the acceleration in net job creation (see Figure 21). Between 2012 and 2017, TFPR
grew by 2.4 percent, while LP measured as sales fell by 1.1 percent, as value added per worker by 1.2
percent, and as sales per worker by 0.5 percent (Figure 29B).
Figure 29.
TFPR, Sales, and Value Added per Worker, 2013–17
a) Annual Percent Change
TFPR 0.7% 0.8%
Value added per worker 0.8%
0.6%
Sales per worker
0.0% 0.2% 0.3%
0.1%
0.1%
-0.3%
0.5% -0.2%
-0.5%
-0.9%
-1.0%
2013 2014 2015 2016 2017
b) Index 2013=100
103 TFPR
Value added per worker
102 Sales per worker
Wage per worker
101
100
99
98
97
2012 2013 2014 2015 2016 2017
Source: KTR; World Bank staff calculations.
Notes: TFPR and value added per worker are adjusted for firm entry and exit (Melitz and Polanec, 2015). TFPR is estimated following
Levinsohn and Petrin (2003). Figures are weighted by firm employment in the entire sample dataset (see Box 1 for details). We include
the total business economy, repair of computers, personal and household goods, except financial and insurance activities.
36 Firm Productivity and Access to Credit
Labor productivity in Kosovo is only one-third that of the EU; and it is much lower than in
aspirational countries. Kosovo’s LP (sales per worker) of 33.5 percent of the EU average compares
badly to Lithuania’s 53 percent and the even higher LPs of Slovakia, Czechia, Estonia, and Slovenia
(Figure 30).
Figure 30.
Labor Productivity Relative to the EU, 2018, Percent
89.3
84.9
81.8
73.4
53.2
31.8
Kosovo Lithuania Slovakia Czechia Estonia Slovenia
Source: KTR; World Bank staff calculations.
Notes: Labor productivity is calculated as the ratio of total turnover (sales) in euros to the total number of employees. We include the
total business economy, repair of computers, and personal and household goods, except financial and insurance activities. The figure
depicts LP in each country relative to the EU average LP as percent.
Between 2012 and 2017 use of capital and intermediate inputs by Kosovar firms declined. Figure
31 suggests that during the period TFPR growth was partly driven by that decline, although firm
output decreased to a much lesser extent; firms were able to produce the same output with less
inputs of production. In parallel, the use of labor fell less than proportionately to the use of capital,
suggesting that firms also became more labor-intensive.
Figure 31.
Labor, Capital, and Intermediate Inputs, 2012–17, Index 100=2012
100
Labor
Sales
80 Intermediate
consumption
Capital stock
60
2012 2013 2014 2015 2016 2017
Source: KTR; World Bank staff calculations.
Notes: Labor is calculated as the number of employees times the average wage (wage bill) in order to control for possible heterogeneities
in worker labor productivity. The figure only comprises the sample of firms with data available to estimate TFPR. Activities: B-N (excluding
K64.2) of NACE Rev. 2.
Firm Productivity and Access to Credit 37
Manufacturing is the activity with the highest TFPR—about 26 percent higher than the average
sector, closely followed by retail (24.7 percent) and wholesale (23.2 percent). Construction and
other services are the industries with the lowest TFPR, about 30 percent lower than the average
sector (Figure 32A). Looking at LP, Figure 32B shows wholesale trade has the highest value added
per worker, follow by construction and manufacturing.
Figure 32.
TFPR and value added per worker by Sector and Average
a) TFPR by Sector and Average in logs
TFPR (log) Average
4.0
3.5
3.0
2.5 2.40 2.34 2.37
2.0
1.5 1.31 1.28
1.0
0.5
0.0
Manufacturing Construction Other services Wholesale trade Retail trade
b) Value Added per Worker by Sector and Average in Euros
Value added Average
per worker
16,000
14,421
14,000
12,854
12,000 11,663 10,903 10,412
10,000
8,000
6,000
4,000
2,000
0
Manufacturing Construction Other services Wholesale trade Retail trade
Source: KTR; World Bank staff calculations.
Notes: Labor productivity is calculated as total value added (total sales minus input purchases) divided by the number of employees. Like
the TFPR calculations, LP estimates were adjusted by for firm entry and exit (Melitz and Polanec 2015). Sector averages were calculated
using firm employment weights (see Box 1 for details). Activities: B-N (excluding K64.2) of NACE Rev. 2.
TFPR growth has been driven by market reallocation toward more productive firms. Box
2 explains how the sources of TFPR growth are decomposed in this note following Melitz and
Polanec (2015). Figure 33 shows that in the 2012–17 period about 80 percent of average TFPR annual
growth rate explained by the increasing productivity of incumbent firms (covariance), reflecting a
reallocation of labor and other inputs to the most productive companies (Foster, Haltiwanger, and
Krizan 2001).
38 Firm Productivity and Access to Credit
Figure 33.
Decomposition of TFPR Growth, 2014–17, Percent Change
-0.4% 0.6% 1.5% 1.8% 0.1%
Entry Within Covariance Aggregate Exit
productivity
growth
Source: KTR; World Bank staff calculations.
Notes: TFPR was calculated following Levinsohn and Petrin (2006), adjusted for firm entry and exit (Melitz and Polanec 2015). Covariance
refers to the change in the share and average productivity of incumbent firms; within denotes the change in average firm productivity;
and entry and exit refers to the contribution to aggregate TFPR of firms exiting and entering the market. Activities: B-N (excluding K64.2)
of NACE Rev. 2.
The contribution of firm productivity, typically related to the adoption of technology,
innovation, and improved management techniques, was much lower in Kosovo than in other
developing countries; firm dynamism had only a marginal role. “Within” firm productivity
growth only accounted for about a third of TFPR growth, far less than in other developing countries
for which there is comparable data.20 Underscoring once again the lack of dynamism of Kosovo’s
economy, exit of less productive firms had only a relatively minor, though positive, role, and entry of
new firms contributed negatively to aggregate TFPR growth (Figure 32).
Box 2. Decomposing the Sources of TFPR Growth21
In this note, TFPR is decomposed into different components following Melitz and Polanec (2015) to gauge how much
surviving, entering, and exiting firms contribute to aggregate productivity changes. The contribution of incumbent
firms is decomposed further into the reallocation of resources across firms by the between component (measured by
the covariance term); and increases in productivity within firms.
The between component is positive when resources are allocated from less productive to more productive firms
and the within component is positive when existing firms become more productive due to technology adoption,
innovation, and better managerial skills. Among barriers to reallocation of resources might be trade barriers, excessive
regulations, and overbearing state-owned enterprises.
Kosovo’s economy is producing low-quality entrants: new firms are smaller, pay lower wages,
and are far less productive than the median company. Figure 34 illustrates22 that entrants’ sales
are only 33 percent of a median company’s sales. Similarly, employment of a Kosovar entrant is only
about 40 percent of that of a median company’s. Moreover, entrants pay 30 percent less in wages,
sales per worker is only about 20 percent, and TFPR is 11 percent lower when compared to the
median firm.
20 According to Iacovone et al. (2018), the “within” component accounted for more than 40 percent of the variation
in TFPR growth in their sample of Latin American economies. Similarly, Cusolito and Maloney (2018) found
that the “within” component accounts for more than 33 percent of TPFR variation in their sample of developing
countries.
21 Based on Cusolito and Maloney (2018)
22 It refers to the median LP and TFPR of incumbents, entrants, and exiting firms.
Firm Productivity and Access to Credit 39
Figure 34.
Characteristics of Entrants Relative to the Median Firm. Percent
Employment 38.8
Turnover 32.3
Wages 71.1
Sales per worker 79.3
TFP 88.7
Source: KTR; World Bank staff calculations.
Notes: TFPR was adjusted for firm entry and exit (Melitz and Polanec 2015). TFPR is estimated following Levinsohn and Petrin (2003).
Figures are weighted by employment in the entire sample dataset (see Box 1 for details). Activities: B-N (excluding K64.2) of NACE Rev.
2. (See Annex 1A for more detail of the definitions of incumbent, entry and exit.
There are large and persistent differences in TFPR and LP by firm and sector. For instance,
median TFPR in the 90th percentile is 1.6 times higher than in the 10th percentile. As for LP, the
differences in productivity are even starker: about 90 times between leading (90th percentile) and
laggard firms (10th percentile). By sector, dispersion in TFPR is largest in wholesale trade (1.9 times)
and manufacturing (1.8), but less marked in construction (1.3) and services (1.34). Similar results
are obtained for LP but with larger differences. By firm size, micro companies (0-9 employees)
exhibit the largest distance between leading and laggard firms in LP (about 98.7 times), although
the difference is smaller in TFPR. By contrast, productivity dispersion is largest in medium (50-249
employees) and large (250+) companies with respect to median TFPR than with respect to median
LP (Table 3).
Table 3. Ratio of Productivity in “Laggard” (90th percentile) to “Frontier” (10th percentile) Firms by Sector
and Size, Average 2012-18
Dimension Category TFPR Sales per worker
Sector Other services 1.445 52.5
Manufacturing 1.797 52.1
Retail trade 1.646 65.9
Wholesale trade 1.878 126.7
Construction 1.328 149.7
Size class Micro (0-9) 1.456 98.7
Small (10-49) 1.540 27.4
Medium and large (50+) 1.879 23.1
Sector average 1.595 89.370
Source: KTR; World Bank staff calculations.
Notes: TFPR was calculated following Levisohn and Petrin (2006), adjusted for firm entry and exit (Melitz and Polanec 2015). LP is also
adjusted by firm entry and exit. TFPR and LP dispersion is calculated as the ratio between mean TFPR/LP in the 90th percentile relative
to mean TFPR/LP in the 90th percentile at NACE 2-digit.
40 Firm Productivity and Access to Credit
Micro firms are much less productive than small, medium, and large firms, especially in
wholesale and manufacturing. Median TFPR in small firms is more than double than median
TFPR in micro firms and is 1.3 times larger in large companies. These differences are even more
pronounced in wholesale, in which median TFPR in medium and large firms is 1.6 times higher than
in micro firms, and in manufacturing, where median TFPR in medium and large firms is 1.5 times
higher than micro firms; the productivity gap between micro and other firm classes is smaller in
construction (1.03) and other services (1.07) (Table 4).
Table 4. TFPR in Micro-Enterprises Relative to Other Firm Classes
Sector Size class Ratio
Manufacturing Small 1.136
Medium and Large 1.467
Construction Small 0.999
Medium and Large 1.038
Services Small 0.976
Medium and Large 1.070
Wholesale Small 1.147
Medium and Large 1.591
Retail Small 1.132
Medium and Large 1.407
Average Small 1.078
Medium and Large 1.315
Source: KTR; World Bank staff calculations.
Notes: TFPR was calculated following Levisohn and Petrin (2006), adjusted for firm entry and exit
(Melitz and Polanec 2015). LP iis also adjusted by firm entry and exit. Dispersion is the average
logged difference between TFPR in micro firms and the other firm classes. The TFPR ratio equals
elnT(FPRm) -ln(TFPRj) where ln(TFPRm) is the logged median TFPR in m micro-firms and ln(TFPTj)
is the logged median in j, the other firm classes.
Higher TFPR is correlated with higher employment, revenues per worker, and to a lesser
extent, value added per worker. The simple correlations plotted in Figure 35 indicate that more
productive firms have far more workers and sales per worker, and to a lesser extent, more value-
added per worker. Table A2.5 in Annex 2 shows that these correlations are robust to controlling for
firm size and capital stock as well as sector, location (municipality), and year fixed effects.
Figure 35.
TFPR, Employment, Revenue, and Value Added per Worker, Logarithms, 2012–18
TFPR and Employment (log) TFPR and Revenue per worker (log) TFPR and VA per worker (log)
TFPR (log) relative to average
1.5 1.5 1.5
1 1 1
.5 .5 .5
0 0 0
-.5 -.5 -.5
-4 -2 0 2 4 6 -5 -2.5 0 2.5 5 -3 -2 -1 0 1 2 3
Employment (log) relative to average
Source: KTR; World Bank staff calculations.
Notes: TFPR, employment, revenue, and value added per worker (in logs) relative to industry-by-year average. Activities: B-N (excluding
K64.2) of NACE Rev. 2.
Firm Productivity and Access to Credit 41
Wholesale was the activity that had the highest TFPR growth rate in 2013–17. Between 2013 and
2017, TFPR in wholesale grew by 6.55 percent, followed by retail (5 percent) and construction (3.26
percent). In contrast, TFPR in manufacturing fell by –2.2 percent, but growth was also low in other
services (0.6 percent). During that period, wholesale also had the highest annual average TFPR
growth, 1.3 percent, almost triple the economy’s average of 0.5 percent, followed by construction (1
percent). Manufacturing had negative average annual TPFR growth of –0.5 percent), and for other
services at 0.7 percent it was just above the average (Figure 36A).
Figure 36.
Total Factor Productivity Revenue (TFPR) by Sector, 2013–17
a) Year-on-year percent change
4.0% Wholesale
Construction
Retail
3.0% Manufacturing
Other services
2.0%
1.0%
0.0%
-1.0%
-2.0%
-3.0%
2013 2014 2015 2016 2017
b) Contribution to TFPR Growth by sector, percent change, 2013–17
1.0% 0.03
0.07 0.82
0.74 0.37 0.75 0.03
0.24 0.10
0.13 0.56 0.06
0.08
0.5%
0.26 0.36
0.26
0.66
0.11 0.07 0.19
0.20 0.27
0.13 0.11
0.0% 0.02
-0.53 -0.24
-0.36 Manufacturing
Construction
-0.50 Other services
-0.5% Wholesale
-0.23 Retail
Aggregate productivity growth
-1.0% 2013 2014 2015 2016 2017
Source: KTR; World Bank staff calculations.
Notes: TFPR is calculated following Levisohn and Petrin (2006), adjusted for firm entry and exit (Melitz and Polanec 2015). See Box 1 and
Annex 1 for details. Activities: B-N (excluding K64.2) of NACE Rev. 2
Wholesale and retail were the main engines of aggregate TFPR growth between 2013–17 (Figure
36B). Wholesale accounted for about 50 percent and retail for 40 percent of yearly aggregate
productivity growth. Construction explained another 17 percent and other services 5 percent, but
manufacturing dragged down aggregate TFPR performance with negative growth in 2013 and 2015
and only marginal positive contributions in the other years.
42 Firm Productivity and Access to Credit
Firms engaged in international trade are larger, more productive, more capital-intensive, and
pay higher wages than domestically oriented companies. Firms that export and import have on
average more employees and sales, higher value-added, and higher capital stock and wages per
worker (Table A2.6 in Annex 2). Similar results, though with less statistically significant coefficients,
are obtained for firms engaged exclusively in either exporting or importing. These results are robust
to controlling for time, municipality, and industry fixed effects.
Firms in which more than 50 percent of shareholders are foreign are also larger, more
productive, more capital-intensive, and pay higher wages (Table A2.8). In contrast, firms with
lower foreign-shareholding are statistically indistinguishable from domestic-owned companies
except for higher wages and capital per employee. It appears that in Kosovo managerial control by
foreign investors rather than just foreign shareholders might be correlated with higher productivity.
Results are robust even when time, municipality fixed effects and firm size are controlled for.
Firms with female decisionmakers, especially in manufacturing and W&R, tend to be smaller,
less productive, and more labor-intensive (Table A2.8). Female-run companies have 20 percent
fewer employees than male-run firms and 57 percent lower total sales. These differences are much
larger for manufacturing and W&R enterprises, though not statistically significant for other sectors.
Sales per worker are on average 10 percent lower in female-run firms in general, and about 30
percent lower in female-run manufacturing companies. Moreover, female-run companies average
30 percent less capital per worker, with larger differences in W&R, manufacturing, and professional
services.
1.4. Access to Credit and Productivity
Limited access to finance can affect firm productivity in several ways. It can prevent hiring
workers, innovation, and investment. On the other hand, it can induce exit of less productive
firms, which may benefit aggregate productivity growth. Credit constraints can discourage entry
of potentially more productive new firms, which may also induce a fall in total productivity growth.
Finally, restricted access to credit may have reallocation effects, by forcing less productive firms to
shrink and by restricting expansion of more productive companies (Linarello et al. 2019)
This section analyzes the links between access to credit and TFPR based on detailed micro data
from the KCR, combined with firm-level information from the KTR (see Box 1 and Annex 1 for
details). It first examines the recent evolution and sector distribution of credit to the private sector.
It then explores two dimensions of the credit-productivity nexus: (1) what determines access to
credit; and (2) the links between access to formal finance and firm TFPR growth.
Between 2010 and 2018, domestic credit to the private sector in Kosovo increased by 24 percent,
from 35.4 to 44.1 percent of GDP. Despite this positive performance, other comparator countries,
such as the Slovak Republic, Estonia, and Croatia, registered much higher and faster growth in their
credit-to-GDP ratios (Figure 37).
Figure 37.
Domestic Credit to the Private Sector, Percent of GDP
% GDP Kosovo
95 Serbia
Estonia
Croatia
85
Slovak Republic
Montenegro
75 Latvia
65
55
45
35
25
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Source: WDI 2021.
Firm Productivity and Access to Credit 43
Less than a third of firms have access to financing in Kosovo through formal channels.
The activities with the highest proportion of firms with that access are mining 47 percent, and
construction 42 percent; the industries with the least access to credit are accommodation and food
services and professional services, all about 20 percent (Table 5).
Table 5. Credit in Kosovo, Summary Statistics, 2018
Sector Firms Accessing Loans per Firm Total Loans Share of Total Nominal Interest
Credit (%) (¤‘000) (¤ million) Loans (%) Rate (pp)
Mining and quarrying 47.7 163.2 10.8 0.9 5.0
Manufacturing 32.5 139.9 198.7 18.5 4.9
Electricity, gas, steam and air-conditioning supply 28.4 449.6 5.0 0.4 2.7
Water supply; sewerage, waste mgmt. and 38.3 59.9 2.9 0.3 5.1
remediation act.
Construction 42.3 195.6 171.0 15.0 3.2
Wholesale and retail trade 25.8 144.4 575.1 52.2 4.3
Transportation and storage 34.1 61.4 24.1 2.1 6.9
Accommodation and food services 20.1 46.9 31.2 2.9 6.1
Information and communication 25.2 154.5 25.2 2.4 5.9
Financial services 34.3 126.1 5.1 0.4 3.0
Real estate activities 14.5 319.7 3.3 0.3 3.5
Professional, scientific and technical activities 21.2 61.2 17.2 1.5 5.4
Administrative and support services 33.6 189.2 31.2 3.1 3.6
All firms 27.7 135.6 1,100.9 100.0 4.3
Source: KCR data; World Bank staff calculations.
Note: Firms not reporting location (municipality) were excluded from the sample.
More than 50 percent of banking credit is directed to W&R, followed by manufacturing (19
percent) and construction (17 percent). No other activity absorbs more than 3 percent of total
financing. Accommodation, food services, and transportation have the lowest share of credit (Table
5). Credit to the private sector is therefore allocated to the economic activities with the largest
shares in total employment (Figure 38). Qualitatively similar results are obtained when plotting the
sectoral distribution of credit against the share of each activity in total sales or value added.23
Figure 38.
Credit Disbursed and Employment by Sector, 2010–18, Percent of Total
60 Wholesale &
retail trade
50
40
Mining and Transportation & Construction
30 quarrying storage
20 Manufacturing
10 Real estate Administrative and support services
Information & communication
0 Accommodation & food services
Water supply; sewerage, Financial services
waste mgmt. and Electricity, gas, steam and rel. act.
remediation act. Professional, sci. & tech. act.
-10 -5 0 5 10 15 20 25 30 35
Source: KCR data; World Bank staff calculations.
23 Results not presented here in lieu of space but are available upon request to the authors.
44 Firm Productivity and Access to Credit
By type of financial instrument, loans account for more than half of credit to the private sector.
Overdrafts account for another 20 percent, followed by guarantees with 17 percent. Other forms
of financing account for the remaining 11 percent, led by other credit assets at about 5 percent and
letters of credit at about 3 percent. (Figure 39). The relative share of guarantees in total credit to the
private sector has been relatively constant since 2015, averaging 15 percent.
Figure 39.
Credit to the Private Sector by Type, Average 2010–18, Euros and Percent of Total
Loan Overdraft Other
52.8 22.3 credit
assets
4.3
Guarantees
16.5
Source: KCR data; World Bank staff calculations.
Note: Firms not reporting municipality were excluded from the sample.
In Kosovo, 70 percent of credit to the private sector consists of short-term loans: 44 percent
have a term of up to one year, 13 percent 1–2 years, and 12.7 percent 2–3 years (Figure 40). Clearly,
long-term finance, which is typically linked to productivity-enhancing investments, is scarce in
Kosovo.
Figure 40.
Loans to the Private Sector by Term (Years), Average 2015–18, Percent of Total
Up to 1 year inclusive 5 years or more 2-3 years 4-5 years
43.9 13.4 12.7 10.4
1-2 years
13.0
3-4 years
6.7
Source: KCR data; World Bank staff calculations.
Note: Firms not reporting municipality were excluded from the sample.
Micro enterprises, firms oriented to the domestic market, and startups are particularly affected
by the lack of access to credit. According to the latest BEEPS, 74 percent of SMEs in Kosovo report
having needed a loan but being either rejected or discouraged from applying— the largest share
among Western Balkans countries.24 Only 24 percent of micro firms have access to credit (Table 6).
Moreover, access to financing is 7 pp lower for small enterprises (63 percent) than for medium and
large firms (about 70 percent).
24 See https://www.beeps-ebrd.com/indicators/
Firm Productivity and Access to Credit 45
Table 6. Access to Credit by Firm Class, Average, 2015–18
Characteristic Type Percent of Firms
Size class Micro (0–9) 24.3
Small (10–49) 63.4
Medium (50–249) 72.8
Large (250 and more) 71.2
Age class Start-ups (0–2 years) 21.5
Young (3–5 years; excludes Start-ups) 28.2
Mature (6+ years) 31.2
Foreign trade Non-exporter 26.6
Exporter 57.8
Non-importer 20.4
Importer 53.9
Foreign shareholding No foreign shareholders 45.2
0.1–24.9% of total assets 42.0
25–49.9% of total assets 36.9
50–100 of total assets 54.2
Gender of decision-maker Male-run companies 45.6
Female-run companies 38.1
Sources: Data from Kosovo Tax and Credit Registries and KPS; World Bank staff calculations
Of the factors that affect creditworthiness, startups have the lowest proportion of firms with
access to credit at about 22 percent, followed closely by firms not engaged in foreign trade: only
20.4 percent of non-exporters and 26.6 percent of non-importers were able to secure financing,
compared to 57.8 percent of exporters and 53.9 percent of importers. Less than 40 percent of
female-run companies are acceptable to the financial system compared to more than 45 percent
of male-fun firms. While firms with some foreign-shareholding seem to have similar or even less
access to credit than firms with only domestic shareholders, companies in which foreigners exert
management control with more than 50 percent of the shares have about 10 pp greater access to
finance, 54 percent, compared to 45 percent for firms that have only domestic shareholders.
High performance and larger firms belonging to more concentrated sectors and located in non-
majority municipalities are more likely to get access to credit.25 The determinants of access to
credit are spelled out econometrically in Table A2.10: Note that the probability of obtaining credit
is higher for firms with higher TFPR growth. As expected, the correlation between productivity
growth and access to credit operates with some lag; financial institutions typically grant financing
based on observed firm performance in the past, not on firm productivity growth at the time of
granting the credit. The probability of getting access to finance is higher the more concentrated the
sector the firm is in as proxied by the normalized HHI. It may be that market power drives access
to credit more than firm efficiency. The results also suggest that larger firms also have a 1 percent
higher probability of gaining access to finance. Enterprises in non-majority municipalities are less
likely to obtain credit. Gender of the firm decisionmaker does not have a statistically significant
correlation with firm access to finance. And while firm age is statistically correlated with access to
finance, the magnitude of the coefficient is not economically significant.
Access to credit is poorly correlated with higher firm productivity, which suggests that in
the local financial system, capital is misallocated. Figure 41 examines the links between credit
allocation and firm productivity in Kosovo by plotting the results of a simple regression analysis of
access to credit with respect to TFPR, controlling for firm size, capital stock, sector, municipality,
sector concentration (proxied by the HHI), and exporter status. The findings suggest that access to
finance is not well correlated statistically with firm productivity.
25 Non-majority municipalities are municipalities with more than 40 percent of Serbian population. See Box 1 for
further details.
46 Firm Productivity and Access to Credit
Figure 41.
Productivity and Access to Credit
60%
Credit access (%) 50%
40%
30%
20%
1 TFPR (log) 1.5 2 2.5 3
Source: KCR and KTR; World Bank staff calculations
Notes: The figures plot access to credit, a dichotomic variable equal to one if the firm was granted any financing and zero otherwise,
against TFPR (in logs) controlling by firm size, capital stock, sector, municipality, and exporter status.
1.5. Barriers to Productivity
Firms reported informality and unreliable electricity supply as the main barriers to
productivity growth (Figure 42 A). One-third of Kosovar firms cited informality as the main
barrier, a considerably higher proportion than in Albania and Croatia. Moreover, 64 percent of
the enterprises in Kosovo cited as a barrier having to compete with an informal firm, the second
largest proportion among Western Balkans countries after North Macedonia. About 25 percent of
companies also cited unreliable electricity supply as a major impediment; that was a problem for
only 8 percent of firms in Albania.26 While tax rates and administration in Kosovo are perceived as
less stringent for investment than in comparator countries, more firms in Kosovo were concerned
about political instability and corruption.
Figure 42.
Barriers to Investment and Exports. Percent of firms
a) Barriers to investment b) Obstacles to exporting to the EU
Informal competitors Lack in international 59
Electricity market linkages
Tax rates Lack of skilled
Political instability 38
workforce
Corruption
Tax administration Negative Kosovo 31
brand reputation
Transport
Access to finance Lack of internal 27
Customs & trade… manufacturing capacity
Poorly educated workers
Licensing and permits Low product quality 17
Labor regulations
Crime, theft and disorder Kosovo Excessive technical 12
Access to land Albania requirements on products
Croatia
Courts Other 8
0% 10% 20% 30% 40% 50%
Source: Enterprise Survey (2019).
Source: Kosovo’s American Chamber of Commerce, “Free Movement of Goods and Conformity Assessment” 2018.
26 See https://www.beeps-ebrd.com/indicators/.
Firm Productivity and Access to Credit 47
Exporters to the EU market cited lack of international links, shortages of skilled workers, and a
negative brand reputation as the main barriers to growth. About 60 percent of respondent firms
identified the lack of international market links as a major obstacle to exporting to the EU, and about
38 percent reported the lack of skilled workers as a major impediment. For instance, the IT sector does
not have enough well-trained programmers, or even graduates with basic IT skills, which prevents
companies from pursuing larger outsourcing contracts (World Bank 2016). Moreover, about 33 percent
of the companies singled out “Negative Kosovo brand reputation” as a formidable barrier to trade with
the EU. Lack of internal manufacturing capacity and low product quality were also named as major
barriers (Figure 41B).
Firms reported not only shortages of skilled workers but also a poor match between training
systems and employer needs. According to STEP 2019, 75 percent of firms that had attempted to fill a
high-skill position and 60 percent firms that tried to fill a medium to lower skill position encountered
problems finding workers with the appropriate skill or experience. Furthermore, nearly 60 percent
of employers think that general education does not give students practical experience, and almost
50 percent think it does not build up-to-date knowledge or soft skills like discipline, timeliness, and
people skills. Although employers are more satisfied with the TVET outcomes, 35–42 percent of
employers think that TVET systems do not provide the skills the market needs.
Infrastructure gaps, too, limit firm productivity growth. Kosovo has invested heavily in new
physical capital. Most main roads and local roads have been upgraded, but many national and regional
roads did not receive much attention. The railway infrastructure is outdated, but only the north-south
section of it merits investment because that connects with European and Asian markets. The mining
sector requires, among other improvements, an upgrade of the railway system (World Bank 2016).
Improving firm productivity will depend on more private investment in R&D, physical capital,
and management training. Firms with higher R&D- and capital stock-to-sales ratios are more
productive in term of sales and value added per worker (Table 7). Investment in training, especially
management techniques, is also correlated with more productive firms. By contrast, firms subsidized
by the government are less productive, as are firms with a higher proportion of unpaid workers in total
workers, though the latter relationship is less statistically robust.
Table 7. Potential Correlates with Productivity in Kosovo
Investment in R&D Capital Stock Unpaid Workers Government Investment in
(Percent of Sales) (Percent of Sales) (Percent of Total Subsidies Training
Workers)
Sales per worker 0.618*** 0.218*** -0.031 -0.978*** 0.293**
(0.147) (0.025) (0.121) (0.233) (0.121)
Value added per worker 0.488*** 0.175*** -0.338* -1.256*** 0.248
(0.182) (0.040) (0.185) (0.342) (0.160)
Source: KPS data; World Bank staff calculations.
Notes: This table shows the results of a cross-sectional regression of sales and value added per worker on each of the covariates,
controlling for firm size, age, gender of the decisionmaker, worker skills, firm capital stock, exporter status, municipality, and sector.
Government subsidies is a dummy variable equal to one if the firm receives any government subsidy and zero otherwise. Investment in
training is a dummy variable equal to one in the firm reports investing in training and zero otherwise.
The Impact of COVID-19 on Firms
This section assesses the economic effects of the COVID-19 shock on Kosovo’s firms based on
the recent World Bank Business Pulse Survey (BPS) and updated administrative data. The BPS
covered more than 100,000 businesses in 49 countries; it was conducted in Kosovo on a sample of
about 1,200 in June and July 2020 to understand the economic impact of the crisis in April 2020
and firm expectations for the next six months.
As in Western Balkans peers, in Kosovo sales dropped in domestic firms affected by lockdown
and MSMEs. Sales eroded by 55 percent in April 2020 relative to April 2019, a slightly larger decline
than in neighbors like Moldova (54.7 percent) and Albania (46.7 percent). For the whole year, based
on tax records, revenues dropped by 9.5 percent. As expected, accommodation and food services,
transportation and storage, real estate, and arts and entertainment suffered the most, and by firm
size, smaller companies had the worst experience based on BPS.
As for jobs, firms largely responded by granting leave and reducing hours and wages, but the
BPS found that MSMEs experienced higher net job losses: About 40 percent of firms granted
leave, 13 percent reduced hours, and 9 percent reduced wages. Only 6 percent fired workers; 2
48 Firm Productivity and Access to Credit
percent hired. Firms in sectors that were shut down experienced net job losses similar to firms
in sectors not shut down and used other means of adjustment, such as reducing hours and wages
less intensively; new job losses were higher in medium and small firms. Consistent with the BPS
findings, tax records also show that formal employment weathered the impact of the downturn. It
appears that wage-subsidy and other policy measures not only helped preserve formal jobs but also
encouraged formalization because employment increased, based on administrative data. However,
unemployment jumped, probably from losses of informal jobs: registering as unemployed was a
prerequisite for benefiting from government social protection measures.
Firms globally integrated through FDI and exports dealt with the shock better than domestic-
oriented firms. Average drops in sales in April 2020 were 43 percent for companies with foreign
shareholders, 50.9 percent for exporters, and 55.2 percent in domestic-oriented firms. The crisis
demonstrated the importance of global integration if Kosovo firms are to become more resilient.
Digital adoption was weaker in Kosovo than in comparator countries, with responses highly
heterogenous by firm size and manager gender. Firm responses focused on increasing use of
platforms (25 percent), and to a lesser extent innovating in products and services (12 percent),
although these reactions were weaker than in Moldova and Albania (Figure 32). Large firms
invested the most in digital solutions and innovation and used digital platforms most intensively. As
did female-run companies.
Figure 43.
Digitalization and Innovation in Response to the COVID-19 Outbreak Compared, Percent of Firms
73
Kosovo
Moldova
Albania
29
25
19 20
12
19
Increased use of Invested in Invested in
digital platforms digital solutions digital solutions
Source: BPS data; World Bank staff calculations.
Note: *Percent of firms that report affirmatively in each category. **Figures may not add up to 100 percent because type of question varies.
More than half of firms in Kosovo are already or are expecting to be in arrears, with SMEs most
financially stressed, slightly more than in Moldova (53 percent) and much more than in Albania (33
percent). Arrears are likely to affect medium-size companies the most. Primary production, services
and construction are the sectors in most financial distress.
More firms in Kosovo reported benefiting from policy measures than in comparator countries;
wage subsidies were the most common measure at end July 2020. At first complex procedures
hindered access to benefits for MSMEs. The number of firms in Kosovo that reported benefiting
from policy measures (65 percent) was much higher than in comparator countries (e.g., 2 percent
in Moldova, and 39 percent in Albania). More than 95 percent of beneficiary firms reported having
received wage subsidies. A smaller proportion of larger firms received wage subsidies than other firm
classes, but large firms benefited more from eased financial conditions (e.g., deferral of payments due
on loans, rent, mortgages, and utilities). Going forward increasing awareness and clarifying eligibility
requirements are likely to increase coverage of future public assistance, especially for MSMEs. Small
firms reported complex procedures as the main barrier to benefiting from policy measures.
Firms in shutdown sectors and companies in majority municipalities received more public
assistance. Of firms in shutdown sectors, 70 percent received public assistance, compared to 63
percent in non-shutdown sectors. While 65 percent of firms located in majority-municipalities
benefited from policy measures, compared to only 50 percent of those in non-majority regions.
Firm Productivity and Access to Credit 49
Conclusions and Tentative
Policy Implications
Kosovo’s economy is dominated by low-productivity micro and small enterprises with low
survival rates. A striking feature is the lack of dynamism in creating new firms: for the last decade
the net rate of firm creation has been close to zero and the country has far fewer firms per million
inhabitants (firm density) than the EU and aspirational peer averages as well as what might be
expected given Kosovo’s income per capita.
Kosovar firms are smaller than EU and aspirational peers, especially medium and large
enterprises. Moreover, Kosovo is also not well integrated into global markets through FDI
and exports. Firms also seem to find it difficult to expand beyond small and medium size, and
survival rates are far lower than in the EU and aspirational economies. Kosovo has a markedly
lower proportion of exporters and firms with foreign shareholders than in the EU, aspirational
peers, and even other Western Balkans economies. The importance of FDI and international trade
for fostering economic growth and employment creation is clear from the fact that importers and
foreign-controlled enterprises on average have more formal employees, are more productive and
capital intensive, and pay higher wages than companies oriented to the domestic market.
Furthermore, in Kosovo productivity trails that of aspirational peers and the EU. For instance,
labor productivity is only one-third of the EU average. In the decade up to the COVID-19 pandemic,
average annual growth rates of labor productivity were negative, and annual growth of TFPR was
relatively low.
Kosovo’s substandard productivity is explained by lower firm productivity and the small and
largely negative contribution of firm dynamics. Confirming the economy’s lack of dynamism, exit
of less productive firms contributed only marginally to aggregate TFPR growth, and entry of new
firms had a negative contribution. Finally, the large dispersion in productivity between “laggard”
and “frontier” firms suggests that there are distortions in product and input markets, especially in
services.
Despite its dreary productivity performance, between 2015 and 2018 Kosovo’s economy
speeded up the creation of jobs, driven by higher job registrations among young SMEs. Higher
firm productivity is associated with higher job creation. In 2015–18, compared to 2011–14 annual
net job creation practically doubled, driven by rising net job creation by incumbent firms, especially
young SMEs. Formal employment growth was also driven by higher worker registration rates:
firms reporting having employees went from about 50 percent in 2010 to about 70 percent in 2018.
Highlighting the centrality of productivity for fostering formal employment growth, higher firm net
job creation was associated with the most productive firms, exporters, and younger and larger firms
with access to credit.
Despite its crucial role in fostering firm growth and productivity and job creation, access to
formal finance is limited to only a third of the companies and mostly to short-term loans. Micro
and small enterprises, startups, domestic-market-oriented firms, companies in less concentrated
economic sectors, and enterprises in non-majority municipalities are hindered by the lack of formal
finance. About 70 percent of loans have a term of no more than three years.
To foster firm productivity, Kosovo needs a multidimensional policy strategy. Based on the main
findings of this and other background notes prepared for the Kosovo CEM, Table 8 organizes this
note’s main policy recommendations along the three main sources of firm TFP growth: (1) growth of
the firm (“within”); (2) market reallocation (“between” firm productivity); and (3) dynamics related
to firm entry and exit.
52 Conclusions and Tentative Policy Implications
Table 8. Main Drivers of Lower TFPR Growth and Recommended Policy Changes
Not Enough “Within” Firm Limited Market Reallocation Lackluster Firm Dynamics
TFPR Sources
Productivity (“Between” Firms) (Entry and Exit)
•• Low private R&D and capital •• Deficient railway and road •• High barriers to entry in
investment infrastructure some services sectors (e.g.,
•• Inadequate firm management •• Informality telecommunications and
practices and room for •• Cumbersome and costly electricity)
improvement in quality procedures for MSMEs to access •• Cumbersome asset liquidation
infrastructure and standards the electricity grid procedures and inadequate
•• Weak numeracy and literacy skills •• Unequal access to financial insolvency laws
Drivers of worker’s not matched with incentives and public •• Lack of access to formal young
employer demands, especially in procurement for publicly owned firms and SMEs access to formal
TVET (Technical and Vocational enterprises (POEs) finance
Educational Training) •• Lack of commercial courts and
•• Poor integration with GVCS via poorly trained legal staff
exports and FDI •• Inefficient, costly, and opaque
•• Limited access to finance for inspection procedures
young firms
•• Devise a strategy to attract •• Prevent allocation of •• Facilitate business registration by:
FDI, strengthen the investment discriminatory financial (a) introducing the silent consent
agency, and create an investor advantages to POEs. principle; and (b) developing an
grievance mechanism •• Ensure equal treatment of effective e-signature system to
•• Foster reforms in TVET and private firms and POEs in public allow full online registration.
general education (e.g., improve procurement processes. •• Streamline asset liquidation
teacher wages and training, and •• Enhance the quality and control procedures.
education infrastructure, and of the regulatory process (e.g., •• Foster the development of
match curricula with firm needs) draft a clear regulation on the commercial courts and provide
•• Strengthen the National interaction between regulators appropriate training to legal staff)
Innovation System (NIS) and and interest groups; and •• Expand the role of microfinance
Proposed Policy
provide management training for strengthen the Regulatory Impact and foster digital-enabled
Changes
MSMEs. Assessment process). business
•• Expand the role of microfinance •• Streamline procedures and •• Promote digitalization and
(e.g., payment services, reduce fees for MSMEs to access technology adoption, especially
remittances, access to equity the electricity grid, draft and in MSMEs and female-run
capital and e-services) and use implement a new and greener companies
of the Kosovo Credit Guarantee energy strategy •• Incentivize piloting programs
Fund. to improve diffusion of good
•• Foster the development of management practices among
Venture Capital Funds targeted MSMEs.
to startups and dynamic young
firms.
Sources: Cusolito and Maloney (2019) and other background notes prepared for the forthcoming Kosovo CEM. See the other background
notes for more detailed actions plans.
Table 8 summarizes some the drivers behind the three sources of firm TFPR growth identified in
this note and the other Kosovo CEM background papers. It also outlines some policy actions to
address the main obstacles to firm productivity in Kosovo. (The background papers provide more
detailed actions plans.)
Conclusions and Tentative Policy Implications 53
This multidimensional pro-productivity strategy should have three main pillars:
• First, the low productivity of Kosovar firms relative to the EU, aspirational economies,
and Western Balkans peers suggest the urgent need to (1) improve firm capabilities related
to human capital skills (e.g., matching TVET curricula to firm demands, strengthen the
general education system to better build basic numeracy and literacy skills), and emphasize
management and organizational practices (e.g., provide training and technical assistance to
MSMEs and strengthen the quality of infrastructure); (2) encourage adoption of technology
and innovation and investment in R&D (e.g., by strengthening the National Innovation
System [NIS]27); (3) facilitate microenterprise access to finance (e.g., among other measures,
expand the role of microfinance institutions and SME guarantee funds); and (4) adopt
strategies to attract FDI and promote exports (e.g., build up the investment agency ) The
country also needs to foster digitally enabled businesses: only 9 percent of firms engage in
digital-intensive activities, and the digital response of Kosovar firms, especially MSMEs, to
the COVID-19 shock trailed other Western Balkans countries.
• Second, the declining contribution to TPFR growth of reallocation of labor to the most
productive firms reinforces the need to remove barriers that prevent movement of capital,
labor, and other inputs and the barriers in product markets related, for instance, to informality
(e.g., the urgency of streamlining registration, inspection, and tax procedures); competition
policy (allowing private firms easier access to financial incentives and public procurement
and improving the quality and control of regulation) unreliable and costly energy supply (e.g.,
streamlining access for MSMEs to the Kosovo Energy Distribution and Supply Company
electricity grid); and inadequate transport infrastructure (e.g., promote investment in some
national and regional roads and upgrade the North-South railway corridor). Kosovo also
needs to improve the gender agenda: only 9 percent of firms are female run and firms with
a female decisionmaker tend to be smaller and less productive, especially in manufacturing
and wholesale and retail.
• Third, the negligible contribution of new firms to TFPR growth illustrates how important
it is to foster entry of highly productive and fast-growing enterprises, just as the negative
contribution of firm exits to productivity growth could be indicating the presence of high
barriers to exit. The barriers already identified (see Table 8) to entry of more productive
enterprises, especially access to formal finance, and the barriers to exit of nonviable firms,
such as simplified and accelerated liquidation procedures, clearly need to be dismantled.
More targeted policies, improved communication, and streamlined application procedures are
necessary to support MSMEs dealing with the economic effects of the COVID-19 economic
crisis. Recognizing the limited fiscal resources, policies should be targeted especially providing
support to displaced workers and minimizing layoffs by firms in financial stress and confronted
by heightened demand or supply shocks. Targeted policies, such as the Kosovo Credit Guarantee
Fund, can be influential in addressing MSME liquidity constraints. Improved communication of
public support measures and simplified application procedures to enhance uptake and reduce the
regulatory administrative burden (especially inspections) not only to encourage local companies
but also to attract FDI are key.
27 See Cirera and Maloney 2019. The NIS incorporates local and international R&D suppliers (universities and
research institutions, among others), government agencies related to R&D and innovation, and private firms
investing in R&D and adopting product, process, and organization innovations.
54 Conclusions and Tentative Policy Implications
References
American Chamber of Commerce in Kosovo. 2018. Free Movement of Goods and Conformity
Assessment. Pristina.
Bloom, Nicholas, Aprajit Mahajan, David McKenzie, and John Roberts. 2010. “Why Do Firms in
Developing Countries Have Low Productivity?” American Economic Review 100 (2): 619–23.
Brown, J. David, Gustavo A. Crespi, Leonardo Iacovone, and Luca Marcolin. 2018. “Decomposing
Firm-level Productivity Growth and Assessing its Determinants: Evidence from the Americas.”
Journal of Technology Transfer 43(6): 1571–1606
Calvino, F., et al. (2018), “A taxonomy of digital intensive sectors”, OECD Science, Technology and
Industry Working Papers, No. 2018/14, OECD Publishing, Paris,
https://doi.org/10.1787/f404736a-en.
Cirera, Xavier, and William F. Maloney. 2017. “The Innovation Paradox: Developing-Country
Capabilities and the Unrealized Promise of Technological Catch-Up.” Overview booklet.
World Bank, Washington, DC.
Cojocaru, A. 2016 Jobs Diagnostic: Kosovo. World Bank. Washington DC.
Criscuolo, C., P. Gal and C. Menon (2014), “The Dynamics of Employment Growth: New Evidence
from 18 Countries”, OECD Science, Technology and Industry Policy Papers, No. 14, OECD
Publishing, Paris, https://doi.org/10.1787/5jz417hj6hg6-en
Cusolito, A. P., and W. F. Maloney. 2018. Productivity Revisited. Shifting Paradigms in Analysis and
Policy. Washington, D.C.: World Bank Group.
Foster, Lucia, John Haltiwanger, and C. J. Krizan. 2001. “Aggregate Productivity Growth: Lessons
from Microeconomic Evidence.” In New Developments in Productivity Analysis, edited by.
Charles R. Hulten, Edwin R. Dean, and Michael J. Harper, 303–63. Chicago and London:
University of Chicago Press.
Haltiwanger, John. 2007. “Measuring and Analysing Aggregate Fluctuations: The Importance of
Building from Microeconomic Evidence.” Federal Reserve Bank of St. Louis Review 79 (3): 55–78.
Hsieh, Chang-Tai, and Peter J. Klenow. 20214. “The Life Cycle of Plants in India and Mexico.” The
Quarterly Journal of Economics 129 (3): 1035–84, https://doi.org/10.1093/qje/qju014
Levinsohn, James, and Amil Petrin. 2003. “Estimating Production Functions Using Inputs to
Control for Unobservables.” Review of Economic Studies 70 (2): 317–41.
https://EconPapers.repec.org/RePEc:oup:restud:v:70:y:2003:i:2:p:317-341
Linarello, Andrea. Andrea Petrella, and Enrico Sette. 2019. “Allocative Efficiency and Finance.
Bank of Italy Occasional Paper No. 487. https://ssrn.com/abstract=3432990 or
http://dx.doi.org/10.2139/ssrn.3432990
Melitz, M. J., and S. Polanec. 2015. “Dynamic Olley-Pakes Productivity Decomposition with Entry
and Exit.” The Rand Journal of Economics 46(2): 362–75.
OECD. 2009. Measuring Capital, OECD Manual 2009: 2nd ed. Paris: OECD Publishing.
https://doi.org/10.1787/9789264068476-en
Syverson, Chad. 2011. “What Determines Productivity?” Journal of Economic Literature
49:2: 326–65.
World Bank. 2018. Country Economic Memorandum: Serbia’s New Growth Agenda. World Bank,
Washington, DC.
———. Forthcoming. “Chapter 4: Human Capital Through Education System Improvement.”
Kosovo: Country Economic Memorandum. World Bank, Washington, DC
55
Annex 1.
Technical Appendix
A. Firm Dynamics
Data sources: Tax Registry from Kosovo Statistics Authority and Kosovo Productivity Survey
Period covered: 2010-2018
Unit-level: Establishment
Methodological definitions and benchmarking: Definitions follow the Eurostat Manual on Business
Demography Statistics (OECD 2007), which ensures international comparability. We also draw on
Eurostat and the OECD, which provide comprehensive and relevant standardized information on
business and employment dynamics: Business Demography Statistics and, the Annual Enterprise
Statistics (Eurostat), and the Structural and Demographic Business Statistics (SDBS, OECD).
Size class: Micro (0–9 employees), small (10–49), medium (50–249), and large (250 and more)
Age class: Entrants (0 years), start-ups (1–2 years), young (1–5 years), and mature and old (6+)
Sector: B-N (excluding K64.2; Activities of Holding Companies in Financial and Insurance Activities)
according to NACE Rev. 2m as followed in Eurostat and OECD manuals unless otherwise stated.
Employed persons: Eurostat defines an employed person as a person aged 15 and over who performs
part-time or full-time work for pay, profit, or family gain.
Active unit: Reported positive turnover in the reference year.
Entry: A unit enters the market in t if it is active in the period of reference but was not active in t-1
and t-2.
Exit: A unit exits the market in t if it is active in t and is not active in t+1 and t+2.
Incumbent: Unit is active in t if it reports activity in t-1, t and t+1.
Table A1.1. Methodological Definitions of Firm Status in Market
Status in t t-2 t-1 t t+1 t+2
Entry No No Yes
Activation No Yes
Exit Yes No No
Inactivation Yes No
Incumbent Yes Yes Yes
Entry rate in t: Sum of all entries in t divided by total active units in the period of reference.
Exit rate in t: Sum of all exits in t divided by total active units in the period of reference.
Turnover rate in t: Sum of entry and exit rates in t.
Incumbent share in t: Difference between gross job creation in t and gross job destruction in t.
Survival unit in t+s: Share of units that entered the market in t and survives s periods ahead,
percent.
Survival rate in t+s: Share of total units that entered the market in t that survived in t+s, percent.
Entry size in t: Number of employed persons of entrant units.
Post-entry growth in t+s: Total employed persons in an active unit in t+s relative to total employed
persons of entrants in t that survived to t+s.
Age: Each firm is given an entry cohort year if the year of entry is known. Age is then calculated as
the difference between the year of reference and the entry year. For firms in the database before
2012, their entry year unknown, that were active throughout the periods of analysis, we assigned a
‘symbolic’ age in 2017 and 2018 in order to carry out the age analysis.
Sector and main sector: Unit activities are classified according to NACE Rev. 2 at the four-digit
level. Because some companies started operations in one sector but then switched to another
activity, we generated the ‘main sector’ variable, which reports the sector in which the firm spends
most of its life cycle.
Gross job creation in t: Sum of all positive unit-level variations of employed persons in active units
in t relative to t-1.
58 Technical Appendix
Gross job destruction in t: Sum of all negative unit-level variations of employed persons in active
units in t relative to t-1.
Net job change in t: The difference between gross job creation and gross job destruction in t.
Total employed persons (total employment): Sum of employed persons of active units.
Gross job creation rate in t: Gross job creation in t divided by total employed persons.
Gross job destruction rate in t: Gross job destruction in t divided by total employed persons,
B. TFPR Estimation
Following Levinsohn and Petrin (2003), we estimate TFPR using semi-parametric methods. We
use this approach as we do not have access to data on firm initial capital stock and only a small
number of firms in the KAS dataset report investment. Hence, we use data on costs of production
(intermediate consumption) reported in the personal income tax and corporate income tax forms
in the KAS dataset.
To estimate TFPR, we take the firm’s total wage bill as the free-standing variable and our estimate of
the capital stock as the state variable. Production costs enter into the production function as a proxy
for capital stock. Because data on worker skills and number of hours worked are unavailable, we
use the total wage bill to control for heterogeneity in labor productivity and labor intensity (hours
worked per year). The total wage bill is the result of multiplying workers’ hourly wages by the
number of hours worked. If production factors are paid according to their marginal productivity,
the total wage bill should account for labor productivity differences across firms (Cusolito and
Maloney 2018).
To estimate the capital stock, we follow OECD methodology (2009) by using data on investment
in durable goods from the KAS dataset, combined with self-reported information on firm assets
provided by the KPS. We compute firm i average investment over time , and estimate initial capital
stock by dividing the average investment by the depreciation rate . Hence, the initial capital stock
of firm i is expressed as follows: . As does OECD (2009), we assume that the depreciation rate is
constant over time and across industries.
Given the relatively small size of our sample, we estimate TFPR for five highly aggregated sectors:
manufacturing (firms classified as letter C in NACE Rev. 2); construction (F and L); other services (H
transport and storage, I: accommodation and food services, J: information and communications, M:
professional, scientific, and technical services, and N: administrative and support service activities);
wholesale trade (division 46: wholesale trade, except of motor vehicles and motorcycles, and class
4531: wholesale trade of motor vehicle parts and accessories); and finally retail trade (division 47:
retail trade, except of motor vehicles and motorcycles, plus activity classes 4511: sale of cars and
light motor vehicles, 4519: sale of other motor vehicles, 4520: maintenance and repair of motor
vehicles, 4531: retail trade of motor vehicle parts and accessories, and 4540: sale, maintenance and
repair of motorcycles and related parts and accessories.
Technical Appendix 59
Annex 2.
Additional Tables and
Figures
Table A2.1. Summary Statistics, Kosovo Statistics Authority
Firms
Year Total (‘000) Micro (%) Exporters (%) Importers (%)
2010 27.918 95.268 - -
2011 30.299 94.789 1.8 18.9
2012 30.866 94.450 1.8 20.5
2013 28.123 93.372 2.3 23.8
2014 28.556 93.059 2.4 24.1
2015 29.041 92.542 2.9 26.3
2016 28.709 91.821 3.4 17.3
2017 30.25 91.537 3.5 17.3
2018 29.441 90.282 4 18.3
Average
29.245 93.011 2.7 20.7
2010-2018
Firm Turnover Wage per worker Sales per worker
Employed persons
(‘000 €) (‘000 €) (‘000 €)
Year Total (‘000) Mean Median Total Mean Median Mean Median Mean Median
2010 100.035 3.583 0 5064265.99 181.398 9.059 4.585 3.521 47.613 16.193
2011 118.313 3.905 1 7967761.229 262.971 10.719 4.763 3.731 61.871 33.341
2012 124.749 4.042 1 8256137.147 267.483 9.818 4.908 3.804 62.089 35.278
2013 130.21 4.630 1 8372110.563 297.696 12.101 4.859 3.809 60.643 35.498
2014 133.268 4.667 1 8028598.981 281.153 12.758 4.738 3.663 58.236 33.589
2015 145.069 4.995 1 8496580.504 292.572 13.642 4.616 3.608 57.606 30.902
2016 155.057 5.401 1 9001941.122 313.558 15.318 4.709 3.757 57.073 31.038
2017 168.821 5.581 1 10042692.799 331.99 16.358 4.854 3.949 57.521 29.667
2018 180.648 6.136 2 10564413.334 358.833 16.989 5.015 3.977 57.526 28.123
Average
139.574 4.771 1 8421611.297 287.517 12.974 4.783 3.758 57.798 30.403
2010-2018
Source: KTR, World Bank staff calculations.
Notes: Active units are those with positive turnover in the relevant period. Activities included: B-N
(excluding K64.2) according to NACE Rev. 2. Wages and sales per worker are weighted by employment.
Monetary values in 2018 constant prices, using Kosovo’s CPI deflator.
62 Additional Tables and Figures
Table A2.2. Summary Statistics. Kosovo Productivity Survey. 2017
Firms Management Workers
Sector Surveyed (units) Share (%) Size (workers) Women (%) W/ tertiary education (%)
Mining and quarrying 25 0.452 9.45 0 18.310
Manufacturing 682 13.960 6.397 8.7 25.114
Electricity, gas, steam and air-conditioning supply 12 0.181 46.077 0 35.733
Water supply, sewerage, waste mgmt. and 33 0.395 30.001 1.9 27.747
remediation act.
Construction 337 8.139 8.095 3.8 28.253
Wholesale and retail trade 1217 50.220 4.248 10.3 31.766
Transportation and storage 154 3.996 7.873 5.6 30.418
Accommodation and food services 366 11.046 6.306 7.7 39.498
Information and communication 149 2.777 11.762 6.9 61.368
Real estate activities 21 0.330 4.198 9.3 36.755
Professional, scientific and technical act. 205 6.183 3.36 10.4 71.657
Administrative and support services 113 2.319 7.094 14.7 42.716
All firms 3314 100 5.654 9 32.925
Foreign
Foreign trade Sales, value added and assets
shareholding
Sector
Unpaid Exporter Importer Share Sales per VA per worker Average assets
(%) (%) (%) (%) worker (‘000 €) stock
(‘000 €) (‘000 €)
Mining and quarrying 0.714 0.6 28.3 0 14.961 4.987 2317.909
Manufacturing 9.231 4.4 36.6 5.29 27.415 21.027 493.623
Electricity, gas, steam 0 1.8 59.2 18.14 31.886 26.562 3555.544
and air-conditioning
supply
Water supply, 1.813 2.4 47.8 3.74 12.973 5.388 2942.46
sewerage, waste
mgmt. and
remediation act.
Construction 3.262 1.4 31.4 10.17 26.576 19.364 307.793
Wholesale and retail 8.650 1 36.4 6.78 25.294 17.705 195.244
trade
Transportation and 5.314 2.4 31.3 9.6 15.281 9.639 155.401
storage
Accommodation and 5.917 0 25.9 3.86 8.378 5.692 88.864
food services
Information and 1.787 2.7 44.7 4.61 28.261 17.696 915.221
communication
Real estate activities 6.977 7.3 11.3 0 28.304 17.164 1244.783
Professional, scientific 8.504 0 29.6 2.36 18.308 14.589 61.055
and technical act.
Administrative and 3.605 0 34.4 2.17 23.84 19.573 353.758
support services
All firms 6.67 1.4 34.4 6.15 22.974 16.526 277.497
Source: KPS, World Bank staff calculations.
Notes: Surveyed units are the effective number of firms that responded the survey. The remaining
variables are weighted using sample weights.
Additional Tables and Figures 63
Table A2.3. Summary Statistics: TFPR Sample, Average 2012–18
Sector Manufacturing Construction Wholesale and retail Services All firms
trade
Firm sin sample 708 622 1,719 676 3,725
Employment 6.6 7.8 4.9 5.3 5.5
Value added 109.2 138.2 90.2 81.5 95.3
Turnover 391.8 526.3 655.3 307.2 513.2
Materials 282.6 388.1 565.1 225.7 417.9
Wage bill 26.1 31.7 20.8 22.5 23.0
Capital stock 715.6 931.4 577.9 358.2 576.3
Investment 84.1 111.4 70.2 42.6 69.4
TFP (log) 2.4 1.3 1.7 1.1 1.6
Value added per 16.2 19.4 17.8 16.9 17.5
worker
Turnover per worker 59.3 83.8 128.4 81.9 99.9
Source: KTR, World Bank staff calculations.
Notes: TFPR is weighted by employment while Employment, VA, Turnover, Materials, Wage bill, Capital
stock and Investment are weighted using firm weights.
64 Additional Tables and Figures
Table A2.4. Net Job Creation and Labor Productivity
Dependent Variable: Net Job OLS FE FE
Creation (1) (2) (3)
Log Turnover 0.399*** 0.684*** 0.684***
(0.066) (0.124) (0.124)
Log Capital stock -0.138** -3.127 -3.127
(0.059) (2.516) (2.516)
Firm age -0.151*** -0.181** -0.181**
(0.018) (0.075) (0.075)
Firm starts to report employees 2.518*** 2.791*** 2.791***
(0.194) (0.265) (0.265)
Exporter status -0.451 -1.078 -1.078
(0.566) (0.896) (0.896)
Access to credit 0.158 0.039 0.039
(0.219) (0.238) (0.238)
Municipality majority -0.877 -0.086 -0.086
(1.178) (6.410) (6.410)
HH Index (four-digit) -1.596 76.464*** 76.464***
(8.660) (22.549) (22.549)
Small (10-49) 1.009*** 1.796*** 1.796***
(0.320) (0.367) (0.367)
Medium (50-249) 7.657*** 13.644*** 13.644***
(1.317) (2.256) (2.256)
Large (250+) 44.678*** 66.635*** 66.635***
(12.921) (20.641) (20.641)
Constant -5.549 30.027 30.027
(9.312) (31.415) (31.415)
Industry effects (four-digit) Yes Yes Yes
Geographic effects (municipality) Yes Yes Yes
Year effects Yes Yes Yes
Firm fixed effects No Yes Yes
R-squared 0.04 0.02 0.02
Number of firms 14,534 14,534 14,534
Observations 66,291 66,291 66,291
Source: KTR, World Bank staff calculations.
Notes: the table presents the results of regressing firm net job creation (log difference between
number of employees in t0 and t1) on a set on covariates, controlling by industry (4 digits), geographic
(municipality) and time (year) fixed effects and using the entire dataset of firms provided by KRT.
Entrants and exiters are excluded from the panel. Micro (0-9 employees) firms are the omitted variable.
Columns (2) and (3) estimate the model using a fixed effects estimator. Robust standard errors. (*),
(**) and (***) indicates significance level at (10), (5) and (1) per cent level.
Additional Tables and Figures 65
Table A2.5. TFPR and Employment, Revenue, and Value Added per Worker (in Logs Relative to Industry-year
Average)
Dependent variable: TFPR (log) OLS OLS OLS
(1) (2) (3)
Log Employment 0.064***
(0.002)
Log Turnover per worker 0.096***
(0.003)
Log Value added per worker 0.064***
(0.003)
Constant 2.223*** 1.383*** 1.779***
(0.017) (0.033) (0.030)
Industry effects (macrosector) Yes Yes Yes
Year effects Yes Yes Yes
Geographic effects (municipality) Yes Yes Yes
R-squared 0.92 0.93 0.91
Observations 6,774 6,774 6,774
Source: KTR, World Bank staff calculations.
Notes: TFPR, employment, revenue and value added per worker (in logs). Robust standard errors. (*),
(**) and (***) indicates significance level at (10), (5) and (1) per cent level.
Table A2.6. Exporting and Importing Premiums, Kosovo
Log Employment Log Turnover per Log Turnover Log Capital per Log Wage per
worker worker worker
(1) (2) (3) (4) (5)
Exporter only 0.115*** 0.558*** 0.838*** -0.085*** 0.043**
(0.024) (0.037) (0.040) (0.024) (0.021)
Importer only 0.196*** 0.423*** 0.764*** -0.180*** 0.110***
(0.005) (0.009) (0.011) (0.007) (0.004)
Importer and exporter 0.353*** 0.621*** 1.155*** -0.320*** 0.141***
(0.014) (0.021) (0.023) (0.015) (0.009)
Constant 0.649*** 9.588*** 9.638*** 11.288*** 7.342***
(0.125) (0.554) (0.473) (0.158) (0.207)
Industry effects (four-digit) Yes Yes Yes Yes Yes
Geographic effects (municipality) Yes Yes Yes Yes Yes
Year effects Yes Yes Yes Yes Yes
Firm fixed effects Yes Yes Yes Yes Yes
R-squared 0.08 0.05 0.08 0.13 0.04
Observations 161,949 161,949 262,231 72,569 160,353
Source: KTR, World Bank staff calculations.
Notes: Robust standard errors. (*), (**) and (***) indicates significance level at (10), (5) and (1) per cent
level. In each column, we regress the dependent variable on foreign trade dummies, which relate to
the type of link that the firm has with foreign markets. “Exporter only” equals 1 if the firm exports but
does not import; “Importer only” follows the same criteria for importing firms. Finally, the “importer and
exporter” dummy values 1 if the firm both imports and exports in the year of reference. Each variable
is constructed based on the VAT forms. Exporting/importing firms are those firms that report a positive
export/import value in the year of reference while importers. Furthermore, we control for industry
effects (four-digit level), geographic effects (municipality), year and firm effects.
66 Additional Tables and Figures
Table A2.7. Foreign-shareholding Premiums in Kosovo
Dependent variables Per worker
(col) /
Employment Turnover Turnover Value added Capital Wages
Independent variables
(row) (1) (2) (3) (4) (5) (6)
Foreign shareholding -0.007 -0.265** -0.207*** -0.103 0.720*** 0.007
(% assets: 0.1%-24.9%)
(0.080) (0.116) (0.063) (0.096) (0.070) (0.028)
Foreign shareholding -0.128** 0.144 -0.099 -0.293*** 0.814*** 0.082***
(% of assets: 25%-
49.9%)
(0.060) (0.111) (0.073) (0.108) (0.075) (0.030)
Foreign shareholding 0.978*** 1.625*** 0.537*** 0.151** 0.860*** 0.425***
(% of assets: 50%-100%)
(0.077) (0.134) (0.076) (0.074) (0.152) (0.041)
Constant 4.501*** 14.670*** 10.242*** 9.261*** 9.383*** 8.336***
(0.109) (0.172) (0.106) (0.158) (0.164) (0.058)
Industry effects (two- Yes Yes Yes Yes Yes Yes
digit)
Geographic effects Yes Yes Yes Yes Yes Yes
(municipality)
Year effects Yes Yes Yes Yes Yes Yes
R-squared 0.20 0.20 0.27 0.17 0.25 0.23
Observations 17,285 20,381 17,285 9,976 16,644 17,227
Source: KTR, World Bank staff calculations.
Notes: Robust standard errors. (*), (**) and (***) indicates significance level at (10), (5) and (1) per
cent level. In each column, we regress the detailed dependent variable on a set of dummies that breaks
down the share of assets owned by foreign shareholders.
Table A2.8. Gender Premiums in Kosovo
Per worker
Employment Turnover Turnover Value added Capital Wages
Sector (1) (2) (3) (4) (5) (6)
All sectors -0.205*** -0.053 -0.053 -0.176** -0.076 -0.017
Manufacturing -0.643*** -0.081 -0.081 -0.445*** -0.221** 0.003
Wholesale and retail -0.224*** -0.029 -0.029 -0.013 -0.296*** 0.038
trade
Professional and -0.122 -0.028 -0.028 0.508*** 0.426** -0.141***
administrative services
Other services -0.109 0.005 0.005 0.194 -0.133 0.030
Industry effecsts (two- Yes Yes Yes Yes Yes Yes
digit)
Geographic effects Yes Yes Yes Yes Yes Yes
(municipality)
Source: KTR, World Bank staff calculations.
Note: Robust standard errors. (*), (**) and (***) indicates significance level at (10), (5) and (1) per
cent level. In each column, we regress the detailed variable on a female-run company dummy variable,
which equals 1 if the establishment is run by a female and 0 otherwise. Each specification is estimated
by OLS and includes geographic (at the municipality level) and sector effects (two-digit level).
Additional Tables and Figures 67
Table A2.9. Access to Finance: Descriptive Statistics, Average, 2015–18
Borrowers Share of disbursed Share of disbursed Approved amount
amount by borrower (%) amount by lender (%) (‘000 €)
Total Legal Physic Non- Banks Insurance MFI Mean Median
Year residents company
2015 11,137 98.3 1.6 0.0 91 7 1 0 0
2016 10,962 98.4 1.6 0.1 88 10 2 0 0
2017 10,516 98.2 1.7 0.1 90 8 2 0 0
2018 10,569 98.0 2.0 0.0 86 13 1 0 0
Disbursed amount Nominal Duration (months) Type of credit
(‘000 €) interest rate (pp) (% of total disbursed amount)
Year Mean Median Mean Median Mean Median Loan Overdraft Guarantee Other
2015 0 0 9.7 8.0 42 24 51 26 17 5
2016 0 0 9.1 7.6 45 13 51 24 17 8
2017 0 0 7.5 7.4 43 21 56 20 14 9
2018 0 0 8.1 7.6 43 24 52 20 17 11
Source: KTR, World Bank staff calculations.
Note: TFPR is weighted by employment while Employment, VA, Turnover, Materials, Wage bill, Capital
stock and Investment are weighted using firm weights.
68 Additional Tables and Figures
Table A2.10. Drivers of Access to Credit in Kosovo
(1)
TFP (log) change 5.591***
(1.235)
TFP (log) change (t-1) 0.167***
(0.043)
Serbian-majority municipality -0.550***
(0.050)
Exporting firm -0.030**
(0.013)
Age 0.003**
(0.001)
Small (10-49) 0.174***
(0.010)
Medium (50-249) 0.244***
(0.020)
Large (250+) 0.326***
(0.043)
Market concentration (HH index) 0.257***
(0.065)
Constant 0.698***
(0.052)
Year effects Yes
Industry effects (four-digit) Yes
R-squared 0.35
Number of firms 1,513
Observations 3,536
Source: KTR and KCR, World Bank staff calculations.
Note: Robust standard errors. (*), (**) and (***) indicates significance level at (10), (5) and (1) per cent
level. Dependent variable is a dummy variable that is 1 if the firms was approved a positive loan amount
during the period of analysis and 0 otherwise.
Additional Tables and Figures 69
WORLD BANK GROUP
Raising Firm Productivity Kosovo Country
November 2021 Economic Memorandum
File and source
- File
- bank-2021-kosovo-country-economic-memorandum-november-2021.pdf
- Size
- 4,052,756 bytes
- SHA-256
- c7b1afae2f4e26760a749a65a062e2b75e530fe6cdb20e3032a94eb51f57889a
- Original
- openknowledge.worldbank.org