JOBS SERIES Issue No. 5 DIAGNOSTIC KOSOVO A l e x an dr u C o j o c ar u © 2017 International Bank for Reconstruction and Development / The World Bank. 1818 H Street NW, Washington, DC 20433, USA. Telephone: 202-473-1000; Internet: www.worldbank.org. Some rights reserved 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 govern- ments 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. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Alexandru Cojocaru. 2017. “Kosovo Jobs Diagnostic.” World Bank, Washington, DC. License: Creative Commons Attribution CC BY 3.0 IGO. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses 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. Images: © World Bank. Further permission required for reuse. ACKNOWLEDGMENTS This report was produced by a team led by Alexandru Cojocaru, and comprised of Mohammed Ihsan Ajwad, Reyes Aterido, Jieun Choi, Agim Demukaj, Francesca Lamanna, David Keith De Padua, Carly Petracco, and Michael Webber. The report benefited from contributions and comments from a number of colleagues, including Bashkim Bellaqa, Ilir Berisha, Simon Davies, Bruce Fitzgerald, Besa Hafiqi, Gabriela Inchauste, Mu- hamet Kastrati, Olga Kupets, Dino Merotto, Gloria La Cava, Marco Mantovanelli, Perihan Ozgye, Elizabeth Ruppert Bulmer, Adrian Scutaru, Bojan Shimbov, Ken Simler, Ifeta Smajic, and Hernan Winkler. Bruce Fitzger- ald made editorial suggestions. The team is thankful to the Kosovo Agency for Statistics for providing the firm and household level data that was employed in the analysis. CONTENTS EXECUTIVE SUMMARY 1 1. INTRODUCTION 4 2. MACROECONOMIC AND SECTORAL CONTEXT 6 Recent macroeconomic and fiscal developments.................................................................................................. 6 Demographic context........................................................................................................................................... 9 3. WHERE ARE THE JOBS AND WHO IS HIRING? DEMAND-SIDE ANALYSIS OF THE FORMAL SECTOR 12 Composition of the formal firm sector................................................................................................................ 12 Job creation and firm growth over the past decade............................................................................................ 14 What constrains firms’ growth?......................................................................................................................... 19 Labor productivity differentials and growth dynamics......................................................................................... 24 4. WHO IS ACTIVE? WHO IS GETTING JOBS? 31 Evolution and distribution of jobs in Kosovo....................................................................................................... 31 Informality – a salient characteristic of employment in Kosovo........................................................................... 41 Size of the informal sector in Kosovo.............................................................................................................. 41 Profile of informal firms and employees.......................................................................................................... 42 Some consequences of informality................................................................................................................. 45 Jobs and poverty................................................................................................................................................ 48 5. SKILL MISMATCHES IN THE LABOR MARKET 54 6. SUMMARY OF FINDINGS AND POLICY CONCLUSIONS 63 Improving the regulatory and business environment........................................................................................... 65 Addressing skill shortages and mismatches......................................................................................................... 66 Enabling women to take advantage of employment opportunities..................................................................... 68 REFERENCES....................................................................................................................................................... 69 ANNEX A............................................................................................................................................................ 71 EXECUTIVE SUMMARY BACKGROUND AND CONTEXT Kosovo’s economy experienced strong growth over the past decade driven by domestic demand largely financed by remittances and aid. GDP growth averaged 3.4 percent during 2008–15. Domestic savings have been negative; consumption and investments were the key contributors to economic growth. Private consump- tion was fueled by both remittances and large increases in public sector wages and transfers (especially pensions). Kosovo’s production capacity is growing but it is narrow, undiversified and uncompetitive. The non-tradable sectors dominate output and employment. Services is the largest sector with value added at over half of GDP, and it contributed heavily to growth over the past decade. Tradables sectors, including agriculture (12 percent) and manufacturing (11.2 percent), are small. Kosovo is one of the youngest countries in Europe, with an increasing share of working age population – a transitory demographic bonus that offers opportunities. Youth (under 15) are 25.7 per- cent of the population, while working age (15–64) account for 67.6 percent. Kosovo’s ratio of working age population to dependents is projected to increase from 1.88 in 2011 to 2.24 in 2021, and will not dip below the 2011 level before 2046. Only eight percent of Kosovo’s population today are elderly, compared to 19 percent in the EU-28, and 22 percent in Italy. Against this background, it is problematic that economic growth over the past decade has not been associated with robust job creation. Data from the registry of formal sector firms over 2005–2014 reveal that the rate of net job creation by formal firms has declined from over 7,000 net new jobs at the beginning of this period, to just over 1,600 in 2014. Commerce and services sectors are the only sectors to have performed consistently well in terms of net job creation over the past decade; the contribution of the manufacturing sector has been negative since 2012, the capacity of the construction sector to add new jobs has also been in decline. UNEMPLOYMENT AND JOBS: LABOR SUPPLY AND DEMAND IN A SMALL, CLOSED ECONOMY Micro firms dominate the private sector, and although they substantially contribute to employment, very few have grown. Established firms have been shedding more jobs than they create. Consequently, the greatest source of formal job creation has been start-up firms, but the rate of new firm entry remains relatively low. Over 90 percent of Kosovo’s formal firms are micro firms with 1–9 employees, and most start-ups are micro but micro firms do not grow. Among firms that start as micro only four percent grow beyond nine employees within five years. The small domestic market and limited integration are important constraints to growth. Firm growth and dynamism, particularly in the non-tradable sector, are constrained by the small domestic market. Proximity to important markets is an asset, but Kosovo is landlocked and depends on transit through other countries. It lags other Western Balkans countries in terms of international financial and trade integration. Improved integration holds the promise of lower transaction costs and bigger markets – and thus greater economies of scale–as well as more competition and greater efficiency. The lack of employment opportunities is reflected in high rates of inactivity rates and unemploy- ment and slim chances of transitioning from unemployment to employment. According to Labor Force Survey data, 58 percent of the working age population in 2015 was inactive. A salient characteristic of inactivity is the high rate among women – 38 percent of working age men were inactive, but the rate among women 1 was more than double that: 78.6 percent. Inactivity among men is principally market related – education/ training or belief that no work is available – whereas the primary reasons among women are family related – over half of inactive women cite family reasons as the main reason for inactivity. Unemployment rates are also stubbornly high – in 2015 almost a third of the labor force was unemployed; among youth the unemployment rate was in excess of 57 percent. Among the unemployed, a large majority (72 percent) reported they had been unemployed for over 12 months. Education is valuable but with few new jobs it is not sufficient for low-income households to improve their labor market outcomes. Low education is strongly associated with inactivity – most inactive men and women have less than secondary education. The overriding problem is that formal job creation has been on a downward trend over the past decade and the number of new formal openings is small. Only a quarter of working age population (and only 12 percent of working age women) were employed in 2015 – considerably fewer than in the EU (64.9 percent for EU-28), or even countries like Greece and FYR Macedonia, where half of the work- ing age adults are employed. Kosovo’s informality is among the highest in Europe with about 35 percent of employees in the informal sector. With high rates of unemployment and inactivity, all jobs – formal and informal – can help to boost the welfare of households below the poverty line or in the bottom forty population. Informal jobs help those with low educational levels who find it harder to compete for formal jobs, but informal sector jobs lack security and exclude workers from many benefits, thus increasing their vulnerability to negative shocks. Unemployment and inactivity are among key risks and they lead to youth disenfranchisement and fragility. The recent Kosovo Risk and Resilience Assessment found that the economic disenfranchisement of youth has led to political tensions and violence. Dissatisfaction with socio-economic conditions and lack of employment opportunities has also fueled emigration – mainly illegal migration to the European Union resulting in applications for asylum. The number of asylum seekers from Kosovo increased from 18,000 in 2013 to 35,000 in 2014, to a peak of 68,000 in 2015, before receding in 2016. In 2015 the number of EU asylum seekers from Kosovo was second only to those fleeing the conflict in Syria. Close to half of Kosovo’s citizens are estimated to have family members living abroad, and family reunification is among key reasons for migration. However, for a large share of migrants the motivation to emigrate related to the lack of economic opportunities at home. According to recent data from the Life in Transition Survey, only 8 percent (20 percent) of Kosovars believed that the political situation (economic situation) today is better than 4 years ago, lower than the average for the transition region (24 percent and 28 percent respectively). Views on social and economic mobility are similarly bleak – 34 percent of Kosovars viewed political connections as the most important factor for success in life, compared to 21 percent on average for the transition period and to 3 percent in Germany. BETTER OUTCOMES WITH BETTER POLICIES This report highlights several constraints to employment creation, focusing on firms’ poor growth prospects, the high degree of informality, and low levels of labor market participation. The World Bank’s recent Systematic Country Diagnostic (SCD) highlights the importance of reducing infrastructure bottle­ necks (energy, connectivity) to improve competitiveness, attract foreign investments, or, more generally, improve macro-financial sustainability. In addition, high levels of public spending on public sector wages and pensions crowd out spending on physical and human capital investment and raise the costs for private firms to attract and retain well-qualified employees. This report highlights two additional constraints that may deter firms from expanding, or discourage informal firms from formalizing: (i) poor business and regulatory environ- ment, and (ii) poor and mismatched skillset of workers and jobs. As new – and larger – cohorts join the labor force, reforms aimed at adopting the right set of rules, and developing the right set of skills, to promote job creation, will help to reduce youth disenfranchisement and productively employ the demographic dividend. In addition, the report analyzes women’s disproportionate labor market difficulties that result from undeveloped systems of family care. First, a poor business and regulatory environment fosters informality and impedes the growth of firms and the productivity of capital and labor resources. Starting a business in Kosovo is easy; operating, growing, or closing it is cumbersome. On a number of Doing Business indicators such as dealing with construction permits, getting credit, enforcing contracts, protecting minority investors, or resolving insolvency Kosovo is far below the 2 best practice frontier or even its regional comparators. Multiple, uncoordinated, and excessive inspections are expensive for firms and create incentives to stay small and informal. Likewise, weaknesses in accreditation and certification systems increase the cost of quality assurance and hamper development of new, exportable products. Second, low levels/quality of education/training and skill mismatches prevent the inactive popula- tion from obtaining and retaining good jobs. Many firms report problems hiring new employees, largely because of insufficient experience or skills. A shortage of qualified specialists in marketing, sales, and design makes it harder to compete in export markets, where new product development and placement are vital. Firms distrust the education and VET systems to train people with the required skills. Hiring is mostly informal and connections based – close to two thirds of firms reported recruiting from informal channels, for professional/ technical jobs. The second most popular hiring channel is making job offers to experienced people from other firms. Only 17 percent of firms reported being in regular contact with educational/technical institutions to hire professional/technical staff, and only 7 percent in the case of non-technical staff. Informal hiring reduces the can- didate pool for firms and disadvantages those without personal connections, particularly the bottom 40. Better alignment between firms’ needs and the skills provided by the education and VET systems, inclusively through greater involvement of the private sector in the planning and provision of training, is a short-term priority. Over the medium term, given the poor performance of Kosovo in the recent PISA exercise, improving the quality and relevance of education at all levels and offering equitable access to educational opportunities at early ages are high priorities with potentially large payoffs. Finally, improving the effectiveness of activation policies can be part of the solution of addressing skills mismatches and inactivity for long-term unemployed and the bottom 40. Finally, existing systems to provide care are not widely available, too expensive, and are embedded with disincentives for labor market participation. Social norms assign child and elder care responsibilities primarily to women, and much of the care activities remain informal and family-based. Over 95 percent of family childcare is performed by family members. The few institutional childcare centers are oversubscribed and not affordable to many, or not open year round, making it difficult to combine care with full-time employment. Flex- ible work arrangements are generally not available, and parental care is minimal, which may encourage exten- sive, rather than intensive margin adjustments on part of women to income shocks by dropping out of the labor market. The design of maternity leave benefits, in terms of length and distribution of financial burden, makes employers favor men over young women in hiring decisions. Kosovo needs more jobs for everyone, but without better institutional care systems, even with more jobs young women would be disadvantaged. 3 1. INTRODUCTION Kosovo is a young country with a potentially large demographic dividend. Close to 40 percent of Kosovo’s population is under the age of 19. The share of working age population will continue increasing in the coming years. This provides Kosovo with an opportunity to build a large and educated workforce that could help attract investments, and support sustained shared prosperity. In addition to its young population, Kosovo benefits from a favorable geographical location and proximity to important markets, notably the European Union (EU). Finally, the prospect and the objective of EU membership, as reflected in the recently signed EU Stabilization and Association Agreement (SAA), is to provide an anchor for the country’s development strategy. Kosovo is also a small country, with a high incidence of poverty. Kosovo’s population of 1.8 million ranks number 146 in the world. Forty eight countries have larger populations in their capital cities alone. Being a small state imposes a number of challenges to growth, such as a small internal market, narrow production base, with limited scope for diversification and vulnerability to economic shocks. Kosovo had a GDP per capita of PPP$ 9,759 in 2015, putting it at number 103 in the world. Its economy has exhibited good growth over the past decade, but job creation has been disappointing, unemployment is high, and large numbers have emigrated to find jobs abroad. Jobs are created in robust, growing economies that attract investment. Globalization has allowed small economies similar to Kosovo’s to improve their growth prospects if they’re competitive and well integrated into world markets so they can benefit from economies of scale and sell their products to affluent foreign buyers. But Kosovo’s economy and its labor market are impeded by persistent, deeply embedded structural problems. High unemployment, underemployment, low labor force participation, labor informality and emigration are legacies of Kosovo’s status as the most disadvantaged area in the Socialist Federal Republic of Yugoslavia. Public finances are burdened by pressures to spend on pensions, categorical benefits and public sector wages, which limits space for investments in the quality of service provision, including in areas such as education that are important for productivity growth. Progress is being obstructed by political instability, uncertain legal and property rights, and weak government institutions. There are doubts as to whether the current model can deliver sustainable growth and shared pros- perity. The World Bank’s recent Systematic Country Diagnostic points out that the pace of public investment in physical infrastructure is slowing as damaged and obsolete infrastructure is replaced, while official development assistance is shrinking. Production for the small internal market draws resources from tradable goods. Creating conditions for a dynamic private sector could become more difficult and social tensions may deepen without more job creation (World Bank, 2016). Kosovo seeks to join the EU but this will require deep structural changes to transform the economy, improve competitiveness, attract investment, and raise productivity. In anticipation of EU accession, the purpose of this report is to provide an integrated analysis of Kosovo’s labor market and to discuss the con- straints to – and priorities for – job creation, and improved labor market and welfare outcomes, particularly for the poor and the bottom forty percent (B40). Kosovo’s economic growth over the past decade has not been associated with robust job creation. Net job creation by formal firms has been declining. Commerce and services are the only two sectors to have performed consistently well in terms of net job creation over the past decade. Start-ups have been the greatest source of (formal) job creation in Kosovo, but the rate of new firm entry remains low. Most start-ups are micro firms which grow slowly or not at all. Firms’ growth prospects are constrained by inadequate hard (energy 4 security, connectivity) and soft (regulatory/business environment, human capital) infrastructure. These obstacles also provide incentives for firms to stay informal – a salient characteristic of the Kosovo labor market. The lack of employment opportunities is reflected in the very high inactivity rates (particularly among women), high unemployment rates, and slim chances of transitioning from unemployment to employment. The report concludes that the right set of rules, and the right set of skills, would be necessary to promote job creation, to prevent youth disenfranchisement, and to productively employ the demo- graphic dividend. The findings of this report are consistent with the broader conclusions of the SCD, and it highlights three key policy areas: (a) improving the regulatory and business environment to aid job creation and firm growth; (b) addressing skill shortages and mismatches that hamper both Kosovo’s ability to harness its competitive advantages and the ability of the unemployed and inactive to transition into employment and (c) improving systems of child and elder care, as well as maternity and care leave to allow those looking for work, particularly women, to take advantage of available opportunities. It should be noted that the data that the jobs diagnostic had at its disposal does not allow us to make more specific recommendations with respect to policy design. Rather the goal of this section is to point to several priorities which can subsequently be explored in more detail as part of the ongoing dialogue on jobs. 5 2. MACROECONOMIC AND SECTORAL CONTEXT Kosovo’s economic and socio-demographic characteristics influence past job creation and future chal- lenges. The small internal market, the proximity to world markets, the economic structure, as well the demographic context provide an important background for understaning both past job performance and future challenges. This chapter is a brief overview of recent macroeconomic, sectoral and fiscal developments, as well as the current demo- graphic factors in order to understand patterns of job creation over the past decade (Chapter 3) and the salient characteristics of employment and its links to poverty (Chapter 4). RECENT MACROECONOMIC AND FISCAL DEVELOPMENTS The economy grew robustly over the past decade. As Figure 2-1 and Figure 2-2 highlight, GDP growth averaged 3.4 percent during 2008–15, driven by consumption and investment that were largely financed by remittances and aid. Domestic savings have been negative, while consumption contributed an average of 2.6 percentage points to annual growth between 2008 and 2015. Private consumption was fueled by both remittances and large increases in public sector wages and transfers (especially pensions). Investments accounted for an average 1.6 percentage points of annual GDP growth. Kosovo’s production capacity is growing but it is narrow, concentrated, and uncompetitive. The agriculture and services sectors are large; manufacturing is small; and little is tradable. As a result has been supplied mainly by imports. Net exports generally reduced growth, except in 2012 and 2013 when imports declined more than exports. There is a perception that there is import substitution in some sectors such as food and construction material, but there is insufficient evidence to verify this conjecture. The dominance of non-tradable sectors in output and employment has increased. Services is the largest sector with value added at 56 percent of GDP, and it contributed heavily to growth over the period, driven by construction, real estate, and retail. Agriculture and manufacturing account for 11 and 11.2 per- cent of GDP respectively. Construction services and real estate are large within the services sector with about Figure 2-1 Real GDP growth 2008–2016 5.0% 4.5% 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 08 09 08 09 10 11 12 3 4 5E F 16 1 1 1 20 20 20 20 20 20 20 20 20 20 20 Source: Statistics Agency of Kosovo and WB staff calculations. 6 Figure 2-2 Figure 2-3 Sources of growth by component of aggregate demand Sources of growth by economic sector 12% 5.0% 10% 4.0% 8% 6% 3.0% 4% 2% 2.0% 0% 1.0% –2% –4% 0.0% –6% –1.0% –8% 2007 2008 2009 2010 2011 2012 2013 2014 2015 CONSUMPTION –2.0% 2011 2012 2013 2014 2015 INVESTMENTS EXPORT AGRICULTURE IMPORT SERVICES INDUSTRY Source: Statistics Agency of Kosovo and WB staff calculations TAXES LESS SUBSID REAL GDP GROWTH Source: Statistics Agency of Kosovo and WB staff calculations 15.5 percent of GDP followed by wholesale and retail with a 12.1 percent share (Figure 2-3). The contribution to growth of industry, including agro-processing, peaked in 2012 when imports declined, a possible indica- tion of some degree of import substitution. Agriculture had slow growth and was affected by weather conditions and climate change. Recent reforms have helped to increase agricultural productivity and output. Non-tradable sectors accounted for 64 percent of Gross Value Added and for 61 percent of employment in 2006, and by 2012 the share of non-tradables in GVA increased to 66 percent, and in employment, to 68 percent (World Bank, 2016). Growth has not translated into robust employment creation, and labor market indicators are among the worst – in the region and in the world. Inactivity and unemployment rates have been very high histor- ically since it was part of the former YU with inactivity and unemployment among women being much higher than among men. Despite economic growth, the private sector has been unable to provide sufficient high quality jobs, particularly in the tradable sector. Net job creation by formal firms has been in decline over the past decade, and very few net new formal sector jobs have been created in recent years (Chapter 3 dis- cusses job creation in the formal sector). Youth unemployment is especially high, over 57 percent, and female unemployment was recently 36.3 percent. High unemployment among women and youth are significant social concerns. Additionally, about 10.7 percent of the working age population were seen as discouraged workers in 2014 (Chapter 4 discusses employment, unemployment and inactivity). Many Kosovars left the country during November 2014–March 2015, driven by high unemployment and poor economic prospects. The number of asylum seekers from Kosovo in the European Union increased from 18,000 in 2013 to 35,000 in 2014, to 68,000 in 2015. Productivity has fallen since 2008, similar to other South East European (SEE) countries. Following the financial crisis, total factor productivity, or efficiency in harnessing human and physical capital, stagnated or declined across the SEE countries. Before the financial crisis, capital accumulation was an important contrib- utor to growth across SEE countries, especially in Kosovo, where the contribution of TFP was negative even during 2000–08. In more recent years capital accumulation was constrained by fiscal pressures, while labor markets remain weak, limiting human capital’s contributions to growth. High growth in the future will require ways of raising productivity, both through higher efficiency and by reducing structural unemployment (World Bank, 2017). 7 Figure 2-4 TFP contribution to growth was negligible or negative after the crisis ALB BIH KSV MKD MNE SRB 2000–08 2000–08 2000–08 2009–14 2009–14 2009–14 2000–08 2009–14 2000–08 2009–14 2000–08 2009–14 10 8 6 4 2 0 –2 CAPITAL STOCK WORKING AGE POPULATION HUMAN CAPITAL TFP GDP Source: World Bank (2017) In recent years, the economy weathered the impact of the global financial crisis of 2009 and the Eurozone crisis of 2012 better than other SEE countries, but stagnated in 2014 due to a domestic crisis. Kosovo’s limited financial and trade linkages together with its strong remittances from migrants damp- ened the impacts of the external financial crises, but the crises arose when Kosovo was recovering from more than a decade of conflict and political uncertainty and the transition from socialist rules to a market econ- omy. In 2014, the combination of the recession in Europe, domestic energy and political/constitutional crises, and slowing domestic credit and exports, reduced GDP growth to 1.2 percent, and unemployment rose to 35 percent. Growth recovered to an estimated 4.1 percent by end-2015, fueled by consumption and private investments (including strong FDI). Economic recovery seems to have helped moderate unemployment to 32.9 percent in 2015. Remittances from abroad continue to boost the economy. In 2014 remittances grew about 12 percent y-o-y and 9 percent y-o-y in 2015. Most migrants work in Germany, Switzerland or Austria, with others spread across Europe and the USA. Since the Eurozone crisis Germany and Switzerland have performed better than other European countries and remittances have supported consumption and investments (mainly housing con- struction). Expansionary fiscal policies helped avoid recession, while revenue and expenditure consolidation in 2015 helped improve fiscal balances. The post-crisis fiscal stance has been marked by deficits and a slow but steady increase in public debt, but the deficit has been within acceptable limits, not exceeding 3 percent of GDP. The 2009–2011 fiscal deficits were financed mainly by surpluses accumulated before 2008. Since 2012 financing has been through domestic debt and drawing from the IMF programs in 2012 and 2015. In 2014, a rule was adopted to limit the overall fiscal deficit at two percent of GDP, with some exceptions for one-off revenues from privatizations that could be spent for capital investments. In 2016 the rule was amended to permit spending on productive capital investments beyond the two percent ceiling, subject to being financed by IFI’s, and not to exceed a public debt to GDP ratio of 30 percent.1 The 2015 fiscal deficit was 1.9 percent of GDP, well below the original projection of 3.5 percent. Consolidation measures in the 2015 mid-year budget review saved about 0.9 percent of GDP, and under-execution of some capital investments provided additional 1 In 2009 Kosovo assumed a public debt to IBRD after secession from former Yugoslavia. By end-2015 Kosovo had a public and publicly guaranteed debt of about 13 percent of GDP, which is low relative to the rest of the region or Europe. 8 savings. Revenue was increased by raising the main VAT rate from 16 to 18 percent (while reducing the rate on food to eight percent), a new VAT of eight percent on medications, higher excise taxes on commodities with environmental and/or health-related externalities, and higher fees for telecom operating licenses. One-off revenues from partial liquidation of SOE privatization fund added 1.1 percent of GDP. Wage and pension increases create extra pressure on the budget and crowd out investments to reduce infrastructure bottlenecks. Kosovo is sensitive to political business cycles. Driven by electoral prom- ises, the wage bill increased by 27 percent in 2011 and by 25 percent in 2014, a notable increase once scaled by a measure of productivity. Pensions also increased by 25 percent in 2014. As a share of government expenses, wages accounted for 46 percent of the total in 2014 – much higher than in other SEE countries or other coun- tries in Eastern and Southern Europe. Social benefits, many categorical, accounted for another 35 percent of current spending (World Bank, 2017). New veteran benefits introduced in September 2015 put further pres- sure on the 2016 budget. The larger-than-expected number of war veteran beneficiaries (29 thousand to date compared to the initial estimate of 13 thousand), led to an increase in spending on this line to over €50 million, more than double original budget. With full implementation of new benefits, the fiscal envelope of war-related transfers reached 1.5 percent of GDP. As a consequence, the IMF kept its SBA in hold during 2016, considering fiscal policy to be unsustainable. The current expenditure growth is crowding out capital spending, which has declined from 10.3 percent of GDP in 2010 to about seven percent in 2015. The recent EU Stabilization Association Agreement (SAA) opens a path towards EU integration. Kosovo is the last Western Balkans country to sign an SAA, in part because five EU states were reluctant to recognize its independence. The SAA opens an opportunity for broader reforms that could accelerate convergence to EU standards: economic, judicial, social, environmental, and quality of life. The SAA, which started implementa- tion in April 2016, is expected to attract new, export-oriented FDI and gains in productivity and competitiveness by attracting fresh capital and know-how. EU accession is protracted and it will require fundamental structural reforms. Over the medium term, domestic demand is expected to drive the economy. Economic growth is projected to remain flat at 3.6 percent in 2016 and increase to 3.9 percent in 2017, supported by consumption growth, larger FDI, and rising exports. Beyond 2016, growth is expected to be supported by increases in: consumption, partly financed by remittances; private investments, especially FDI; and exports of goods and services. The gov- ernment’s program to improve the business environment and foster growth based on tax policy changes and other structural changes should attract new investments. SAA with EU should be an impetus to further improve Kosovo’s image of stability and to attract higher FDI especially from the EU. Productivity gains are necessary to increase future growth. Adding labor and capital resources is essen- tial for growth, but raising productivity – producing more from new and existing inputs – is of particular importance. Generating productivity gains faster than competitors would attract investment and increase exports to better integrate Kosovo into the world economy. This is particularly important in light of the deteriorating productivity dynamics since the war in Kosovo and, especially, since the global financial crisis. To reach a higher growth path Kosovo must overcome structural constraints. Addressing its low insti- tutional, physical and human capital and moving to sustained growth can help to reduce the constraints and to catch up with neighbors. Improving the rule of law and building a better business environment are essential measures to improve growth and prosperity. DEMOGRAPHIC CONTEXT The population is one of the youngest in Europe. According to the 2011 Population Census, Youth (0–14) were 25.7 percent of the population, while working age (15–64) group accounted for 67.6 percent and the elderly (65+) were 6.8 percent of the population. The ratio of working age population to dependents is pro- jected to increase from 1.88 in 2011 to 2.24 in 2021, and will not dip below the 2011 level before 2046. Only eight percent of Kosovo’s population today is elderly, compared to 19 percent in the EU-28, and 22 percent in Italy (Figure 2-5 and Figure 2-6). 9 Pre-war population growth has resumed, but at a slower pace. The population increased steadily until 1997, having doubled between 1961 and 1997. As a result of the war, the population fell by some 20 percent between 1997 and 2000. Over the past decade, population growth resumed, but at a slower pace – the total population grew by about seven percent between 2000 and 2013 according to the Population Census, although this is still a faster pace of growth than elsewhere in Europe. Figure 2-5 Figure 2-6 Population structure, Western Balkans countries, 2015 Population projections by age cohort through 2061 100.0 100% 80.0 90% 80% 60.0 70% 40.0 60% 20.0 50% 40% 0.0 KSV ALB BIH BGR MDA MKD ROM SRB HRV MNE 30% 20% POPULATION AGES 0–14 POPULATION AGES 15–64 10% POPULATION AGES 65+ 0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 Source: World Development Indicators World Bank, Kosovo AGE 0–14 Agency of Statistics AGE 15–64 AGE 65+ Source: Kosovo Agency for Statistics. Lower fertility rates will lead to an aging population in the medium-term. With the fertility rate declining (estimated at 2.1—roughly replacement rate—in 2014), population growth slowed to 0.6 percent (2004–2014). As highlighted in Figure 2-7, in comparison, the average population growth for other lower middle income countries for this period was 1.6 percent. Figure 2-7 Kosovo has the highest population growth in the Western Balkans 2004−2014 POPULATION GROWTH AND GDP PER CAPITA 4 POPULATIONGROWTH % 2 KSV MKD 0 MNE BIH ALB BGR –2 6 8 10 12 LOG GDP PER CAPITA, 2014 HIGH INCOME LOWER MIDDLE INCOME UPPER MIDDLE INCOME LOW INCOME Source: World Development Indicators, World Bank With a young population, Kosovo can benefit from investing in skills. Educational attainment remains very low. According LFS data, more than 30 percent of males had no education beyond primary; among women 10.1 percent had no education in 2014, and around two third had no education beyond primary. While the 10 educational profile of the population has been improving – the share of those with secondary and post- secondary education among the 20–29 cohort is considerably higher than for the population in the 40+ age group – achievement based on objective indicators remains low. The 2015 PISA exercise (the first in Kosovo) revealed that more than two thirds of 15 year olds lack basic proficiency in sciences, and over three quarters are below basic proficiency in reading and mathematics. Public spending on education grew steadily over the past decade, and is catching up with other SEE6 countries, but because of its youthful demo­ graphic struc- ture, Kosovo spends only 13 percent of per capita income per student in primary and secondary education— considerably less than upper-middle income countries (16.2 and 17.8 percent, respectively primary and second- ary education), or average EU15+3 countries (19.5 and 22.0 percent) (World Bank, 2016). Figure 2-8 Educational attainment by age cohort and gender HIGHEST LEVEL OF EDUCATION COMPLETED BY GENDER OVER YEARS HIGHEST LEVEL OF EDUCATION COMPLETED ACROSS AGE COHORTS 2014 LFS 100 100 13.6 13.8 14.3 7.0 7.1 8.1 80 23.5 27.0 26.0 80 60 51.0 52.1 52.6 60 40 55.2 55.5 55.8 40 20 31.0 30.8 29.9 20 14.2 10.4 10.1 4.5 3.3 3.2 0 0 12 13 14 12 13 14 9 4 9 4 9 4 9 4 9 44 −1 −2 −2 −3 −3 −4 −4 −5 −5 20 20 20 20 20 20 0− 15 20 25 30 35 40 45 50 55 .6 1. 2. 3. 4. 5. 6. 7. 8. 9. MALE FEMALE 10 1. NO EDUCATION 1. NO EDUCATION 2. PRIMARY 2. PRIMARY 3. SECONDARY 3. SECONDARY 4. POST-SECONDARY 4. POST-SECONDARY Source: Staff estimates based on LFS data. 11 3.  WHERE ARE THE JOBS AND WHO IS HIRING? DEMAND-SIDE ANALYSIS OF THE FORMAL SECTOR What have been the patterns of formal sector job creation over the past decade? This chapter con- siders, based on data for all formal firms, the patterns of (net) job creation over the past decade, including the pace with which new jobs have been created and the distribution of new jobs across sectors and types of firms. This chapter also describes firms’ growth as well as constraints to growth and the links between firms’ growth and productivity.2 COMPOSITION OF THE FORMAL FIRM SECTOR Micro firms dominate the formal private sector landscape. Data from the registry of formal firms cov- ering 2005–2014 (Box 1) reveal almost all firms are micro firms (size 1–9). In 2005, 95 percent of firms were micro firms; by 2014 the share of micro firms had fallen to 91 percent on account of the increase in the share of firms with 10–99 employees from 4.7 percent to 7.7 percent. Compared to other countries with recent jobs diagnostics, only Zambia has levels similar to Kosovo’s; in Moldova the share of micro firms was below 60 percent, and in Peru less than 50 percent. Almost the entire spectrum of firms is micro, but jobs are disproportionally offered by larger and older firms. Large firms started large at the beginning of the period, following privatization, and appear, for BOX 1: MEASURING FORMAL FIRM DYNAMICS IN KOSOVO Firm-level data is not readily available in Kosovo. The analysis in this chapter relies on unique firm level panel data collected by the Kosovo Agency of Statistics during 2005 to 2014. The data covers formal firms with one or more wage employee, which facilitates investigation of the dynamics of the formal private sector and individual firms over a period of almost a decade. This data includes location, sector of the main economic activity at the 4-digit level, the year the firm began operations, number of workers, total sales, ownership, and labor costs. It does not distinguish whether a firm is government-owned, but it denotes foreign-owned firms. A firm’s entry is determined from the year it started operations. Exit is not directly reported, but is inferred from the last year a firm is in the data. The analysis focuses on the formal private sector, excluding self-employment. The number of firms in the dataset increases every year from 12,753 in 2005 to almost double that number in 2014. This was mostly due to expanding the coverage of firms in the dataset. The distribution of firms is slightly skewed towards micro firms as a result of additional (not new) firms in the data every year. The panel was merged with firm-level customs data (only available for 2011–2014) to obtain export and import data. 2 Chapter 3 of this report is based on Aterido and Petracco (2016). 12 the most part, to have stayed large over the past decade. The persistence of micro firms irrespective of the firms’ age is indicative of their lack of growth. We return to this below. Figure 3-1 Figure 3-2 Total number of firms by firm size and firm age, 2014 Total employment by firm size and firm age, 2014 KOSOVO: TOTAL NUMBER OF FIRMS 2014 KOSOVO: TOTAL EMPLOYMENT 2014 9,000 30,000 8,000 25,000 NUMBER OF WORKERS 7,000 NUMBER OF FIRMS 6,000 20,000 5,000 15,000 4,000 3,000 MISSING 10,000 MISSING 2,000 10+ 10+ 6 –10 5,000 6–10 1,000 0 2–5 0 2–5 1–9 1–9 10–19 1 10–19 1 20–99 20–99 100+ 100+ SIZE SIZE Source: Staff estimates based on firm registry data. Kosovo stands out in international comparisons in terms of the greater contribution of micro firms to employment. In most countries large firms provide most jobs, but in Kosovo both large and micro firms equally account for most employment. Large firms (100 or more employees) account for 35 percent of employment on average but represent only 0.5 percent of the firms. Micro firms provide 36 percent of jobs and comprise 91 percent of the firms. In other countries large firms provide a higher proportion of all jobs because they are proportionally more numerous and because they are larger. In Kosovo, there are fewer large firms and they are relatively small. Zambia, by contrast, has a similar proportion of large firms, but they represent a much higher share of jobs (Figure 3-3 and Figure 3-4). Firms are concentrated in a few sectors. Commerce and services dominate in terms of number of firms, number of entrant firms and number of employees. In 2005 commerce accounted for 55 percent of all firms, Figure 3-3 Figure 3-4 Firm distribution by size, cross-country comparison Employment share by size, cross-country comparison FIRMS OF SHARE SHARE OF EMPLOYMENT BY SIZE BY SIZE 100 100 % OF TOTAL EMPLOYMENT 80 80 % OF TOTAL FIRMS 60 60 40 40 20 20 0 0 KSV-2014 MAR-2006 MDA-2014 PER-2012 VTN-2014 ZMB-2010 KSV-2014 MAR-2006 MDA-2014 PER-2012 VTN-2014 ZMB-2010 1−9 1−9 10−19 10−19 20−99 20−99 100+ 100+ Notes: KSV – Kosovo; MAR – Morocco; MDA – Moldova; PER – Peru; VTN – Vietnam; ZMB – Zambia. Source: Staff estimates. 13 declining to 45 percent over time. The share of firms in the services sector increased from 21 percent to 28 percent and for mining, utilities, and construction from 7 percent to 10 percent between 2005 and 2014. Jobs shifted mainly away from manufacturing towards commerce or, to a lesser extent, services. The contraction of sectoral investments into industry and mining (between 1980 to 1990 sectoral investment flows into industry and mining dropped 90 percent3) resulted in the contraction of the industrial sector. Furthermore, employment in some sectors is dominated by a few large industries. For example, the 4-digit industry level growing of vegetables and melons, roots and tubers (ISIC rev. 4 code 113) employs 35 percent of the agricultural labor force. Electric power generation, transmission and distribution (ISIC rev. 4 code 3510) accounts for 21 percent of employment in indus- try. The commerce and services sectors are less concentrated in the wholesale of food, beverages and tobacco (ISIC rev. 4 code 4630) at the top accounting for just 12 percent of the labor market. Geographic concentration of firms is high. Roughly 35 percent of firms are located in Pristina, a trend that has increased over time. Kosovo’s economy, like that of many other countries, is concentrated along regional and sectoral lines. The next largest regional concentration is in Prizren, but it has less than 16 per- cent of firms. Furthermore, 54 percent of jobs are in the capital, which is similar to comparator countries like Moldova. The largest firms are in Pristina, but even these firms are relatively small. The 99th percentile in Pristina hardly reaches the 100-employee threshold, while the distribution of employment across the other regions varies only slightly. Few firms are exporters, even fewer are foreign-owned. Only four percent of firms exported in 2014; these are larger firms. Of those exporting, the average share of exports was only five percent of sales. Since 2011, there has been little increase in either number of firms exporting or in share of exports relative to total sales. Only 0.7 percent4 of firms are foreign-owned and they are only slightly larger than domestic firms. Kosovo’s firms, constrained by infrastructure, regulatory, and human capital bottlenecks, are not taking full advantage of the EU export market. Kosovo is one of the least well-integrated countries in the region, and exports remain considerably lower than for all other countries in Southeastern Europe and comparable small states. Export markets are not diversified, with the top five export destinations accounting for almost 61 percent of Kosovo’s exports of merchandise in 2014 (with the top client, Italy, alone accounting for 15.3 percent). Product con- centration is also high—Kosovo’s exports are concentrated on low value-added items, such as ferro-nickel, which accounts for more than half of total exports. A number of factors contribute to the poor export performance. The limited transport infrastructure increases the cost of accessing international markets, and the cost of intermediate inputs and imports. Business environment constraints and difficulties with accessing workers with the right qualifi- cations also reduce competitiveness. For instance, exporters are 76 percent more likely to report that customs or trade regulations are a constraint.5 About 32 percent of firms employing 10 to 249 workers report “finding suitable personnel and filling vacancies” as a major problem.6 JOB CREATION AND FIRM GROWTH OVER THE PAST DECADE Most firms are small, but do these firms grow? Create jobs? This section explores the dynamics of job creation and job destruction, and hence net job creation, by formal firms over 2005–2014. Net job creation has declined. From 2005 to 2014, roughly 63,000 jobs were created by formal firms (18 per- cent due to startups). At the start of the period, an average of over 7,000 net jobs were created, peaking in 2010. Annual net job creation decreased after 2010 to just over 1,600 net jobs in 2014. During this period, job creation remained steady but job destruction increased, mainly due to downsizing, not firm exit. Sectors such as commerce and services have done consistently well in terms of net job creation over the past decade, and the case of commerce, there has been no decline in net jobs created during 2010–2014; the contribution of agriculture to net job creation has been minimal. Mining, manufacturing and utilities experienced a deep 3 Jacobs, D., Herderschee, H., and Demekas, D. 2002. 4 Moldova for example has eight percent of private formal firms being foreign owned. 5 Our analysis based on BEEPS data reveals that, even after controlling for other firm characteristics, exporters are more likely than other firms to report that transportation, the working of the courts and the customs/trade regulations are constraints to business. 6 KOSME SME Survey. 14 decrease in net job creation in 2008–9, recovered soon after, but have been in rapid decline since 2010 attain- ing net job destruction of almost four thousand jobs by 2014. This is highlighted in Figure 3-5 and Figure 3-6. Figure 3-5 Figure 3-6 Job flows 2005–2014 Net job creation by sector KOSOVO: JOB FLOWS (000) KOSOVO: NET JOB CREATION (000) 20 BY SECTOR 4 10 2 TOTAL (000) TOTAL JC(000) 0 0 –10 −2 –20 −4 2006 2008 2010 2012 2014 2006 2007 2008 2009 2010 2011 2012 2013 2014 YEAR YEAR CREATION AGROFORESTFISH NET CONSTRUCTION DESTRUCTION SERVICES MINIMUM THRESHOLD: 1 EMPLOYEES MINMANUUTIL COMMERCE Source: Staff estimates based on firm registry data. Source: Staff estimates based on firm registry data. The greatest source of net job creation is startups, but the rate of entry is relatively low. Between 2005 and 2014, 18 percent or new jobs were due to entry. Most jobs in Kosovo and in other countries are created, or destroyed, by incumbent firms, but in Kosovo – except in 2011 – incumbent firms have been shed- ding more jobs than they are creating. Unlike incumbent firms, entrants were responsible for positive net job creation during 2005–14. However, at rates of five percent to seven percent, entry is low compared to other counties, though the exit rate is also one of the lowest. Kosovo’s economy is relatively static in terms of entry/ exit, lacking the churning process that, in more developed markets, reallocates labor from less efficient to more efficient firms (Figure 3-7 and Figure 3-8).7 Figure 3-7 Figure 3-8 Job flows in Kosovo: Incumbents, entering & exiting Entry and exit rates across countries INCUMBENTS−ENTERING−EXITING COUNTRIES ENTRY AND EXIT RATES 20 20 10 15 TOTAL (000) 0 RATE % 10 −10 5 −20 2006 2008 2010 2012 2014 0 YEAR DEN EST FRG HUN ITA KSV-2005-08 KSV-2005-14 KSV-2009-14 MDA MEX ROM SVN USA JC BY INCUMBENTS JD BY INCUMBENTS NET BY INCUMBENTS JC BY ENTRY JD BY EXIT ENTRY RATE MINIMUM THRESHOLD: 1 EMPLOYEES; OUTLIERS EXCLUDED: OUTL_LNLPQ EXIT RATE Source: Staff estimates based on firm registry data. Source: Staff estimates for Kosovo and Moldova. Bartelsman et al. (2009) for all other countries 7 Jovanovic (1982), Hopenhayn (1992), and Asplund and Nocke (2006). 15 Survival rates increase with firm size and are higher for firms created in later years. On average, 48 percent of Kosovo’s firms survive beyond the first five years of operation. Survival rates are considerably higher among larger firms – only 45 percent of micro firms do not exit in the first five years, compared to more than three quarters of firms in the 20+ size category. Encouragingly, survival rates have been increasing over time – only 17 percent of firms that started operations in 2005 made it into the 6th year, compared to 31 percent of firms that started operations in 2009. This is the more remarkable given the fact that this last firm cohort spans the 2009–14 period, when many of the countries in Western Balkans and in Europe more broadly were suffering in the aftermath of the financial crisis. The data show little variation in survival rates across regions in Kosovo, whereas across sectors the highest survival rates are in business/finance, as well as mining/quarrying (although the latter have very few firms overall), while in hotels/restaurants and construction survival rates appear to be relatively lower (Figure 3-9). Figure 3-9 Firm survival rates (% of firms in category) FIRM SIZE 100% 87% FIRM COHORT 90% 78% 78% 81% 35% 80% 31% 70% 67% 30% 26% 60% 25% 22% 23% 48% 20% 50% 45% 20% 17% 40% 15% 30% 10% 20% 10% 5% 0% 0% 1–9 10 –19 20– 49 50 –99 100 – 499 500+ TOTAL 2005 2006 2007 2008 2009 TOTAL REGION 60% 56% SECTOR 49% 50% 49% 48% 50% 47% 47% 48% 80% 72% 40% 70% 66% 60% 53% 53% 52% 30% 46% 48% 50% 43% 40% 43% 40% 35% 20% 30% 20% 10% 10% 0% 0% PRISTINA FERIZAJ GJAKOVA GJILAN MITROVICA PEJAPEC PRIZREN TOTAL E NG NG ES N CE T TS E ES L TA UR IC NC O TI AN IC ER YI RI TI TO E, LT ILI A RV UC TU RR M UR G IN UT CU SE M RA UA AC ,F TR TA RI CO ER TO SS NS ,Q UF ES AG NE TH ,S CO ,R AN G SI O IN RT LS M BU IN O TE M SP HO AN TR Source: Staff estimates based on firm registry data. Most new entrants are micro firms that create jobs, but they do not add as many jobs as large firms. Micro firms contribute 34 percent of job creation and 46 percent of job destruction while the few large firms create and destroy jobs at rates of 27 percent and 33 percent, respectively. However, the average large firm, if it survives, adds 400 employees over a 10 year span while the average micro firm adds one (Table A-1 shows the relationship between employment and firms’ characteristics). Although firms tend not to create jobs beyond four years, the average surviving large firm starts at 300 employees and grows to almost 800 while the average micro firm stays almost the same. Firms that enter the market small, remain small. More than half of micro firms stay in the same size category from the time of entry to age five. More generally, comparing the age of a firm at any time (t) and three years later (t+3), more than three-quarters of firms with fewer than ten employees stay in the same size bracket (Box 2). The next section examines some of the constraints to the ability of firms to grow over time. 16 BOX 2: A COMPARATIVE VIEW OF FIRM DYNAMICS IN KOSOVO AND MOLDOVA A comparison of firm dynamics in Kosovo and Moldova suggests different patterns – in Kosovo small firms survive but do not grow, while in Moldova fewer small firms survive, but the ones that do become larger. The transition matrix that tracks how size category changes over time, including exit, implies more than half of micro firms in Kosovo stay in the same size category from startup to age 5, while 36 percent exit. Compared to Moldova where 63 percent of micro firms have exited by age 5, survival of micro firms in Kosovo is high but they do not grow much. Larger firms in Kosovo have both greater survival and growth potential in comparison with Moldovan firms, however few firms are these size categories. Transition Matrix: Kosovo versus Moldova KOSOVO 2005–2014 MOLDOVA 2003–2014 SIZE AT AGE 5 SIZE AT AGE 5 SIZE 1–9 10–19 20–99 100+ EXIT 1–9 10–19 20–99 100+ EXIT 1–9 59% 3% 1% 0% 36% 30.84 3.52 1.99 0.22 63.43 10–19 24% 26% 28% 2% 20% 21.70 24.89 16.08 1.82 35.51 SIZE AT BIRTH 20–99 0.07 0.14 0.49 0.17 0.13 10.30 11.24 37.24 8.67 32.55 100+ 0.04 0.00 0.15 0.67 0.15 6.90 1.72 15.52 48.28 27.59 SIZE AT T + 3 SIZE AT T + 3 1–9 77% 3% 1% 0% 19% 44.60 3.53 0.93 0.07 50.86 SIZE AT T 10–19 21% 46% 23% 0% 10% 23.90 37.26 12.14 0.27 26.42 20–99 6% 12% 66% 7% 8% 7.37 13.61 61.43 3.57 14.02 100+ 2% 2% 14% 72% 10% 3.09 1.43 22.96 66.24 6.27 Figure A and B show the growth rates of the average surviving firm in each size category in Kosovo and Moldova, normalizing the average size at startup to 1. By 10 years, both a micro firm and a large firm are approximately 2.5 times larger in Kosovo than when they started. Though the growth rates are similar, the starting point is what makes the difference in terms of absolute number of jobs added. By comparison, in Moldova micro firms grow more robustly: a micro firm at 10 years will be more than 6 times larger, while a large firm may even become smaller. Figure A Figure B Average size by size at startup and age, Kosovo Average size by size at startup and age, Moldova AVERAGE SIZE BY SIZE AT BIRTH AND AGE AVERAGE SIZE BY SIZE AT BIRTH AND AGE 3 8 2.5 6 AVERAGE SIZE AVERAGE SIZE 2 4 1.5 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 AGE AGE 1−9 1−9 10−19 10−19 20−99 20−99 100+ 100+ 17 Figure 3-10 Figure 3-11 Only large firms grow with age Most firms create new jobs only in the first two years of operation AVERAGE SIZE BY SIZE AT BIRTH AND AGE NET JOB CREATION SINCE BIRTH BY AGE 600 25 500 20 NET JOB CREATION (000) 400 AVERAGE SIZE 15 300 10 200 5 100 0 0 1 2 3 4 5 6 7 8 9 10 11 12 −5 AGE 1 2 3 4 5 6 7 8 9 10 11 12 1−9 1−9 10−19 10−19 20−99 20−99 100+ 100+ Source: Staff estimates based on firm registry data. In line with the sectoral and geographic concentration of firms, most jobs were created in Pristina, in the services sector. Beyond entrance, large firms in the commerce and services sector created jobs during the 2005–2014 period, while manufacturing sector firms shed them (Figure 3-12 and Figure 3-13). Geographically, Pristina accounted for 58 percent of net job creation; the other regions had job creation rates in single digits. Figure 3-12 Figure 3-13 Net job creation: manufacturing, mining, utilities & construction sectors Net job creation: commerce & services sectors KOSOVO 2005–2014: NET JOB CREATION KOSOVO 2005–2014: NET JOB CREATION MANUFACTURING COMMERCE AND SERVICES 8,000 16,000 6,000 14,000 12,000 NET JOB CRETION NET JOB CRETION 4,000 10,000 MISSING 2,000 10+ 8,000 6–10 6,000 MISSING 0 2–5 4,000 10+ 1–9 1 10–19 2,000 6–10 –2,000 20–99 100+ 2–5 0 –4,000 –2,000 1– 9 1 10 –19 20 – 99 BASE SIZE 100+ BASE SIZE Source: Staff estimates based on firm registry data. Small, but not micro firms, especially in Pristina, have shown most growth potential. Data show that only eight percent of firms overall and only 6 percent of micro (size 1–9) firms transitioned into the next size category over a three year period. Firms in the 10–19, 20–49 and 50–99 size categories, while many fewer, are much more likely to transition into the next size category (e.g. 31 percent of firms in the 10–19 employees size category did so). It is likely that the nature of micro firms, most of them in the services sector, may not be conducive to growth. Note, for instance, low transition rates in sectors such as hotels/restaurants as well as commerce, manufacturing and other types of services. Firm growth rates appear to be the highest in business/ finance and in construction sectors (Figure 3-14). Geographically, consistent with the evidence that most jobs were created in Pristina, firms in Pristina are considerably more likely to transition into a higher size category than firms from elsewhere. Finally, across cohorts, growth rates appear to have been higher in earlier years, and somewhat lower for firms that started operations during 2008–10. 18 Figure 3-14 Share of firms that grew in 3 years (% of category) FIRM SIZE FIRM COHORT 35% 14% 31% 12% 30% 28% 12% 11% 10% 10% 25% 24% 10% 20% 8% 7% 7% 6% 15% 6% 10% 8% 8% 4% 6% 5% 2% 0% 0% 0% 1–9 10–19 20–49 50–99 100–499 500+ TOTAL 2005 2006 2007 2008 2009 2010 TOTAL REGION 12% 11% 10% SECTOR 8% 8% 25% 23% 6% 7% 7% 7% 21% 6% 6% 20% 6% 14% 15% 10% 10% 4% 9% 10% 10% 8% 6% 7% 4% 2% 5% 0% 0% PRISTINA FERIZAJ GJAKOVA GJILAN MITROVICA PEJAPEC PRIZREN TOTAL E NG NG ES N CE T TS CE ES L TA UR IC O TI AN IC ER AN YI RI TI TO E, LT ILI RV UC TU RR M UR G IN UT CU SE M RA UA AC ,F TR TA RI CO ER TO SS NS ,Q UF ES AG NE TH ,S CO ,R AN G SI O IN RT LS M BU IN O TE M SP HO AN TR Source: Staff estimates based on firm registry data. WHAT CONSTRAINS FIRMS’ GROWTH? Modern businesses require reliable social and physical infrastructure to compete and grow; Kosovo falls short of its neighbors in a number of key dimensions. High cost and low quality of infrastructure, both social (regulatory environment, human capital), as well as physical (power supply, physical and IT connectivity) undermine firms’ competitiveness, particularly in export markets. Clear rules to govern trade and investment can help improve productivity, innovation, and help firms connect to global value chains. Energy insecurity is a notable cost to busi- ness and it stymies FDI. Connectivity is vital for successful integration in a small landlocked economy. Improvements in human capital are necessary both to increase productivity, and for low-income households to benefit from growth. (World Bank, 2016). While Kosovo has made progress in some of these areas, notable challenges remain. Reliable, competitively priced energy is central to growth and investment. Energy security in Kosovo remains at risk. The Government has committed to closing the 50-year-old, highly inefficient and polluting Kosovo A Power Station in an economically, environmentally, and socially acceptable manner, but this will increase the existing energy supply gap. Without reliable energy, long-term investments in other sectors may be postponed. The looming energy crisis is a major concern, particularly given that most neighboring countries are also in energy deficit. Meeting demand growth would require increasing generation capacity, achieving renewable energy and energy efficiency targets, implementing the market reforms outlined by the Energy Community8, and improving the transmission network and distribution system. Kosovar firms perceive unreliable electricity supply as the second most significant obstacle. In 2013, 81 percent of Kosovar firms experienced an average of 13.6 power outages per month, the worst performance 8 The Energy Community is an international organization dealing with energy policy. The organization was established by an international treaty in Oct 2005 in Athens, Greece. The Treaty entered into force in Jul 2006. The Treaty establishing the Energy Community brings together the European Union, on one hand, and countries from the South East Europe and Black Sea region. https://www.energy-community.org/ 19 in SEE.9 A recent study of the U.S. Agency for International Development found that, when all costs and losses are considered, Kosovo businesses realize €359 million in additional annual costs and losses because of unreli- able power supply and that SMEs were the most affected.10 The fiscal burden of unreliable supply is higher for microenterprises (44.5 percent of average annual turnover) and small businesses (21 percent of average annual turnover) relative to medium enterprises (3.9 percent) and large businesses (1.8 percent). Improving connectivity – roads and communication networks – is also important for future growth. Ade- quate balancing of new investments and good quality maintenance of existing assets is the key. Kosovo has invested heavily in new physical capital. Most high-level main roads and local roads have been developed and upgraded, but many national and regional roads did not receive much attention. Kosovo’s average investment in maintenance and rehabilitation has dropped significantly in the past few years. The railway infrastructure is outdated, but only the north-south section of it is considered to be worth investing, connecting with European and Asian markets. A better utilization of mining sector requires – among other improvements – an upgrade of the railway system. Figure 3-15 Infrastructure remains a constraint to growth ELECTRIC POWER CONSUMPTION (MRV YEAR) ROADS, PAVED (MRV YEAR) MNE (2012) BIH (2010) SRB (2012) MDA (2011) BIH (2012) MNE (2011) KSV (2012) SRB (2011) ALB (2012) ALB (2002) MDA (2012) KSV (2010) 0 1,000 2,000 3,000 4,000 5,000 6,000 0.0 20.0 40.0 60.0 80.0 100.0 KWH PER CAPITA % OF TOTAL ROADS Source: 2016 Doing Business Report, World Bank Reducing bottlenecks in broadband electronic communications is strategic for competitiveness. ICT provides income to high-skilled and lower-skilled labor, and it also supports economic growth through creating direct, indirect, and induced jobs. The underdeveloped telecommunications infrastructure hinders individual and firm productivity, poses a constraint to FDI and to the regional economic integration. Kosovo’s wireless market is underdeveloped with a per capita penetration rate of only 89.1 percent in end-2013, in contrast to the average of 157 percent for the western Balkan region (World Bank, 2016). Governance indicators have improved, but Kosovo ranks lower in most categories compared to neigh- boring countries, particularly in terms of political stability and rule of law. The shadow economy is more likely to be small where a government can control corruption, maintain the rule of law, and promote regulatory qual- ity (Packard, Koettl and Montenegro, 2012; Perry, 2007). Worldwide Governance Indicators reveal some improve- ments across several dimensions such as political stability, government effectiveness and rule of law between 2010 and 2015. However, Kosovo continues to perform poorly compared to the ECA average or other Western Balkans countries (Figure 3-16). These institutional challenges are intrinsically linked to labor market concerns. Political instability discourages workers and has been one of the important push factors contributing to outmigration. The quality of public services, including education, remains poor, which creates incentives for firms to remain small and informal. Across the ECA countries in the LITS survey, Kosovo is the third worst per- former in terms of perceptions of overall performance of local governments. Public schools are perceived to 9 See “Kosovo,” Business Environment and Enterprise Performance Survey (BEEPS) V Country Profile, European Bank for Reconstruction and Development and World Bank, London, http://ebrd-beeps.com/wp-content/uploads/2015/04/kosovo.pdf. 10 U.S. Agency for International Development, “The Effect of Unreliable Power Supply and Quality on Kosovar Businesses,” Presentation, December, 2012, USAID, Pristina, Kosovo. 20 be overcrowded, and hospitals in poor condition, relative to other ECA countries (Figure 3-17). Furthermore, Riinvest (2013) found that representatives of over 40 percent of firms feel tax evasion is justified under some cir- cumstances, and, when asked about the fairness of the tax system “given the level of public services,” 60 percent believe they should pay less in taxes. The recent Kosovo Risk and Resilience Assessment notes that the low quality of public services is due, in part, to the clientelistic nature of public administration, which have poorer capacity and a lower degree of accountability. Furthermore, a decreased level of satisfaction with public services has also been associated with a decrease in the satisfaction with political institutions, which, according the UNDP Public Pulse polls has been declining over the 2009–2016 period, with current approval ratings of central level institutions at a low 20 percent, compared to over 50 percent in 2009 (World Bank, 2017). Figure 3-16 Government performance: percentage of respondents who think the overall performance of local government at level is bad or very bad 60 50 40 30 20 10 0 BOSNIA SERBIA KOSOVO CROATIA KYRGYZSTAN ITALY ROMANIA MACEDONIA BULGARIA POLAND AZERBAIJAN MOLDOVA UKRAINE RUSSIA TURKEY SLOVAKIA MONTENEGRO LITHUANIA MONGOLIA ARMENIA LATVIA SLOVENIA CZECH REPUBLIC ALBANIA KAZAKHSTAN TAJIKISTAN GEORGIA GERMANY BELARUS GREATBRITAIN HUNGARY UZBEKISTAN ESTONIA SWEDEN FRANCE Source: Life in Transition Survey II (2011). Figure 3-17 Quality of public education: percentage of respondents who found overcrowded classrooms in public schools in the last year 40 35 30 25 20 15 10 5 0 KOSOVO ITALY FRANCE TURKEY MACEDONIA ALBANIA TAJIKISTAN GERMANY ROMANIA KYRGYZSTAN HUNGARY MONGOLIA UKRAINE KAZAKHSTAN CZECH REPUBLIC POLAND CROATIA LITHUANIA MOLDOVA SLOVENIA RUSSIA ARMENIA SERBIA BOSNIA UZBEKISTAN MONTENEGRO BELARUS SLOVAKIA GREATBRITAIN ESTONIA LATVIA SWEDEN GEORGIA AZERBAIJAN BULGARIA Source: Life in Transition Survey II (2011). 21 Figure 3-18 Quality of public health: percentage of respondents who found frequent and unjustified absence of doctors in public health clinics or hospitals in the last year 40 35 30 25 20 15 10 5 0 KOSOVO MACEDONIA MONTENEGRO UKRAINE BOSNIA KYRGYZSTAN BELARUS SERBIA KAZAKHSTAN UZBEKISTAN RUSSIA ALBANIA SLOVENIA HUNGARY LITHUANIA TURKEY MOLDOVA BULGARIA POLAND TAJIKISTAN SLOVAKIA ARMENIA ITALY ROMANIA GERMANY MONGOLIA SWEDEN CZECH REPUBLIC ESTONIA AZERBAIJAN CROATIA GREATBRITAIN FRANCE LATVIA GEORGIA Source: Life in Transition Survey II (2011). Improving the business climate may help to attract new, job-creating foreign investment. The World Bank’s Doing Business 2017 ranks Kosovo as 60th among the 190 economies it evaluated. This is worse most other SEE6 countries: FYR Macedonia (10), Serbia (47), and Montenegro (51) and Albania (58) are each rated higher; Bosnia and Herzegovina (81) is lower. Kosovo trails other Western Balkans and ECA countries in resolv- ing insolvency.11 On other important indicators – construction permits, access to electricity and credit, protecting minority investors, enforcing contracts – Kosovo (and its comparators) are far from the frontier (Figure 3-20). These drawbacks deter new firms from entering the market and existing firms from expanding. Larger, productive firms are typically the leading providers of formal, private sector jobs (La Porta and Shleifer, 2014) and improving their returns from investment is a policy instrument to create employment and increase the formal sector. Figure 3-19 Figure 3-20 Governance indicators, global rank (100 = best practice) 2015 Doing business indicators (100 = best practice) 2017 CONTROL OF CORRUPTION STARTING A BUSINESS 80 100 DEALING WITH RESOLVING INSOLVENCY 90 60 80 CONSTRUCTION PERMIT 70 VOICE AND 40 GOVERNMENT 60 ACCOUNTABILITY EFFECTIVENESS ENFORCING CONTRACT 50 GETTING ELECTRICITY 20 40 30 0 20 TRADING ACROSS BORDERS REGISTERING PROPERTY RULE OF LAW POLITICAL STABILITY PAYING TAXES GETTING CREDIT PROTECTING MINORITY REGULATORY QUALITY INVESTORS ALB MNE KOSOVO BIH SRB WESTERN BALKANS MDA KOS EUROPE AND CENTRAL ASIA MKD Source: 2017 Doing Business Report, World Bank Source: Kaufmann D., A. Kraay, and M. Mastruzzi (2010); http:// info.worldbank.org/governance/wgi/index.aspx#home Each bar measures the distance to frontier index, where 0 represents the worst performing country and 100 is the best. 11 22 The tax burden is relatively low. The recent SCD synthesis report (World Bank, 2017a) notes that in Western Balkans countries high labor taxes hinder job creation and work incentives by decreasing the appeal of creating (and accepting) formal sector jobs. Compared to other countries in the ECA region and a number of Western European countries, the tax wedge – the difference between the cost of hiring an employee and the wage – tends to be high in the Western Balkans region, particularly in countries like Bosnia and Herzegovina and Serbia. On this issue, Kosovo is an exception, with a low tax wedge, on account of low social security contributions and personal income taxes. Employers face high direct and indirect costs from hiring young women because of maternity leave provisions. Kosovo ranks on the high end of ECA countries in terms of the total duration of paid maternity leave. Women are entitled to nine months of paid family leave and three months of unpaid family leave. In addition to the long duration, the cost of maternity leave is borne largely by employers, who are responsible for two thirds of leave allowance – this is at variance with the ILO recommendations on maternity leave that payments be made through compulsory social insurance or public funds in order to protect the women’s position in the labor market (World Bank, 2015a). Employers report that the costs of a long maternity leave are not only financial, but also in terms of productivity – time and resources need to be invested in hiring and training replacement staff. This situation leads to bias against women of child-bearing age in the labor market, with employers favoring men both in hiring, and in terms of providing the newly hired with longer-term contracts. Focus group participants revealed that employers ask candidates during job interviews about their family status and childcare obligations (World Bank, 2015a). While the tax burden is low, firms face costly regulatory burdens, particularly in the form of inefficient and extensive inspection system. The inefficiencies are due to several factors: (i) multiple inspecting agencies, inspectors, and uncoordinated visits; (ii) insufficient risk-focus – many businesses get inspected despite a low or moderate risk level; (iii) inconsistency, poor coordination and incoherence among inspection agencies; and (iv) focus on finding violations rather than improving compliance and outcomes. Almost all firms are visited by tax authorities, and “gift requests” are commonplace – tax officials inspected more than 90 percent of formal firms over the previous year, the highest in the region according to BEEPS 2013 data. An analysis done by the WB Busi- ness Environment Technical Assistance Project in 2013 found approximately 40 inspectorates at the central level in Kosovo and four to five at the local level. These inspectorates are regulated by approximately 80 to 90 different laws and many more bylaws. Most laws have few or no details on the responsibilities and rights of inspectors and businesses subject to inspection. This makes the recourse or appeal process unclear and uncertain. Previous research has shown this to be problematic. The regulatory burden stunts the growth of small firms and displaces economic activity from SMEs and large firms toward micro firms (Aterido and Hallward-Driemeier, 2009; 2010). Difficulties in developing new products, and in certifying the quality of existing products, constrain firms’ ability to compete in export markets. New product development is essential, especially in export markets. There are insufficient designers and production engineers in Kosovo. Most companies do not have dedicated product research and development functions. In the food processing sector, fragmented land ownership limits the ability of processors and aggregators to deliver sufficient quantities of primary agricultural products at uniform quality standards. Inputs vary greatly in quality, and suppliers to processors and aggregators are not adequately monitored. Furthermore, the standardization, accreditation, and certification systems are weak and increase the cost of quality assurance for firms. Domestic testing facilities are inadequate, and certification bodies cannot offer internationally recognized certification. In the absence of benchmarks, firms do not have standards that they aspire to reach to improve their competitiveness. Also, export markets do not have confidence in Kosovar goods as they lack adequate certification and do not meet required standards. Interviews with firms in the food processing and tourism sectors reveal the lack of appropriate certification and standardization infrastructure among factors that inhibit their growth. In the wood processing sector many companies have certified quality assurance systems (such as ISO 9001), but lack important sector-specific standards such as the chain of custody certification from certified and sustainably managed forests. More broadly, exporters are not able to use testing reports from domestic laboratories or certificates from domestic product certification bodies when they export their products abroad (to the EU market or elsewhere), since the domestic laboratories and certification bodies are not recognized internationally. Weak skills among the workforce penalize companies as they try to grow. Companies express difficulties in finding workers who are ready-to-go upon graduation from vocational education and training (VET) schools. Com- panies also report specific and continuing needs for higher-value and higher-level skills, such as machine oper- ators, design and graphics experts, and marketing and sales professionals. A number of sectors have expressed 23 dissatisfaction with the VET system; additionally, there are inadequate collaboration mechanisms between technical schools, universities, and businesses. The IT sector does not have an adequate number of well-trained programmers or even graduates with basic IT skills, which prevents companies from pursuing larger outsourcing contracts. (Skill mismatches are analyzed in Chapter 5). However, the results of the recent PISA exercise that revealed that more than two thirds of 15 year olds lack basic proficiency in sciences, and over three quarters are below basic proficiency in reading and mathematics highlights the fact that weak workforce skills are not just the outcome of inefficiencies in the design pf vocational training programs, rather, and in particular in light of the young population, improve- ments in general education, and in early education, are a priority. SMEs in all sectors cite issues with marketing and sales as a hindrance to access to export markets. Most companies do not have dedicated sales management and lack knowledge of export markets and how to effectively promote their products. Export promotion is usually done through fairs, and these costs are challenging for com- panies. Companies also lack sustainable connections to export buyers and wholesale and retail chains in the EU, resulting not only in weak sales opportunities as well as limited knowledge of product demand trends. Even in the internal market, local companies find it difficult to compete with imports, in part due to poor design and quality of packaging, even if domestic product quality is equal or superior. Lack of access to finance is one of the main problems for SMEs. Working capital finance and long-term equipment finance are limited and expensive. To secure export contracts and deliver larger quantities of goods, SMEs need growth capital for additional equipment and to expand operations. Investment in R&D, inclusively test- ing and development of new products, is almost nonexistent due to the large up-front investment requirements. Aterido and Hallward-Driemeier (2009) find that improved access to finance has a positive effect on firm growth, especially for firms with more than 10 employees. In Kosovo, lack of access to financing is an important barrier to growth of micro enterprises. Experience from World Bank Kosovo Youth Development Project (KYDP) showed that micro firms had the desire to invest and expand, but were unable to access affordable financing. The EIB’s assess- ment of financing needs of SMEs in Kosovo (EIB, 2016) found that while bank loans are available for small and medium enterprises, micro-enterprises are significantly underserved. Major banks are not serving Kosovo’s micro enterprises, such that micro firms are much more likely to be credit constrained than other types of firms. Constraints to firms’ growth vary across sectors, but issues such as insufficient skills, regulatory defi- ciencies, or lack of finance affect all sectors. Interviews conducted by the World Bank in collaboration with the Jobs and Competitiveness project, as well as the analysis of firm growth constraints under the auspices of the USAID Empower Private Sector Project reveal many common concerns (Table 3-1).12 LABOR PRODUCTIVITY DIFFERENTIALS AND GROWTH DYNAMICS Successful businesses in efficient market economies add inputs to expand production. Low-productivity businesses are more likely to shrink or to close (Haltiwanger, 2011). To gain insights into productivity dynamics in Kosovo over the past decade, this section examines three aspects of allocative efficiency: (i) static relation- ships between productive firms and labor shares (Olley and Pakes, 1996)13; (ii) dynamic allocative efficiency which includes changes in productivity associated with firms’ expansion or contraction and with entry and exit (Foster et al. 2006); and (iii) multivariate (regression) analysis of productivity using sector and location effects. Over the past decade, labor has been allocated to more productive firms, although efficiency remains rel- atively low by international standards. The method of static productivity analysis due to Olley and Pakes (1996) decomposes an index of weighted aggregate productivity into a covariance term and unweighted productivity. The covariance term, if positive, implies that productive firms have larger shares. The analysis found that the efficiency of labor allocation has been on the rise over the past decade (Figure 3-21). Covariance can be decomposed into four components: positive effect of (i) larger firms being more productive or (ii) smaller firms being unproductive, and 12 The team conducted structured interviews with a small set of 14 firms in four tradeable sectors: wood and furniture (3), textile and garments (5), information technology (4), and food-processing (2). The survey was structured in four parts: (i) general information regarding firm and management; (ii) exports section with basic information, trends, and constraints; (iii) employment section, which included trends on employment, types of employment, and constraints; (iv) future projections and their relation to the business environment. 13 The method proposed by Olley and Pakes (1996) decomposes aggregate productivity into an unweighted aggregate productivity and a covariance term. The covariance term measures allocative efficiency, or the extent to which firms with more productivity receive greater labor shares. 24 Table 3-1 Firms express similar concerns to growth in some key sectors in Kosovo Major Product development constraints ã Marketing and sales and quality Workforce skills Finance Wood processing - Limited promotion;  ack of designers and -L Lack of readiness of - -L  ack of access to -L ack of dedicated production engineers; VET school graduates; working capital sales management; Most companies do -  ompanies feel dis- -C finance; -L ack of knowledge not have dedicated connected from VET - Long-term equipment of export markets product R&D functions system; finance is limited and or export promotion  ifficulties in filling -D expensive strategies. high-skill positions -E  xport finance is (CNC machine opera- limited. tors, design/graphics experts, marketing professionals). Apparel and  romotion activities -P Lack of good design- -  ack of access to -L Leather are not systematic ers for updating/ working capital and oftentimes expanding the finance; informal (through product line;  imited equipment -L personal contacts); finance;  romotion costs con- -P  ack of expansion/ -L sidered to be high by growth capita. most companies;  ack of dedicated -L sales management. Food processing -L  imited access to Land ownership frag- - L -  ack of expertise  ack of working -L wholesale/retail food mentation limits the for higher-value capital finance; chains in the EU; ability of processors production;  ack of risk/growth -L -L  ack of knowledge to deliver sufficient nadequate collabora- -I capital for expanding of product design quantities of primary tion with agriculture/ operations. trends; agricultural products technical schools.  oor design and -P at uniform quality quality of packaging, standards; even for high quality Lack of adequate - products. monitoring of suppli- ers for product quality; Contractual relation- - ships between grow- ers and processors not well established; Limited investment - in new product development; Lack of international - food safety and qual- ity certification. ICT Lack of dedicated - - Lack of proactive - Inadequate supply  ack of risk/growth -L marketing resources product development and high turnover of capital that hampers or connections with and R&D good programmers; R&D and product Western markets. - Underdeveloped development. cooperation between universities and VET schools in the ICT field. Tourism Lack of overall tourism - -A  doption of inter­ Insufficient workforce -  ifficulty attracting -D strategy; national standards for training; bank finance due to  o national tourism -N sustainable tourism/ No education/training - high risk and small organization (NTO) in accommodation has institution devoted to size of firms. charge of marketing; been slow. hospitality/tourism.  oor infrastructure -P (roads, maps, signs, park/monument designation). Source:  Results based on structured team interviews with firms in Kosovo Figure 3-21 Figure 3-22 Percentage of firms that were inspected by tax officials over the year, 2013 Percentage of firms where an informal gift was requested by tax officials, 2013 100 90 14 80 12 70 60 10 50 8 40 30 6 20 4 10 2 0 KOSOVO ALBANIA MONTENEGRO MACEDONIA BOSNIA AND HERZEGOVINA BULGARIA ROMANIA HUNGARY SERBIA CROATIA SLOVENIA 0 ALBANIA KOSOVO BOSNIA AND HERZEGOVINA BULGARIA ROMANIA MONTENEGRO SERBIA MACEDONIA SLOVENIA CROATIA HUNGARY Source: BEEPS 2013 data. Source: BEEPS 2013 data. negative effect of (iii) larger firms being less productive or (iv) smaller firms being more productive (Figure 3-22). In Kosovo, while large productive firms account for much of this positive variance, the contribution of small unproductive firms has also been important. There are sectoral differences, but this trend has been consistent. For instance, in both manufacturing and services larger labor shares are consistently associated with more productive firms (Figures 3-23 and 3-24). In 2014, the covariance term is 0.08 in manufacturing and 0.26 in services, indicating that the services sector is more efficient in its labor allocation. In comparative terms, though, efficiency in both sectors is low – for U.S. manufacturing the covariance terms were 0.5 in the 1990s. Entry and within-firm effects had the strongest effects on productivity increases. The dynamic allocative efficiency analysis reveals that firms can increase productivity by shedding workers, but the more productive firms expanded (Figure 3-25). Firms that entered the market at the beginning of the period (2005), whether as new companies or spinoffs from state owned enterprises, contributed the largest productivity gains in the early years. This is especially pronounced in sectors such as transport, storage and communications (Figure 3-26). After the financial crisis productivity changes came as large “within” (i.e., holding shares constant) firm productivity losses in 2010, recovering in 2011. Firms that became more productive reduced their employment shares, as indicated by the negative “cross term”, but the effect was small. Productive firms increased employment shares, as “between” effects were mostly positive, but small. On the whole, firms’ exit brought negligible gains to productivity: ineffi- cient firms exited, but not enough to make a difference. The large “within” effects related to the financial crisis14 Figure 3-23 Figure 3-24 Static decomposition of output per worker index in all sectors Component of variance by firm size KOSOVO: OUTPUT PER WORKER (LOG) KOSOVO: COMPONENTS OF COVARIANCE ALL SECTORS ALL FIRMS 3.40 0.80 1.00 OUTPUT PER WORKER (LOG) AGGREGATE PRODUCTIVITY 3.20 0.60 COVARIANCE 3.00 0.50 0.40 2.80 0.20 0.00 2.60 2.40 0.00 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 −0.50 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 AGGREGATE PRODUCTIVITY COVARIANCE LARGE-PRODUCTIVE WEIGTHS EMPLOYMENT LARGE-UNPRODUCTIVE SMALL-PRODUCTIVE SMALL-UNPRODUCTIVE 14 Being a closed economy, the demand shock after the financial crisis is likely to have come from lower remittances. 26 Figure 3-25 Figure 3-26 Static decomposition of output per worker index in manufacture Static decomposition of output per worker index in services KOSOVO: OUTPUT PER WORKER INDEX KOSOVO: OUTPUT PER WORKER INDEX MANUFACTURE SERVICES 0.40 0.48 0.20 0.75 AGGREGATE PRODUCTIVITY AGGREGATE PRODUCTIVITY 0.46 0.25 0.10 0.70 COVARIANCE COVARIANCE 0.44 0.65 0.10 0.00 0.42 0.60 −0.05 0.40 −0.10 0.38 0.55 −0.20 −0.20 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 AGGREGATE PRODUCTIVITY AGGREGATE PRODUCTIVITY COVARIANCE COVARIANCE WEIGTHS EMPLOYMENT WEIGTHS EMPLOYMENT Source: Staff estimates based on firm registry data.  Figure 3-27 Figure 3-28 Dynamic decomposition of changes in productivity (2005 to 2014) Dynamic decomposition of changes in productivity by sector KOSOVO KOSOVO CHANGE IN OUTPUT PER WORKER (LOG) CHANGE IN OUTPUT PER WORKER (LOG) 0.30 0.04 0.03 0.20 0.02 0.01 0.10 0.00 0.00 −0.01 AGRICULTURE MININGQUARRYING MANUFACTURING UTILITIES CONSTRUCTION COMMERCE TRANSPORTSTORAGECOMM HOTELSRESTAURANTS BUSINESSFINANCE OTHERSERVICES −0.10 −0.20 2006 2007 2008 2009 2010 2011 2012 2013 2014 WITHIN FIRM BETWEEN FIRMS CROSS ENTRY WITHIN FIRM EXIT BETWEEN FIRMS WEIGHT EMPLOYMENT CROSS ENTRY EXIT Source: Staff estimates based on firm registry data. WEIGHT EMPLOYMENT PERIOD 2005−2014 suggest that the productivity changes resulted from demand shocks, especially in sectors such as commerce, and not from expansion or contraction. Larger, foreign-owned and exporting firms are more productive. Regressions on determinants of productivity, measured as the logarithm of revenue per worker, and accounting for year, location (prov- ince level), and industry (2-digit level) effects, show a positive relationship between productivity and firm characteristics such as firm size, foreign ownership and exporter status (Box 3), as the share of sales from exports is also a significantly positive determinant of productivity (Annex Table A-2). Across sectors, relative to manufacturing, productivity is higher in commerce, construction, and agriculture, while lower in services. Finally, there is evidence that sectors with higher labor concentration are neither more nor less productive, though the ones with more revenue concentration (a proxy for less competition) are less productive. A high degree of concentration in sales markets, as measured by the Herfindahl index, is correlated with lower productivity. Monopolies in Kosovo’s small market would be normal, however, their low productivity likely indicates that these firms may be not investing or innovating. 27 BOX 3: EXPORTERS: LIVING LONGER & DOING MORE Since 2011, around 10 percent of firms in the data set have exported, with an average of 23 percent of revenue com- ing from exports. These firms are mostly in two sectors: manufacturing (~32 percent) and commerce (~37 percent), and largely in the capital (53 percent) and to a lesser degree in Prizren (12.5 percent). On average exporting firms are slightly older and significantly larger than non-exporters. Regressions also show that exporters generate significantly more employment than non-exporters. There are other important differences. Wages at exporters are significantly higher, though there is no significant differ- ence in wage growth. Exporting firms enjoy a higher rate of survival than other firms. Regression results indicate that the probability of a firm exiting the market is significantly lower – nearly 28 percent – for an exporting firm. Exporting firms are significantly more productive even when controlling for sector and region. The literature suggests two reasons for this: more productive firms self-select into exporting and/or learning by exporting. The dataset does not allow for disentangling the cause of this relationship, and both may be applicable. Small, exporting, and more productive firms are more likely to grow. Regression analysis of incumbents’ employment growth15 (post-entry growth excluding exit) reveal that small firms16 have higher growth rates17 relative to micro firms, holding constant other variables: age, ownership and firm location, year and industry (Annex Table A-2). However, the higher growth rate of small firms disappears when a measure of productivity is accounted for. As firms age, their rates of growth tend to fall. In developing economies small firms typically have higher employment growth (see Ayyagari et al.) Foreign firms grow slower, which could be a symptom of business environment constraints18. The second and third regressions add an exporter dummy and dum- mies indicating the lagged productivity quartile, respectively19. These characteristics are associated with higher employment growth. The firms in the 25–75th percentile of the firm productivity distribution are likely to grow 15 percent more than those in the bottom 25th percentile. This is consistent with “creative destruction”, i.e., with more productive firms expanding and less productive ones contracting. Productivity increases are linked to contracting employment in the earlier years and with expanding employment toward the end of the period. Regression estimates that relate average effects of within firm changes in productivity and with changes in employment reveal that, overall, a 10 percent increase in produc- tivity is associated with a small contraction in employment of 0.6 percent. This effect is not linear, becoming smaller as firms increase productivity. Firms’ age is positively correlated with employment growth. When the relationship between productivity and employment growth is examined separately for two periods (2005–2010 and 2011–2014), the estimates point to a negative relationship between productivity and employment growth during the earlier years, and an expansion of employment in more productive firms in the later years. Larger, more productive, and exporting firms are less likely to exit. Examining the number of firms and workers by cohort suggests that more firms entered Kosovo’s private sector in the later cohorts, but they have been adding fewer workers. The number of firms has declined rapidly and evenly in all cohorts, but less so the 15 Here the measure of employment growth is the change in the enterprise’s number of workers divided by the firm’s simple average of workers. The measure is symmetric around zero and bounded by values -2, and +2. While monotonically related to the conventional growth rate and a second order approximation of the logarithmic first difference, this measure allows computing meaningful growth rates for firms suffering sharp expansions or contractions, avoiding any arbitrary treatment of outliers. This measure has been extensively used to measure job creation and job destruction (see Davis and Haltiwanger,1992, and Davis, Haltiwanger and Schuh, 1999) 16 The dummy variables for size refer to the average size between the two years of growth period. This avoids the upward bias of coefficients estimated using the base year or the downward bias that results from using the end year of the growth period. This is done to account for the inverse correlation between firm size and growth or “reversion to the mean effect” that would yield spurious results. 17 Growth rates are likely to be higher the smaller the firm is, although at the same rate a larger firm will add more jobs. Micro firms in this case do not grow. As we have seen in unconditional analysis, large firms, controlling for productivity have a negative growth compared to micro firms. 18 Foreign firms and exporting firms are not the same. Only 10 percent of foreign firms are exporters, though that is significantly more than the 4.2 percent of domestically owned firms that export. Note that both foreign and exporting firms tend to be larger, on average, but the employment growth regressions are conditional on firm size. 19 The omitted category is the bottom 25 percentile less productive firms. LPQq2 includes firms in the middle (from 25 to 75 percentile) and LPQq3 is the top 25 percentile. 28 Figure 3-29 Productivity and employment in Kosovo EMPLOYMENT (LOG) - FIXED EFFECTS (1) (2) (3) (4) (5) 2005-2010 2011-2014 2011-2014 AGE 0.0524*** 0.0525*** 0.0593*** 0.0384*** 0.0383*** (0.00134) (0.00133) (0.00177) (0.00195) (0.00194) Y/L -LOG –0.0652*** –0.0381*** –0.0660*** 0.0367*** 0.0363*** (0.00329) (0.00790) (0.00962) (0.00805) (0.00802) Y/L SQUARED -LOG –0.00526*** –0.00712*** –0.0252*** –0.0253*** (0.00143) (0.00154) (0.00151) (0.00150) EXPORT SHARE 0.201*** (0.0437) CONSTANT 0.723*** 0.695*** 0.727*** 0.793*** 0.794*** (0.0106) (0.0125) (0.0179) (0.0211) (0.0211) OBSERVATIONS 142,102 142,102 73,987 68,115 68,115 R-SQUARED 0.107 0.107 0.101 0.057 0.058 NUMBER OF ID 33,526 33,526 22,684 25,154 25,154 R2 0.107 0.107 0.101 0.0567 0.0575 R2-ADJUSTED 0.106 0.107 0.100 0.0566 0.0575 BETWEEN R2 0.00527 0.00610 0.0299 0.0156 0.0138 OVERALL R2 7.58E-05 0.000289 0.0185 0.0159 0.0141 WITHIN R2 0.107 0.107 0.101 0.0567 0.0575 ROBUST STANDARD ERRORS IN PARENTHESES *** p<0.01, ** p<0.05, * p<0.1 Figure 3-30 Figure 3-31 Number of firms (‘000) by cohort Number of workers (‘000) by cohort KOSOVO: SURVIVAL KOSOVO: SURVIVAL N. OF FIRMS (000) BY COHORT N. OF WORKERS (000) BY COHORT 1.2 10 N. WORKERS (000) N. FIRMS (000) 1 8 .8 6 .6 4 .4 2 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 YEAR YEAR COHORT2005 COHORT2006 COHORT2005 COHORT2006 COHORT2007 COHORT2008 COHORT2007 COHORT2008 COHORT2009 COHORT2010 COHORT2009 COHORT2010 COHORT2011 COHORT2012 COHORT2011 COHORT2012 COHORT2013 COHORT2013 MINIMUM THRESHOLD: 1 EMPLOYEES MINIMUM THRESHOLD: 1 EMPLOYEES Source: Staff estimates based on firm registry data. 29 number of workers. Instead, the surviving firms have grown (Annex Table A-3). Estimated survival rates20, show that larger and productive firms are less likely to exit (Annex Table A-4). Exporters survive for longer periods but foreign firms are more likely to shut down. Across sectors, firms in utilities and hotels/restaurants are more likely to exit than firms in manufacturing. Average wages in all sectors increased over the period, but increases were slowing at the end of the period. Real wages rose through the past decade with no region or sector left behind (Figure 3-32). Pristina had the highest real wages throughout the period, but all regions enjoyed real wage increases. Wage increases can also be observed across all sectors, though the commerce and services sectors paid the highest real wages. There is an observable jump in wages from 2010 to 2011, associated with the adoption of the first minimum wage law. The law required that employees aged less than 35 years old must be paid at least €130 per month, while those over 35 must be paid €170.21 No distinctions were made based on experience, type of position, or sector of activity. Figure 3-32 Average real wages by region KOSOVO: AVERAGE REAL WAGE BY REGION 1 LN OF REAL WAGE .8 .6 .4 .2 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 YEAR PRISTINA FERIZAJ GJAKOVA GJILAN MITROVICA PEJAPEC PRIZREN Source: Staff estimates based on firm registry data. Larger firms pay more; younger firms have the highest wage increases. Regression analysis reveals that wages are positively and monotonically correlated with size (Annex Table A-6) and that, other things being equal, large firms pay higher wages. Foreign firms pay almost 80 percent higher wages than domestic firms. More productive firms and exporters also pay more – the wage premium for exporting firms is 15 percent. Firm age does not appear to have a clear association with wages, however, young firms (less than five years) are the ones where employees experience the most rapid wage growth. This result is robust to various specifications, includ- ing interactions with sector, region and ownership. Furthermore, more productive firms – especially those in the top 25th percentile of productivity – have more rapid wage growth than less productive firms (Annex Table A-7). Productivity increases have been associated with expanding employment in recent years, but who gets these jobs? This chapter has considered questions of job creation and destruction in the formal sector – which types of firms create jobs, how has formal sector job creation fared overtime, can small firms grow in Kosovo, and the links between firm growth and productivity. The next chapter turns to the supply side analysis of the labor force to consider who has jobs, what types of jobs, and what factors may impede greater labor market participation in Kosovo. 20 The survival rates are inferred from probit estimates and hazard ratios from a Cox regression. The Cox proportional-hazards regression model is a survival analysis that examines the relationship between survival distributions and covariates over some period of time. Breslow method is used to break ties. 21 Republic of Kosovo, Ministry of Labor and social Welfare. http://mpms.rks-gov.net/News/PublicationNews/tabid/134/articleType/ArticleView/ articleId/753/language/en-US/The-minimum-wage-for-2011-in-Kosovo-will-be-170-euros-at-the-national-level.aspx 30 4.  WHO IS ACTIVE? WHO IS GETTING JOBS? Jobs, particularly good jobs, are critical for shared prosperity, but relatively few people, especially women, are employed in Kosovo. This chapter relies on data from the Labor Force Survey to describe the pat- terns of employment, unemployment and inactivity, as well as links between employment and poverty. In addition, this chapter begins an analysis of informality, its patterns and implications. EVOLUTION AND DISTRIBUTION OF JOBS IN KOSOVO Labor force participation is very low, especially among women. Around 57 percent of the working age population was inactive in 2014. The level of inactivity is higher in rural than in urban areas. A salient charac- teristic of inactivity is the high rate among women – 37 percent of working age men were inactive, but the rate among women was more than double that: 77.2 percent (Figure 4-1). Figure 4-1 Key labor market indicators LFP SHARES OVER YEARS, AGE 15−64 LFP SHARES OVER YEARS, AGE 15−64 100 100 80 80 60 60 40 40 20 20 0 0 2012 2013 2014 2012 2013 2014 2012 2013 2014 2012 2013 2014 URBAN RURAL MALE FEMALE EMPLOYED EMPLOYED UNEMPLOYED UNEMPLOYED NEITHER AT WORK NOR SCHOOL NEITHER AT WORK NOR SCHOOL OTHER INACTIVES OTHER INACTIVES Source: Staff estimates based on LFS data. Notes: Other inactive refers to the inactive population net of NEET. Family responsibilities are the main reason for women’s inactivity. Inactivity among men is princi- pally market related – education/training or belief that no work is available – whereas the primary reasons among women are family related. Almost half of the inactive working-age men cite education/training as the primary reason, but only one fifth of inactive women are out of labor force for those reasons. On the other hand, over 50 percent of inactive women cite personal/family responsibilities, including look- ing after children or incapacitated adults as primary reasons for inactivity, compared to only six percent of men. Family responsibilities are less of a driver of inactivity for women with more education or those in younger cohorts. Across age cohorts, the share of combined child care and other family responsibilities appears 31 to be stable, but in households without children (Figure 4-2), the share of women in the 35–44 cohort who are inactive for family responsibilities is almost 10 percentage points lower than in the 45–54 cohort. It is also the case that women with incomplete secondary education or less are much more likely to be inactive due to family responsibilities than women in the same age cohort with upper secondary education or above. In the latter group, the share of women who are inactive due to beliefs that no work is available is also higher by the same order of magnitude (about 10 percentage points or more). Figure 4-2 Reasons for inactivity by age cohort OVERALL HOUSEHOLDS WITHOUT CHILDREN 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 25–34 35–44 45–54 55–64 25–34 35–44 45–54 55–64 CARE (CHILDREN/FAMILY CARE (CHILDREN/FAMILY EDUCATION/TRAINING EDUCATION/TRAINING NO WORK AVAILABLE NO WORK AVAILABLE OTHER OTHER INCOMPLETE SECONDARY OR LOWER UPPER SECONDARY OR ABOVE 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 25–34 35–44 45–54 55–64 25–34 35–44 45–54 55–64 CARE (CHILDREN/FAMILY CARE (CHILDREN/FAMILY EDUCATION/TRAINING EDUCATION/TRAINING NO WORK AVAILABLE NO WORK AVAILABLE OTHER OTHER Source: Staff estimates based on LFS 2014 data. Inactivity among women is associated with family formation. Among young men, employment starts to rise around the age of 18, as the number in school falls, but for women the primary transition after school appears to be into inactivity, rather than employment. Late teens is also the beginning of family formation among women in Kosovo, with the percentage rising throughout their twenties and thirties. During these ages, the per- cent who are not in education, employment, or training (NEET rate) among men is either flat or decreasing, whereas for women it continues increasing until almost the age of 40 (Figure 4-3). Availability of childcare is oftentimes problematic. According to nationally representative data from the Life in Transition Survey, in over 95 percent of households with children under the age of six, childcare is pro- vided by household members only, and only two percent rely on either public or private institutional childcare. This is the second lowest incidence of institutional childcare in the ECA region (second only to Czech Republic), and well below the regional average (Figure 4-4). While there is a positive relationship between the share of children in institutional care and a country’s GDP per capita, the incidence of formal care in Kosovo is even below what would be predicted given its PPP-adjusted GDP per capita (Figure 4-5). A recent study of childcare in the Western Balkans (World Bank, 2015b) confirms that childcare in Kosovo is almost universally informal, and the availability of institutional childcare is very low, and practically nonexistent in rural areas. The existing childcare centers are usually at capacity – only 22 percent of providers could accept new clients without putting 32 Figure 4-3 School to work transitions and family formation in Kosovo FROM SCHOOL TO JOBS TRANSITION, MEN, AGED 10−35, 2014 LFS FROM SCHOOL TO JOBS TRANSITION, WOMEN, AGED 10−35, 2014 LFS 1 1 .8 .8 .6 .6 .4 .4 .2 .2 0 0 10 15 20 25 30 35 10 15 20 25 30 35 AGE AGE NO WORK NO SCHOOL NO WORK NO SCHOOL SCHOOL ONLY SCHOOL ONLY NON-FARM WAGE NON-FARM WAGE NON-FARM NON-WAGE WORKER NON-FARM NON-WAGE WORKER FARM WORK FARM WORK NOT IN EMPLOYMENT, EDUCATION, OR TRAINING BY AGE AND GENDER, 2014 LFS FAMILY FORMATION BY AGE AND GENDER, 2014 LFS 1 1 PERCENTAGE OF MARRIED .8 .8 .6 .6 .4 .4 .2 .2 0 0 10 15 20 25 30 35 40 12 17 22 27 32 37 AGE AGE FEMALE FEMALE MALE MALE Source: Staff estimates based on LFS data. Notes: Schooling information is not available for ages 0–14. Figure 4-4 Provision of childcare in the ECA region (2016) CZECH REP. KOSOVO AZERBAIJAN TURKEY MOLDOVA HUNGARY TAJIKISTAN ESTONIA FYR MACEDONIA CROATIA UZBEKISTAN GREECE BOSNIA AND HERZ. BULGARIA ARMENIA GEORGIA MONTENEGRO ROMANIA LITHUANIA UKRAINE KYRGYZ REP. SERBIA GERMANY ALBANIA ITALY POLAND SLOVAK REP. KAZAKHSTAN MONGOLIA SLOVENIA CYPRUS BELARUS LATVIA RUSSIA TOTAL 0 10 20 30 40 50 60 70 80 90 100 ONLY HOUSEHOLD MEMBERS ONLY INSTITUTIONAL CARE (PRIVATE) ONLY INSTITUTIONAL CARE (PUBLIC) INSTITUIONAL CARE AND/OR OTHER TYPE OF CARE OTHER Source: Staff estimates based on Life in Transition 2016 data. Figure 4-5 Provision of institutional care and GDP per capita 70 RUSSIAN FEDERATION 60 % OF POPULATION IN HOUSEHOLDS USING INSTITUTIONAL CARE WHEN LATVIA BELARUS 50 SLOVENIA SLOVAK REPUBLIC THERE IS A CHILD NEEDING CARE MONGOLIA POLAND CYPRUS 40 KAZAKHSTAN SERBIA UKRAINE LITHUANIA KYRGYZ REPUBLIC GERMANY 30 ROMANIA ITALY GEORGIA MONTENEGRO ALBANIA BULGARIA 20 UZBEKISTAN BOSNIA AND HERZEGOVINA CROATIA GREECE ESTONIA ARMENIA MOLDOVA FYR MACEDONIA 10 HUNGARY TURKEY TAJIKISTAN KOSOVO AZERBAIJAN CZECH REPUBLIC 0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 50,000 GDP PER CAPITA, PPP (CONSTANT 2011 INTERNATIONAL $) 2015 Source: Staff estimates based on LiTS 2016 and WDI data. them on a waitlist. Childcare centers in Kosovo also rarely offer their services year round – in urban areas only 25 percent of facilities in the study were open during the summer months. Even when institutional childcare is available, there are important issues of affordability. Kindergar- tens are perceived to be rather expensive. Private childcare providers, who are less likely to suffer from over- capacity problems are deemed inaccessible to the majority of the population. For rural households, the high cost of urban childcare is augmented by the cost of transportation into town, since childcare centers are not available in many rural communities. Lack of institutional childcare is likely also influenced by prevailing social norms. In Kosovo, 69 percent of respondents in the childcare study agreed with the statement “a pre-school child is likely to suffer if his/her mother works,” and only three percent of respondents disagreed with that statement (World Bank, 2015b). In focus groups, men and women of reproductive age and formal and informal employers believe that women are primary, natural caregivers, whereas men have the role of supporting the family financially (World Bank, 2015a). Eldercare is limited by social norms, low quality and narrow provision. In an eldercare study 43 percent of respondents agreed that “when parents are in need, daughters should take more caring responsibility than sons.” In comparison, only 13 percent of respondents agreed with this statement in FYR Macedonia. Partici- pants in focus groups stated that the preferred format for care of elderly is informal care at home. Availability of institutional eldercare is minimal, and, when centers are available, the quality of service is perceived to be low (see World Bank, 2015a and references therein). The design of maternity benefits – both length and distribution of financial burden – make it harder for women enter the labor market. The length of paid maternity leave (nine months) and the high share of employer responsibility for providing maternity benefits is concerning in light of the findings from the literature. Research shows that paid maternity leave may have beneficial effects on women’s labor force participation, but 34 extended leave (beyond six months) can create disincentives to hiring women and also lead to skill deterioration and missing opportunities for mothers’ training and promotions. Family responsibilities limit work activities even among employed women. Ten percent of working women and eight percent of working men reported working part-time in 2014. Among men, 85 percent claimed that inability to find full-time employment is the main reason, with fewer than two percent citing family responsibilities as contributing to partial employment. Among part-time women, almost half reported part-time work to be due to personal/family reasons, or looking after children or incapacitated adults. Women are also less likely than men to work on week-ends, in the evenings or at night, and are more likely to work at home. The childcare study revealed that the majority of women with children in institutional childcare centers have a work intensity of over 60 percent (most work for at least 80 percent of their time). As the result, the fact that many childcare centers are closed during the summer months would be a burden for those with full-time employment (and, in the summer months, even for those with part-time employment). Lack of flexible work arrangements and limited availability of parental care is a burden for working women. Parental leave for child care in Kosovo is scarce. Mothers with children under three benefit from two days of paid annual leave for childcare (not transferable to fathers). In addition, parents do not have rights to flexible working hours, except for civil service employees (World Bank, 2015a). Unavailability of adequate paren- tal care and of limited job flexibility may imply that income shocks force women who are providing care (e.g. due to child illness) to adjust their labor market participation on the extensive margin (by dropping out to provide care), rather than the intensive margin (by resorting to flexible work arrangements or reduced workload). Activity is strongly correlated with education for women and to a lesser extent for men. Over three quarters of inactive women and around 41 percent of inactive men have less than secondary education (Figure 4-6). Among men and women, the share of those with no education is equally low. Among women, a larger share of those with primary education is inactive, as compared to men, who have to a larger extent at least secondary educa- tion. The differences for post-secondary education are narrower, and the share of inactive with post-secondary education is lower for women than for men. Figure 4-6 Composition of inactive population SHARE OF INACTIVES (NEITHER WORK NOR SCHOOL), AGE 15−64 SHARE OF INACTIVES (NEITHER WORK NOR SCHOOL), AGE 15−64 0. MALE 1. FEMALE 100 2.5 2.5 2.0 4.4 5.0 5.8 21.1 25.4 26.5 80 47.4 50.3 53.3 60 66.9 40 64.6 67.0 41.6 40.9 36.6 20 6.7 9.4 6.4 3.8 4.3 5.6 0 2012 2013 2014 2012 2013 2014 1. NO EDUCATION 2. PRIMARY 3. SECONDARY 4. POST-SECONDARY Source: Staff estimates based on LFS data. 35 Even among active individuals, more than a quarter are unemployed. The unemployment rate in 2014 for the 15–64 age group was 35 percent and it was considerably higher among women (41.6 percent) than among men (33.1 percent). Unemployment is particularly prevalent among youth – 56 percent of active men and 72 percent of active women in the 15–24 age group were unemployed in 2014. Unemployment is also higher among those with low education – the unemployment rate among men in the 15–64 age group with tertiary education was 14.7 percent, compared to 70 percent among those with no education and 34 percent among those with secondary education (Figure 4-7). Among women, differences in unemployment rates by educational attainment are narrower and, in fact, unemployment is highest among women with tertiary education, but this has to do with much higher levels of economic activity among women with high education levels; women with low educational attainment are largely inactive. Figure 4-7 Unemployment rate by age and education 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 15–24 25–34 35–44 45–54 55–64 NO SCHOOL I-IX CLASSES SECONDARY SECONDARY TERTIARY VOCATIONAL GYMNASIUM MALE FEMALE MALE ALL FEMALE ALL Source: Staff estimates based on LFS 2014 data. The chances of transitioning from unemployment into employment are slim. Among the unemployed, 69 percent reported to have been unemployed for over 12 months. The difficulty of re-employment is present even among prime age adults – while in the 15–19 age cohort among men only 40 percent of unemployed were without employment for over 12 months, this share reaches close to 80 percent already in the 25–29 age group, and remains high throughout the entire 25–59 age range. Among women, by age 30 almost 90 percent of unemployed have been so for over 12 months, and this share exceeds 90 percent in the 40 – 54 age cohort. The prospects of unemployment/inactivity to employment transitions are limited by job scarcity and by skill mismatches. The firm-level data in Chapter 3 showed the decreasing rate at of job creation. The Ministry of Labor and Social Welfare, found that in 2015 there was one vacancy for about 10 registered unemployed (Box 4). In comparison, in the United Sates there were 1.3 unemployed per job opening in 2016, and even at the height of the financial crisis in 2009 this ratio was below seven (BLS, 2017). This can also be due, in part, to the lack of trust of job seekers in the Public Employment Service (PES). Furthermore, BEEPS data shows that the percentage of firms identifying an inadequately educated workforce as a major constraint (26 percent in 2013) is considerably higher than in other Western Balkans countries. Data on placements by PES also show that the placement rate among those with tertiary education is more than 10 times higher than the average, suggesting that unemployment to employment transitions are particularly difficult for those in the B40 group with lower levels of education. Economic disenfranchisement, particularly among youth, is driving social fragility. The recent risk and resilience assessment (World Bank, 2017b) associates the high levels of unemployment and inactivity among youth, among with other factors, with rising political tensions and emigration. Interest in migrating has been consistently high in recent years – according to the UNDP Public Pulse poll during 2010–2014 more than a quarter of Kosovo’s population had plans to migrate, and at its apex in late 2012 – early 2013 the share of pop- ulation planning to migrate was close to 40 percent. Intentions to migrate have fell from 2015 onward, in part 36 BOX 4: JOB SEEKERS AND VACANCIES IN KOSOVO The Ministry of Labor and Social Welfare (MLSW) provides information on the performance of services offered by public employment offices (PEOs), including statistics on job seekers, vacancies and placements. The information below gives some figures from the data collected by the Employment Management Information System (EIMS) from officers of PEOs and their respective regional employment centers for 2015. Jobseekers: The total number of registered unemployed in public employment services in Kosovo until end 2015 was 112,179 (48,960 females and 63,219 males). Most registered unemployed (48 percent) belonged to the 25 to 39 age group, and over 10 percent to the 15 to 24 age group. Over half of the jobseekers were unskilled, and 18.7 percent had a secondary-level education. Vacancies: In 2015 a total of 11,506 vacancies were registered with public employment services. Over a 12-month period, there was one vacancy registered with employment services for about 10 registered unemployed. It should be noted, however, that registered vacancies are only a subset of the total number of vacancies available at any given moment as employers oftentimes do not trust the ability of employment services to provide effective inter- mediation, or simply prefer other sources of advertisement, like the internet. In terms of economic sectors, most vacancies in 2015 were in the tertiary sector (services, 68 percent), followed by the secondary sector (production, 29 percent), and the primary sector (agriculture, 3 percent). The private sector accounted for 67 percent of the overall vacancies. Placement: PES offices enabled 6,174 job placements in 2015 out of which 3,574 were through job matching ser- vices and 2,600 from ALMPs that facilitated direct job placement such as employment subsidies or public works. Among those placed, 33.7 percent were women (women accounted for 43 percent of job-seekers). Youth (age 15–24) accounted for almost 23.2 percent of those placed (recall they accounted for 10 percent of job-seekers). The overall placement rate from job matching services remains low (2.5 percent for females, 3.7 percent for males), although there is considerable heterogeneity – for university graduates the placement rate was 20.5 percent. Over a quarter of placements were among the unskilled, even though this group accounted for over half of job-seekers. because many migrants to the EU were forced to return. But even in 2015, 15 percent of the population was estimated to have plans to migrate, perhaps suggesting long-standing tensions of social exclusion and distrust of institutions (World Bank, 2017b). Just over a quarter of working-age adults are employed and most of the employed have at least secondary education. The employment rate in 2014 was 26.9 percent. Over 40 percent of working age men were employed, but only 12.5 percent of working age women. The level of employment is considerably lower than in the EU (64.9 percent for EU-28) or even Greece and FYR Macedonia, where half the working age adults were employed. Only five percent of working age adults with no education were employed in 2014; employ- ment rose to 15 percent with primary education, to 41 percent with secondary education, and to 73 percent with post-secondary education. The increase in the employment with education is particularly notable among women, both in rural and urban areas – very few women with no primary education are employed (Figure 4-8). Other statistical techniques confirm the importance of education for employment. Using multivariate estimates, women are estimated to be 30 percent less likely to be employed compared to comparable males, controlling for characteristics such as age, education, and residence. Primary school increases the probability of employment by ten percent while secondary and postsecondary education increases the probability by 25 percent and 58 percent respectively. Likewise, the probability of inactivity decreases by 62 percent with post-secondary education, and by 33 percent for secondary education (Figure 4-9). Links between education and employment are stronger for women. Educational attainment is higher, on average, among men than women in Kosovo, but the reverse is the case among the employed population. In this latter group the share of those with tertiary education is almost twice as high among women (Figure 4-10). There is a premium to public sector employment. About a third of those employed are wage employees in the public sector, and another third are wage employees in the private sector. There is a clear education gradient among sectors – almost two thirds of those with post-secondary education were in the public sector, compared to only four percent in self-employment. Self-employment appears to be dominated by those with no education 37 Figure 4-8 Labor force participation by gender in urban and rural areas LFP BY EDUCATION, AGE 15−64, 2014 LFS 100 80 60 40 20 0 Y RY N Y N Y Y RY Y N Y Y Y N RY Y AR AR AR AR AR AR AR AR AR IO IO IO IO DA A A AT AT AT AT ND ND ND M ND IM IM ND ND IM ND N I UC UC UC UC PR PR PR PR O CO O CO CO CO O O EC EC EC EC ED ED ED ED 2. 2. 2. 2. SE SE SE SE -S -S -S -S NO NO NO NO 3. ST 3. 3. 3. ST ST ST PO PO PO PO 1. 1. 1. 1. 4. 4. 4. 4. RURAL−MALE RURAL-FEMALE URBAN-MALE URBAN-FEMALE 1. EMPLOYED 2. UNEMPLOYED 3. NON-LF Source: Staff estimates based on LFS 2014 data. or with primary education. Public sector wages are higher than those in the private sector (Figure 4-11). Multi­ variate regressions also confirm the higher wages in the public sector and in formal employment with contracts. Public employment pays seven percent higher wages controlling for individual, household, and regional factors. Informal workers (no social security) earn 18 percent less than formal workers. Workers with contracts earn 28 percent more than comparable workers without contracts. There are no significant regional differences except for Mitrovica where an average worker earns eight percent less than an average worker in Pristina. Figure 4-9 Labor market status, multinomial logit marginal effects MULTINOMIAL LOGIT: MARGINAL EFFECTS, KOSOVO 2014 LFS MULTINOMIAL LOGIT: MARGINAL EFFECTS, KOSOVO 2014 LFS EMPLOYED .034 EMPLOYED −.058 INDIVIDUAL AGE RURAL .026 −.3 PRIZREN FEMALE .1 −.014 GJAKOVA −.017 .056 PRIMARY .25 MITROVICA SECONDARY FERIZAJ −.055 .58 PEJA −.064 POST-SECONDARY GJILAN UNEMPLOYED .0021 UNEMPLOYED .021 INDIVIDUAL AGE RURAL .027 −.14 PRIZREN .0082 FEMALE .048 GJAKOVA .096 PRIMARY .073 MITROVICA .095 FERIZAJ .21 SECONDARY −.027 .042 PEJA POST-SECONDARY GJILAN OUT OF LF OUT OF LF .037 −.036 −.053 RURAL INDIVIDUAL AGE .44 PRIZREN .0062 FEMALE −.15 GJAKOVA −.079 PRIMARY −.33 MITROVICA –.15 SECONDARY FERIZAJ −.16 −.62 PEJA .091 POST-SECONDARY GJILAN −1 −.5 0 .5 –.2 –.1 0 .1 .2 AV. MARGINAL EFFECTS AV. MARGINAL EFFECTS Source: Staff estimates based on LFS data. Note: Base region is Pristina. 38 Figure 4-10 Education and employment overall and among those employed HIGHEST LEVEL OF EDUCATION COMPLETED BY GENDER OVER YEARS HIGHEST LEVEL OF EDUCATION COMPLETED BY GENDER OVER YEARS. EMPLOYED (15−64) 100 7.0 7.1 8.1 100 13.6 13.8 14.3 23.2 19.7 21.8 23.5 36.5 34.7 40.2 80 27.0 26.0 80 60 51.0 52.1 52.6 60 59.7 58.6 59.5 41.6 40 55.2 40 39.4 55.5 55.8 51.8 20 31.0 30.8 29.9 20 20.1 19.3 23.0 19.3 14.2 16.8 11.4 4.5 10.4 10.1 0 3.3 3.2 0 0.5 0.4 0.3 0.3 0.7 1.1 2012 2013 2014 2012 2013 2014 2012 2013 2014 2012 2013 2014 MALE FEMALE MALE FEMALE 1. NO EDUCATION 1. NO EDUCATION 2. PRIMARY 2. PRIMARY 3. SECONDARY 3. SECONDARY 4. POST-SECONDARY 4. POST-SECONDARY Source: Staff estimates based on LFS data. Figure 4-11 Public sector employment is associated with higher educational attainment and higher wages EMPLOYMENT TYPES BY EDUCATION, AGE 15−64, 2014 LFS MEAN WAGE PER MONTH: PUBLIC VS PRIVATE − ONLY WAGE WORKERS 100 400 7.3 21.0 17.8 4.4 0.2 27.3 80 24.0 17.3 29.5 1.3 300 60 53.0 2.3 40.5 40 MEAN 64.2 200 38.6 20 13.4 23.1 6.3 8.6 0 100 Y N Y Y AR AR AR IO AT IM ND ND UC PR O CO EC ED 2. SE −S NO 3. ST 0 PO 1. 2012 2013 2014 4. PUBLIC 1. WAGE PUBLIC 2. WAGE PRIVATE PRIVATE 3. SELF−EMPLOYED AGRICULTURE 4. SELF−EMPLOYED NON−AGRICULTURE 5. OTHER Source: Staff estimates based on LFS data. Higher wages are related to education, in part due to occupational and sectoral sorting. Earnings are positively correlated with educational attainment for men and women (Figure 4-12). Returns to post-secondary education are particularly high, with salaries being a third higher on average among those with post-secondary education compared to those with secondary education. Among men, there are few differences in wages among the no education and primary education groups, with somewhat higher salaries for those with second- ary education and a notable jump in monthly earnings for those with post-secondary degrees. Among women, wages also increase for higher education levels, with a somewhat more pronounced increase in average wages for secondary education relative to primary. Multivariate regressions confirm and explain the high returns to post-secondary education. Post-secondary education increases earnings by 41 percent compared to no education. Notably, primary and secondary school education do not have significantly higher returns in wages compared to no schooling, when other characteristics are controlled. Much of the returns to post-secondary education is explained by sector and occupation differences – once we consider differences within sectors, the returns to postsecondary education decreases from 41 percent in the baseline specification to 27 percent, and once 39 Figure 4-12 Educational attainment and earnings in Kosovo MEAN EARNINGS PER MONTH BY EDUCATION, MEAN EARNINGS PER MONTH BY EDUCATION AND GENDER, EXCLUDING UNPAID WORKERS 2014 LFS EXCLUDING UNPAID WORKERS 2014 LFS 1. NO EDUCATION NO EDUCATION 2. PRIMARY MALE 3. SECONDARY PRIMARY 4. POST−SECONDARY SECONDARY 2. PRIMARY FEMALE 3. SECONDARY POST-SECONDARY 4. POST-SECONDARY 0 100 200 300 400 0 100 200 300 400 MEAN MEAN MEAN WAGE PER MONTH BY SECTOR AND GENDER, ONLY WAGE WORKERS 2014 LFS 1. AGRICULTURE 5. CONSTRUCTION 2. MINING 3. MANUFACTURING 6. COMMERCE MALE 8. FINANCIAL AND BUSINESS SERVIC 10. OTHER SERVICES, UNSPECIFIED 9. PUBLIC ADMINISTRATION 4. PUBLIC UTILITIES 7. TRANSPORT AND COMNUNICATIONS 1. AGRICULTURE 5. CONSTRUCTION 2. MINING 3. MANUFACTURING 6. COMMERCE FEMALE 8. FINANCIAL AND BUSINESS SERVIC 10. OTHER SERVICES, UNSPECIFIED 9. PUBLIC ADMINISTRATION 4. PUBLIC UTILITIES 7. TRANSPORT AND COMNUNICATIONS 0 100 200 300 400 500 MEAN Source: Staff estimates based on LFS survey data. occupation is included returns drop to 22 percent. Accounting for differences in education, experience also matters – a one-year increase in age is associated with a wage that is about four percent higher. Women receive lower wages, on average, at each level of education, and in each industry. Regression estimates reveal that when educational and regional differences are accounted for, there is an 11 percent wage gap between men and women, and very little of this gap is explained by occupational or sectoral differences (i.e., the coefficient for females changes very little across specifications). Note, however, that wages are expressed in monthly equivalents, thus not accounting for differences in the number of hours worked, other than the dif- ferences in gender composition and hours worked across sectors, geographic space, or other characteristics controlled for in regressions. In hourly terms the gap between male and female wages would likely decrease on account of fewer hours worked among women. This report does not present direct evidence but the literature has found that maternity leave provisions may be among reasons why lower wages are offered to women, as employers avoid the increased cost of paid maternity leave. This is confirmed by studies in a number of Euro- pean countries, although studies from Australia and the United States do not find long-term wage effects (for a review of the literature, see World Bank, 2015a and references therein). 40 BOX 5: CHALLENGES OF DEFINING INFORMALITY IN KOSOVO An informal sector, also grey or shadow economy, exists in every country, yet, despite its ubiquity, there is no single definition. In some countries, informality refers to the workforce not protected by labor laws and not entitled to social insurance benefits, notably social security. In others, informality is broader encompassing activities such as microenterprises, street-corner vendors, house cleaners and babysitters, small firms not registering their employees nor complying fully with regulations, or small or large firms evading taxes. Here we define informality as market-based production of goods or services that are legal under prevailing laws, and that is concealed either to avoid paying taxes or insurance contributions or to avoid regulation (Packard and Nguyen 2014, OECD 2008). Agents may interact with the state along some dimensions but not others (Perry et al., 2007). On some dimensions, there might be full compliance, while on others noncompliance. A small firm may register its business, but may not pay social security contributions for some workers. We consider three alternative definitions: Definition Categories of workers included  npaid family workers, self-employed workers in small firms, self-employed workers in unprofes- U Main sional occupations, workers in small firms (5 or fewer employees). Legal Workers without a labor/employment contract.  mployees working in units of size 10 or less, self-employed workers with or without employees E Broad in non-professional occupations, and unpaid family workers. When estimating the size of the informal sector the advantage of including small firms is that the extent of employ- ment in small firms is highly correlated with the size of the shadow economy and a significant share of employees without contracts work in small firms. In Kosovo, salaried workers in large firms are more likely to have contracts than those in small firms and, further, firm size is negatively correlated with informality (Era Dabla-Norris et al, 2008). The main disadvantage of classifying small firms as informal is that many small firms are in the formal sector and, conversely, there are large firms that use unregistered workers. Using firm size to measure informality will miss each of these. Notes: For details, see Ajwad et al., 2016. INFORMALITY – A SALIENT CHARACTERISTIC OF EMPLOYMENT IN KOSOVO 22 Labor markets in Kosovo and other SEE6 countries have high degrees of informality. There is general agreement with respect to the importance of informality in Kosovo, but there are no systematic estimates of the size of the informal sector. This chapter develops a profile of Kosovo’s informal employment, defines and estimates the size, and benchmarks informality against other countries. Size of the Informal Sector Measuring informality is challenging, but over a quarter of Kosovo’s workers are estimated to be in the informal sector. Defining informality is problematic and there are multiple conceptions (Box 5). The main definition in this section is that it consists of unpaid family workers, self-employed workers in small firms, self-employed workers in unprofessional occupations, workers in small firms (five or fewer employees). By this definition, just over a third of workers would be classified as informal. The other 65.2 percent of employed workers are counted as formal sector workers, but data may be unreliable because there is substantial under- reporting of incomes to avoid taxes – a phenomenon known as “envelope wages” in Eastern Europe. Kosovo’s informality is among the highest in ECA. In order to make cross-country comparisons, other defini- tions need to be applied. Based on the legal definition, i.e., defining informality more narrowly as the percentage This section is based on Ajwad, Vazquez and Winkler (2016). 22 41 of wage workers who do not have labor/employment contracts, about 15 percent of Kosovo’s wage workers are informal. Comparable estimates from the 2010 Life in Transition Survey (LiTS), using the legal definition of informality, shows that Kosovo has one of the largest informality rates in ECA (Figure 4-13). Note, however, that the level of informality in Kosovo based on the LiTS data is higher than what is obtained from the LFS.23 Figure 4-13 Informality rates in ECA countries using a legal definition of informality, 2010 60% 50% 40% 30% 20% 10% 0% TURKEY HUNGARY KYRGYZSTAN KOSOVO ALBANIA MOLDOVA TAJIKISTAN GEORGIA GREAT BRITAIN ARMENIA MACEDONIA UZBEKISTAN MONTENEGRO SERBIA AZERBAIJAN UKRAINE BOSNIA KAZAKHSTAN CZECH REPUBLIC SLOVENIA POLAND RUSSIA FRANCE BULGARIA LATVIA ROMANIA ESTONIA CROATIA SLOVAKIA LITHUANIA ITALY SWEDEN GERMANY BELARUS Source: World Bank staff calculations using LITS 2010. Informal employment in Kosovo is higher than in most EU states. Alternatively, informality in Kosovo is behind only Greece and Cyprus if informality is defined as “employees in units of 10 or fewer, self- employed workers with or without employees in non-professional occupations, and all the unpaid family workers.” Informality rates in Kosovo are about twice as high as rates in Belgium and the Netherlands under this definition (Figure 4-14). Profile of Informal Firms and Employees Informal employment encompasses diverse sectors, workers with differing socioeconomic characteris- tics, and a wide array of skills and tasks. Informality in Kosovo is particularly high in agriculture and construc­ tion, sectors where jobs tend to require less skilled personnel, casual labor and – in the case of agriculture – unpaid family workers (Figure 4-15). High degrees of informality in these sectors have also been observed in other countries, inclusively in the ECA region (e.g. Bulgaria, the Czech Republic, Estonia, Latvia, Poland, and Slovakia).24 Based on the legal definition, half of the construction workers in Kosovo work without a labor contract. Even in manufacturing and services, about 16 percent of employees are informal. Informality is more common in the tradable than in the non-tradable sector. Informality rates are usu- ally assumed to be higher in the tradable sector as a way of reducing costs to compete (Ajwad et al. 2016 discuss the literature). In Kosovo, based on the traditional AMECO categorization, the proportion of informal 23 The estimate of informality using the LITS 2010 is not directly comparable to that of the LFS 2012, as the sample and questionnaire design are different. 24 Perry et al. (2007) reveal similar findings for Argentina and the Dominican Republic. 42 Figure 4-14 Informality rates in European countries using a broad definition of informality 60 50 40 30 20 10 0 GREECE CYPRUS KOSOVO PORTUGAL CROATIA ROMANIA SPAIN ITALY FINLAND IRELAND ICELAND POLAND HUNGARY SLOVAKIA AUSTRIA FRANCE SLOVENIA LATVIA CZECH REPUBLIC SWITZERLAND UNITED KINGDOM SWEDEN BULGARIA NORWAY LUXEMBOURG LITHUANIA ESTONIA DENMARK BELGIUM NETHERLANDS Source: EU Labor Force Survey Database and Kosovo Labor Force Survey (2012). Note: Informal workers = employees working in units of size 10 or less, self-employed with or without employees in non-professional occupations (isco codes >=400), and unpaid family workers. Figure 4-15 Informality rates vary significantly by sector of activity SHARE OF WORKERS WHO ARE INFORMAL UNDER SHARE OF WORKERS WHO ARE INFORMAL UNDER MAIN DEFINITION LEGAL DEFINITION AGRICULTURE 78.8 AGRICULTURE 22.7 MANUFACTURING 42.8 MANUFACTURING 16.0 WATER AND ELECTRICITY 6.4 WATER AND ELECTRICITY 1.1 CONSTRUCTION 54.7 CONSTRUCTION 52.1 SERVICES 41.3 SERVICES 16.9 PUBLIC ADMINISTRATION 1.1 PUBLIC ADMINISTRATION 1.8 EDUCATION, HEALTH AND SOCIAL WORK 1.1 EDUCATION, HEALTH AND SOCIAL WORK 1.1 ARTS AND OTHER SERVICES 29.1 ARTS AND OTHER SERVICES 12.6 HOUSEHOLDS AS EMPLOYERS 47.5 HOUSEHOLDS AS EMPLOYERS 47.2 EXTRATERRITORIAL ORGANIZATIONS 19.0 EXTRATERRITORIAL ORGANIZATIONS 6.8 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 Source: Labor Force Survey (2012). Note: Workers in low productivity activities: workers in small firms, unpaid family workers and self-employed in either small firms or not professional occupations. workers in the tradable (NACE categories A-E, G-I) sector (48.8 percent) is more than double the proportion in the non-tradable (NACE categories F, J-P) sector (21.6 percent) under the main definition of informality. However, informality rates are not very different under the legal definition (12.5 percent in the tradable sector and 17.9 percent in non-tradable), or when social security benefit incidence is compared, or when temporary work contract incidence is compared across the sectors, suggesting that the higher degree of informality in the tradable sector is likely driven by factors other than those affecting labor, such as social security benefits and permanent worker contracts. Farming activities are prominent in the share of informal employment in Kosovo’s tradable sectors. There notable differences in the types of employment between the informal and formal sectors. More than a third of the informal sector (36 percent) are employees in small firms; 28 percent are self-employed 43 Figure 4-16 Informal sector workers in Kosovo are predominately employed in small firms COMPOSITION OF INFORMAL SECTOR COMPOSITION OF FORMAL SECTOR 40% 60% 54% 36% 35% 50% 41% 28% 30% 40% 25% 30% 19% 20% 17% 20% 15% 10% 3% 2% 10% 0% EMPLOYEES IN LARGE FIRMS EMPLOYEES IN PUBLIC SECTOR SELF-EMPLOYED IN PROFESSIONAL EMPLOYERS IN LARGE FIRMS 5% OCCUPATIONS 0% EMPLOYEES IN SMALL FIRMS PROFESSIONAL OCCUPATIONS EMPLOYERS IN SMALL FIRMS UNPAID FAMILY WORKERS SELF-EMPLOYED IN NON- Source: Authors’ calculations using Labor Force Survey (2012). in non-professional occupations; 19 percent are employers in small firms; and the remaining 17 percent are unpaid family workers (Figure 4-16).25 On the other hand, the public sector provides the bulk of the formal sector jobs (54 percent) together with large firms, where 41 percent of workers are employed. Informality is higher among youth, males, the less-educated, and rural residents. Almost half (48.2 per- cent) of employed youth (15–24 year olds) are in the informal sector, while the proportion of older workers (45–64 year olds) who are informal is about a quarter (25.9 percent). This is consistent with the informal sector serving a gateway to paid employment for poorly educated youth, and similar patterns have been observed in Mexico (Maloney, 2004), as well as a number of European countries where younger workers are more likely to be in jobs with no labor contracts (Packard, et al., 2012). Informality also declines steadily with educational attainment, from over two thirds of workers with less than primary education to 11 percent for those with tertiary education, similar to patterns observed for other countries in Europe and Latin America (Packard et al., 2012; Perry et al. 2007). Men are also more prevalent in Kosovo (Figure 4-17), consistent with evidence from other European countries (Packard et al., 2012), but unlike some of the countries in Latin America (Argentina, Brazil, Chile and Mexico) where informality rates are higher among women, which is attributed to greater location and time of day flexibility (Cunningham, 2001; Perry et al., 2007). Employees in small firms are workers who report themselves as an employee in a private firm, NGO or other organization, and where 25 there are five or fewer employees. Self-employed in non-professional occupations are those who report themselves as self-employed without employees (own-account worker or free-lancer), whose occupations fall under ISCO-88 codes 411-933 (clerks, service workers and shop and market sales workers, skilled agricultural and fishery workers, craft and related trade workers, plant and machine oper- ators and assemblers, elementary occupations). Employers in small firms are those workers who report themselves as self-employed with employees in a private firm, NGO or other organization, where the total number of persons working at the local unit is five or less than five. Unpaid family workers are people whose work contributes directly to a business, farm, or professional practice, owned or operated either by themselves or by a relative, but who receive no pay or profits. In all cases, the main job/activity is considered for classification. 44 Figure 4-17 Informal sector workers are more likely to be young, unskilled, males, who live in rural areas BY AGE GROUP BY EDUCATION 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 NO SCHOOL PRIMARY SECONDARY TERTIARY 0 15–24 25–44 45–64 BY GENDER BY AREA 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 FEMALES MALES URBAN RURAL BY REGION 70 60 50 40 30 20 10 0 GJAKOVA GJILANI MITROVICA PEJA PRIZRENI PRISHTINA FERIZAJI Source: Staff estimates based on LFS 2012 survey data. Notes: (1) Productive definition of informality used: a person is informal if he works as an employee or employer in a small firm, or as self-employed in non-professional occupations, or as an unpaid family worker. (2) Informality rates computed over the employed population in working age (15-64) in Kosovo Some Consequences of Informality Competition between formal and informal firms can lead to undesirable outcomes. Both consume public services, but only formal firms pay the required taxes. This may lead to suboptimal service levels for all firms and reduce productivity. And when faced with competition from informal firms, it could create incentives for formal firms to evade taxes or ignore regulations to reduce costs and become more competitive. 45 Almost six out of ten firms classify informal competition as major or severe obstacle to growth. This is one of highest in the world and higher than in neighboring and comparator countries (Figure 4-18). In FYR Macedonia 35 percent of firms complain about major or severe competition, in Bulgaria and Romania it is 33 percent, and in Albania it is 20 percent. More than a quarter of formal firms in Kosovo (26 percent) identified informal competition as the single biggest impediment to growth. They ranked it ahead of access to finance, corruption, customs and trade regulations, electricity, skills of the workforce, and other factors. Constraints posed by informality appear to be most commonly expressed by small (<20) and large (100+) firms, as well as by older firms and those with lower levels of productivity. Figure 4-18 Percentage of firms in Kosovo that complain that they are “majorly” or “severely” affected by competition from informal sector firms is one of the highest in the world 10 0 70 58.8 60 80 50 40 34.6 32.9 32.9 KOSOVO 60 30 19.9 18.8 20 16.0 12.6 11.2 40 10.1 8.6 MKD BGR ROM 10 AL B HRV 0 20 BIH KOSOVO M ACE DO NI A BUL G ARI A RO M ANI A AL BANI A CRO AT I A BO S NI A AND HE RZ E G O V I NA S E RBI A M O NT E NE G RO HUNG ARY S L O V E NI A SRB MNE HUN SVN 0 0 10,000 20,000 30,000 GDP PER CAPITA Source: Authors’ calculations using BEEPS (circa 2013) and WDI (GDP per capita in 2013, PPP (constant 2011 international $)). Labor taxes and contributions are relatively low but informal firms are likely to derive an advantage by avoiding taxes and regulations. The share of sales not reported to tax authorities is higher for smaller firms than it is for larger firms (Riinvest, 2013). Because smaller firms are more likely to be informal, we infer that informal firms derive an advantage from underreporting sales and skirting regulation. Similarly as firms grow, demand for services associated with formality increases, as does the probability of detection by authori- ties. However, when evasion is measured as a percentage of firm sales, firms with sales of EUR 50,000–100,000 are the group with the highest proportion of sales not reported for taxes. It is possible that these firms have a higher incentive to under-report their sales to avoid being registered for VAT. Formal firms that compete with informal firms have incentives to take steps that affect workers directly. These steps include offering temporary, rather than permanent, work contracts more frequently. In Kosovo, temporary contracts are used by a larger share of firms and affect a larger share of firm employees in comparison with other ECA countries (Figure 4-19), and this is true even when adjusting for countries’ levels of development (Figure 4-20). Despite their shorter term, temporary contracts are better than no contracts: they are legal, they provide social security rights, and they give employers flexibility, so they may help to spur hiring. Young and unskilled workers are more likely to be on temporary contracts than older and better-educated workers. Informal employment usually excludes workers from benefits. Kosovars are entitled to a basic old-age pension that may protect them from poverty, but contributory pensions are offered to the formally employed and they are substantially higher.26 Informal workers do not receive a number of additional benefits such as severance pay, maternity leave, or private health insurance (in some firms). A broad swath of the population does not buy private insurance (often health, disability, etc.) and some in ECA may offset unexpected expenses 26 See IMF 2016. 46 Figure 4-19 Average share of temporary workers employed in a firm (Kosovo and neighboring countries, 2013) PERCENTAGE OF FIRMS THAT THAT USE TEMPORARY AVERAGE SHARE OF TEMPORARY WORKERS CONTRACTS EMPLOYED IN A FIRM 60 25 50 20 40 15 30 10 20 5 10 0 0 KOSOVO ALBANIA BOSNIA & HERZ. MACEDONIA HUNGARY MONTENEGRO ROMANIA SLOVENIA CROATIA BULGARIA SERBIA KOSOVO MONTENEGRO SLOVENIA CROATIA HUNGARY BOSNIA & HERZ. SERBIA ALBANIA BULGARIA ROMANIA MACEDONIA Source: BEEPS. Figure 4-20 Percentage of firms that employ temporary workers in a country, circa 2013 100 30 MNG 80 BOL UKR UKR 20 UGA BOL COL 60 SLV MNG COL KOSOVO UGA CHL SLV SWE KOSOVO SWE 40 CHL 10 BLZ ALB ALB 20 ARM ARM BLZ 0 0 0 10,000 20,000 30,000 40,000 0 10,000 20,000 30,000 40,000 GDP PER CAPITA GDP PER CAPITA Source: BEEPS circa 2013 and WDI (GDP per capita in 2013, PPP (constant 2011 international $)) by working more hours, sending nonworking household members into the labor market, or by reducing health care. Kosovar employees without labor contracts work significantly more hours than those with labor contracts – 47.9 vs. 41.1 hours a week.27 Informal work and temporary contracts provide lower benefits. In the 2012 labor survey, informal work- ers (using the main definition of informality) are less than half as likely as formal workers to be entitled to social security or to receive half their wages during their absence. Only 10 percent of informal workers are entitled to social security, compared to almost 19 percent of formal workers (Figure 4-21). This disparity is greater under the legal definition of informality – only three percent of workers without a labor contract are eligible for social security, while almost 20 percent of workers with a labor contract are eligible. Note however, that even in the formal sector the social security entitlement is low, which would reduce the benefits of formalization. During 27 See also World Bank (2011) and Dasgupta and Ajwad (2011). 47 absence from work 82.5 percent of formal workers receive at least half their wages and salary, but only 30.8 percent of informal workers receive that much. Absence from work may be due to annual leave (57 percent), weather (16 percent), firm closing (five percent), or maternity leave (four percent). As with social security, when labor contracts are used to define formality and informality, the difference becomes starker. Only 86 percent of workers with contracts and only 12.3 percent of workers without contracts receive at least half their wages and salary during their absence. Figure 4-21 Informal workers are half as likely as formal workers to be entitled to social security and to receive at least half their wages during absence PERCENTAGE OF WORKERS ENTITLED TO SOCIAL SECURITY BENEFITS, BY INFORMALITY STATUS 20 18.7 PERCENTAGE OF WORKERS WHO CONTINUE RECEIVING 50% OR MORE OF THEIR WAGE EVEN DURING ABSENCE, BY INFORMALITY STATUS 18 90 82.5 16 80 14 70 12 60 10.1 10 50 8 40 30.8 6 30 4 20 2 10 0 0 FORMAL INFORMAL FORMAL INFORMAL Source: Labour Force Survey 2012. Notes: Main definition of informality: employees and employers in small firms (size<=5), self-employed in non-professional occupations, and unpaid family workers. A significantly larger proportion of employed informal sector workers are searching for jobs with better working conditions: nine percent of the informal sector vs. 4.1 percent of the formal sector. The fac- tors they are trying to improve are pay, commute times, quality of work, and benefits. JOBS AND POVERTY Labor market activity is the principal channel to reduce poverty. The World Bank’s Systematic Country Diag- nostic found that the main channels for the poor to increase consumption or income are (i) employment, (ii) increased returns to employment, (iii) private transfers such as remittances, and (iv) public transfers. Kosovo’s income data is incomplete, but increased returns to employment have been the biggest factor in raising consumption and income. This resulted in part from workers moving from lower-paying to higher-paying sectors. Wage increases raised the returns to employment, especially in the public sector (World Bank, 2016).28 Other income components such as remittances, social assistance, and pensions contributed little to income gains among the bottom 40 percent. Data limitations make it difficult to link labor market characteristics to poverty. The main source of labor market information is the Labor Force Survey (LFS) which does not track household expenditures which is necessary to relate poverty to employment activity and characteristics. This section relies on the Household Budget Survey (HBS), the survey on which the national definition of poverty is based. The HBS has a short set of questions on labor market participation, but it is not as precise the LFS and the two surveys cannot be linked. Poverty is much higher among the unemployed. In 2015, 17.6 percent of the population had expenditures below the poverty threshold. Employment meant notable differences in poverty. Among working age adults, 28 Although public sector wages are higher than average, a significant share of poor households in Kosovo depend on public sector wages, with one public sector employee supporting a large household, often as the only employed member of the household. 48 the poverty rates were 11 percent among the employed, 23.3 percent among the unemployed, and 17.1 percent among the inactive. There is a link between unemployment and poverty despite the differing characteristics of the employed and unemployed. The probability of being employed depends on many factors: education, urban or rural resi- dence, gender, or age. With multivariate regressions we estimated conditional correlations between labor market participation and poverty while holding some characteristics constant (Appendix Table 4). Column (I) shows that the poverty rate among the employed (baseline) was 11 percent (constant term), while the poverty rate among the unemployed (inactive) was 12.1 percent (5.8 percentage points) higher in 2015. Columns (II) – (IV) show that rural poverty is higher, that poverty differs across regions, and that it depends on household composition. How- ever, even accounting for these differences, poverty is more common among the inactive and, particularly, the unemployed. Educational differences account for some welfare improvements. The incidence of poverty declines among the more educated. Relative to those with primary education or less, the poverty rate is seven percentage points lower among those with secondary education, and 14 percentage points lower among those with tertiary educa- tion (conditional on other observables, Table A-8). However, education is strongly correlated with the probability of being employed. So, when educational differences are held constant, the difference between poverty rates among the employed and the inactives falls notably, and that between employed and unemployed also becomes smaller. When we account for the differences across age cohorts, the difference in poverty rates between the employed and the inactives disappears, and that between the employed and the unemployed falls to nine per- centage points, but is still quite large. The relationship between employment and poverty differs between men and women. Column (VIII) of Table A-8 shows the interaction between gender and labor market status, controlling for background char- acteristics. Relative to a baseline of employed men, the probability of being poor is 3.9 percentage points lower among employed women, whereas the probability of falling into poverty among inactive men and inactive women is no different from that of employed men. The probability of poverty is higher for unemployed men: It increases by 13.3 percentage points relative to employed men, and by 5.1 percentage points for unemployed women. Employment improves welfare, but there are considerable differences in poverty even among those who are employed. In the B40, working-age population, 10 percent were employed in executive/professional occupations, while 55 percent were either laborers or trades workers. Among the T60 group occupations differ considerably – 28 percent have executive/professional occupations, while 41 percent are laborers/trade workers. Poverty outcomes also differ across occupational groups – among professionals, fewer than four percent were below the poverty level in 2015, compared to 15 percent among trades workers and 18 percent among laborers (Figure 4-22). These patterns are strongly linked to education (Figure 4-23) – about two thirds of those in executive/professional occupations have tertiary education; among laborers or trades workers this is five per- cent or less, and 30–33 percent in these occupations have at most completed primary education. The sectoral patterns of employment are similar for higher- and lower-income households. The nota- ble difference between the B40 and T60 groups is that a higher share of the B40 population is employed in con- struction, and a much lower share is in public administration/social services (Table 4-1). This is consistent with the lower share of professional occupations in the B40 group and among those below the poverty threshold. Men and women work in different sectors. A large share of women work either in public administration or wholesale/retail trade – these categories account for two thirds of female employment. Among men, these sectors provide less than 30 percent of employment. Males are employed more heavily in agriculture/fishing, construction, manufacturing. There are recent improvements in education and occupations for the B40 group. Over 2012–15 there is a trend of improving educational attainment, and the pace of improvement is faster in the B40 group. The proportion of working age adults with no more than primary education fell by nine percentage points from 58 percent in 2012 to 49 percent in 2015, compared to a four percentage points decrease in the T60 group 49 Figure 4-22 Figure 4-23 Poverty headcount across occupations Educational attainment across occupations PERCENT PERCENT 0 5 10 15 20 0 20 40 60 80 100 EXECUTIVE AND PROFESSIONAL EXECUTIVE AND PROFESSIONAL CLERKS CLERKS TRADES WORKERS AND MACHINISTS SKILLED AGRICULTURAL WORKERS LABORERS TRADES WORKERS AND MACHINISTS SKILLED AGRICULTURAL WORKERS LABORERS PRIMARY OR LESS SECONDARY TERTIARY Source: Staff estimates based on HBS 2015 data. Table 4-1 Distribution of working age population by sector of employment and B40/T60 groups Sector (NACE 2) Top 60 Bottom 40 Male Female Overall 15–64 Agriculture and fishing 11.7 11.7 13.6 2.8 11.7 Mining 1.1 3.2 1.8 1.1 1.7 Manufacturing 9.3 10.5 10.5 6.0 9.7 Utilities 2.8 2.1 3.0 0.5 2.6 Construction 12.1 21.1 17.7 1.4 14.8 Wholesale, retail trade 15.3 17.7 14.5 23.1 16.0 Transportation 3.7 3.2 4.1 0.9 3.6 Food and accommodation 4.3 5.3 4.8 3.5 4.6 Information and communication 1.7 1.0 1.7 0.9 1.5 Finance and real estate 2.5 1.7 2.3 2.4 2.3 Professional and scientific activities 2.5 2.4 2.3 3.3 2.5 Administrative and support activities 2.1 0.6 1.8 1.2 1.7 Pub. admin. and social services 22.7 10.6 14.1 42.4 19.1 Recreation and other services 6.1 7.5 6.1 8.5 6.5 Household activities 2.1 1.3 1.8 2.2 1.9 Total 100 100 100 100 100 Notes: Staff estimates based on HBS 2015 data. 50 (Figure 4-24). The share of the B40 group with a secondary education increased by seven percentage points. In terms of occupations, working age adults with professional/clerical occupations also increased faster in the B40 group – by nine percentage points in the B40 group, compared to a five percentage points increase overall. Despite recent improvements, there are notable educational and occupational differences between the B40 and T60 groups. In 2015, 10 percent of employed adults were in professional occupations among the B40, compared to 28 percent in the T60 group. Across sectors, 2012–15 is associated with a reduction of the share of B40 employed in agriculture and construction, and a higher share of B40 moved into manufacturing, wholesale and retail trade, as well as accommodation/food and other recreational services. One implication is that despite better educational outcomes, these workers remain in the B40 group, i.e., there are still not enough good jobs in Kosovo. The B40 population can benefit from recent patterns of job creation. Commerce and services are dom- inating in terms of number of firms, number of entrant firms and number of employees. Data from the labor force survey confirm that during the 2012–2015, despite an overall decrease in the number of employed per- sons, sectors such as accommodation and food services, wholesale and retail trade, but also public admin- istration have expanded in absolute numbers (Table 4-2). These jobs draw on employees with secondary education to a large degree, and the recent educational improvements for the B40 population may qualify them for a greater share of jobs in these sectors. The overriding problem, as was pointed out in the previous section, is that formal job creation has been on a downward trend over the past decade and the number of new formal openings is small and the number of registered unemployed per vacancy is large. Given the current high levels of unemployment and inactivity, all jobs – formal and informal – can help to boost the welfare of households below the poverty line or in the B40 group. Informal jobs can also help those with low educational levels, who may find it harder to compete for formal jobs. The informal sector dynamics in Kosovo are explored in detail in Chapter 5. Figure 4-24 Educational attainment of B40 and T60 groups, 2012 and 2015 70 60 50 40 30 20 10 0 2012 2015 2012 2015 B40 T60 PRIMARY OR LESS SECONDARY TERTIARY Source: Staff estimates based on HBS data. Low labor force participation and high unemployment, particularly among women and youth, are also the results of structural (cultural and policy) constraints. Weaknesses in the education system, as evidenced by the weak PISA scores, may constrain the ability of formal educational institutions to impart the skills necessary in the job market. (Skills mismatches are analyzed in Chapter 5). Furthermore, while education is valuable, with few new jobs, it is not sufficient for low income households to improve their labor market outcomes. Cultural attitudes to female employment and biases in employers’ assessments of skills for men and women (Chapter 5) also make it more difficult to increase labor force participation among women. The capacity of Public Employment Services (PES), which may be particularly beneficial for activating disadvantaged groups who are less able to take advantage of the internet or connections to search for jobs, remains constrained and uneven, despite some recent and ongoing reforms (Box 6). 51 Table 4-2 Changes in employment 2012–2015 by main economic activity Main economic activity Total employment 2015 (‘000) 2012–2015 change (‘000) Agriculture, forestry and fishing 6.7 -7.2 Mining 2.5 -1.1 Manufacturing 43.9 0.5 Electricity, gas, steam, AC 5.3 -1.7 Water supply, sewerage and waste management 4.4 0.8 Construction 28.2 -0.8 Wholesale and retail trade, repair of motor vehicles 42.9 2.1 Transportation and storage 8.4 -2.2 Accommodation and food services 19.8 5.5 Information and communication 9.6 2.7 Financial/insurance activities 5.9 -0.9 Real estate 0.3 -0.4 Professional, scientific and technical activities 5.5 -0.1 Administrative and support activities 10.7 -3.7 Public administration 22.1 6.9 Education 36.4 -0.3 Human health and social work 18.6 -4.2 Arts, entertainment and recreation 4.4 -2.3 Other service activities 12.5 0.6 *Activities of households as employers 7.2 2.2 Activities of extraterritorial organizations 3.5 -2.5 Total 298.8 -5.9 Source: LFS 2012 and 2015 surveys. BOX 6: PUBLIC EMPLOYMENT SERVICE IN KOSOVO – RECENT REFORMS AND REMAINING CHALLENGES The Public Employment Service (PES) in Kosovo was established in 2001 as part of the Ministry of Labor and Social Welfare and consists of 7 regional PES centers, 23 municipal Employment Offices (Eos), and 8 Vocational Training Centers (VTCs). Following a 2010 UNDP review of PES performance, an “integrated concept of employment services” and a new registry of the unemployed were introduced to address multiple inefficiencies in the PES system. The new management information system (MIS) makes it possible to distinguish between active and passive job seekers, and also reorganized the function of counselors, allowing an unemployed person to interact with a single case manager for counseling, training or placement. The introduction of the MIS led to a drop in the number of registered unemployed from 328,000 in December 2011 to 100,000 in December 2012, helping to reduce EO caseloads. In 2012 there were 573 caseloads per EO employee, above EU and OECD average, and with considerable variation across municipalities, from 208 in Pristina to 3,038 in Kacanik. The PES reforms are ongoing, and a number of outstanding issues remain. Recent reforms include the adoption of individual plans, as well as a differentiated approach to clients, and better trained personnel for counseling. However, one challenge has to do with the fact that training of counselors has been undertaken by donors in its entirety, raising questions about longer term sustainability. Source: World Bank (2013). 52 Active labor market programs have shown some effectiveness, but they are only part of the solu- tion and the composition of programs matters. A recent meta-evaluation of the effectiveness of active labor market policies (ALMPs) by Card et al. (2015) shows that average impacts of ALMPs are close to zero over the short run but become positive after 2–3 years. Furthermore, they find that (i) impacts are larger for women and long-term unemployed; (ii) gains are larger for programs focusing on human capital accumulation (training); and (iii) ALMPs are more effective during recessions in a depressed labor market. The combination of high, and long-term, unemployment, particularly among women, and a depressed labor market, may suggest greater scope for ALMPs effectiveness, albeit against the evidence of limited impact of ALMPs elsewhere, pilot approaches with rigorous impact evaluations may be needed. The provision of activation measures in Kosovo has expanded in recent years (although ALMP financing is still comparatively low, at less than 0.5 percent of budget expenditure in 2012), and recent evaluations of several on the job training programs have shown a positive effect on employability (Box 7). BOX 7: ALMPS IN KOSOVO – SOME RESULTS OF PAST INTERVENTIONS UNDP Active Labour Market Programme for Youth in Kosovo (2007) The objective of the 2007 ALMP for the youth project in Kosovo was to (i) Strengthen the capacity of labor market institutions to provide individualized services to clients and to sequence active labor market measures; and (ii) pro- vide direct assistance to registered young jobseekers through on-the-job training (OJT), pre-employment training, employment subsidies, and internship schemes by partnering with enterprises that require additional workforce. Impact: Impact evaluation of the program showed a higher rate of employment in the treatment group (46 percent) compared to the control group (20 percent). Among beneficiaries, 38 percent secured employment immediately after the training, which suggests that they continued in the same enterprises where they received OJT. Contrib- uting success factors included (i) development of individual employment plans for the beneficiaries, (ii) careful preselection of the enterprises, (iii) good matching of the candidates with the profession or area of training, and (iv) quality of the training. The cost-benefit analysis suggested that the project generated a positive benefit of just over 1.42 times of the costs incurred. This is consistent with findings of ALMP evaluations in a number of other countries (Betcherman et al. 2004: 27). Second evaluation of UNDP ALMP projects (2012) The second evaluation of UNDP ALMPs was conducted in 2012, examining mainly on-the-job training (OJT), com- bined Institutional and Workplace Training (INST), and Internships (INT) implemented during the period 2008–10 by UNDP. Impact: During the three-year period subject to evaluation, 1,903 jobseekers took part in the Programs and 377 were in the control group. In the treatment group, 38 percent had a job, compared to 19 percent in the control group, indicating a positive effect of ALMPs on employment outcomes at the 3 year mark. Among those employed, 81 percent stated that they were using the skills learned during the program in their present job. Almost half of those employed were working in the enterprise where they underwent the training. The degree of informality was also lower in the treatment group. There are significant differences in outcomes across interventions. While a majority of participants undertook OJT (984 participants overall), employment outcomes for this group (between 25–35 percent during 2008–10) were lower than for INST training (37 to 42 percent), or among internship program participants (44 to 59 percent). A sig- nificant aspect of the results is the measure of sustained employment. The Interns showed a high level of sustain- able employment effect, in that 54 percent of the 2008 cohort, were in employment on the survey date (three years later). 51 percent of the 2009 cohort were in employment on the survey date (two years later) and 37 percent of the 2010 cohort were in employment one year later. For INST the rates are 43/33/20 percent for 2008/9/10 respectively. The lowest sustainable employment rate was for OJT at 25/24/26 percent for 2008/9/10. 53 5.  SKILL MISMATCHES IN THE LABOR MARKET Many people remain unemployed while job vacancies are unfilled, due to skills mismatches. The labor market is characterized by high rates of inactivity and unemployment. This may be due to a number of reasons including few jobs being created (Chapter 3), family-related or other reasons for inactivity, or the work in the informal market which is oftentimes difficult to measure (Chapter 4). Another contributing factor is mismatches between the types of education and skills that job seekers have to offer, and the education/skills that employers are seeking at any given point in time. Inadequate of skills is not the main business constraint, but it is more stringent than in other coun- tries in ECA. In the 2013 Enterprise survey, inadequately educated workforce was seen as a top business obstacle by seven percent of surveyed firms.29 This is low compared to 26 percent of firms in that saw informal practices as the top business environment obstacle. However the inadequate workforce appears to be more important in Kosovo than in ECA on average or in neighboring Albania. Moreover, while not a top constraint, 25 percent of firms saw it as a “major constraint”, compared to 14 percent in ECA and 21 percent across all countries where Enterprise Surveys were fielded during 2010–2016. Inadequate education is a greater constraint for larger firms. Among small firms (5–19 work­ ers), 19 percent identified inadequate education as a major constraint, whereas among medium size firms (20–99 employees) this increased to 32 percent, and among large firms (100+) employees 43 percent saw it as a constraint (Figure 5-1). This pattern is not unique to Kosovo; in other countries, and across the ECA region on average, the education is more important in larger firms. What is different is a steeper gradient in Kosovo: The difference between small and large firms in Kosovo is 24 percentage points, while it is only four percentage points for the ECA region on average, and this is also the case when averaging across all enterprise survey, not restricted to the ECA region. Figure 5-1 Share of firms identifying inadequately educated workforce as a major constraint 43.2 32.5 24.7 23.6 20.1 19.4 17.7 16.6 13.1 KOSOVO ECA ALL COUNTRIES SMALL (5–19) MEDIUM (20–99) LARGE (100+) The estimates are based on responses of business owners and top managers in 202 firms that were interviewed from January 2013 29 through November 2013. More details are available at http://www.enterprisesurveys.org/ 54 BOX 8: KOSOVO EMPLOYER STEP SURVEY To deepen the understanding of the nature of skills mismatches and the links between skills and labor market out- comes among the working-age population, 500 small, medium and large firms were interviewed in Kosovo during November/December 2015. The STEP survey sampling frame was based on a database of firms compiled by the Kosovo Agency of Statistics, and a small complementary frame of Serb enterprises in Northern Kosovo, obtained independently. Two geographic domains were defined: Pristina and Other (remaining regions of Kosovo). The sam- pling frame was stratified by geographic domain and four size strata in reference to the number of employees: 5–9, 10–15, 16–50 and 51+ employees. A separate frame of 39 Serb firms in Northern Kosovo was combined with this large sample of firms for the first phase. A total of 500 firms were interviewed as part of the STEP survey (351 from the Target sample and 149 from the reserve sample). Of these, 43 percent of firms were from Pristina and the remainder from outside Pristina. Economic activity by sector Code Frequency Percent Agriculture, forestry and fishing A 4 0.8 Mining and quarrying B 2 0.4 Manufacturing C 61 12.2 Electricity, gas, steam and air conditioning supply D 10 2.0 Water supply; sewerage, waste management and remediation activities E 17 3.4 Construction F 68 13.6 Wholesale and retail trade; repair of motor vehicles and motorcycles G 90 18.0 Transportation and storage H 23 4.6 Accommodation and food service activities I 29 5.8 Information and communication J 14 2.8 Financial and insurance activities K 10 2.0 Real Estate activities L 0 0.0 Professional, scientific and technical activities M 4 0.8 Administrative and support service activities N 5 1.0 Public administration and defense; compulsory social security O 3 0.6 Education P 6 1.2 Human health and social work activities Q 16 3.2 Arts, entertainment and recreation R 8 1.6 Other service activities S 124 24.8 Activities of households as employers; undifferentiated goods T 6 1.2 Total 500 100% Source: Enterprise survey data. Kosovo’s Department of Statistics conducted a special survey (STEP) on skills mismatches. Neither the firm census data used to assess the patterns of job creation (Chapter 3), nor the Labor Force Survey used to assess employment patterns (Chapter 4), offered a direct way to assess the alignment of skills demand with skills supply. However, the STEP survey addressed this directly and provided the data that is discussed in this chapter. The STEP survey verified other findings on the constraints that inadequate skills pose to businesses. Overall, 27 percent of firms reported labor factors of different types to be major or severe constraints to doing 55 business. Among labor factors, the most constraining are finding workers with previous experience (36 per- cent), labor availability and “technical and vocational education and training” (22 percent each), high job turn- over (21 percent) and general education of workers (18 percent). On the other hand, factors such as overall wage level or employment protection legislation are cited less frequently. There is some variation across indus- tries; overall, firms in agriculture/industry appear to be more constrained by inadequate education/experience than firms in sectors such as business services, public services or other activities. Not surprisingly, in sectors such as construction and trade/transport/accommodation, adequacy of vocational and technical training is more constraining that adequacy of general education. Firms have tried to hire employees in recent years, across multiple occupational categories. Over a quarter of firms tried hiring in managerial, professional and sales occupations, over one fifth of firms attempted to hire technicians/associate professionals or service workers (Figure 5-2). Hiring attempts were heavier in sectors such as agriculture/industry, construction as well as in trade/transport/accommodation. In some other sectors such as public services quite a few firms reported trying to hire employees into professional occupations (47 percent of firms), but, at the same time, very few managerial hires. Figure 5-2 Occupational categories of new hires (10) ELEMENTARY OCCUPATIONS 11.1 (9) DRIVERS, PLANT & MACHINE OPERATORS, & ASSEMBLERS 16.9 (8) CONSTRUCTION, CRAFTS & RELATED TRADES WORKERS 13.6 (7) SKILLED AGRICULTURE, FORESTY AND FISHERY WORKERS 0.6 (6) SALES WORKERS 25.4 (5) SERVICE WORKERS 24.0 (4) CLERICAL SUPPORT WORKERS 14.3 (3) TECHNICIANS AND ASSOCIATE PROFESSIONALS 22.1 (2) PROFESSIONALS 26.9 (1) MANAGERS 26.5 Source: Staff estimates based on the Employer STEP survey. Large firms were by far the employers most active in hiring. Half of large firms attempted to hire pro- fessionals over the past three years, and 56 percent tried to hire sales workers, compared to 20 percent in both categories among micro firms (1–10 employees). Young firms (up to five years) are particularly active in hiring service workers and sales workers (32 percent and 28 percent respectively) while well-established firms (16+ years) are the most active in hiring professionals. Hiring is linked to innovation, investment in R&D and international business contacts. Firms that invested in research and development, as well as those that introduced innovative products during the previous three years are considerably more likely to have tried hiring managers, professionals as well as technicians/associate profes- sionals than firms that are traditional/do not invest in R&D. These differences are significant – 38 percent of firms with R&D investments tried to hire technicians, compared to 15 percent of firms that had no R&D investments. Among innovative firms, 30 percent tried hiring processionals, compared to 15 percent among traditional firms. Firms with international business contacts were also a lot more active in hiring in higher-skill occupations (man- agers, professionals as well as technicians). There are few differences between firms that differ along the above characteristics, on the other hand, in terms of their hiring intentions in occupations such as service workers and sales workers. Many firms had problems hiring new employees related largely to the lack of skills or prior work experience. Overall, more than three quarters of firms had problems when hiring managers, professionals 56 Figure 5-3 Type of problems when hiring (share of firms reporting problems) (10) ELEMENTARY OCCUPATIONS (9) DRIVERS, PLANT & MACHINE OPERATORS, & ASSEMBLERS (8) CONSTRUCTION, CRAFTS & RELATED TRADES WORKERS (7) SKILLED AGRICULTURE, FORESTY AND FISHERY WORKERS (6) SALES WORKERS (5) SERVICE WORKERS (4) CLERICAL SUPPORT WORKERS (3) TECHNICIANS AND ASSOCIATE PROFESSIONALS (2) PROFESSIONALS (1) MANAGERS 0 10 20 30 40 50 60 70 APPLICANTS DID NOT LIKE WORKING CONDITIONS THERE WERE NO OR FEW APPLICANTS APPLICANTS EXPECTED WAGES HIGHER THAN WE CAN OFFER APPLICANTS LACKED REQUIRED SKILLS APPLICANTS LACKED REQUIRED WORK EXPERIENCE Source: Staff estimates based on Employer STEP survey. or technicians (referred in the survey as Type A workers). Problems with hiring other types of employees (Type B workers) are somewhat less pronounced, but even for these types of jobs the share of firms reporting problems is over 40 percent of total. Multiple problems were cited, but the inadequacy of skills or experience dominates (Figure 5-3). Firms face greater hiring challenges if they are innovative, invest in R&D, are foreign owned or have international business contacts. This is more pronounced for Type A workers (managers, professionals, technicians), but even for Type B occupations that would generally require lower education levels. Insufficient skills are also more constraining for larger firms – 72 percent of large firms that had problems hiring Type A workers cited insufficient skills of the candidates, compared to 51 percent of micro firms. The same patterns apply to Type B occupations, but the differences are narrower. There are few differences between firms of different age, either in the case of Type A or Type B workers (Table 5-1). Table 5-1 Problems when hiring Type A and Type B workers due to insufficient skills among firms with hiring problems, by firm type Type A worker Type B worker Size (based on information about total permanent employment, full-time and part-time, all with social security contributions) Micro (1–10 employees) 51.2 46.6 Small (11–50 employees) 56.6 45.6 Large (51+ employees) 72.5 58.6 Age (based on information about the year when the workplace began operations in the country; year of fieldwork is 2015) Up to 5 years 56.6 50.2 6 to 15 years 56.7 47.1 16 years and more 51.9 44.0 (continued on next page) 57 Table 5-1 (continued) Problems when hiring Type A and Type B workers due to insufficient skills among firms with hiring problems, by firm type Type A worker Type B worker Ownership (based on information about the major shareholder/owner) Private domestic 55.8 46.0 Foreign 83.6 84.6 State 32.6 67.6 Innovation (whether firms introduced new technology, process, products or services in the past 3 years) Traditional 39.2 29.9 Innovative 58.5 51.2 International business contacts (whether firms have international business contacts with entities in other countries) Have international business contacts 61.1 46.3 Do not have international business contacts 48.7 48.8 Investment in R&D (based on question 5.06 “During the last three years, did this establishment spend on formal research and development?”) Invested 65.4 56.4 Didn’t invest 51.9 45.7 Source: Staff estimates based on Employer STEP survey. English and computer skills are deficient. Employers were asked to evaluate, for a set of skills, whether there are differences between what is required for the job and the current level of this skill in a typical worker. On average, less than five percent of firms thought there were differences between the skills of typical Type A workers and job requirements (Figure 5-4). For Type B workers the gap is somewhat wider, but still less than 10 percent for most skills. Insufficient skills are most common with respect to knowledge of foreign languages (especially English, and also Serbian for Type B workers), as well as insufficient computer skills (basic word pro- cessing skills for Type B; presentations and/or other advanced purposes like creating and managing data- bases, or using specialized computer programs for Type A). Hiring is largely informal and connections-based. Close to two thirds of firms recruited from informal chan- nels (personal contacts, people recommended by others) for Type A occupations; almost 60 percent of firms relied on the same for Type B jobs (Figure 5-5). The second most popular recruitment channel for both types of jobs was also based on offering jobs to individuals known from other firms. Both formal and informal firms hire from similar channels to fill most of their vacancies.30 But, there are some differences in recruitment strategies. Formal sector firms depend more on the internet and the public employment service to fill job vacancies than informal firms. Contacts between employers and employment services appears to be sporadic. Only one fifth of firms reported relying on public employment services for type A workers, and only a quarter relied on Private Employ- ment Services, while the share of firms relying on direct contact with educational institutions, schools, training centers and universities was even lower. Only 17 percent of firms reported being in regular contact with educational/training institutions for hiring Type A workers (6.7 percent for Type B workers), although there is considerable variation across firms of different sizes; among large firms more than half reported having regular contact with the education/training institutions for Type A occupations (although only eight percent report this for Type B), compared to only 11 percent among micro firms. However, that recruitment is not the only reason for being in contact with the education/training institutions; other reasons include providing work experience to students (internships), training, as well as testing students or providing feedback on curricula. Results from STEP survey. 30 58 Figure 5-4 Correspondence between employee skills and employer needs TYPE A OCCUPATION CAN DEMONSTRATE SPECIFIC TECHNICAL SKILLS CAN USE A COMPUTER FOR MAKING PRESENTATIONS AND/OR… CAN EASILY ADAPT TO NEW TASKS OR CHANGES IN THE WORKPLACE CAN CONTINUE IN THE FACE OF CHALLENGING SITUATIONS AT WORK CAN WORK WELL IN VERY BUSY OR DIFFICULT SITUATIONS CAN WORK WELL WITH OTHERS AND LISTENS TO OTHERS' VIEWS CAN BE RELIED ON TO GET THINGS DONE CAN STAY ON A LONG AND DIFFICULT TASK UNTIL IT IS FINISHED CAN FIND NEW AND BETTER WAYS TO DO THINGS CAN READ AND WRITE IN A FOREIGN LANGUAGE (FOR EXAMPLE, TURKISH, GERMAN, AND OTHER EU LANGAUGES) CAN READ AND WRITE IN OTHER (NON MATERNAL) OFFICIAL LANGUAGE CAN READ AND WRITE IN ENGLISH CAN DO CALCULATIONS AND WORK WITH NUMBERS 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% YES, THERE IS A DIFFERENCE NO, THERE IS NO DIFFERENCE THIS SKILL IS NOT REQUIRED FOR THE JOB TYPE B OCCUPATION CAN DEMONSTRATE SPECIFIC TECHNICAL SKILLS CAN USE A COMPUTER FOR BASIC WORD PROCESSING TASKS, EMAIL AND INTERNET SEARCHES CAN EASILY ADAPT TO NEW TASKS OR CHANGES IN THE WORKPLACE CAN CONTINUE IN THE FACE OF CHALLENGING SITUATIONS AT WORK CAN WORK WELL IN VERY BUSY TIMES OR DIFFICULT SITUATIONS CAN WORK WELL WITH OTHERS AND LISTENS TO OTHERS' VIEWS CAN BE RELIED ON TO GET THINGS DONE CAN FIND NEW AND BETTER WAYS TO DO THINGS CAN READ AND WRITE IN A FOREIGN LANGUAGE (FOR EXAMPLE, TURKISH, GERMAN, AND OTHER EU LANGUAGES) CAN READ AND WRITE IN ENGLISH CAN READ AND WRITE IN SERBIAN CAN READ AND WRITE IN ALBANIAN CAN DO CALCULATIONS AND WORK WITH NUMBERS 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100% YES, THERE IS A DIFFERENCE NO, THERE IS NO DIFFERENCE THIS SKILL IS NOT REQUIRED FOR THE JOB Source: Staff estimates based on Employer STEP survey. Figure 5-5 Channels of recruitment (% of firms reporting Type A or type B occupations) OTHER INFORMAL CHANNELS (PERSONAL CONTACTS, PEOPLE RECOMMENDED BY OTHERS) OTHER MEDIA ADVERTISEMENTS/POSTINGS (NOT INCLUDING INTERNET) INTERNET POSTINGS DIRECT CONTACT WITH EDUCATIONAL INSTITUTIONS, SCHOOLS, TRAINING CENTERS, UNIVERSITIES, ETC. OFFERS TO EXPERIENCED PEOPLE IN OTHER FIRMS JOB FAIRS PRIVATE EMPLOYMENT SERVICES PUBLIC EMPLOYMENT SERVICES 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 TYPE A WORKER TYPE B WORKER Source: Staff estimates based on Employer STEP survey. 59 Informal hiring is related, in part, to a distrust of the education and training systems to supply employees with appropriate skills. One third of firms disagreed that the general education system pro- duced people with the practical experience that can be applied to their work; 16 percent of firms expressed similar views about the vocational training system (Figure 5-6). A notable share of firms was also pessimistic in assessing the ability of the education system to train people with up-to-date knowledge of methods, materials and technology. Overall, the employers’ views of the general education system are more negative than their views of the vocational training system. The discontent with the education system tends to be higher among larger firms, or among firms that innovate; it is also more negative when judging the ability of formal education/training systems to equip people with the required skills for Type A occupations in comparison with Type B occupations. Figure 5-6 Share of firms which disagree with the following statement about the education/training system PRODUCES PEOPLE WITH PERSONAL SKILLS LIKE TIME-MANAGEMENT, RELIABILITY, ABILITY TO WORK WITH OTHERS, ETC. PRODUCES PEOPLE WITH PRACTICAL EXPERIENCE THAT CAN BE APPLIED TO THEIR WORK PRODUCES PEOPLE WITH THE UP-TO-DATE KNOWLEDGE OF METHODS, MATERIALS AND TECHNOLOGY PRODUCES PEOPLE WITH THE SKILLS YOU RANKED AS IMPORTANT MEETS THE SKILL NEEDS OF YOUR BUSINESS ADEQUATELY 0 5 10 15 20 25 30 35 GENERAL EDUCATION SYSTEM TRAINING SYSTEM Source: Staff estimates based on the Employer STEP survey. Some firms use on-the-job training to compensate for deficiencies in the education and training institutions. On-the-job training and other external or internal training for Type A workers is more common among firms that experienced problems hiring due to insufficient skills among applicants than for other firms that did not have skill-related hiring problems (Figure 5-7). Thus larger firms, firms that invested in R&D, or foreign owned firms were more likely to provide training for Type A workers. This does not always hold, though, for Type B workers. For instance, more of the large firms provide OJT for Type B workers, but micro firms had the greatest incidence of providing internal training. Likewise, foreign firms were much less likely to provide training of any kind for Type B workers than state firms. Finally, firms that did not agree that the education system provided the skills necessary for their businesses were actually less likely to provide additional training than the firms that had more positive views. It may be that such views on their own, not supported by actual difficulties of hiring because of skills deficiencies, are not sufficient to prompt the firms to address skills shortages through in-house training. Employers assess the skills of men and women differently. More than 30 percent of employers thought men to be better qualified at demonstrating specific technical skills, continuing in the face of challenges, working well in busy/difficult conditions, or staying on a long and difficult task until finished. Men are also perceived to be better at adapting to new tasks or finding new and better ways of doing things by over a fifth of the firms in the sample. The only skills where women were viewed more favorably than men are foreign languages (Figure 5-8). Despite the less favorable assessment of women’s skills, men are more likely to receive training. Among firms that provided training for type A workers during the pervious year, 47 percent reported that males 60 Figure 5-7 Self-provision of training by firms in Kosovo TRAINING AMONG FIRMS WITH PROBLEMS HIRING DUE TO SKILLS TRAINING IN FIRMS WITH CRITICAL ASSESSMENT OF THE ED. SYSTEM 35.3 28.8 23.7 23.9 27 21.7 21.5 21.2 20.6 20.2 19.4 18.5 17.4 22 17.8 16.3 13.4 12.3 11.7 6.5 5.8 5.2 4.2 3.5 OJT INTERNAL EXTERNAL OJT INTERNAL EXTERNAL OJT INTERNAL EXTERNAL OJT INTERNAL EXTERNAL TYPE A WORKER TYPE B WORKER TYPE A WORKER TYPE B WORKER SOME PROBLEMS IN HIRING BECAUSE OF SKILLS FIRMS WITH CRITICAL ASSESSMENT OF THE EDUCATION AND TRAINING SYSTEM OTHER FIRMS WITH RESPECT TO GENERATED SKILLS OTHER FIRMS Source: Staff estimates based on Employer STEP survey. Figure 5-8 Gender assessment of skills DEMONSTRATING SPECIFIC TECHNICAL SKILLS 31.7 4.3 COMPUTER SKILLS 11.5 6.1 EASILY ADAPTING TO NEW TASKS OR CHANGES IN THE WORKPLACE 23.3 10.2 CONTINUING IN THE FACE OF CHALLENGING SITUATIONS AT WORK 35.3 4.5 WORKING WELL IN VERY BUSY OR DIFFICULT SITUATIONS 33.9 4.6 WORKING WELL WITH OTHERS AND LISTENING TO OTHERS' VIEWS 19.5 9.4 BEING RELIED ON TO GET THINGS DONE 25.5 8.7 STAYING ON A LONG AND DIFFICULT TASK UNTIL IT IS FINISHED 36.6 3.4 FINDING NEW AND BETTER WAYS TO DO THINGS 21.5 7.7 READING AND WRITING IN (ANOTHER) FOREIGN LANGUAGE 4.9 14.3 READING AND WRITING IN ENGLISH 7.5 14.0 DOING CALCULATIONS AND WORKING WITH NUMBERS 10.3 9.1 MEN ARE BETTER EQUIPPED WOMEN ARE BETTER EQUIPPED Notes: The omitted category is “men and women are equally equipped”. Source: Staff estimates based on Employer STEP survey. received more training than females (30 percent reported equal training for men and women and an additional 10 percent reported that women received more training). Similarly, for Type B workers, a third of the firms that provided training said that men received more training, compared to 13 percent providing more training to women and 36 percent reporting no gender differences. These patterns are consistent with the broader attitudes regarding women’s commitment to the workplace. More than 40 percent of firms viewed women as having competing time demands given family responsibility, and a similar share stated that women are costlier to hire than men given regulations such as maternity leave. Women were also viewed by over 30 percent of employers to have expectations of higher benefits (such as flexible hours) than firms can offer (Figure 5-9). 61 Figure 5-9 Expectations of female applicants WOMEN HAVE COMPETING TIME DEMANDS GIVEN FAMILY RESPONSIBILITIES WOMEN ARE COSTLIER TO HIRE THAN MEN GIVEN LABOR REGULATIONS (SUCH AS MATERNITY LEAVE PROVISIONS) WOMEN LACK THE REQUIRED SKILLS WOMEN LACK THE REQUIRED WORK EXPERIENCE WOMEN EXPECT HIGHER BENEFITS THAN FIRMS CAN OFFER (SUCH AS FLEXIBLE HOURS) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 SHARE OF FIRMS THAT AGREE Source: Staff estimates based on Employer STEP survey. 62 6.  SUMMARY OF FINDINGS AND POLICY CONCLUSIONS Kosovo is a young, low-income country with weak job creation. Its growth averaged 3.4 percent over 2008–2015, and the economy weathered the financial crisis and recession that affected many other Euro­ pean nations. However, growth did not reduce unemployment which stood at 32.9 percent in 2015. Higher unemployment among women (about 36 percent) and youth (over 57 percent) are significant social concerns. Additionally, about 10.7 percent of the working age population were seen as discouraged workers in 2014. Micro firms are important creators of jobs in the formal sector, but few of them grow so their capacity to expand jobs is limited. Over 90 percent of all formal firms in 2014 were micro and they provided 36 per­ cent of jobs, but large firms (100 employees or more) account for 35 percent of employment despite being only 0.5 percent of firms. Startups – an important source of job creation – are low by international standards and typically don’t grow, a symptom of a static economy. As a result, large enterprises are the strongest job creators – the average surviving large firm starts at 300 employees and grows to almost 800, while the average micro firm remains almost the same. Future job creation and productivity growth and should be linked. Productivity increases were linked with contracting employment in early years, and with expanding employment in recent years. On balance, over the past decade labor has been allocated to more productive firms, although efficiency remains low by inter­ national standards. Large, foreign-owned and exporting firms are more productive, and are also less likely to exit. Exporting and more productive firms are also more likely to grow. Overall, the firm-level evidence is consis­ tent with “creative destruction”, i.e., with more productive firms expanding and less productive ones contract­ ing, but allocative efficiency is low by international standards. Growth has been driven largely by remittances and has led to rising imports. The Systematic Country Diagnostic (SCD) found that limited job creation and business expansion in the tradables sector are byproducts of the current growth process. Domestic savings are negative, economic growth was driven by remittances and aid. Because of the narrow, undiversified, and uncompetitive productive base (the agriculture and services sectors are large; manufacturing is small; and little is tradable), the steady consumption demand fueled by remittances has been met mainly by imports. The supply response has been limited to nontradable sectors such as construction, real estate, and retail and other services, where there has been considerable dynamic growth. Infrastructure shortages prevent Kosovo from capitalizing on its comparative advantages. Several goods (garments, textiles, and food) and services (travel and communications) are among Kosovo’s comparative advantages. Manufacturing has the potential to spur income and exports and to generate skilled and unskilled employment. Modern services are already a major source of income and they have the potential to transform the economy by their impacts on other sectors and on human capital formation. However, manufacturing and modern services have been held back by regulatory constraints and infrastructure gaps. An unfavorable business environment, weak infrastructure, and firms’ poor access to technology and finance reduce growth prospects, profitability, and competitiveness and they encourage investments in nontradable sectors (World Bank, 2016). Kosovo’s informality is among the highest in ECA due to some of the factors that prevent Kosovo from exploiting its comparative advantages – inefficient regulations and poor public services. About 36 percent of employees are in the informal sector. Informal workers are in small firms (five or fewer employees), unpaid family members, or self-employed in small firms or unprofessional occupations. Informality 63 is particularly high in agriculture and construction, sectors where jobs tend to require less skilled personnel, casual labor and – in the case of agriculture – unpaid family workers. More than a quarter of formal firms (26 percent) identified informal sector competition as the single biggest impediment to their growth. They ranked it ahead of access to finance, corruption, customs and trade regulations, electricity, skills of the work­ force, and other factors. At the same time, informal employment – more common among youth, males, the less-educated, and rural residents – is also associated with reduced job security vis-à-vis formal sector jobs, and excludes workers from benefits. Inactivity is high, particularly among women and youth, and leads to social fragility. High informality makes it harder to measure economic activity accurately, and some informal employment cannot be recorded. Nonetheless, inactivity is high, particularly among women owing to family responsibilities, poor provision of child/ elderly care service, and also of employers’ biases. Inactivity is also related to employers’ perceptions of the low quality of formal education and vocational training institutions that are seen as not equipping employees with skills that are in demand. One outcome of low skills/skills mismatches is that hiring remains highly informal and reliant on personal contacts. High levels of economic disenfranchisement (together with political disenfranchisement) are highlighted in the recent risk and resilience assessment as being associated with political instability, low institutional trust, widespread desire to migrate. Youth disenfranchisement is a key driver of fragility in Kosovo today. Table 6-1 Kosovo SCD prioritization Time horizon of impacts Poverty and Priority Growth shared prosperity Sustainability Short Medium Long Reducing energy bottlenecks Improving governance, the rule of law, and the business climate Improving the allocation and efficiency of public expenditure Increasing employment and labor productivity through education Ensuring the sustainable management of natural resources Strengthening tax policy and administration Increasing productivity in agriculture Raising employment and labor productivity through labor policies Reducing ICT bottlenecks Raising the quality and equality of opportunity through social protection Raising the quality and equality of opportunity through health care Maintaining financial stability and deepening financial intermediation Reducing transport bottlenecks High priority Medium priority Low priority Color key Short term Medium term Long term Source: World Bank (2016) 64 Creating more and better jobs, particularly for the bottom 40 percent, requires diverse, improved economic policies. The SCD identified the infrastructure bottlenecks in energy and building up governance and the rule of law as top priority areas for creating a more attractive environment necessary to reap the ben­ efits of European integration, stimulate productivity gains, and create quality jobs and inclusion. Among other high priorities are improving the allocation and efficiency of public resources and addressing skills mismatches to increase employment and labor productivity. The analysis and findings in this Jobs Diagnostic reinforce the priorities of the SCD. The SCD’s policy agenda had a broader scope, and the policy section in this report focus more narrowly on creating formal employ­ ment, reducing informality, and raising labor force participation. Policy recommendations fall into three categories: (a) improving the regulatory and business environment to aid job creation and firm growth; (b) addressing skill shortages and mismatches that hamper both Kosovo’s ability to harness its competitive advantages and the ability of the unemployed and inactive to transition into employment; and (c) adopting reforms facilitate the creation of more jobs and improve skills matches, accompanied by measures for women in particular to take advantage of employment opportunities. It should be noted that the data that the jobs diagnostic had at its disposal does not allow us to make specific recommendations with respect to policy design. Rather the goal of this section is to point to several priorities which can subsequently be explored in more detail as part of the ongoing policy dialogue. IMPROVING THE REGULATORY AND BUSINESS ENVIRONMENT A transparent and efficient regulatory and legal framework is essential for a well-functioning, com- petitive, and innovative market economy. Clear rules to govern trade and investment would strengthen businesses to withstand competitive pressures and connect with global value chains, thereby generating jobs and spurring competition, innovation, and productivity increases. They would make Kosovo more attractive to investors – domestic and foreign. Inefficient and costly government inspections are among factors responsible for a large informal sector. Policies to improve the quality of institutions and of service delivery would create the environment needed to encourage greater formality and give incentives for citizens to join the formal economy and pay taxes. Starting a business is easy, but operating, growing, or closing it is cumbersome. According to the Doing Business indicators, it is very difficult to deal with construction permits, to get electricity and credit, to protect minority investors, to enforce contracts and to resolve insolvency. These issues can decrease the incentives for firms to invest and create jobs. Professionalizing the tax and labor inspections agencies could encourage firms to operate formally. The current system of inspections creates incentives for firms to stay small and informal. Lessons from other tran­ sition economies (see Box 9) show that making the inspectors more accountable, increasing the transparency of the inspections (by establishing time limit to the visits and writing a summary of the visits), and randomizing the assignment of inspectors to firms can help improve compliance. Better directing inspections toward unregistered businesses is an additional option. Improving the hiring process for inspectors and investing in training to shift their work towards a more client-oriented service could lead to better cooperation with tax authorities. Finally, coordina­ tion with other inspection agencies (such as food safety, building safety, etc.) may bring additional benefits. Improvements in the accreditation and certification systems would make it easier for firms to develop, test and certify new products. The standardization, accreditation, and certification systems are weak and increase firms’ costs of quality assurance. Domestic testing facilities are inadequate, and certification bodies cannot offer internationally recognized certification. In the absence of benchmarks, firms do not have stan­ dards to improve their competitiveness. Also, export markets do not have confidence in Kosovar goods as they lack adequate certification and do not meet required standards. Interviews with firms in the food processing and tourism sectors reveal the lack of appropriate certification and standardization infrastructure among factors that inhibit their growth. Improving certification/accreditation infrastructure will help Kosovo to prepare for EU accession and to comply with WTO rules. Given the small domestic market, establishing specialized laboratories might not be cost-effective; however, in some sectors – e.g., agriculture, livestock, and energy, may be econom­ ical and using specialized laboratories in the region may be a feasible option. Also, without harmonization of 65 BOX 9: REGIONAL APPROACHES TO REFORMING INSPECTIONS Bosnia and Herzegovina (including both entities Federation BiH and RS) had originally 26 inspections per entity prior to the reform, with the National level having 2 inspections. Following the reform, the Federation BiH now has 15 inspections in total, out of which 11 are part of Central inspectorate and 4 (tax, customs, education and fire) are still within relevant ministries. In the RS, there are now 15 inspections, out of which 11 are part of General inspec- torate. The rest are within ministries[2]. In Bosnia and Herzegovina prior to the reform (in both entities) inspections were controlled by some 7–9 ministries and 2–3 agencies[3]. In Croatia, 12 inspections from 4 large ministries have been incorporated under the umbrella of Central Inspec- torate, which managed to reduce the number of inspection units from 100 to 49 and the number of branches from 22 to 5[4]. Croatia had some 27 inspections prior to the reform[5]. However, Croatia disbanded the Central inspec- torate in 2013, and inspections returned to sectoral inspectorates within ministries. Serbia has 36[6] different inspections now, controlled by 13 ministries and 7 agencies, and they had no plans to create central inspectorate or to merge inspections. Serbia adopted single procedural (not organizational) inspec- tion law and created an inspection coordinating body. Serbia did not reduce the number of inspection agencies as a result of reform, but they’ve improved inspection procedures in all 36 inspection departments. Slovenia had 27 inspections agencies before the reform. Following the reform, Slovenia brought down the number of inspections to 21, controlled by 14 ministries or agencies. Slovenia created a single procedural inspection law and an inspection coordinating council. Albania reformed inspections using combination of the models presented. A single inspection law was created that covered both procedural aspects and some organizational/governance aspects of the inspection organiza- tion and works in Albania. Inspections were merged from 33 to 15 sectoral inspectorates in ministries, while the Central Inspectorate was created as a stand-alone institution. In addition to coordinating the work of inspectors and E-Inspector system, the Central Inspectorate has the right to oversee the inspectors and in cases of proven infringements proposes disciplinary actions. References: [2] Sahovic, Tarik, draft Working Paper on Inspection Reform in Western Balkans, January 2008; updated 2014 [3] Sahovic, Tarik, draft Working Paper on Inspection Reform in Western Balkans, January 2008; updated 2014 [4] Mr. Branko Jordanic, former Inspector General of Croatian Inspectorate [5] Sahovic, Tarik, draft Working Paper on Inspection Reform in Western Balkans, January 2008: updated 2014 [6] Serbia, USAID BEP Project official data, and Tarik Sahovic own research 2014 Source:  Cordova and Sahovic, “Inspection Reform - Do Models Exist”, Research and Experience from 25 countries, The World Bank Group – Investment Climate border inspections on technical and customs regulations there are avoidable delays to obtain certifications for imports and exports. Improving public services may increase citizens’ willingness to pay taxes and formalize. Low willing­ ness to pay taxes is often associated with lack of satisfaction with public services. Abdih and Medina (2013) find institutional quality (measured by way of the control of corruption, rule of law, regulatory quality and gov­ ernment effectiveness components of the World Bank’s governance indicators) to be among the top two deter­ minants (alongside regulatory burden) of the size informal sectors in a number of ECA countries, even more important than other factors such as tax burden and labor rigidities. Improving these perceptions would be a slow process while new measures to raise transparency and accountability are adopted. However, improving education is a more pressing concern, given the poor quality of education, and mismatches with labor market demands. Addressing skills shortages and mismatches is the second area of policy focus detailed below. REDUCING SKILLS SHORTAGES AND MISMATCHES Policies to improve human capital can increase productivity and provide incentives to expand formal sector employment. Moreover, given the youth of the population, the time to invest in education and skills is now. Educational attainment has risen over time – a much larger share of younger age cohorts (ages 20–29) 66 have post-secondary education compared with the 40–49 or older cohorts. However, in the 2015 Programme for International Student Assessment (PISA) Kosovo’s students performance was low, even by regional stan­ dards. Improving the quality and relevance of education at all levels and offering equitable access to educa­ tional opportunities at early ages are high priorities with potentially large payoffs. Educational achievement is strongly correlated in Kosovo with the probability of employment. A young, skilled, and competent workforce will help attract investment to expand opportunities for employment and can lead to sustained income gains. Better education will lead to higher employment only if it provides skills that employers demand. Neither higher education nor VET programs have been particularly effective in equipping students with the competencies required by employers or in responding quickly to labor market demands. The numerous voca­ tional training centers (VTC) are underutilized by Public Employment Offices (PEOs), due to the limited visibility, credibility, and pertinence of the training. Coordination between the private sector and the government has failed, contributing to the large gap between employers’ needs and workforce development. Specifically, the education system’s weak relevance to labor market needs and the limited private sector involvement with the intermediation services have contributed to weak market-oriented skills formation and job placement. Graduates of training programs would benefit from greater involvement of the private sector in shaping the curricula and delivery of VTCs’ training programs. Skills shortages are not only subject-specific; inter-personal and other business skills should also be improved. More than a quarter of firms disagreed with the premise that the general education system produces people with the up-to-date knowledge of methods, materials and technology; 15 percent had the same view of the training system. The SCD recommended to upgrade vocational schools with state of the art teaching methods on agricultural topics relevant for commercially oriented farmers and agroprocessors and to stimulate adoption of innovative technologies. However, an even-larger share of firms disagreed with the premise that the education system produced people with practical experience that can be applied to their work, and over a fifth of firms thought personal skills like time-management, reliability, and ability to work with others were deficient among graduates of the general education system. Given the importance of the services sector in Kosovo, these skills will likely be of increasing importance going forward. In addition, given the importance of start-ups and micro firms in the private sector, skills related to business planning and development, accounting/ finance, product/service development, would be an important skill shortage to address. Better skills in marketing, sales, and design, as well as computer skills and foreign languages, are important to compete in export markets. Most companies in Kosovo do not have dedicated sales manage­ ment expertise to promote their products, and they lack knowledge of export markets. Export promotion is usually done through fairs, and the costs are challenging for companies. Companies have few viable connec­ tions to export buyers and wholesale and retail chains in the EU. This results in a lack of sales opportunities and limited knowledge of international sales trends. Even in the internal market, local companies find it difficult to compete with imports. Poor design and quality of packaging is a negative in competing with imports, even if domestic product quality is equal or superior. Lack of qualified specialists in marketing, sales and design is cited by firms among constraints to growth and also constrains export potential. To address these needs, the VET system needs to be better aligned with the skill demands in these areas. While commerce and services are the two sectors that accounted for most of the new jobs over the past decade, the sectoral profile of VET students is heavily skewed toward business/law, as well as medicine and computer science, whereas the share of VET students in the services skill sector (based on Eurostat categorization) is very low relative to the share of jobs created in services (MEST, 2015). Effective intermediation and activation policies can be part of the solution of addressing skills mis- matches and inactivity for the long-term unemployed and the B40. More than a third of out-of-work individuals who are looking for jobs contact PEOs. About 64 percent of the jobs found through PEOs are in the private sector. Given that the poor and B40 households are marked by both low educational attainment, and greater degree of inactivity, unemployment and underemployment, it is more difficult for these groups both to rely on the internet and connections to secure employment, and also to take advantage of formal vocational training options. While the effectiveness of activations measures may be limited in the general population, it is likely to be higher among these hard to reach populations who could benefit from access to counselors 67 provided by EOs, individualized employment plans, and direct links with employers that EOs can provide. This is supported by the evidence of the effectiveness of some of the earlier activation programs in Kosovo that focused on human capital upgrading. Kosovo could build on past success and ongoing reforms by: (i) upgrad­ ing PES’s information system to update the unemployment registry, to offer matching services, and to enable interoperability between employment offices and centers for social works; (ii) better collaborating with the private sector to assess market needs and the supply of skills; and (iii) developing the online job platform to attract more youth by improving job matching. PES has been establishing monitoring and control mechanisms to set annual targets and measure progress towards them. Women appear to be facing additional constraints in comparison with male jobseekers, such that gender-disaggregated targets, for women in particular, should be a priority. Upgrading the Employment Management Information System would allow monitoring the PES staff and would provide information to manage the training system and organize individualized training of employment counselors. ENABLING WOMEN TO TAKE ADVANTAGE OF EMPLOYMENT OPPORTUNITIES Creating more jobs and improving skill matches are vital, but it is also important to enable youth and women to compete more effectively for new employment opportunities and to offer better skills. This report highlighted the limited capacity of the formal private sector to generate new jobs in Kosovo. Lack of new jobs leads to high rates of inactivity and unemployment. Kosovo needs to increase labor force participation for both men and women. This report argues that relaxing key constraints related to the regulatory and busi­ ness environment, as well as upgrading the skills of the workforce and resolving skill mismatches are important. But even with more jobs and better skills, women face additional constraints. Developing systems of care for children and elderly can have many advantages. The analysis of childcare concludes that expanding formal care services would not only improve labor market opportunities for women, but would also enhance school readiness of children through better early childhood education. Given Kosovo’s poor performance in the recent PISA, and the wide recognition of early childhood education for longer-term human capital development, this is an important collateral advantage. Similarly, improvements in eldercare can improve the health of the elderly, and thus enable savings in the health care sector (World Bank, 2015b). Reforms should address the availability, affordability and quality of care. To improve the availability and use of childcare services, particularly in rural areas, priorities should include the expansion of publicly provided childcare centers, as well as implementation of public subsidies to private childcare provision and use, and investments in the quality of childcare through creating education and certification programs for caregivers. Address the disincentives embedded in the maternity leave policies. Improving the maternal leave poli­ cies can increase opportunities for women by reducing incentives that disadvantage women in hiring and career advancement. Long absences from jobs for childbirth, financed by employers, reduce skills and threaten job security and make employers reluctant to hire women. If these costs were shifted toward the state, it would improve women’s job prospects. There is no single solution to address the dimensions of the length, distribu­ tional of financial burden, and low benefit take-up, and a number of alternative options can be considered (see World Bank, 2015a for some possible alternatives). Policies aimed at increasing the take up of paternal leave should be part of the solutions to be considered. Greater availability of care leave and wide-spread flexible working arrangements are important for employment opportunities of young women. The discussion in Chapter 4 noted that flexible work arrangements are virtually non-existent, while parental care, unlike maternal leave provisions, is minimal. Offer­ ing effective care leave arrangements can increase protection against income shocks and, if properly designed, reduce the length of detachment from jobs. Improving the availability of flexible work arrangement enables care givers to reduce labor supply along the intensive instead of extensive margin, thus guarding against lengthy breaks that can lead to human capital depletion (World Bank, 2015a). 68 REFERENCES Abdih, Yasser, Leandro Medina. 2013. “Measuring the informal economy in the Caucasus and Central Asia.” IMF Working Paper No. 13/137. Ajwad, M. I., Vazquez, E., and Winkler, H. 2016. “European informality: A study of Kosovo’s shadow economy”, Background paper for the Kosovo Jobs Diagnostic (mimeo). Washington, DC: The World Bank. Aterido, R. and Petracco, C. 2016. “Kosovo jobs diagnostic: potential for more and better jobs”, Background paper for the Kosovo Jobs Diagnostic (mimeo). Washington, DC: The World Bank. Ayyagari, M., Demirguc-Kunt, A. and Maksimovic, V. 2011. “Small vs. Young Firms across the World Contribution to Employ- ment, Job Creation, and Growth”. Bartelsman, E., Haltiwanger, J., and Scarpetta, S., 2009. “Measuring and Analyzing Cross Country Differences in Firm Dynamics.” In Producer Dynamics: New Evidence from Micro Data, (Dunne, Jensen and Roberts, eds.) NBER/University of Chicago Press. Bartelsman, E., Haltiwanger, J., Scarpetta, S., 2013. “Cross-country differences in productivity: the role of allocation and selection”. Am. Econ. Rev. 103 (1), 305–334. Bernard, Andrew B. and J. Bradford Jensen. 1999. “Exceptional exporter performance: cause, effect, or both?” Journal of International Economics 47, 1–25. Bernard, Andrew B. and Joachim Wagner. 1997. “Exports and Success in German Manufacturing.” Weltwirtschaftliches Archiv/Review of World Economics 133, 134–157. Betcherman, G., et al., 2004, ‘Impacts of Active Labor Market Programs: New Evidence from Evaluations with Particular Attention to Developing and Transition Countries’, Social Protection Discussion Paper Series no. 0402, World Bank, Washington, D.C. Borensztein, Eduardo, José De Gregorio, and Jong-Wha Lee, 1998, “How Does Foreign Direct Investment Affect Growth?” Journal of International Economics, Vol. 45 (June), pp. 115–35. Card, David, Jochen Kluve, and Andrea Weber, What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations, NBER Working Paper No. 21431 July 2015 JEL No. J08,J24. Cunningham, W. (2001) “Breadwinner Versus Caregiver: Labor Force Participation and Sectoral Choice over the Mexican G. Katz and M.  Business Cycle.” In The Economics of Gender in Mexico: World, Family, State, and the Market, ed. E.  C. Correia, 85–132. Washington, DC: World Bank. Dabla-Norris, Era, Mark Gradstein, Gabriela Inchauste (2008) “What causes firms to hide output? The determinants of informality” Journal of Development Economics 85 (2008) 1–27. Dasgpta, Basab and Mohamed Ihsan Ajwad (2011) “Income Shocks Reduce Human Capital Investments: Evidence from Five East European Countries” World Bank Policy Research Working Paper 5926, Washington, DC. Davis, S. J., and J. Haltiwanger. 1992. “Gross Job Creation, Gross Job Destruction, and Employment Reallocation.” Quarterly Journal of Economics 107(3): 819–863. Davis, S. J., J. C. Haltiwanger and S. Schuh. 1999. “On the Driving Forces Behind Cyclical Movements in Employment and Job Reallocation.” American Economic Review 89(5):1234–1258. European Investment Bank. 2016. “Kosovo: Assessment of financing needs of SMEs in the Western Balkans countries.” Fialová, K., and O. Schneider. 2011. “Labor Institutions and Their Impact on Shadow Economies in Europe.” Background paper for “In from the Shadow: Integrating Europe’s Informal Labor.” Policy Research Working Paper 5913, World Bank, Washington, DC. Haltiwanger, John C. 2011. Firm dynamics and productivity growth, EIB Papers 16(1). Hazans, M. 2011. “Informal Workers across Europe: Evidence from 30 European Countries.” Background paper for “In from the Shadow: Integrating Europe’s Informal Labor.” Policy Research Working Paper 5912, World Bank, Washington, DC. International Labour Organization (ILO) (2002) Resolution and conclusions concerning decent work and the informal economy. International Labour Conference, 90th session (Geneva). Available online at http://www.ilo.org/public/english/ standards/relm/ilc/ilc90/pdf/pr-25res.pdf.OECD (2008). 2008 OECD Employment Outlook. Paris: OECD. 69 International Monetary Fund (IMF) (2016), IMF Country Report No. 16/123, Kosovo Technical Assistance Report—Enhancing Social Protection Cash Benefits Hsieh, Chang-Tai, and Benjamin Olken (2014) “The Missing” Missing Middle”. Journal of Economic Perspectives 28.3: 89–108. La Porta, Rafael, and Andrei Shleifer (2014). “Informality and Development.” The Journal of Economic Perspectives 109–126. Li, Y. and Rama, M. 2015. “Firm Dynamics, Productivity Growth, and Job Creation in Developing Countries: The Role of Micro and Small Enterprises”. Maloney, William (2004) “Informality Revisited” World Development, Vol. 32, No. 7. Ministry of Education, Science and Technology. 2015. “Evaluation report: Kosovo education strategic plan 2011–2016.” Government of Kosovo. Packard, Truman G., Johannes Koettl, and Claudio E. Montenegro (2012). In from the Shadow: Integrating Europe’s Informal Labor. Washington, DC: World Bank. Packard, Truman G., and Trang Van Nguyen (2014). “Work in the informal economy”. In East Asia Pacific at work: employment, enterprise, and well-being. Washington, DC: World Bank. Perry, Guillermo E., William F. Maloney, Omar S. Arias, Pablo Fajnzylber, Andrew D. Mason, and Jaime Saavedra-Chanduvi (2007). Informality: Exit and Exclusion. Washington, DC: World Bank. Riinvest (2013). “To Pay or Not to Pay: A Business Perspective of Informality in Kosovo.” Pristina: Riinvest Institute and FES. Sahovic, Tarik, draft Working Paper on Inspection Reform in Western Balkans, January 2008; updated 2014 Schneider, Friedrich, Andreas Buehn, and Claudio E. Montenegro (2010) “Shadow Economies all over the World: New Esti- mates for 162 Countries from 1999 to 2007.” World Bank Policy Research Working Paper Series. Schneider, Friedrich (2009) “The Shadow Economy in Europe. Using Payment Systems to Combat the Shadow Economy.” Commissioned by AT Kearney. Sleuwaegen, L., Goedhuys, M., 2002. Growth of firms in developing countries: evidence from Côte D’Ivoire. J. Dev. Econ. 68, 117–135. Seetharam Mukkavilli in June 2008 on UNDP interventions “Evaluation of Active Labour Market Programme for Youth in Kosovo.” Torgler, Benno, and Friedrich Schneider (2007). “Shadow Economy, Tax Morale, Governance and Institutional Quality: A Panel Analysis.” IZA Discussion Papers 2563. Bonn: Institute for the Study of Labor (IZA). Torgler, Benno and Friedrich Schneider (2009). “The impact of tax morale and institutional quality on the shadow economy.” Journal of Economic Psychology. Vol. 30(2): 228–245. World Bank. 2011. “The Jobs Crisis: Household and government responses to the great recession in Eastern Europe and Central Asia” Directions in Development, Washington, DC. World Bank. 2013. “Activation and Smart Safety Nets in Kosovo: Constraints in Beneficiary Profile, Benefit Design, and Institutional Capacity.” Social Protection and Labor Global Practice, World Bank. World Bank. 2014. “Activation and Smart Safety Nets in Kosovo: Constraints in Beneficiary Profile, Benefit Design, and Institutional Capacity.” Mimeo. World Bank. World Bank. 2015a. “Maternity leave and women’s labor market status in Kosovo: Five key messages.” Washington, DC: The World Bank. World Bank. 2015b. “Why should we care about care? The role of childcare and eldercare in Kosovo.” Washington, DC: The World Bank. World Bank. 2016. Systematic Country Diagnostic. Washington, DC: The World Bank. World Bank. 2017a. Western Balkans Systematic Country Diagnostic regional synthesis report. (forthcoming). Washington, DC: The World Bank. World Bank. 2017b. “Risk and Resilience Assessment: Republic of Kosovo.” Washington, DC: The World Bank. 70 ANNEX A. 71 72 Table A-1 Employment determinants (1) (2) (3) (4) (5) (6) (7) (8) (9) age_6to9 0.0546*** 0.0729*** 0.0599*** 0.0771*** 0.0771*** 0.0723*** 0.112*** 0.0699*** 0.0700*** (0.00711) (0.00654) (0.00767) (0.00710) (0.00710) (0.00693) (0.0104) (0.00729) (0.00729) age_10to19 0.0543*** 0.0873*** 0.0636*** 0.0947*** 0.0947*** 0.0860*** 0.133*** 0.0751*** 0.0754*** (0.0124) (0.0116) (0.0138) (0.0129) (0.0129) (0.0124) (0.0149) (0.0139) (0.0138) Foreign 0.163** 0.163** 0.273*** 0.152* 0.294*** 0.295*** (0.0806) (0.0806) (0.0847) (0.0922) (0.105) (0.105) Ferizaj -0.193*** (0.0459) Gjakova -0.173*** (0.0445) Gjilan -0.0922** (0.0445) Mitrovica -0.186*** (0.0446) PejaPec -0.138*** (0.0426) Prizren -0 147*** (0.0432) (continued on next page) Table A-1 (continued) Employment determinants (1) (2) (3) (4) (5) (6) (7) (8) (9) AgricForestFish -0.209*** (0.0599) MiningQuarrying 0.322*** (0.0942) Utilities 0.606*** (0.189) Construction 0.117* (0.0671) WholesaleRetail -0.305*** (0.0425) TransportStorageComm 0.0217 (0.0653) HotelsRestaurants -0.0962** (0.0432) BusinessFinance 0.474** (0.223) OtherServices -0.212*** (0.0595) (continued on next page) 73 74 Table A-1 (continued) Employment determinants (1) (2) (3) (4) (5) (6) (7) (8) (9) Exporter 0.254*** (0.0193) Sales per worker (log) -0.0180*** -0.0131* (0.00479) (0.00788) Sales per worker squared -0.000938 (0.00142) Constant 0.486*** 0.424*** 0.601*** 0.541*** 0.541*** 0.764*** 0.731*** 0.585*** 0.581*** (0.0198) (0.0478) (0.0519) (0.0564) (0.0564) (0.0500) (0.0671) (0.0574) (0.0583) Observations 157,786 157,786 157,786 157,786 157,786 157,786 77,128 142,102 142,102 R-squared Sector dummies NO YES NO YES YES NO YES YES YES Location dummies NO NO YES YES NO YES YES YES YES Year Dummies YES YES YES YES YES YES YES YES YES Number of id 40,473 40,473 40,473 40,473 40,473 36,637 30,654 36,790 36,790 Between R2 0.000663 0.0903 0.00921 0.100 0.100 0.0528 0.130 0.111 0.111 Overall R2 0.00722 0.111 0.0193 0.121 0.121 0.0665 0.146 0.131 0.131 Within R2 0.0798 0.0788 0.0797 0.0787 0.0787 0.0810 0.0146 0.101 0.101 With respect to micro (1–9), young (1–5) and domestic. Robust standard errors in parentheses. Significance: ***p<0.01, ** p<0.05, * p<0.1 With respect to Pristina (5) and Manufacturing (6) Random Effects. Clustered at the year-region-sector5-size3 level Table A-2 Determinants of productivity (log) (1) (2) (3) (4) (6) (7) sz_10to19 0.114*** 0.114*** 0.126*** 0.180*** 0.110*** 0.111*** (0.0178) (0.0178) (0.0183) (0.0240) (0.0174) (0.0174) sz_20to49 0.132*** 0.132*** 0.156*** 0.233*** 0.131*** 0.133*** (0.0341) (0.0341) (0.0377) (0.0480) (0.0426) (0.0425) sz_50to99 0.0964** 0.0964** 0.122** 0.287*** 0.0893 0.0937 (0.0476) (0.0476) (0.0516) (0.0610) (0.0591) (0.0590) sz_100to499 -0.0329 -0.0329 0.00315 0.334*** -0.0417 -0.0327 (0.0737) (0.0737) (0.0754) (0.0947) (0.0824) (0.0824) sz_500plus 0.432*** 0.432*** 0.482*** 0.361*** 0.416*** 0.437*** (0.137) (0.137) (0.138) (0.125) (0.139) (0.138) age_6to9 -0.0159 -0.0159 -0.0303*** 0.0522*** -0.0144 -0.0148 (0.00974) (0.00974) (0.00993) (0.0110) (0.0126) (0.0125) age_10to19 -0.150*** -0.150*** -0.175*** -0.00561 -0.146*** -0 147*** (0.0178) (0.0178) (0.0179) (0.0165) (0.0267) (0.0265) Foreign 0.818*** 0.818*** 1.118*** 0.795*** 1.073*** 1.083*** (0.0959) (0.0959) (0.109) (0.107) (0.116) (0.116) Ferizaj -0.306*** (0.0317) Gjakova -0.275*** (0.0412) Gjilan -0.518*** (0.0354) Mitrovica -0.198*** (0.0377) PejaPec -0.150*** (0.0421) Prizren -0.0924*** (0.0319) Agriculture 0.158** (0.0661) MinUtilConstr 0.245*** (0.0548) Commerce 0.401*** (0.0324) Services -0.410*** (0.0372) (continued on next page) 75 Table A-2 (continued) Determinants of productivity (log) (1) (2) (3) (4) (6) (7) Exporter 0.335*** (0.0169) Hefindahl_L -0.0229 (0.0510) Herfindahl_S -0 174*** (0.0376) Constant 2.326*** 2.326*** 2.199*** 2.822*** 2.300*** 2.315*** (0.0679) (0.0679) (0.0418) (0.0782) (0.0763) (0.0761) Observations 143,325 143,325 143,325 68,356 143,325 143,325 Number of id 36,858 36,858 33,588 27,664 36,858 36,858 Sector dummies YES YES NO YES NO NO Location dummies YES NO YES YES YES YES Year Dummies YES YES YES YES YES YES Between R2 0.214 0.214 0.128 0.206 0.0493 0.0488 Overall R2 0.211 0.211 0.125 0.211 0.0449 0.0447 Within R2 0.0478 0.0478 0.0524 0.00286 0.0479 0.0485 Random Effects. Robust standard errors in parentheses; clustered at the year-region-sector5-size3 level *** p<0.01, ** p<0.05, * p<0.1 76 Table A-3 Employment growth (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) xvar = xvar = xvar = xvar = VARIABLES Basic Region Sector exporter exports_h herfindahl_L herfindahl_S InLPQ InLPQ InLPQq sza_10to19 0.0335*** 0.0335*** 0.0354*** 0.0234** 0.0284*** 0.0349*** 0.0303*** -0.00646 -0.00656 -0.000197 (0.00769) (0.00769) (0.00782) (0.0103) (0.0101) (0.00786) (0.00801) (0.00793) (0.00794) (0.00769) sza_20to49 0.0216 0.0216 0.0261* 0.0164 0.0258 0.0255 0.0214 -0.0223 -0.0225 -0.0128 (0.0157) (0.0157) (0.0158) (0.0201) (0.0199) (0.0158) (0.0153) (0.0158) (0.0158) (0.0154) sza_50to249 0.0114 0.0114 0.0155 0.00145 0.0139 0.0142 0.0138 -0.0190 -0.0192 -0.0134 (0.0129) (0.0129) (0.0127) (0.0169) (0.0167) (0.0126) (0.0144) (0.0144) (0.0145) (0.0142) sza_250to499 0.0180 0.0180 0.0255 0.0299 0.0373 0.0232 0.0283 0.000525 0.000496 0.00606 (0.0425) (0.0425) (0.0410) (0.0569) (0.0574) (0.0410) (0.0317) (0.0347) (0.0347) (0.0352) sza_500plus -0.0188 -0.0188 -0.00419 -0.0366 -0.0194 -0.00850 -0.0308 -0.0572 -0.0576 -0.0867* (0.0402) (0.0402) (0.0386) (0.0479) (0.0473) (0.0389) (0.0406) (0.0451) (0.0451) (0.0465) age_6to9 -0.0331*** -0.0331*** -0.0331*** -0.0394*** -0.0394*** -0.0323*** -0.0365*** -0.0399*** -0.0398*** -0.0380*** (0.00434) (0.00434) (0.00433) (0.00655) (0.00656) (0.00429) (0.00442) (0.00419) (0.00418) (0.00413) age_10to19 -0.0495*** -0.0495*** -0.0501*** -0.0544*** -0.0543*** -0.0487*** -0.0572*** -0.0532*** -0.0531*** -0.0509*** (0.00615) (0.00615) (0.00610) (0.00725) (0.00726) (0.00602) (0.00660) (0.00651) (0.00649) (0.00653) Foreign -0.0710*** -0.0710*** -0.0585** -0.0885*** -0.0879*** -0.0609*** -0.0380 -0.136*** -0.136*** -0.126*** (0.0234) (0.0234) (0.0235) (0.0256) (0.0258) (0.0234) (0.0317) (0.0307) (0.0307) (0.0308) Ferizaj -0.00771 (0.00614) Gjakova -0.00364 (0.00586) Gjilan -0.00928 (0.00572) (continued on next page) 77 78 Table A-3 (continued) Employment growth (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) xvar = xvar = xvar = xvar = VARIABLES Basic Region Sector exporter exports_h herfindahl_L herfindahl_S InLPQ InLPQ InLPQq Mitrovica -0.00234 (0.00788) PejaPec 0.00172 (0.00610) Prizren -0.00395 (0.00535) Agriculture -0.0207 (0.0157) MinUtilConstr -0.0101 (0.0110) Commerce -0.00195 (0.00445) Services -0.00779 (0.00555) xvar 0.0495*** 0.0491** 0.00484 0.0108 (0.00696) (0.0233) (0.0132) (0.0119) IInLPQ 0.0581*** 0.0546*** (0.00183) (0.00421) IInLPQsq 0.000631 (0.000688) (continued on next page) Table A-3 (continued) Employment growth (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) xvar = xvar = xvar = xvar = VARIABLES Basic Region Sector exporter exports_h herfindahl_L herfindahl_S InLPQ InLPQ InLPQq IInLPQq2 0.0641*** (0.00287) IInLPQq3 0.157*** (0.00459) Constant 0.0356** 0.0356** 0.0570*** 0.0557** 0.0580** 0.0519*** 0.0658*** -0.141*** -0.137*** -0.0413** (0.0179) (0.0179) (0.00780) (0.0232) (0.0233) (0.00694) (0.00750) (0.0199) (0.0205) (0.0189) Observations 117,204 117,204 117,204 62,208 62,208 117,204 108,727 108,623 108,623 108,276 Number of id 30,557 30,557 27,691 24,610 24,610 30,557 28,661 28,868 28,868 28,786 Sector dummies YES YES NO YES YES NO NO YES YES YES Location dummies YES NO YES YES YES YES YES YES YES YES Year Dummies YES YES YES YES YES YES YES YES YES YES Between R2 0.00710 0.00710 0.00260 0.0202 0.0190 0.00247 0.00574 0.0216 0.0215 0.0228 Overall R2 0.00757 0.00757 0.00603 0.00984 0.00912 0.00602 0.00653 0.0301 0.0301 0.0294 Within R2 0.00715 0.00715 0.00714 0.00307 0.00266 0.00701 0.00655 0.0509 0.0512 0.0376 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 79 Table A-4 Probability of exit31 (1) (2) (3) (4) (5) (6) (7) Exporter- Exporter- VARIABLES Basic Region Sector lnLPQ Exporter Sector Region sz_10to19 -0.0495*** -0.0424*** -0.0426*** -0.0253*** -0.0551*** -0.0551*** -0.0550*** (0.00194) (0.00258) (0.00258) (0.00205) (0.00291) (0.00291) (0.00291) sz_20to49 -0.0500*** -0.0430*** -0.0433*** -0.0263*** -0.0585*** -0.0587*** -0.0585*** (0.00220) (0.00314) (0.00317) (0.00247) (0.00252) (0.00244) (0.00250) sz_50to249 -0.0522*** -0.0449*** -0.0447*** -0.0278*** -0.0539*** -0.0544*** -0.0537*** (0.00377) (0.00407) (0.00421) (0.00396) (0.00536) (0.00514) (0.00539) sz_250to499 -0.0500*** -0.0427*** -0.0402*** -0.0293*** -0.0579*** -0.0582*** -0.0580*** (0.00521) (0.00504) (0.00639) (0.00537) (0.00472) (0.00458) (0.00468) sz_500plus -0.0615*** -0.0527*** -0.0531*** -0.0334*** -0.0551*** -0.0563*** -0.0553*** (0.00557) (0.00527) (0.00546) (0.0104) (0.0163) (0.0149) (0.0158) age_6to9 -0.0293*** -0.0181*** -0.0195*** -0.0168*** (0.00160) (0.00162) (0.00167) (0.00149) age_10to19 -0.0411*** -0.0373*** -0.0387*** -0.0255*** (0.00158) (0.00222) (0.00218) (0.00127) Foreign 0.0199* 0.0211** 0.0179* 0.0600*** 0.0463** 0.0455** 0.0464** (0.0108) (0.00995) (0.0101) (0.0176) (0.0201) (0.0200) (0.0201) Ferizaj -0.00597 (0.00673) Gjakova -0.0123** (0.00627) Gjilan -0.00336 (0.00696) Mitrovica -0.00835 (0.00636) PejaPec -0.00696 (0.00660) Prizren -0.0140** (0.00624) Agriculture 0.0516*** -0.0368* (0.0129) (0.0199) MinUtilConstr 0.0274*** 0.0457 (0.00831) (0.0684) (continued on next page) Probit (reporting marginal effects) 31 80 Table A-4 (continued) Probability of exit31 (1) (2) (3) (4) (5) (6) (7) Exporter- Exporter- VARIABLES Basic Region Sector lnLPQ Exporter Sector Region Commerce 0.00207 -0.0782** (0.00530) (0.0363) Services 0.0261*** -0.0772 (0.00769) (0.0487) Sales per worker-log -0.0255*** (0.000807) exporter -0.0335*** -0.0134 -0.0360*** (0.00477) (0.00925) (0.00713) exporter_Agriculture -0.0179 (0.0272) exporter_MinUtilConstr -0.0136 (0.0165) exporter_Commerce -0.0392*** (0.00970) exporter_Services -0.0272** (0.0113) exporter_Ferizaj 0.0115 (0.0208) exporter_Gjakova 0.0196 (0.0274) exporter_Gjilan -0.0282 (0.0190) exporter_Mitrovica 0.0283 (0.0413) exporter_PejaPec 0.0161 (0.0348) exporter_Prizren 0.00897 (0.0251) Observations 137,562 159,040 159,065 124,869 56,327 56,327 56,327 Sector dummies YES YES NO YES YES YES YES Location dummies YES NO YES YES YES YES YES Year Dummies NO NO NO NO NO NO NO Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 81 Table A-5 Determinants of wages (log) (1) (2) (3) (4) (5) sz_10to19 0.178*** 0.178*** 0.188*** 0.136*** 0.205*** (0. 00867) (0.00867) (0.00886) (0.00774) (0.0104) sz_20to49 0.279*** 0.279*** 0.297*** 0.231*** 0.330*** (0.0212) (0.0212) (0.0229) (0.0162) (0.0239) sz_50to99 0.370*** 0.370*** 0.398*** 0.336*** 0.404*** (0.0282) (0.0282) (0.0293) (0.0226) (0.0374) sz_100to499 0.340*** 0.340*** 0.389*** 0.397*** 0.482*** (0.0526) (0.0526) (0.0543) (0.0376) (0.0588) sz_500plus 0.332*** 0.332*** 0.417*** 0.438*** 0.525*** (0.0937) (0.0937) (0.0939) (0.0565) (0.0950) age_6to9 0.0256*** 0.0256*** 0.0204** 0.0179*** 0.104*** (0.00833) (0.00833) (0.00838) (0.00640) (0.00704) age_10to19 -0.0543*** -0.0543*** -0.0646*** -0.0358*** 0.100*** (0.0160) (0.0160) (0.0162) (0.0123) (0.0105) Foreign 0.788*** 0.788*** 0.932*** 0.508*** 0.793*** (0.0737) (0.0737) (0.0886) (0.0712) (0.100) Ferizaj -0.201*** (0.0202) Gjakova -0.189*** (0.0199) Gjilan -0.231*** (0.0223) Mitrovica -0.169*** (0.0205) PejaPec -0.0863*** (0.0203) Prizren -0.114*** (0.0197) Agriculture -0.129*** (0.0417) MinUtilConstr 0.0574** (0.0238) Commerce -0.0293 (0.0180) Services -0.0744*** (0.0224) (continued on next page) 82 Table A-5 (continued) Determinants of wages (log) (1) (2) (3) (4) (5) Sales per worker-log 0.236*** (0.00497) Exporter 0.152*** (0.0113) Constant 0.0934** 0.0934** 0.212*** -0.545*** 0.568*** (0.0448) (0.0448) (0.0252) (0.0452) (0.0395) Observations 145,952 145,952 145,952 131,878 76,418 R-squared Sector dummies YES YES NO YES YES Location dummies YES NO YES YES YES Year Dummies YES YES YES YES YES Number of id 39,678 39,678 35,909 36,251 30,357 Between R2 0.159 0.159 0.131 0.354 0.169 Overall R2 0.166 0.166 0.136 0.360 0.179 Within R2 0.0969 0.0969 0.0947 0.312 0.0321 With respect to micro, young, domestic, Pristina and Manufacturing Random Effects. Robust standard errors in parentheses; clustered at the year-region-sector5-size3 level *** p<0.01, ** p<0.05, *p<0.1 83 84 Table A-6 Wage growth (1) (2) (3) (4) (5) (6) (7) (8) xvar= xvar= xvar= xvar= VARIABLES Basic Regions exporter exports herfindahl_L herfindahl_rS lnLPQ lnLPQq sza_10to19 -0.00671 -0.00671 -0.0225** -0.0213** -0.00667 -0.0147** -0.0444*** -0.0367*** (0.00754) (0.00754) (0.00904) (0.00899) (0.00714) (0.00716) (0.00738) (0.00761) sza_20to49 -0.00284 -0.00284 -0.0259** -0.0236** -0.00197 -0.0115 -0.0478*** -0.0370*** (0.0104) (0.0104) (0.0102) (0.0101) (0.00995) (0.0103) (0.0104) (0.0104) sza_50to249 -0.0365** -0.0365** -0.0435** -0.0404** -0.0331** -0.0513*** -0.0825*** -0.0681*** (0.0173) (0.0173) (0.0207) (0.0206) (0.0159) (0.0167) (0.0171) (0.0171) sza_250to499 -0.0168 -0.0168 -0.0472 -0.0446 -0.00746 -0.0250 -0.0510* -0.0358 (0.0435) (0.0435) (0.0534) (0.0534) (0.0442) (0.0294) (0.0293) (0.0299) sza_500plus -0.0431 -0.0431 -0.0596 -0.0549 -0.0163 -0.0118 -0.0742 -0.0565 (0.0592) (0.0592) (0.0541) (0.0540) (0.0584) (0.0510) (0.0509) (0.0535) age_6to9 -0.143*** -0.143*** -0.137*** -0.137*** -0.143*** -0.147*** -0.145*** -0.146*** (0.00439) (0.00439) (0.00627) (0.00628) (0.00442) (0.00446) (0.00440) (0.00435) age_10to19 -0.124*** -0.124*** -0.128*** -0.128*** -0.126*** -0.134*** -0.125*** -0.128*** (0.00684) (0.00684) (0.00730) (0.00733) (0.00700) (0.00735) (0.00691) (0.00685) Foreign -0.0181 -0.0181 -0.0547 -0.0544 -0.00617 0.0504 -0.0451 -0.0284 (0.0401) (0.0401) (0.0403) (0.0404) (0.0395) (0.0459) (0.0478) (0.0466) Ferizaj 0.0165 (0.0145) Gjakova -0.00333 (0.0175) Gjilan -0.0177 (0.0137) Mitrovica -0.0143 (0.0168) (continued on next page) Table A-6 (continued) Wage growth (1) (2) (3) (4) (5) (6) (7) (8) xvar= xvar= xvar= xvar= VARIABLES Basic Regions exporter exports herfindahl_L herfindahl_rS lnLPQ lnLPQq PejaPec -0.0176 (0.0153) Prizren 0.00607 (0.0169) xvar 0.0135 0.0157 0.0135 0.0317 (0.00876) (0.0319) (0.0267) (0.0207) Sales per worker-log 0.0675*** (0.00369) lnLPQq2 0.113*** (0.00661) lnLPQq3 0.164*** (0.00951) Constant 0.0952*** 0.0952*** 0.0957*** 0.0962*** 0.101*** 0.128*** -0.0780** 0.0252 (0.0344) (0.0344) (0.0298) (0.0298) (0.0124) (0.0118) (0.0383) (0.0372) Observations 97,539 97,539 61,548 61,548 97,539 90,330 88,797 88,582 Number of id 28,615 28,615 24,426 24,426 28,615 26,908 26,607 26,548 Sector dummies YES YES YES YES NO NO YES YES Location dummies YES NO YES YES YES YES YES YES Year Dummies YES YES YES YES YES YES YES YES Between R2 0.000400 0.000400 0.000365 0.000372 1.64e-06 0.00623 0.0168 0.0202 Overall R2 0.0287 0.0287 0.0349 0.0349 0.0279 0.0345 0.0412 0.0439 Within R2 0.0475 0.0475 0.0878 0.0878 0.0473 0.0449 0.0678 0.0585 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 85 Table A-7 Mincer wage regressions Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Female -0.112*** -0.103*** -0.122*** -0.088*** -0.120*** -0.114*** -0.119*** (-6.45) (-5.89) (-6.94) (-4.71) (-6.81) (-6.63) (-7.00) Individual age 0.036*** 0.037*** 0.035*** 0.035*** 0.035*** 0.035*** 0.032*** (5.96) (6.52) (6.52) (6.05) (5.63) (5.72) (5.43) Age Squared -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** (-5.55) (-6.00) (-6.10) (-5.68) (-5.45) (-5.35) (-5.17) Primary -0.042 -0.209 -0.118 -0.195 -0.026 -0.027 -0.112 (-0.30) (-1.26) (-1.34) (-1.26) (-0.17) (-0.19) (-0.62) Secondary 0.134 -0.019 0.022 -0.004 0.140 0.134 0.025 (0.95) (-0.11) (0.27) (-0.03) (0.92) (0.96) (0.14) Post-secondary 0.410*** 0.271* 0.216** 0.281* 0.395*** 0.413*** 0.273 (2.90) (1.65) (2.55) (1.86) (2.59) (2.94) (1.51) Rural -0.027 -0.030* -0.021 -0.032* -0.035* -0.028 -0.024 (-1.50) (-1.69) (-1.25) (-1.80) (-1.93) (-1.57) (-1.38) Prizren -0.027 -0.026 -0.043* -0.014 -0.031 -0.070*** -0.001 (-1.12) (-1.04) (-1.83) (-0.58) (-1.28) (-2.68) (-0.03) Gjakova -0.029 -0.035 -0.033 -0.029 -0.033 -0.027 -0.038* (-1.27) (-1.49) (-1.49) (-1.21) (-1.46) (-1.20) (-1.77) Mitrovica -0.079** -0.072** -0.080** -0.087*** -0.091*** -0.105*** -0.060* (-2.33) (-2.21) (-2.58) (-2.58) (-2.60) (-3.11) (-1.82) Ferizaj -0.019 -0.018 -0.037 -0.005 -0.018 -0.027 -0.026 (-0.54) (-0.51) (-1.08) (-0.14) (-0.51) (-0.78) (-0.78) Peja 0.006 0.012 0.044 0.016 0.003 0.012 0.050* (0.19) (0.41) (1.64) (0.57) (0.11) (0.43) (1.75) Gjilan 0.006 0.003 -0.015 0.004 0.003 0.004 -0.022 (0.21) (0.11) (-0.61) (0.14) (0.11) (0.15) (-0.88) industry 0.619*** 0.709** (2.81) (2.44) services 0.525** 0.587** (2.41) (2.05) Senior officials 0.029 (0.12) Professionals -0.091 (-0.38) Technicians -0.034 (-0.14) Clerks -0.104 (-0.42) (continued on next page) Table A-7 (continued) Mincer wage regressions Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Service and market sales workers -0.258 (-1.07) Craft workers -0.184 (-0.76) Machine operators -0.151 (-0.63) Elementary occupations -0.343 (-1.41) Armed forces -0.187 (-0.77) Mining 0.873*** (3.79) Manufacturing 0.561** (2.55) Public utilities 0.581*** (2.65) Construction 0.644*** (2.92) Commerce 0.468** (2.12) Transport and Comnunications 0.660*** (2.99) Financial and Business Services 0.505** (2.30) Public Administration 0.547** (2.51) Other Services Unspecified 0.509** (2.35) Private -0.067*** (-2.68) informal -0.179*** (-5.39) Without Contract -0.275*** (-9.65) Constant 4.826*** 4.414*** 4.554*** 4.458*** 4.923*** 5.036*** 5.092*** (26.26) (17.60) (26.63) (18.23) (25.00) (27.32) (24.27) Observations 2243 2243 2243 2243 2243 2243 2243 ll -819.834 -787.998 -693.780 -757.360 -812.760 -797.669 -748.143 t statistics in parentheses = “* p<0.1 ** p<0.05 *** p<0.01” 88 Table A-8 Poverty regressions, 2015 Dep. var: Poor = 1 (I) (II) (III) (IV) (V) (VI) (VII) (VIII) Employed (baseline) Unemployed 0.121*** 0.120*** 0.115*** 0.117*** 0.092*** 0.089*** 0.110*** (0.011) (0.011) (0.011) (0.011) (0.011) (0.011) (0.012) Inactive 0.058*** 0.056*** 0.053*** 0.060*** 0.014 -0.001 0.024* (0.009) (0.009) (0.009) (0.009) (0.010) (0.012) (0.013) Urban (baseline) Rural 0.026*** 0.029*** 0.024*** 0.01 0.01 0.007 0.007 (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) (0.008) Gjakove (baseline) Gjilan 0.081*** 0.082*** 0.088*** 0.088*** 0.089*** 0.089*** (0.014) (0.013) (0.013) (0.013) (0.013) (0.013) Mitrovice 0.118*** 0.119*** 0.131*** 0.132*** 0.130*** 0.132*** (0.014) (0.014) (0.014) (0.014) (0.014) (0.014) Peje 0.226*** 0.228*** 0.242*** 0.245*** 0.248*** 0.247*** (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) Prizren -0.008 -0.016 -0.014 -0.011 -0.01 -0.013 (0.010) (0.011) (0.010) (0.010) (0.010) (0.011) Pristina 0.079*** 0.077*** 0.091*** 0.092*** 0.093*** 0.095*** (0.012) (0.012) (0.012) (0.012) (0.012) (0.012) Ferizaj 0.270*** 0.266*** 0.278*** 0.280*** 0.282*** 0.281*** (0.016) (0.016) (0.016) (0.016) (0.016) (0.016) Share of children under age 15 0.186*** 0.154*** 0.175*** 0.179*** 0.178*** (0.025) (0.025) (0.027) (0.027) (0.027) Share of elderly aged 65+ -0.113*** -0.112*** -0.089*** -0.088*** -0.087*** (0.033) (0.033) (0.033) (0.033) (0.033) (continued on next page) Table A-8 (continued) Poverty regressions, 2015 Dep. var: Poor = 1 (I) (II) (III) (IV) (V) (VI) (VII) (VIII) Primary or less (baseline) Secondary -0.068*** -0.073*** -0.079*** -0.081*** (0.010) (0.010) (0.010) (0.010) Tertiary -0.146*** -0.142*** -0.143*** -0.145*** (0.013) (0.013) (0.013) (0.013) 15–24 0.053*** 0.044*** 0.048*** (0.013) (0.014) (0.014) 25–34 (baseline) 35–44 0.016 0.018 0.017 (0.013) (0.013) (0.013) 45–54 0.045*** 0.047*** 0.044*** (0.014) (0.014) (0.014) 55–64 -0.025* -0.028** -0.033** (0.014) (0.014) (0.014) Female -0.044*** (0.010) (continued on next page) 89 90 Table A-8 (continued) Poverty regressions, 2015 Dep. var: Poor = 1 (I) (II) (III) (IV) (V) (VI) (VII) (VIII) Employed # female = 0 (baseline) Employed # female = 1 -0.039*** (0.013) Unemployed # female = 0 0.133*** (0.017) Unemployed # female = 1 0.051*** (0.013) Inactive # female = 0 -0.000 (0.018) Inactive # female = 1 -0.011 (0.012) Constant 0.113*** 0.098*** 0.005 -0.021* 0.057*** 0.036** 0.047*** 0.049*** (0.006) (0.007) (0.010) (0.011) (0.014) (0.017) (0.017) (0.017) R-squared 0.016 0.017 0.074 0.083 0.095 0.1 0.102 0.104 Obs 8275 8275 8275 8275 8275 8275 8275 8275 Notes: OLS estimates. Taylor–linearized standard errors, clustered at PSU level in parentheses. Sample restricted to the 15–64 age cohort. Significance levels: * 0.1, ** 0.05, *** 0.01. Source: Staff estimates based on HBS 2015 data. Address: 17 76 G St NW, Washington, DC 20006 Website: ht tp://w w w.worldbank .org/en/topic/jobsanddevelopment Twit ter : @WBG_ Jobs Blog: ht tp://blogs.worldbank .org/jobs