Abstract
In the past 25 years, exports have contributed strongly to growth and economic convergence in many small open economies. However, the Western Balkan (WB) region, consisting of small emerging market economies, has not fully availed itself of this driver of growth and convergence. A lack of openness, reliance on low value products, and weak competitiveness largely explain the insignificant role of trade and exports in the region’s economic performance. This paper focuses on how the countries in the WB could lift exports through stronger integration with global value chains (GVCs) and broadening of services exports. The experience of countries that joined the European Union in or after 2004 shows that participation in GVCs can help small economies accelerate export and income growth. WB countries are not well integrated into Europe’s vibrant GVCs. Trade within the region is also limited—it tends to be bilateral and not cluster-like. Our analysis shows that by improving infrastructure and labor skills and adopting trade policies that ensure investor protection and harmonize regulations and legal provisions, the region can greatly enhance its engagement with GVCs. Services exports are an increasingly important part of global trade, and they offer an untapped source of growth. The magnitude of services exports from the WB region compares favorably with that of peers in Europe, particularly in travel services where several of these countries have a revealed comparative advantage. But there is significant room for growth in tourism exports and an untapped potential in business and information technology services exports that these countries can materialize through policy efforts that increase openness and enhance connectivity and labor skills. Serbia offers a good example of how decisive efforts, including education policies to ensure a sustained supply of skilled labor, can help information technology services exports to take off.
Chapter 1 Motivation
In the past 25 years, exports have strongly contributed to growth and economic convergence in small open economies (Figure 1). During this period, in real terms, the annual average growth of exports, at 5.6 percent, was about 1½ times the rate of world GDP growth. The gains from exports have been significant in the emerging market economies in Europe, particularly those that joined the European Union (EU) in 2004 or later—the new member states (NMS).1 Much of the strong income convergence witnessed in Europe prior to the global financial crisis came from these countries, aided by large intra-Europe investment and trade flows (IMF 2011). Despite some slowdown in the years following the global crisis, exports remain a robust avenue for growth for developing countries.
The Western Balkan (WB) region, consisting of small emerging market economies, is aspiring to embark on an export-led growth and convergence path.2 After a late start in transition due to civil wars, the region experienced notable income growth nearly doubling its real GDP level in the last two decades (Figure 1). However, income convergence with advanced Europe has been less impressive when compared to NMS, partly reflecting lower exports. Although the contribution of net exports to growth turned positive in most countries in the post-crisis years, the region is yet to experience a meaningful boost from exports. A lack of openness, reliance on low value products, and weak structural competitiveness largely explain the subdued role of exports in the region. WB countries are less open than the NMS (Figure 2). Our estimates of trade openness—which control for price level and size bias—show that the WB countries are only about one-third as open as the NMS.3 Although trade openness in the WB region has increased over time, this has largely been driven by Serbia and Republic of North Macedonia. Labor-intensive products—which tend to generate low value—account for a higher share of merchandise exports in WB countries than in NMS. Meanwhile, a late start in transition and slow progress in key structural reforms have held the region back in competitiveness despite low wages relative to European peers.
Exports, Real GDP Growth and Income Convergence
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics; World Economic Outlook; WDI; and IMF staff calculations.Note: NMS-7 includes small economies that are members of the European Union: Bulgaria, Croatia, Estonia, Latvia, Lithuania, Slovakia and Slovenia.1 Small countries are those with population below 10 million.Exports, Real GDP Growth and Income Convergence
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics; World Economic Outlook; WDI; and IMF staff calculations.Note: NMS-7 includes small economies that are members of the European Union: Bulgaria, Croatia, Estonia, Latvia, Lithuania, Slovakia and Slovenia.1 Small countries are those with population below 10 million.Exports, Real GDP Growth and Income Convergence
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics; World Economic Outlook; WDI; and IMF staff calculations.Note: NMS-7 includes small economies that are members of the European Union: Bulgaria, Croatia, Estonia, Latvia, Lithuania, Slovakia and Slovenia.1 Small countries are those with population below 10 million.
Trade Openness and Labor Intensity of Exports
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: UN Comtrade; World Development Indicators (WB); and IMF staff calculations.Note: Correcting for price level and country size distortions, Trade openness = (X + M )/((GDPi(–0.2)) + (GDProw)(–0.2))(–5), where, X and M are real exports and import of goods and services; GDPi and GDProw are a country’s own GDP and rest of the world’s GDP, respectively (in constant PPP). See Tang, 2011 for coefficient estimates.*Data point is 2006 due to lack of available data for 2000.1 Include textiles and textile products, leather and leather products, and footwear.Trade Openness and Labor Intensity of Exports
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: UN Comtrade; World Development Indicators (WB); and IMF staff calculations.Note: Correcting for price level and country size distortions, Trade openness = (X + M )/((GDPi(–0.2)) + (GDProw)(–0.2))(–5), where, X and M are real exports and import of goods and services; GDPi and GDProw are a country’s own GDP and rest of the world’s GDP, respectively (in constant PPP). See Tang, 2011 for coefficient estimates.*Data point is 2006 due to lack of available data for 2000.1 Include textiles and textile products, leather and leather products, and footwear.Trade Openness and Labor Intensity of Exports
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: UN Comtrade; World Development Indicators (WB); and IMF staff calculations.Note: Correcting for price level and country size distortions, Trade openness = (X + M )/((GDPi(–0.2)) + (GDProw)(–0.2))(–5), where, X and M are real exports and import of goods and services; GDPi and GDProw are a country’s own GDP and rest of the world’s GDP, respectively (in constant PPP). See Tang, 2011 for coefficient estimates.*Data point is 2006 due to lack of available data for 2000.1 Include textiles and textile products, leather and leather products, and footwear.This paper focuses on how exports and trade can play a stronger role in WB countries by drawing on the experience of the NMS. To make comparisons relevant, we use a subset of NMS, consisting of smaller economies. This group, which we refer to as NMS-7, excludes Czech Republic, Hungary, Poland, and Romania. With WB countries’ average per capita income at half of that in NMS, and a quarter of the EU-15 level (on purchasing power parity [PPP] basis), there is significant room for income convergence. However, to attain this goal sustainably, the WB countries will need to move away from their domestic demand–driven growth model and rely more on exports. In this paper, we explore the following three avenues on how this can be achieved.
Enhancing integration with global value chains. For countries in the WB, linking up with global value chains (GVCs)—which account for nearly 80 percent of global trade— could be greatly beneficial for promoting exports. In western Europe—the main destination for WB exports today—GVCs are predominantly regional, located around a few hub countries, most notably Germany, and concentrated in a few sectors, namely automotive, electrical equipment, machinery, chemicals, and metals (Figure 3). Although GVCs are not the only means to raise exports, they have proven to be an efficient way for smaller economies, particularly in Europe, to integrate into global trade without having to build expertise in all aspects of production of a good. We explore the following questions in Chapter 2. How are WB countries linked to the European supply chains? What can these countries do to enhance their links? Would enhancing labor skills, improving trade logistics, and strengthening infrastructure sufficiently compensate for the diseconomies arising from their small size?
Expanding services exports. The services sector dominates the economies of the WB — generating 66–79 percent of gross value added and 60–82 percent of employment. However, its role in exports is limited, with the exceptions of Albania and Montenegro where tourism accounts for a large share of exports. In Chapter 3 we explore the types of services in which WB countries have a comparative advantage and how this can be enhanced. Is the tourism industry competitive relative to regional peers? What are the prospects for boosting exports in newer services products, such as IT and business services?
Raising imports’ contribution to export growth. Structurally, imports in WB countries are skewed toward consumer goods, reflecting their domestic demand–driven growth model. The share of intermediate and capital goods—which are deemed beneficial to investment and long-term growth—is substantially lower than in the NMS. With GVCs dominating the global trade scene, imports are increasingly viewed as a means for firms to access most efficient inputs. In Chapter 4 we approach competitiveness through import composition by asking how imports can play a more enabling role for exports.
GVC Participation Index: Western Balkans and New Member States
(Sum of forward and backward GVC linkages; percent of GDP)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; WEO; WDI; and IMF staff calculations.Note: The diamond indicates median. WB = ALB, BIH, MKD, NME, and SRB; NMS = CZE, EST, HUN, LTU, LVA, POL, SVK, and SVN.GVC Participation Index: Western Balkans and New Member States
(Sum of forward and backward GVC linkages; percent of GDP)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; WEO; WDI; and IMF staff calculations.Note: The diamond indicates median. WB = ALB, BIH, MKD, NME, and SRB; NMS = CZE, EST, HUN, LTU, LVA, POL, SVK, and SVN.GVC Participation Index: Western Balkans and New Member States
(Sum of forward and backward GVC linkages; percent of GDP)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; WEO; WDI; and IMF staff calculations.Note: The diamond indicates median. WB = ALB, BIH, MKD, NME, and SRB; NMS = CZE, EST, HUN, LTU, LVA, POL, SVK, and SVN.Chapter 2 Enhancing Integration with Global Value Chains
Participation in GVCs can help developing economies fast track export and income growth. Since the mid-1990s, production of tradable goods has become increasingly organized around GVCs, globally, as well as in Europe. Advances in information technology and transportation and falling trade barriers have allowed firms to unbundle production into tasks performed across international borders, thus maximizing gains from specific comparative advantage (Feenstra and Hanson 1997, Grossman and Rossi-Hansberg 2008). From an individual country perspective, participation in GVCs can enhance productivity and income through multiple channels: a fner cross-border division of labor, greater availability of input varieties, increased competition, and technology spillovers (Halpern, Koren, and Szeidl 2015). Indeed, recent studies conclude that developing economies can obtain significant growth and productivity gains from engaging in GVC-related exports (Constantinescu, Mattoo, and Ruta 2017; Raei, Ignatenko, and Mircheva 2019; Figure 4).
GVC Participation and Income
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics and IMF staff calculations.Note: GVC trade represents the sum of backward and forward linkages.1 Change in GVC intensity is defined as the GVC trade in 2013 relative to the US (1993 = 1 normalized); and income convergence (1993–2013) is the income per capita in 2013 relative to the US (1993 = 1 normalized).GVC Participation and Income
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics and IMF staff calculations.Note: GVC trade represents the sum of backward and forward linkages.1 Change in GVC intensity is defined as the GVC trade in 2013 relative to the US (1993 = 1 normalized); and income convergence (1993–2013) is the income per capita in 2013 relative to the US (1993 = 1 normalized).GVC Participation and Income
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics and IMF staff calculations.Note: GVC trade represents the sum of backward and forward linkages.1 Change in GVC intensity is defined as the GVC trade in 2013 relative to the US (1993 = 1 normalized); and income convergence (1993–2013) is the income per capita in 2013 relative to the US (1993 = 1 normalized).However, being able to successfully link with GVCs is not a given, nor are the potential gains that can materialize from such links. Empirical studies show that to establish or harness these links, countries need to improve human and physical capital, institutional quality, and trade logistics/connectivity (Rahman and Zhao 2013, World Bank 2017b)—all of which take effort, time, and resources. In addition, competition from other countries is typically strong, particularly at the lower end of production where competitive wages matter most. Unless countries pursue reforms and invest in education to upgrade skills and productivity, there is a risk that participation in GVCs may not help (Rodrik 2018). They may find themselves stuck in a low-wage, low-productivity trap possibly losing long run comparative advantage to automation or other countries with lower labor costs. In addition, GVC links are not without risks as they tend to accentuate spillovers from negative shocks.
Background
Most WB countries are not heavily integrated into GVCs, although there has been an increase over time. Following the literature, GVC participation is defined as the sum of backward linkages (foreign inputs that are embodied in a country’s exports) and forward linkages (a country’s exports of intermediate goods that are embodied in another country’s exports). The size of GVC links in WB countries—measured as a share of total exports— is comparable to that in NMS-7 with a notable improvement since 2000 (Figure 5). However, when measured as a share of GDP, WB countries are significantly less GVC-linked than NMS-7, thus reflecting the limited role of exports in these economies. Among WB countries, Serbia, Montenegro, and North Macedonia have experienced the largest increases in GVC links since 2000 (Figure 5).
GVC Participation: Western Balkan Countries and New Member States
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Cheng et al. (2015); Eurostat; World Development Indicators (The World Bank); WEO; and IMF staff calculations.Note: Low value manufacturing includes food, beverage, textiles, apparel, wood and paper products. High value manufacturing comprises the rest of manufacturing products. See Annex Table I.1 for a complete list.GVC Participation: Western Balkan Countries and New Member States
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Cheng et al. (2015); Eurostat; World Development Indicators (The World Bank); WEO; and IMF staff calculations.Note: Low value manufacturing includes food, beverage, textiles, apparel, wood and paper products. High value manufacturing comprises the rest of manufacturing products. See Annex Table I.1 for a complete list.GVC Participation: Western Balkan Countries and New Member States
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Cheng et al. (2015); Eurostat; World Development Indicators (The World Bank); WEO; and IMF staff calculations.Note: Low value manufacturing includes food, beverage, textiles, apparel, wood and paper products. High value manufacturing comprises the rest of manufacturing products. See Annex Table I.1 for a complete list.Compared to NMS-7, the GVC links of WB countries are more concentrated in services and low value manufacturing products (Figure 5). Given the relatively low level of development and sophistication of countries in the WB region, backward linkages are more common—these countries are mostly assembly centers in the GVCs in which they operate. For NMS-7, GVC links are also predominantly backward linkages but concentrated in high value manufacturing products, reflecting these countries’ well-established positions in machinery and automotive and chemical industries. However, it is important to note that although automotive and machinery may be considered high value as a product group, it is possible that the specific part of production taking place in a country may not be of high value.
Compared to NMS-7, WB countries are less linked with Germany, the most important GVC hub in Europe. The composition of products and partners matters in determining gains from GVC links (Arora and Vamvakidis 2004). The NMS-7 countries’ strong GVC ties with Germany—a competitive and dynamic export hub—likely contributed to generating sustained gains for these countries (Ward and Weber 2011). On the other hand, several WB countries’ strong GVC links with Italy—a country caught up in the euro area crisis experiencing slower exports growth than Germany—may explain the sluggish export growth in these countries. Two WB countries, Bosnia and Herzegovina and North Macedonia, also show strong links with NMS-7 countries, a possible indication of mature investors in NMS-7 outsourcing some of the low value activities (Figure 6).
GVC Links to European Hubs, WB and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (WB); WEO; and IMF staff calculations.GVC Links to European Hubs, WB and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (WB); WEO; and IMF staff calculations.GVC Links to European Hubs, WB and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (WB); WEO; and IMF staff calculations.Intraregional GVC links in the WB region are generally limited, except in Montenegro. The WB region’s potential to be viewed as a common economic space by European investors, who can bring significant capital and trade to the region, is underutilized largely because of how regional trade agreements were implemented. WB countries are members of the Central European Free Trade Agreement (CEFTA) on goods trade and investments, under which members enjoy zero tariff on trade. However, individual countries have imposed bilateral tariffs on each other covering all merchandise trade. In addition, other rigidities created by a lack of recognition of production standards and agreed customs procedures may be contributing to low levels of intraregional trade.
Empirical Results
Our empirical analysis aims to identify factors that are important in explaining GVC links. Following the literature, we control for standard gravity variables and focus on the relative importance of costs, skills, quality of infrastructure and institutions, and the presence and depth of preferential trade agreements (PTAs) in explaining GVC links. A novelty in our empirical approach is that we include a variable that captures the depth of trade agreements. Using a new database that documents the content and legal enforceability in 52 policy areas of PTAs around the world, we employ a dummy variable that captures the depth of PTA in addition to a dummy that captures the PTA itself (World Bank 2017b; Box 1).
We find that GVC links and good policies are positively associated. As expected, gravity factors have an overwhelming role in explaining GVC links in Europe and elsewhere (Annex Table II.2), However, good policies also matter. The quality of infrastructure and institutions, quality of education, and the presence of PTAs and their depth are positively associated with GVC links. Our findings—showing the importance of infrastructure, institutions, and human capital in trade and investments—are consistent with empirical studies covering the region (Rahman and others 2015, World Bank 2017b, Jirasavetakul and Rahman 2018). A large set of robustness tests provide assurance about the significance of the variables used in our analysis (Annex II). Nonetheless, it is important to stress that our results establish associations rather than causality.
Although the small size of the WB economies puts them at a disadvantage, the experience of NMS-7, which are also small economies, shows that good policies can help overcome gravity-related disadvantages. For example, for NMS-7, more than half of the explained variation in GVC links is from non-gravity variables, highlighting the role of better infrastructure, and the existence of deep PTAs that ensure institutional and regulatory harmonization (Figure 7). GVCs require cross border investments to allow multinational firms, often based in advanced economies, to break up their production chain along efficiency lines. For such firms to invest and produce in foreign countries, a certain level of institutional maturity, regulatory predictability, and contract enforceability is needed. “Deep” trade agreements are one way to provide that (Box 1).
Determinants of GVC Participation
(In percent of total)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (WB); WEO; and IMF staff calculations.Note: Contributions of each explanatory variable are calculated based on the estimated coefficients (as in column 11 of Annex Table II.3) and observed values as a share of total predicted value of the GVC linkages. The regional contributions are the average of the calculated country estimates. WB = ALB, BIH, MKD, NME, SRB; NMS-7 = BGR, EST, HRV, LTU, LVA, SVK, SVN.Determinants of GVC Participation
(In percent of total)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (WB); WEO; and IMF staff calculations.Note: Contributions of each explanatory variable are calculated based on the estimated coefficients (as in column 11 of Annex Table II.3) and observed values as a share of total predicted value of the GVC linkages. The regional contributions are the average of the calculated country estimates. WB = ALB, BIH, MKD, NME, SRB; NMS-7 = BGR, EST, HRV, LTU, LVA, SVK, SVN.Determinants of GVC Participation
(In percent of total)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (WB); WEO; and IMF staff calculations.Note: Contributions of each explanatory variable are calculated based on the estimated coefficients (as in column 11 of Annex Table II.3) and observed values as a share of total predicted value of the GVC linkages. The regional contributions are the average of the calculated country estimates. WB = ALB, BIH, MKD, NME, SRB; NMS-7 = BGR, EST, HRV, LTU, LVA, SVK, SVN.The WB countries have significantly fewer and shallower trade agreements compared to NMS-7 (Figure 8). The number of countries that have trade agreements with the WB economies (38, all European) is less than half the number for NMS-7 (83). The larger number of trade agreements that NMS-7 are signatories to reflects their membership in the EU which enables them to be part of all EU-level trade agreements. Not only are the agreements to which WB countries are signatory to fewer in number but they are not as deep—they have fewer legally enforceable provisions than those held by NMS-7.
How Deep are Western Balkans Trade Agreements?
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: World Trade Organization and IMF staff calculations.Note: For panel 4, dark green = provision covered and legally enforceable; light green = provision covered but not legally enforceable; red = provision not covered. The areas shown are a subset of 52 policy areas described in Box 2.1. These selected areas are considered to be more economically pertinent (Hofman and others, 2017).How Deep are Western Balkans Trade Agreements?
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: World Trade Organization and IMF staff calculations.Note: For panel 4, dark green = provision covered and legally enforceable; light green = provision covered but not legally enforceable; red = provision not covered. The areas shown are a subset of 52 policy areas described in Box 2.1. These selected areas are considered to be more economically pertinent (Hofman and others, 2017).How Deep are Western Balkans Trade Agreements?
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: World Trade Organization and IMF staff calculations.Note: For panel 4, dark green = provision covered and legally enforceable; light green = provision covered but not legally enforceable; red = provision not covered. The areas shown are a subset of 52 policy areas described in Box 2.1. These selected areas are considered to be more economically pertinent (Hofman and others, 2017).The WB region can potentially reap significant gains from targeted reforms. Although the WB countries’ performance in quality of education, infrastructure, and institutions is broadly in line with peers of similar per capita income, they lag that of NMS-7. To attract investment from European countries, they will however need to compete with other smaller countries in the region and, hence, raise the level of institutional readiness and labor skills. We try to estimate the gains in the trade/GDP ratio if WB countries were to achieve this. Using our estimation results (Annex Table II.2), we try to gauge potential gains in GVC links for WB countries, if these countries were to close the policy gaps relative to the top NMS-7 performer. The largest gains come from closing infrastructure and education quality gaps— where WB countries have, on average, 70 and 90 percent of the quality of top NMS-7, respectively. This is followed closely by deepening trade agreements (Figure 9). The gains from closing all policy gaps could imply an increase of about 6–12 percentage points in GVC trade flows. These gains in trade, using the typical range of elasticities found in the literature, would translate into a 3- to 10-percent gain in GDP level (Raei, Ignatenko, and Mircheva 2019).
Gains from Bridging the Policy Gaps
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (The World Bank); IMF WEO database; and IMF staff calculations.Note: The boxes represent 25–75 percentile range; the line and cross represent median and mean, respectively. Simulated gains are calculated by applying the estimated regression coefficients (column 22, Annex Table II.2) to the change in dependent variable (maximum value of variable in NMS-7 less actual value of variable in the WB country) where the actual observed value is less than the top value in NMS-7 (i.e. where there is scope for policy improvement).Gains from Bridging the Policy Gaps
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (The World Bank); IMF WEO database; and IMF staff calculations.Note: The boxes represent 25–75 percentile range; the line and cross represent median and mean, respectively. Simulated gains are calculated by applying the estimated regression coefficients (column 22, Annex Table II.2) to the change in dependent variable (maximum value of variable in NMS-7 less actual value of variable in the WB country) where the actual observed value is less than the top value in NMS-7 (i.e. where there is scope for policy improvement).Gains from Bridging the Policy Gaps
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Eurostat; World Development Indicators (The World Bank); IMF WEO database; and IMF staff calculations.Note: The boxes represent 25–75 percentile range; the line and cross represent median and mean, respectively. Simulated gains are calculated by applying the estimated regression coefficients (column 22, Annex Table II.2) to the change in dependent variable (maximum value of variable in NMS-7 less actual value of variable in the WB country) where the actual observed value is less than the top value in NMS-7 (i.e. where there is scope for policy improvement).Survey Findings
We conducted a survey of the largest export firms operating in the WB region to complement our empirical analysis of determinants of GVC links. The key aim of the survey was to examine the characteristics of companies, factors that affect participation in GVCs, and the key obstacles that companies encounter while engaging in such trade. Our survey respondents comprise 66 firms that operate in five WB countries (Albania, Bosnia and Herzegovina, North Macedonia, Montenegro, and Serbia),1 about half of which participate in GVCs. Considering the limited scope of GVCs in the region, the survey appears to capture a critical mass. The majority of respondents were established after these countries shifted to market economy and operate at the lower end of export supply chains, primarily in labor-intensive products. Members of their labor force hold a secondary education degree, on average (Figure 10). Since we cover the largest exporters, the survey has a representation and selection bias, thus its results should be interpreted with caution.
Summary of Findings from Survey of Export and GVC Companies in WB
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: Survey conducted by IMF resident representative office of export firms in Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia.1 Low value manufacturing comprises food and beverage, textiles and wearing apparel, wood and paper. High value manufacturing includes the rest. See Annex Table 1 for a complete list.Summary of Findings from Survey of Export and GVC Companies in WB
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: Survey conducted by IMF resident representative office of export firms in Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia.1 Low value manufacturing comprises food and beverage, textiles and wearing apparel, wood and paper. High value manufacturing includes the rest. See Annex Table 1 for a complete list.Summary of Findings from Survey of Export and GVC Companies in WB
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: Survey conducted by IMF resident representative office of export firms in Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia.1 Low value manufacturing comprises food and beverage, textiles and wearing apparel, wood and paper. High value manufacturing includes the rest. See Annex Table 1 for a complete list.Consistent with the empirical findings reported in this chapter and Chapter 4, respondent firms see skills shortage and low institutional quality as key constraints for exports and GVCs. GVC participating firms have diverse and developed backward linkages with the EU and Organisation for Economic and Co-operation and Development (OECD) countries (Figure 11). However, their forward linkages are limited, indicating that these firms mostly assemble final products. The major reason why these firms were attracted to the region are cheap labor, geographic proximity to investors, favorable taxation, and low tariffs (Figure 10; see Annex V for survey questionnaire). In contrast, skills shortage, political instability, poor trade logistics, and institutions are viewed as key obstacles. Since WB countries have slashed tariff rates over time to the levels comparable to NMS-7, respondents deem non– tariff barriers—such as unequal treatment at borders, time spent at customs, poor technical facilities for standard compliance,2 and a lack of preferential agreements—to be constraints to trade.
GVC Connectivity in the WB
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: Survey conducted by IMF resident representative office of export firms in Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia. For Serbia, survey respondents did not fill out this part.GVC Connectivity in the WB
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: Survey conducted by IMF resident representative office of export firms in Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia. For Serbia, survey respondents did not fill out this part.GVC Connectivity in the WB
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: Survey conducted by IMF resident representative office of export firms in Albania, Bosnia and Herzegovina, Macedonia, Montenegro, and Serbia. For Serbia, survey respondents did not fill out this part.The findings of our survey are also consistent with those of several other surveys conducted by international financial institutions, United Nations institutions, and investment banks.3 The sample size in other surveys varies from about 1,000 (European Investment Bank, UNESCAP) to 16,000 (EBRD) firms, with the issues covered ranging from setup logistics to sector-specific policies (Mekong Development Research Institute 2017). Despite differences in the sample size and coverage of issues and sectors between our and earlier surveys, several common themes emerge. GVCs are more likely to innovate than local firms and are an important source of knowledge and technology transfer to the local economy and exporters. Key factors for successful functioning of GVC firms are establishment of quality control institutions and standards, improvement in governance, and amelioration of labor skills. Specifically, the OECD and UNSCAP surveys identified that non–tariff barriers, including a lack of transparency in rules and regulations, an absence of preferential market access schemes, and a lack of technical facilities for standards compliance, are among the most important obstacles to operations of GVCs. Political stability is of lesser importance compared to these factors.
Deep Trade Agreements: Why and How They Matter
All World Trade Organization (WTO) members commit to tariff and non–tariff reforms, but there are significant differences in coverage and enforceability. The newer agreements provide greater coverage to dispute resolution and legal enforceability. Such deep agreements promote trade more effectively and confer benefits akin to good institutions.
All members of the WTO have signed at least one preferential trade agreement (PTA) through which they commit to cut tariffs and undertake additional obligations in policy areas covered by the WTO, such as customs administration or contingent protection. However, not all trade agreements are created equal. They differ in coverage and enforceability of provisions and have also evolved over time. Prior to the 1990s, PTAs focused on a narrow set of policy areas—mostly commitments to reduce tariffs on industrial and agricultural goods and remove other constraints that directly affected market access. The newer agreements tend to be deeper and have wider reach. They include trade dispute remedies (that is, countervailing measures and antidumping duties) and subsidies, and a broader set of more legally binding behind-the-border measures covering services trade, investment, intellectual property rights, and domestic regulations (Lawrence 1996).
Deep trade agreements promote trade but also benefit an economy in ways akin to good institutions. Production in complex GVCs is spatially spread out and tends to involve many cross-border exchanges among firms, each facing a risk of contract breach. In this context, rule of law, protection of property rights, and enforcement of contracts irrespective of residency status can help promote cross border investment and GVC trade. Deepening of trade agreements is one way to fast track the development of key institutions needed for investment and GVC trade (World Bank 2017b). Because deep integration often involves leveling the playing field for investment, intellectual property, and competition policy, participation in deep agreements is an effective way for countries to fast track key institutional reforms and expand GVC involvement. Also, as these measures are nondiscriminatory, they have positive externalities on the overall economy as well (Mattoo, Mulabdic, and Ruta 2017).
Policy Areas in Preferential Trade Agreements1
There may be some overlap in the scope between the two categories. For example, certain GATS obligations (WTO+) are closely linked to FDI and investment in services, whereas Investment is a WTO-X category.
Policy Areas in Preferential Trade Agreements1
| WTO1 | WTO-X | |
|---|---|---|
| Provisions falling under the current mandate of the WTO | Obligations that are outside the current mandate of the WTO | |
| Tariffs on Industrial Goods | Anti-Corruption | Health |
| Tariffs on Agriculture Goods | Competition Policy | Human Rights |
| Customs | Environmental Laws | Illegal Immigration |
| Export Taxes | Intelectual Property Rights (IPR) | Illicit Drugs |
| Sanitary and phytosanitary measures (SPS) | Investment | Industrial Cooperation |
| Technical Barriers to Trade (TBT) | Labour Market Regulation | Information Society |
| State Trading Enterprises (STE) | Movement of Capital | Mining |
| Antidumping (AD) | Consumer Protection | Money Laundering |
| Counterveiling Measures (CVM) | Data Protection | Nuclear Safety |
| State Aid | Agriculture | Political Dialogue |
| Public Procurement | Approximation of Legislation | Public Administration |
| Trade Related Investment Measures (TRIMs) | Audio Visual | Regional Cooperation |
| General Agreement on Trade in Services (GATS) | Civil Protection | Research and Technology |
| Agreement on Trade Related Aspects of Intelectual | Innovation Policies | SME |
| Property Rights (TRIPs) | Cultural Cooperation | Social Matters |
| Economic Policy Dialogue | Statistics | |
| Education and Training | Taxation | |
| Energy | Terrorism | |
| Financial Assistance | Visa and Asylum | |
There may be some overlap in the scope between the two categories. For example, certain GATS obligations (WTO+) are closely linked to FDI and investment in services, whereas Investment is a WTO-X category.
Policy Areas in Preferential Trade Agreements1
| WTO1 | WTO-X | |
|---|---|---|
| Provisions falling under the current mandate of the WTO | Obligations that are outside the current mandate of the WTO | |
| Tariffs on Industrial Goods | Anti-Corruption | Health |
| Tariffs on Agriculture Goods | Competition Policy | Human Rights |
| Customs | Environmental Laws | Illegal Immigration |
| Export Taxes | Intelectual Property Rights (IPR) | Illicit Drugs |
| Sanitary and phytosanitary measures (SPS) | Investment | Industrial Cooperation |
| Technical Barriers to Trade (TBT) | Labour Market Regulation | Information Society |
| State Trading Enterprises (STE) | Movement of Capital | Mining |
| Antidumping (AD) | Consumer Protection | Money Laundering |
| Counterveiling Measures (CVM) | Data Protection | Nuclear Safety |
| State Aid | Agriculture | Political Dialogue |
| Public Procurement | Approximation of Legislation | Public Administration |
| Trade Related Investment Measures (TRIMs) | Audio Visual | Regional Cooperation |
| General Agreement on Trade in Services (GATS) | Civil Protection | Research and Technology |
| Agreement on Trade Related Aspects of Intelectual | Innovation Policies | SME |
| Property Rights (TRIPs) | Cultural Cooperation | Social Matters |
| Economic Policy Dialogue | Statistics | |
| Education and Training | Taxation | |
| Energy | Terrorism | |
| Financial Assistance | Visa and Asylum | |
There may be some overlap in the scope between the two categories. For example, certain GATS obligations (WTO+) are closely linked to FDI and investment in services, whereas Investment is a WTO-X category.
Chapter 3 Expanding Services Exports
Services exports are an increasingly important part of global trade and they offer new sources of growth and stability (Figure 12). The share of services exports in world GDP has more than doubled in the last 25 years, and the sector accounted for nearly a quarter of global trade in 2016. The share of developing countries in global services exports has also grown, reaching 25 percent in 2016. Services trade is found to be less volatile than goods trade, and it is thus seen as an important source of stability during periods of macroeconomic and financial turbulence (Loungani and others 2017).
Exports of Services
(Percent of world GDP, LHS; percent of EU-28 GDP, RHS)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Papageorgiou et al 2017, WEO, and IMF staff calculations.Exports of Services
(Percent of world GDP, LHS; percent of EU-28 GDP, RHS)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Papageorgiou et al 2017, WEO, and IMF staff calculations.Exports of Services
(Percent of world GDP, LHS; percent of EU-28 GDP, RHS)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Papageorgiou et al 2017, WEO, and IMF staff calculations.Background
The overall size of services exports in the WB countries is comparable with that in NMS-7. It amounted to 15 percent of WB region’s GDP in 2016 compared to 17 percent in NMS-7. Apart from Bosnia and Herzegovina, all countries in the region have experienced an increase in services exports in recent years (Figure 13), primarily driven by tourism, which dominates services exports, particularly in Albania and Montenegro. The near tripling of travel-related exports in Kosovo since 2008 is notable, partly reflecting home visits of emigrants. In Serbia, the increase in services exports has been driven by newer sectors, namely information technology and communications (ITC), financial services, and business services, benefiting from policy efforts (Box 2). Studies show that services exports from newer products can provide a useful and more productive employment vehicle for middle-income countries that may need to shed labor from low productivity manufacturing sectors due to automation (Flaaen, Ghani, and Mishra 2013). These sectors have seen large increases globally in recent years and are also prominent in NMS-7.
Services Exports in the Western Balkans and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database and IMF staff calculations.Note: For panel 1, regional averages for 2016 are GDP weighted. “NMS” refers to NMS-7. For panel 2, averages are GDP-weighted.Services Exports in the Western Balkans and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database and IMF staff calculations.Note: For panel 1, regional averages for 2016 are GDP weighted. “NMS” refers to NMS-7. For panel 2, averages are GDP-weighted.Services Exports in the Western Balkans and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database and IMF staff calculations.Note: For panel 1, regional averages for 2016 are GDP weighted. “NMS” refers to NMS-7. For panel 2, averages are GDP-weighted.Most WB countries have a comparative advantage in travel services. A calculation of the revealed comparative advantage (RCA) index for travel services—that is, the share of travel services exports in a country’s total exports relative to the share of such exports in total world exports—shows that, apart from North Macedonia and Serbia, WB countries exhibit an RCA in this category (Figure 14). However, in all other main services categories, WB countries appear to lack a comparative advantage, particularly in fast-growing sectors such as ITC and business services. Serbia, which has benefited from policy support and recent success, now has a small RCA in ITC services. NMS-7 countries also lack comparative advantage in most services products except for transportation, possibly reflecting their stronger GVC links that yield transport services exports as a by-product.
Revealed Comparative Advantage in Services Exports: WB and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments database; and IMF staff calculations.Revealed Comparative Advantage in Services Exports: WB and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments database; and IMF staff calculations.Revealed Comparative Advantage in Services Exports: WB and NMS-7
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments database; and IMF staff calculations.Despite their revealed comparative advantage in travel exports, WB countries seem to be less competitive than regional peers (Figure 15). A comparison across a range of competitiveness indicators with other countries in southern Europe that are renowned tourist attractions—Cyprus, Malta, and Greece— shows that the WB region lags significantly in several aspects. These include government prioritization of tourism, international openness, tourist services and air transport infrastructure, and safety/security. The only category where WB region ranks better than European comparators is price competitiveness, likely reflecting low wages. A more granular comparison of Albania with Croatia also suggests the former’s weaker competitiveness arising from less developed infrastructure (Box 3).
Travel and Tourism Competitiveness Index Ranking
(Rank out of 136 countries)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: World Economic Forum Travel and Tourism Competitiveness Report, 2017.Note: Higher ranking indicates a worse score.Travel and Tourism Competitiveness Index Ranking
(Rank out of 136 countries)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: World Economic Forum Travel and Tourism Competitiveness Report, 2017.Note: Higher ranking indicates a worse score.Travel and Tourism Competitiveness Index Ranking
(Rank out of 136 countries)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: World Economic Forum Travel and Tourism Competitiveness Report, 2017.Note: Higher ranking indicates a worse score.Empirical Results
We conduct empirical analysis to identify factors that are associated with better performance in two services products: tourism and ITC/business services. Using a global sample of 184 countries and the period 2000–2015, we run panel fixed effects regressions to analyze variations in tourism performance and pooled regression analyses to explain variations in exports of ITC/business services (see Annex III for details). As suggested by the fixed effects regressions, a country’s fixed characteristics— including geography—explain the bulk of tourism performance, with limited role for policies. That said, we do find evidence that openness to trade, political stability, and the quality of infrastructure (electricity/telephone) matter for tourism performance (Figure 16). There appears to be a role for policy in ITC/business services exports, where the quality of math/science education, openness to FDI and trade, broadband Internet penetration, and lower unit labor costs are positively associated with exports (Figure 16).
Contributions to Services Exports1
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database; WEF Global Competitiveness Index; WB World Governance Indicators; ILO; Hofman, Osnago, and Ruta (2017); and IMF staff calculations.1 Contributions of each explanatory variable are calculated based on the estimated coefficients and observed values as a share of total predicted value of the export. These estimates do not take into account possible interactions between variables. The regional contributions are the average of the calculated country estimates.Contributions to Services Exports1
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database; WEF Global Competitiveness Index; WB World Governance Indicators; ILO; Hofman, Osnago, and Ruta (2017); and IMF staff calculations.1 Contributions of each explanatory variable are calculated based on the estimated coefficients and observed values as a share of total predicted value of the export. These estimates do not take into account possible interactions between variables. The regional contributions are the average of the calculated country estimates.Contributions to Services Exports1
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database; WEF Global Competitiveness Index; WB World Governance Indicators; ILO; Hofman, Osnago, and Ruta (2017); and IMF staff calculations.1 Contributions of each explanatory variable are calculated based on the estimated coefficients and observed values as a share of total predicted value of the export. These estimates do not take into account possible interactions between variables. The regional contributions are the average of the calculated country estimates.There could be significant gains in travel and ITC/other business services exports if the WB countries were to bridge policy gaps. Our simulation results suggest that policy improvements could boost, over time, exports of travel and ITC/other business services within a range of ¼ to 4 percentage points of GDP in WB countries (Figure 17).1 For travel services, further trade openness and improvements to electricity/telephone infrastructure would have the largest gains. For ITC/other business services, an expansion of high-speed Internet, improvements in math and science education quality, and greater openness to FDI would yield the largest gains.2
Simulated Gains to Services Exports from Policy Improvements1
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database; and IMF staff calculations.1 Benchmarks for policy improvements are based on the threshold value for the upper quartile of the observations for each indicator. Simulated gains are calculated by using this threshold value and the estimated coefficient where the actual observed value is less than the threshold value (i.e. where there is scope for policy improvement).Simulated Gains to Services Exports from Policy Improvements1
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database; and IMF staff calculations.1 Benchmarks for policy improvements are based on the threshold value for the upper quartile of the observations for each indicator. Simulated gains are calculated by using this threshold value and the estimated coefficient where the actual observed value is less than the threshold value (i.e. where there is scope for policy improvement).Simulated Gains to Services Exports from Policy Improvements1
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments Database; and IMF staff calculations.1 Benchmarks for policy improvements are based on the threshold value for the upper quartile of the observations for each indicator. Simulated gains are calculated by using this threshold value and the estimated coefficient where the actual observed value is less than the threshold value (i.e. where there is scope for policy improvement).Serbia’s IT Success
Serbia’s success as an IT hub can be attributed to successful public intervention at elementary school level and a supportive regime for investment and taxation.
Serbia is fast becoming a hub for IT services in the region. The value of such exports grew to 2 percent of GDP in 2016, from ½ percent in 2007, a level now only exceeded by exports of electric machines and road vehicles. The pioneering launch of the Microsoft Development Center Serbia (MDCS) in 2005 was a turning point. The center has been involved in key development activities for Microsoft, including Live Search, handwriting recognition, image analysis, and database management. A positive externality of the center is that other international IT companies have entered the market by buying out Serbian firms that were providing coding services, software development and testing, and web design.
A multipronged policy approach has helped ensure the IT sector’s success. The Serbian government has introduced “informatics” as a mandatory subject in elementary schools and raised enrollment quotas in IT departments in universities in 2017 by 20 percent. More than a quarter of university graduates each year major in technical subjects, more than in any other Western Balkan country, creating a strong computer science talent pool (Box Figure 2.1). The government also launched an Innovation Fund in 2011, with support from the EU and World Bank, to provide risk financing to Serbian businesses to support innovation in science and technology (Sanfey, Milatovic, and Kreisc 2016). Many IT professionals in Serbia also appear to benefit from a government tax regime that allows for presumptive taxation for entrepreneurs and sole proprietors, thus effectively taxing income at a lower rate.
Tourism Performance: A Comparison of Albania and Croatia
Croatia and Albania have experienced fast growth in tourism, although the two are far apart. Albania lags Croatia in infrastructure and tourism development. Sand and sun tourism imposes challenges of seasonality on both countries.
Both Croatia and Albania have experienced significant tourism growth over the past two decades, although the two countries are at different stages of tourism development. Tourism growth in Albania took off only around 2000 once post-war economic transformation had taken root. Croatia is a more established market, with a high ratio of tourist beds to per capita income. A higher ratio of travel services exports to foreign tourist arrival in Croatia points to its status as a more upmarket destination than Albania. On the other hand, Albania has seen a faster increase in bed capacity.
Croatia is perceived to be more competitive in the tourism sector relative to Albania (Box Figure 3.1). The World Economic Forum’s Travel and Tourism Competitiveness Report ranks Croatia and Albania as the 32nd and 98th most competitive in the world, respectively. Relative to Albania, Croatia stands out for marketing and branding, greater international openness (defined by the openness of visa regimes and the number of bilateral air service and regional agreements), better perceived safety and security, and a much higher perceived quality of tourist service and air transport infrastructure. However, Albania is more competitive on prices. This apparent paradox—that Albania has seen more rapid growth compared to Croatia despite being perceived as less competitive—can be understood in light of their different stages of development, with Croatia a more established, upmarket destination and Albania an emerging destination that competes well on price but lacks the overall development of Croatia.
Croatia and Albania’s Tourism Sectors
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments, WEO, Croatian Bureau and Statistics, Albanian Institute of Statistics, UN WTO, World Economic Forum Travel and Tourism Competitiveness Report 2017, and IMF staff calculations.Croatia and Albania’s Tourism Sectors
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments, WEO, Croatian Bureau and Statistics, Albanian Institute of Statistics, UN WTO, World Economic Forum Travel and Tourism Competitiveness Report 2017, and IMF staff calculations.Croatia and Albania’s Tourism Sectors
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: IMF Balance of Payments, WEO, Croatian Bureau and Statistics, Albanian Institute of Statistics, UN WTO, World Economic Forum Travel and Tourism Competitiveness Report 2017, and IMF staff calculations.The “sand and sun” tourism of Albania and Croatia comes with the challenge of seasonality. The bulk of tourists come in the summer, a phenomenon that is more pronounced in Croatia. Strong seasonality limits the backward linkages of tourism to the rest of the economy, as demand for labor and other inputs spikes on a seasonal basis. To build stronger year-round tourism the authorities of both countries are developing cultural, natural, gastronomic, nautical, and health tourism.
Albania’s past airport development strategy may not have been well suited for tourism. The 2005 concession for Tirana Airport that gave an international consortium monopoly rights to all commercial air traffic in Albania, resulted in relatively high landing fees. This made the country unattractive for low-cost carriers, which tend to prefer regional airports (typically closer to resort areas) with lower landing fees. Since 2016, the ownership and terms of the concession have changed, and low-cost carriers are becoming more common in Albania. A new airport has been built in the north, with plans to add an airport in the south near tourist areas.
In contrast, Croatia embraced the regional airport model much earlier, attracting low-cost carriers. Low-cost carriers have been operating in Croatia since 2004 and currently operate in five coastal and two inland airports, bringing tourists much closer to resort areas. Research also suggests that low-cost carriers can play a significant role in attracting visitors from emerging, relatively far markets, including during shoulder (late spring/early fall) seasons (Mandić, Teklić, and Petrić 2017).
Chapter 4 Raising Imports’ Contributions to Growth
Economic integration through trade offers significant opportunities for domestic manufacturing to access and adopt imported technologies in the production process. In this context, imports play an important supporting role through technology transfer, thus offering the possibility of increasing value added for domestic firms. Although the direct causality is difficult to ascertain, an analysis of the composition of imports can provide useful insights into a country’s growth model, the growth and exports structure, and country’s labor skill mix (Box 4).
Background
Studies find that GVC trade can play a significant role in lifting growth not just through higher exports but also through transfer of technology embodied in imported capital goods (Mazumdar 2000; Herzer, Nowak-Lehmann, and Siliverstovs 2006; Halpern, Koren, and Szeidl 2015; Taglioni and Win-kler 2016; Rodrik 2018). Export firms that operate at the higher end of the supply chain tend to be more imports intensive (Herzer, Nowak-Lehmann, and Siliverstovs 2006; Organisation for Economic Co-operation and Development 2007). Specifically, exports with high industrial value-added demand technologically advanced inputs, which are associated with capital goods imports. By facilitating transfer of foreign technology and associated knowledge and skills, GVCs may help boost the quality of exports, hence export competitiveness.
However, such a positive role of capital goods imports in enhancing exports and growth is conditional upon country’s domestic absorptive capacity, particularly the quality of its labor, institutions, and infrastructure (Glas, Hübler, and Nunnenkamp 2016). Among other things, the presence of highly skilled labor augments the positive association between imported capital goods and firm productivity (Hümmels, Ishii, and Yi 2001; Yi 2003; Bas and Strauss-Kahn 2014). Therefore, having an adequate pool of skilled labor can facilitate the absorption of capital goods and create a conducive framework for higher production and economic growth. For example, if a country lacks sufficient number of engineers, then importing high-tech equipment will not add to productivity and output growth because the equipment cannot be utilized effectively in production.
In Western Balkan countries, the structure of imports broadly reflects a consumption-based growth model. Consumption goods account for a sizable part of imports, ranging from 33 percent of total imports in Serbia to about 45 percent in Albania. The share of manufacturing in gross value added is small—ranging from about 4 percent in Montenegro to about 15 percent in Serbia compared to an average of 16 percent or more in NMS-7 (Figure 18). In Albania and Bosnia and Herzegovina, the manufacturing sector also consists of mostly labor-intensive products, namely leather, apparel, and unprocessed foods. Among other things, the composition of manufacturing is reflective of limited GVC links, or GVC links that are predominantly backward, concentrated in labor-intensive products, as discussed in Chapter 2. Consequently, the reliance on technologically advanced imports is limited with capital goods accounting for less than a fifth of total imports—considerably less than the share in NMS-7 (Figure 19). Following methodology proposed by Raveh and Reshef (2015) and Caselli and Wilson (2004), we decompose the capital goods into research and development (R&D)-intensive and R&D-nonintensive components. Our decomposition shows that the latter occupy a significantly greater share in Albania, Bosnia and Herzegovina, and North Macedonia compared to NMS-7 countries, whereas Montenegro and Serbia resemble NMS-7 (Coe and Helpman 1995; Coe, Helpman, and Hoffmaister 1997).
Gross Value Added – Manufacturing
(In percent of GDP)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics and IMF staff calculations.Gross Value Added – Manufacturing
(In percent of GDP)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics and IMF staff calculations.Gross Value Added – Manufacturing
(In percent of GDP)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: Haver Analytics and IMF staff calculations.
Composition of Capital Goods Imports
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: World Integrated Trade Solutions, UN Comtrade, and IMF staff calculations.1 A lower ratio means more R&D intensive capital goods imports.2 R&D intensive goods include electrical machinery and equipment, aircraft, vehicles other than railway, optical equipment, and chips/semiconductors. R&D unintensive goods include tools, nuclear reactors/boilers, railways, mechanical parts, and articles of iron ore/steel.Composition of Capital Goods Imports
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: World Integrated Trade Solutions, UN Comtrade, and IMF staff calculations.1 A lower ratio means more R&D intensive capital goods imports.2 R&D intensive goods include electrical machinery and equipment, aircraft, vehicles other than railway, optical equipment, and chips/semiconductors. R&D unintensive goods include tools, nuclear reactors/boilers, railways, mechanical parts, and articles of iron ore/steel.Composition of Capital Goods Imports
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: World Integrated Trade Solutions, UN Comtrade, and IMF staff calculations.1 A lower ratio means more R&D intensive capital goods imports.2 R&D intensive goods include electrical machinery and equipment, aircraft, vehicles other than railway, optical equipment, and chips/semiconductors. R&D unintensive goods include tools, nuclear reactors/boilers, railways, mechanical parts, and articles of iron ore/steel.Empirical Results
We explore WB countries’ ability to benefit from higher capital goods imports by empirically testing whether imported capital goods matter for growth beyond their direct impact on investment. We start with a simple ordinary least squares (OLS) panel data regression for a sample of 45 OECD and European countries (that include WB and NMS-7), with country fixed effects. The results point to a statistically significant association between growth, capital goods imports, labor skill mix, and investments. Further, to control for endogeneity and potentially missing variables, we use a fixed effect instrumental panel specification with labor skill and investments as instruments. The results are broadly consistent with the OLS panel estimates. We then complement our analysis with a dynamic panel data model, which, in addition, contains lags of the dependent variable as regressors, and accounts for issues such as momentum and inertia, by using Arellano-Bond dynamic panel data estimation. The results broadly resemble our instrumental panel regression estimates (see Annex IV for details on methodology and estimates).
After establishing a simple positive association between capital goods imports and growth, we explore the necessary conditions that may amplify the impact of capital goods imports on growth. Although this is a somewhat unconventional approach for looking at the association of imports and growth, we develop this idea based on existing studies that show that the absorption of high-tech imports of capital goods is conditional upon skilled labor (Glas, Hübler, and Nunnenkamp 2016). Using a panel threshold model for our sample of 45 countries, we estimate the impact of capital goods imports on growth, using labor skill as the threshold variable (Annex IV).1 We then test the hypothesis whether imports of high-tech capital goods are positively associated with growth in the absence of high skilled labor. We find that capital goods imports are absorbed more efficiently when the share of unskilled labor in total labor force is lower than an estimated threshold of 61 percent; that is, other things being equal, the imports of capital goods will have statistically significant association with growth if the share of low skilled labor is below 61 percent. Although our findings are consistent with earlier studies, the level of the estimated threshold is not precise, and the results have to be interpreted with caution, given potential issues of endogeneity and missing variables that we tried to address through several techniques (Annex IV).
Upgrading skills through better education and labor market policies would allow the Western Balkan countries to more efficiently utilize and absorb the higher technology embodied in imported capital goods. Our findings suggest that when the share of low-medium skilled labor is below 61 percent, the annual growth rate (proxied by the growth of GDPPC) could be higher by an additional 0.5 percentage points,2 other things being equal (Figure 20). When this share exceeds 61 percent, the magnitude of the estimated effect is much lower and statistically insignificant (Annex IV), implying that in that case an increase in capital goods imports would not impact economic growth.
Gains from Higher Labor Skills and Capital Goods Imports in the Western Balkans
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: ILOSTAT, World Integrated Trade Solutions, World Development Indicators (WB), WEO, and IMF staff calculations.Note: Additional gains are estimated conditional upon the threshold value, as follows: the coefficients of each explanatory variable in the threshold regression (see Annex IV) were multiplied by the observed value of that variable for each of the countries and for the average of NMS-7, proxied by the average value for 2010–2015, then all the components were dded together given the signs of their coefficients. The difference between the individual country value and the average of NMS-7 is used to estimate the additional gains to growth from closing the gap between the individual country and NMS-7.Gains from Higher Labor Skills and Capital Goods Imports in the Western Balkans
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: ILOSTAT, World Integrated Trade Solutions, World Development Indicators (WB), WEO, and IMF staff calculations.Note: Additional gains are estimated conditional upon the threshold value, as follows: the coefficients of each explanatory variable in the threshold regression (see Annex IV) were multiplied by the observed value of that variable for each of the countries and for the average of NMS-7, proxied by the average value for 2010–2015, then all the components were dded together given the signs of their coefficients. The difference between the individual country value and the average of NMS-7 is used to estimate the additional gains to growth from closing the gap between the individual country and NMS-7.Gains from Higher Labor Skills and Capital Goods Imports in the Western Balkans
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: ILOSTAT, World Integrated Trade Solutions, World Development Indicators (WB), WEO, and IMF staff calculations.Note: Additional gains are estimated conditional upon the threshold value, as follows: the coefficients of each explanatory variable in the threshold regression (see Annex IV) were multiplied by the observed value of that variable for each of the countries and for the average of NMS-7, proxied by the average value for 2010–2015, then all the components were dded together given the signs of their coefficients. The difference between the individual country value and the average of NMS-7 is used to estimate the additional gains to growth from closing the gap between the individual country and NMS-7.What Explains a Country’s Import Structure?
There is sizeable literature on factors that explain a country’s import structure contingent on its growth model, exports and industrial structure, and presence of GVCs.
The literature to date has tried to establish interrelations between a country’s import structure and growth. As such, the structure of imports in a country is endogenous—a reflection of a country’s economic growth model, industrial structure and labor skills, patterns in domestic and foreign investments, and composition of exports.
Growth model: Kharroubi and Kohlscheen (2017) and Klemm (2013) find that a consumption-driven growth model is associated with the prevalence of consumer goods imports, whereas investment- and export-driven growth is associated with greater imports of capital goods. Countries with high levels of remittances tend to be prone to consumption-driven growth, and hence their imports are likely to be more consumption-heavy (Barajas and others 2009; Chami, Hakura, and Montiel 2009; Fullenkamp 2015). Emigration of skilled labor in its turn, reduces a country’s human capital and thus the absorptive capacity of imported capital goods.
Export structure: Export firms tend to be more imports-intensive than domestically oriented firms (Goldar 2013). A more diversified export structure tends to be associated with a greater variety of imported intermediate inputs through direct and indirect channels (Chuang 1998; Hümmels, Ishii, and Yi 2001; Lewer and Van den Berg 2003; Yi 2003; Herzer, Nowak-Lehmann, and Siliverstovs 2006; Bas and Strauss-Kahn 2014). For instance, capital goods imports may help enhance productivity and reduce exporters’ fixed costs (the indirect channel). If imported capital goods’ prices are low, they may help increase the revenue flow of exporters by lowering costs (the direct-cost channel).
Industrial structure and labor skills: A country’s industrial structure and the skills of its labor have a significant impact on the composition of its imports. Countries that rely more on manufacturing (including R&D intensive sectors) tend to import more capital goods, specifically technologically advanced goods that can have knowledge spillovers. Lee (1994), Romer (1990), De Long and Summer (1991, 1993), Veermani (2009), and Cavallo and Laundry (2009) find a strong causal link between equipment investment and growth, also showing that investment-specific technological change stimulates the growth rate of the output by raising efficiency.
GVC links: Taglioni and Walker (2016) explore the role and benefits of imports of capital goods for GVCs. They find that GVCs boost the competitiveness of the exports of low- and middle-income countries by facilitating imports of foreign technology in combination with enhancing countries’ labor skills. De Backer and Miroudot (2014) and Garetto (2013) find a strong impact of imports of capital goods on firm competitiveness and participation of GVCs.
Chapter 5 Conclusions and Policy Recommendations
The slowdown in income convergence of WB countries with the EU over the past decade calls for a reconsideration of policies and structural reforms. The small economies of the WB region lie on the southeast edge of mainland Europe—the largest trading bloc in the world—but their exports have played a subdued role in economic growth. Exports remain a small share of output, after adjusting for size and domestic price level. This paper focuses on the need to reorient policies to make exports in goods and services an engine of growth convergence. It asks for reigniting exports to European neighbors through better integration into GVCs, reorienting the approach to services exports, and enabling imports to play a more supportive role in growth.
To integrate into GVCs and reap benefits, WB countries will need to improve infrastructure and labor skills, as well as deepen trade agreements. The findings in the paper show that by increasing GVC links, WB countries could raise GDP level by 3–10 percent. Of course, integration into GVCs requires foreign investment, which, in turn, necessitates regulatory reforms, improvement in institutions, and investment facilitation. As the results in Chapter 2 and 3 show, these deep trade reforms are important ingredients along with the widely understood need for improvements in infrastructure and education for the region to better integrate into Europe’s sophisticated production and services networks (Figure 21). In addition, WB’s poor infrastructure poses a serious physical constraint to the region’s ability to integrate into GVCs and should be upgraded promptly (Atoyan and others 2018).
It is critical to deepen trade agreements by including legally enforceable provisions. The WB states have carried out significant trade reforms over the past two decades, including through the adoption of the Central European Free Trade Agreement (CEFTA), by either already having joined WTO or being at an advanced stage of membership negotiations and having signed stabilization and association agreements (SAAs) with the EU. A visible impact has been a substantial reduction in tariffs. The priority is now to pursue deep PTAs—ones that have provisions that are legally enforceable so that they can be successfully invoked by a nonresident complainant in a dispute settlement proceeding. Although de jure changes to laws are a necessary first step, improvements in de facto enforcement is key, and is perhaps more challenging than the former.
The WB countries should seek to frontload adoption of key provisions of the EU accession process. Although the cross-border provisions of the chapters of the acquis communautaire will generally only apply once accession occurs, domestic reforms will enter into effect once a country has completed and implemented required legislation, which can happen years before accession. To facilitate integration into GVCs, the WB countries should seek to frontload adoption of legislations in the areas that will strengthen domestic institutions and competitiveness, such as company law; intellectual property law; competition policy; judiciary and fundamental rights; and justice, freedom, and security. Countries with SAAs that are already in place benefit from a privileged trade access to the EU market (Figure 22). But they should seek to align domestic legislation with that of the EU to allow them to improve trade facilitation (this was the case for some NMS that also received donor financial and capacity building assistance). These steps would also help address the current weaknesses in trade logistics.
Infrastructure Gaps
(In percent)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: WDI, EIA, IRF, Eurostat, Atoyan and others 2018, and IMF staff calculations.Note: Infrastructure gaps are measured against the EU-28 average.Infrastructure Gaps
(In percent)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: WDI, EIA, IRF, Eurostat, Atoyan and others 2018, and IMF staff calculations.Note: Infrastructure gaps are measured against the EU-28 average.Infrastructure Gaps
(In percent)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Sources: WDI, EIA, IRF, Eurostat, Atoyan and others 2018, and IMF staff calculations.Note: Infrastructure gaps are measured against the EU-28 average.
Trade Facilitation Indicators
(Index)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: OECD Trade Facilitation Indicator Simulator.Note: NMS-7 = BGR, EST, HRV, LTU, LVA, SVK, SVN; WB = ALB, BIH, MKD, NME, SRB.Trade Facilitation Indicators
(Index)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: OECD Trade Facilitation Indicator Simulator.Note: NMS-7 = BGR, EST, HRV, LTU, LVA, SVK, SVN; WB = ALB, BIH, MKD, NME, SRB.Trade Facilitation Indicators
(Index)
Citation: Departmental Papers 2019, 017; 10.5089/9781498314916.087.A001
Source: OECD Trade Facilitation Indicator Simulator.Note: NMS-7 = BGR, EST, HRV, LTU, LVA, SVK, SVN; WB = ALB, BIH, MKD, NME, SRB.Links within the WB are weak and deepening regional integration could also bring about development of local GVC networks. Empirical evidence suggests that deep agreements would be more effective in promoting trade and investment if a group of neighboring economies joined together (World Bank 2017b). WB countries can deepen the CEFTA in the areas of capital movement and foreign investment–related trade measures (TRIMS). Removal of non–tariff barriers would also help facilitate intraregional trade within the WB (Kaloyanchev, Kusen, and Mouzakitis 2018). Improvements in connective infrastructure, particularly one that lowers the time it takes to cross WB borders—a key weakness—can position the region as an attractive foreign direct investment (FDI) and GVC destination.
The gap in services exports between WB countries and European neighbors is narrower than that in goods exports, but there is room for further growth. There is scope for quality enhancement in tourism and scaling up of IT and business services exports. The findings in Chapter 3 suggest that policies to increase openness, enhance skills, and improve various aspects of infrastructure could significantly boost services exports and increase real GDP by up to 4 percent.
Further openness in services trade in the WB region would also require fast-tracking reforms associated with regional initiatives. WTO member countries typically use the General Agreement on Trade in Services (GATS) to progressively remove barriers that foreign suppliers face in accessing home services markets. However, because not all WB countries are WTO members, it limits the scope of such global initiatives in the region. The WB countries could begin to include commitments on services trade liberalization in their existing trade agreements (which generally do not include GATS commitments or legal enforcement mechanisms; Hofmann, Osnago, and Ruta 2017). As in the case of fast-tracking provisions related to adoption of GVCs, front-loading the services reform agenda—including those related to the EU services directive1—should also result in positive domestic spillovers.
Without an improvement in labor skills, the agenda to upgrade WB exports and integrate production into the European GVCs would remain incomplete. As the results in Chapters 2 and 4 suggest, high quality and skilled labor is a critical requirement for enhancing GVCs—firms that work seamlessly across international borders by splitting production along optimal value-added lines—and increasing the domestic absorptive capacity of capital investments, including imports of technologically advanced capital goods. To the extent that the WB economies lag significantly in skilled manpower, there is a need to address this key weakness. The agenda for upgrading education and labor training is discussed in detail elsewhere (Rahman and others 2015, Jirasavetakul and Rahman 2018). It is critical that the WB countries undertake GVC-related deep trade reforms in tandem with the upgrading labor skills. An improvement in the latter without improving employment prospects for skilled workers could further exacerbate the region’s endemic emigration problem and the attendant negative consequences for these countries (Atoyan and others 2016). This underscores the need for comprehensive labor market and education reforms to address chronic informality, emigration, and high level of inactivity that plague all countries in the region. To the extent GVC firms are essentially an extension of a single firm that exists over multiple borders, efficient management and organization of such firms requires the presence of appropriate top tier management in the various locations to coordinate activities. Constraints on the employment of such talent, if any, should be eased. Investing in and harnessing such talents and removing remaining constraints on employment can greatly enhance the region’s human capital.
