This Selected Issues paper on the Republic of Kosovo’s 2013 Article IV Consultation highlights growth and Kosovo’s external environment. In the wake of the global financial crisis, Kosovo’s economic growth slowed but remained positive, while most other Western Balkans slipped into recession. Moreover, the annual average growth rate has been among the highest in the Western Balkans since the onset of the financial crisis in 2007. Kosovo’s tax-to-GDP ratio is comparable to the average of Southeastern Europe, although its tax system relies significantly more on indirect taxation—including a high share of trade taxes. Kosovo’s reliance on trade taxes may create budgetary pressures in the event of further trade liberalization.

Abstract

This Selected Issues paper on the Republic of Kosovo’s 2013 Article IV Consultation highlights growth and Kosovo’s external environment. In the wake of the global financial crisis, Kosovo’s economic growth slowed but remained positive, while most other Western Balkans slipped into recession. Moreover, the annual average growth rate has been among the highest in the Western Balkans since the onset of the financial crisis in 2007. Kosovo’s tax-to-GDP ratio is comparable to the average of Southeastern Europe, although its tax system relies significantly more on indirect taxation—including a high share of trade taxes. Kosovo’s reliance on trade taxes may create budgetary pressures in the event of further trade liberalization.

Benchmarking Financial Development1

In the latter half of the past decade, private credit-to-GDP grew rapidly. Similar to most other Western Balkan countries, Kosovo may have “overshot” its structural benchmark on credit deepening. In contrast, other dimensions of development, such as efficiency and access still tend to lag. The impact of rapid credit deepening on asset quality in Kosovo has not been as bad as in some of the other Western Balkan countries. Nonetheless, going forward, monitoring closely and assessing the driving factors of credit-to-GDP growth would be an important task of macroprudential policy.

A. Introduction

1. Over the last decade, Kosovo’s banking system has been one of the fastest growing in the Western Balkans, although it still remains the least developed. Since the end of the war in 1999, the banking system has grown from four banks (half of which were new foreign banks) to nine banks (of which seven are foreign). Kosovo’s growth prospects, in particular from a relatively young and fast growing population, have attracted foreign banks, first from the Euro Area and more recently from near-by countries, in particular Turkey. Assets of the banking sector as a share of GDP have increased more than six-fold over the last decade, whereas that of other Western Balkan countries have increased less than three-fold. On the other hand, non-bank financial institutions and in particular securities markets have not seen much growth and remain relatively small or nonexistent.

2. This paper seeks to put Kosovo’s banking sector development in perspective, in particular with respect to the country’s structural characteristics and policy impacts. Given the dominant position of banks in the Kosovo financial system, this paper focuses on indicators of banking sector development. First, it compares Kosovo’s banking sector growth and indicators of development to that of the other middle-income Western Balkan countries (i.e., Albania, Bosnia-Herzegovina, Macedonia FYR, Montenegro, and Serbia, referred to as the SEE5). Second, it looks at select banking development indicators controlling for the level of economic development and structural characteristics to provide a benchmark of Kosovo’s progression. Lastly, it reviews how Kosovo’s business enabling environment may have had an impact on banking sector development.

3. To track banking sector development, a few indicators of depth, access, efficiency, and soundness are used. While ideally financial development indicators would measure how well the financial sector functions (e.g., how well savings are allocated to productive projects), direct measures are lacking. Instead, the literature has focused on a set of indicators that measure dimensions of financial intermediation, such as size as measure of depth, reach as measure of access, efficiency, and soundness as a measure of stability (Appendix). For parsimony and data availability reasons, this paper focuses on a subset of the banking development indicators used in the literature: depth is measured by bank credit to the private sector and broad money deposits relative to GDP; reach or access is measured by the number of bank deposits accounts and share of firms with credit; efficiency is measured by the bank lending-deposit interest rate spread; and stability or soundness is measured by the ratio of non-performing loans (NPL) to total loans.

Size of Financial Institutions and Markets in Kosovo, SEE5 and ECA

(Percent of GDP)

article image
Sources: Central Bank of Kosovo; World Bank Global Financial Development Database; and IMF staff calculations.SEE5 comprise Albania, Bosnia-Herzegovina, Macedonia, Montenegro, and Serbia.ECA comprise middle-income countries in Europe and Central Asia.N/A is not available.

4. Further, to better evaluate Kosovo’s banking sector development, indicators are benchmarked(i.e., structural characteristics are taken into account). This follows the methodology developed by the World Bank (Appendix) that benchmarks development indicators by controlling for economic-structural characteristics (e.g., income per capital, population characteristics, etc.). Cross-country differences arising from policies are thus partially captured in the residual. In benchmarking, what is relevant is not so much the level of the indicator, but the gap between the benchmark and actual level of the indicator. While Kosovo is currently not included in the Finstat database, its benchmark can be proxied by that of Albania, which has the most similar level of income per capita and population structural characteristics. 2

Select SEE and ECA Countries Level of Economic Development and Structural Characteristics

article image
Source: World Bank Development Indicators.

SEE=South Eastern Europe; ECA = Europe and Central Asia (World Bank grouping); LMI=Lower Middle Income; UMI=Upper Middle Income; and HI=High Income.

People per sq. km of land area

Percent of working-age population (i.e., 15-64 years). Young are under 15 and old are over 64.

For Kosovo data are for 2010.

B. Banking Sector Growth and Development

5. Kosovo’s banking sector has grown quickly but is still the least developed in the Western Balkans, reflecting in large part its relatively lower level of economic development. Kosovo’s fast growth rate of private credit and deposits over the last decade appears in line with its lower starting level of depth. Despite its fast growth, however, Kosovo’s banking sector depth, in particular for private sector credit, is still lower than that of its neighbors. This positioning is largely expected, reflecting largely its lower level of economic development (i.e., GDP per capita).3 Further, up until the global financial crisis of 2008-2009, the Western Balkan countries had experienced a credit boom, in several cases financed by foreign funding.4 In contrast, Kosovo’s banking sector, although dominated by foreign banks, is largely domestic deposit funded.

uA04fig01
Sources: IMF International Financial Statistics (IFS) and staff calculations.

6. Nonetheless, Kosovo’s financial depth now appears to be above its structural benchmark—as is the case with many of its neighbors—which may indicate some overshooting of credit growth. Whereas in 2003 Kosovo appeared to have been below its proxy benchmark (i.e., Albania’s expected median) for private sector credit and deposit depth, by 2005 its banking depth appeared above where Kosovo’s structural characteristics would indicate.5 Indeed, by 2008 Kosovo’s private sector credit-to-GDP ratio was close to its expected 75th percentile. This could be approaching the zone for an excessive credit boom (i.e., assuming this is not a result of a better enabling environment or more macro stability).6 However, the private credit-to-GDP gap in Kosovo still is lower than gaps in most other neighboring SEE countries, suggesting that high credit depth is a regional characteristic.

uA04fig02
Sources: IMF IFS; World Bank Finstat; and IMF staff calculations.

7. While depth may have overshot, the NPL ratio—an indicator of banking stability—is still relatively low. Since 2006, the NPL ratio in Kosovo has been on an increasing trend, coinciding with the increasing credit depth gap. Nonetheless, NPLs in Kosovo have remained mostly below its benchmark for NPLs, which suggests that banks in Kosovo may have been relatively more cautious and that credit quality compares favorably to that in other comparable countries with similar structural characteristics. Furthermore, the gap between the actual NPL ratio and the NPL benchmark, while narrowing in recent years, still compares very favorably with the situation in the other SEE countries, which have consistently been above their threshold, in particular following the global financial crisis of 2008.

uA04fig03
Sources: Central Bank of Kosovo; World Bank Finstat Database; and IMF staff calculations.

8. In terms of reach, in particular access to credit by a broader segment of firms, Kosovo needs to make further progress. Reach, as measured by the number of bank accounts has been improving, closing the gap with the structural benchmark. However, reach as measured by percent of firms with credit access has not improved since 2007, when it was close to its benchmark. Indeed the gap has since widened.

uA04fig04
Sources: IMF Financial Access Survey (FAS); World Bank Finstat; and IMF staff calculations.

9. Kosovo’s banking efficiency, as measured by the interest rate spread, is also below its statistical benchmark (i.e., the spread is higher than expected). This may reflect the lack of sufficient progress on creditor rights as well as financial structure characteristics (e.g., bank concentration or extent of competition). There is some limited evidence with regard to the former factor from the World Bank Doing Business indicator on the cost of enforcing a claim that Kosovo’s that has remained elevated, while it has been mostly declining in the SEE5 (see below).

uA04fig05
Sources: IMF IFS; World Bank Doing Business Surveys and Finstat; and staff calculations.

10. Kosovo has made some progress on credit information, but the enabling environment for credit enforcement remains quite weak. Data for Kosovo is very limited—in particular Doing Business (DB) indicators on the ease of getting credit, enforcing contracts, and resolving insolvency for Kosovo are only available from 2009 (i.e., DB2010 survey). Nonetheless, Kosovo has made some improvements in the enabling environment for banking. In particular the coverage of the CBK credit registry has improved,7 and Kosovo has a similar ranking to that of the SEE5 median and a higher ranking than the ECA median.8 However, Kosovo still lags most SEE5 and ECA countries in terms of enforcing contracts and resolving insolvencies, both of which would tend to reduce credit expansion (as they add to interest rate spreads and thus the quantity of credit demanded). This would tend to be consistent with Kosovo not being able to expanding credit as much as the SEE5.

11. Going forward, the envisaged development of a macroprudential policy framework that includes monitoring closely and assessing credit-to-GDP growth would be an important task. The authorities are working on improving the enabling environment, which should allow for further reach and efficiency of the Kosovo banking sector. However, sustainable financial development also requires financial stability. In this regard, monitoring the evolution of the credit-to-GDP ratio would be important task. The Central Bank of Kosovo is currently developing a macroprudential framework, in line with one of the key recommendations of the 2012 Financial Sector Stability Assessment (FSSA).

Indicators of the Business Environment

article image
Source: World Bank Doing Business Surveys.

References

  • Barajas, A., T. Beck, E. Dabla-Norris, and R. Yousefi (2013): Too Cold, Too Hot, Or Just Right? Assessing Financial Sector Development Across the Globe, IMF WP/13/81.

    • Search Google Scholar
    • Export Citation
  • Čihák M., A. Demirgüç-Kunt, E. Feyen, and R. Levine (2012): Benchmarking Financial Systems around the World, World Bank Policy Research Working Paper 6175.

    • Search Google Scholar
    • Export Citation
  • De la Torre, A., E. Feyen and A. Ize (2011): Financial Development: Structure and Dynamics, World Bank Policy Research Paper 5854, 2011).

    • Search Google Scholar
    • Export Citation
  • Feyen, E. and K. Kibuuka (2013): FinStats 2013: A ready-to-use tool to benchmark financial sectors across countries—User Guide and Benchmarking Methodology, October 2012 (Version 3.0).

    • Search Google Scholar
    • Export Citation
  • IMF (2012): Republic of Kosovo: Financial System Stability Assessment, IMF Country Report No. 13/99, April.

  • IMF (2013): The Boom-Bust of Foreign Bank Flows into CESEE, Central, Eastern and Southeastern Europe Regional Economic Issues, April.

Appendix. Measuring and Benchmarking Financial Development

Measuring Financial Development

Financial development occurs when agents mitigate the effects of imperfect information, limited enforcement, and transactions costs through the use of financial instruments, intermediaries, or markets in order to efficiently allocate financial resources and diversify risks.1 Financial development indicators would aim to measure such financial functioning. However, direct measures are mostly lacking. Thus the literature typically uses intermediate indicators of dimensions of financial intermediation such as of the size, efficiency, reach, and soundness of financial institutions and markets. Specifically, for banks the following indicators are often used:

article image
Source: Beck et al. (2008)

Work by the World Bank has evaluated the above list based on specific selection criteria: (i) empirical evidence of direct link with final welfare (e.g., indicators that perform best in predicting economic growth, such as private credit-to-GDP); (ii) goodness of fit (i.e., R2) in structural regression estimates; (iii) factor loading along the principal component that measures the specific dimension of intermediation; (iv) broadest coverage of countries and over time; and (v) stability (i.e., the ratio of within sample variance to between sample variance).

Although not included in the study above, financial benchmarking studies (discussed below) also include measures of stability or soundness as this is an important dimension of sustainable financial development. Measures of banking sector soundness include indicators on capital adequacy, liquidity, and non-performing loans. Recent work has also included the z-score (a proxy for the distance to default).

Benchmarking Financial Development

When comparing levels of financial development across countries, it is useful to take into account the level of economic development and structural characteristics that impact development regardless of the financial sector policies employed by a country. In particular, structural characteristics such as population size, density and age composition along with income per capita and whether the country is a transition economy, fuel exporter, or offshore financial center have been found to statistically impact the indicators of financial development.2

Controlling for these structural, or policy-invariant, variables in a regression provides a country’s structural “benchmark”, i.e., the expected average (if using least-squares regressions) or median (or other percentile if using quantile regressions). Thus, it may be possible to say if a country is over-performing even though it has a lower absolute level of an indicator (e.g., point A in the figure below), or underperforming even if it has a higher absolute level (e.g., point B).

As long as the impact of financial policies on economic development is lagged, policy should at least be partially embedded in the residual. The residual (or gap) between actual performance and a country’s expected average or median can therefore indicate whether a country is performing below its potential—perhaps due to weaknesses in its enabling environment, e.g., due to weak information and legal infrastructure—or above, perhaps due to financial innovation, which if combined with weaknesses in the regulatory and governance frameworks, however, can lead to overshooting or a boom-bust episode (e.g., point C in the figure below).

uA04fig06

Stylized Financial Possibility Frontier

Citation: IMF Staff Country Reports 2013, 223; 10.5089/9781484307854.002.A004

Source: Barajas et al. (2013)

With regard to depth indicators, regression results show a positive relationship between income per capita (which tends to level off at higher income levels) as well as size of the market (i.e., population size). In contrast, age dependency (i.e., in particular relatively more young people that are below 15) and being a transition economy tend to reduce depth. Population density tends to increase deposit depth but reduce credit depth.

Benchmarked financial development indicators are available from the World Bank’s Finstat database, which covers 183 countries and estimates benchmarks based on data from 2002-2011. The panel regression includes year-fixed effects, but not country fixed-effects (i.e., a global shock can lift or lower all countries, and a common path of development is assumed).

1

Prepared by Pamela Madrid Angers.

2

Adjusting for Kosovo’s denser and younger population would indicate that it’s expected quantiles should be somewhat lower than Albania’s.

3

As noted in the Appendix, Kosovo’s relatively younger population (below 15 years of age) also would tend to lower its level of depth relative to other Western Balkans, all else equal.

4

Foreign funding was particularly relevant for Montenegro and Bosnia-Herzegovina during the boom years of 2003-2008. See IMF (2013).

5

As noted in the Appendix, the benchmark of the indicator is estimated from a quantile regression of a panel of countries, controlling for their GDP per capita, population characteristics, and dummies for other structural variables (e.g., if the country is a transition country, fuel exporter, or offshore financial center). It also includes time dummies.

6

Barajas et al. (2013) find that the probability of a banking crisis surpasses 10 percent when a country exceeds its benchmark by 50 percent.

7

Kosovo’s public credit registry has information on borrowers going back to 2003. Its information content was recently expanded. Banks cite it as a very useful tool to mitigate credit risk.

8

This is consistent with recent research on the structure and dynamics of financial development that finds that informational frictions are easier to overcome than contractual frictions (see de la Torre et al., 2011).

1

See M. Čihák, A. Demirgüç-Kunt, E. Feyen, and R. Levine, “Benchmarking Financial Systems around the World”, World Bank Policy Research Working Paper 6175, 2012, for a fuller discussion.

2

Beck, Feyen, Ize, and Moizezowicz, “Benchmarking Financial Development”, World Bank Policy Research Working Paper 4638 (2008). Maximizing model fit was used as a criterion to select the final set of controls from the large set of potential controlling factors.