Journal Issue

Azerbaijan Republic: Selected Issues

International Monetary Fund
Published Date:
January 2005
  • ShareShare
Show Summary Details

II. Policies to Promote Regional Economic Convergence in Azerbaijan11

A. Introduction

36. Azerbaijan’s macroeconomic performance has been strong since 1996, with high real GDP growth and low inflation. Initially, the economic recovery benefited mostly the areas in and around Baku, leading to growing regional income inequalities. While more recently economic growth has started to spill over to other regions of Azerbaijan, income levels in the Baku region remain significantly above the national average. Together with a better provision of key public services in Baku—including health, education, and infrastructure services—this income differential has been responsible for continuing migration from other regions. Concerned that large gaps in income and livings standards can act as a barrier to the social cohesion necessary for successful economic modernization, the Azerbaijan authorities have adopted a regional economic development program for the period 2004–08, aimed at promoting balanced regional economic development.

37. This paper summarizes the international experience with regional policies and analyzes the Azerbaijan government program of regional development in light of this experience. The remainder of this paper is organized as follows. Section B provides indicators on Azerbaijan’s regional distribution of incomes and poverty. Section C presents the main elements of the authorities’ regional development program. Section D summarizes the theoretical and empirical economic literature on regional development policies. Finally, Section E draws lessons for Azerbaijan from the international experience.

B. Regional Income and Poverty Indicators in Azerbaijan

38. Azerbaijan has a total resident population of about 8.3 million. Administratively, the territory of the country is divided into 59 districts, which are grouped into 9 regions: Nakhchivan (southwest),12 Absheron-Guba (northeast), Mugan-Salyan (central), Ganja-Gazakh (west), Sheki-Zagatala (northwest), Lankaran-Astara (southeast), Shirvan (central), Karabagh-Mil (southwest), and Baku city (east). Following the armed conflict with Armenia during 1988–94, about 1 million people (12 percent of total population) became refugees and internally displaced persons.

39. There is significant income disparity among regions, with Baku having the highest per capita income—despite continued, significant migration from other regions. The income level in Baku is 30 percent higher than in Nakhchivan that has the lowest income (Table 2.1). The eastern regions including Baku are, in general, better off than the rest of the country, benefiting from the massive foreign direct investment (FDI) for oil in recent years and associated spillover effects. The western regions, on the other hand, where a significant part of the old industrial structure is located (Ganja-Gazakh), tend to be the “poorest” in terms of income, and also have higher unemployment rates relative to the rest of the country. This is mainly due to existing barriers to the integration of products, capital, and labor markets, such as an underdeveloped transportation network, unreliable provision of utility services, insufficient presence of financial institutions, and illiquid housing markets.

Table 2.1.Azerbaijan: Per Capita Monthly Income by Geographical Zones, 2003(In thousands of Manat)
Total income167194177172176175179181216
Income from agriculture30193437574546384
Income from property110101001
Current transfers212125202320192820
Benefits and social contributions238322362
Social transfers in kind012100051
Other income223430392538223638
o/w: Income from other households192826312129182629
Memorandum items:(in percent)
Regional income relative to Baku7790828081818384100
Share of income from employment313222211822182648
Share of income from agriculture18101921332625212
Source: 2003 Household Budget Survey.
Source: 2003 Household Budget Survey.

40. Baku, in addition to having the highest average income relative to the rest of Azerbaijan, also has the lowest incidence of poverty (40 percent compared to a national average of 46.7 percent). However, the share of people living in extreme poverty in Baku is higher than the national average (9.1 percent compared to a national average of 8.8 percent, Table 2.2), reflecting difficulties to absorb the rapid migration from other regions.

Table 2.2.Azerbaijan: Poverty Incidence by Regions
 Poverty levels
CategoryPoverty line

(175,000 Manat)
Extreme poverty line

(125,134 Manat)
Total population46.78.8
Of which: Urban47.89.6
Source: 2003 Household Budget Survey.
Source: 2003 Household Budget Survey.

41. Local governments in Azerbaijan have limited fiscal autonomy. Tax revenue accruing directly to local governments (excluding Baku) accounts for only about 15 percent of their spending, with the remainder covered by transfers from the State Budget.13 There are no established rules or formulas on how to distribute such transfers, and thus decisions on allocating them to individual regions are largely made on an ad hoc basis.

C. Summary of the Regional Economic Development Program

42. The program of regional development lays out regional policies, as well nationwide measures aimed at promoting the mobility of factors of production and improving institutions. Regional policies include increasing public investments in regions outside Baku for improvements in health, education, and infrastructure (such as roads—particularly in rural areas and small districts—power stations, and water supply and irrigation facilities), enhancing the regions’ fiscal autonomy, creating local employment information and labor training centers, establishing rural credit unions, using government lending to the private sector in poor regions via commercial banks, and establishing special economic zones (SEZs). The nationwide measures aim at promoting competition, improving the business environment, and deepening financial intermediation, with a view to helping reduce existing bottlenecks to interregional labor and capital mobility.

43. A Secretariat has been formed in the Ministry of Economic Development to oversee the implementation of the regional development program. The 2005 budget will finance a number of initiatives envisaged in the program, and some steps have already been taken to ensure its consistency with other government initiatives, particularly the State Program for Poverty Reduction and Economic Development. However, interagency coordination remains weak.

D. Economic Theory and Experience with Regional Development Policies

Review of Economic Theory on Convergence and Divergence

44. The theoretical literature on economic growth can be grouped under two competing hypotheses. The first one, the convergence hypothesis, argues that technological improvements lead to economic convergence among regions and countries, provided there is free trade and relatively unrestrained competition. The second one, the divergence hypothesis, asserts that market forces, in the presence of free trade and factor mobility, lead almost unavoidably to income inequality and divergence in economic growth rates among regions and countries.

45. The key assumption of convergence theories is that all factors of production (e.g., labor, as well as human and physical capital) have decreasing returns to scale.14 According to these theories, free trade and open competition encourage movement of mobile factors toward regions where they have a higher marginal product, that is, capital to the poorer regions, and labor to the richer regions. This, in turn, leads to a higher per capita growth rate in poorer regions, and thus convergence in per capita incomes until an approximately uniform distribution of these mobile factors is reached across regions. These models therefore recommend a reduction in barriers to trade and factor mobility as a way to promote regional convergence.

46. Conversely, the key assumption of divergence theories is the presence of increasing returns to scale in at least some sectors of the economy.15 According to these theories, increasing returns, in the presence of high fixed costs, are the engines of economic progress, while comparative advantages and competition play a secondary role. These theories suggest that, due to the presence of increasing returns, any increase in the degree of trade openness and factor mobility is likely to send the most productive factors flowing toward the advanced regions, where their return is higher, leaving the disadvantaged areas further behind. Thus, a “core-periphery” structure may emerge with trade integration.

47. An extreme version of divergence theories predicts that the construction of infrastructure for transportation and communication may be harmful to the poorest areas, as it facilitates migration of their most productive factors (Martin, 1997). A weaker version of divergence theories generates divergence because the poor regions have not managed to cross a threshold level in their endowment of strategic inputs, such as the following: human capital, public infrastructures, research and development activity, and financial deepening. According to these theories, regions will cluster within different clubs, which are determined by upper and lower bounds in the endowments of the strategic factors. While there could be convergence among regions of the same cluster, the income disparity between regions in different clusters might actually continue to increase.

48. The impact of open trade, factor mobility, and public investment policies has varying effects under these two competing groups of growth theory models. Liberalizing trade and factor mobility (across both regions and countries) will lead to convergence in per capita income under convergence theories, but will have the opposite or mixed effects under divergence theories.

49. The impact of income-equalizing grants on convergence also differs under the two competing extreme versions of these theories. If the objective is convergence in incomes, such grants are discouraged under extreme convergence theories, as they undermine economic growth in poorer regions, but they are encouraged under extreme divergence theories in order to reduce incentives for the most productive factors to move from poorer to richer regions. However, under the weaker versions of convergence and divergence theories there seems to be more or less agreement regarding the desirability of income-equalizing grants, largely on equity grounds. Since not all factors of production are equally mobile, even if there is free mobility, many people will stay in the poorer regions due to, for example, age or ethnicity considerations. Hence, income-equalizing grants may be called for on welfare grounds because transfers to the poor regions will have a direct positive welfare impact through income transfer effects. If such grants finance spending that lowers transactions costs between regions, they can also induce firms in labor intensive industries to relocate to poor regions.

Convergence or Divergence? A Review of International Experience

50. While a few developing countries appear to have been able to reduce regional income inequalities (e.g., India and Pakistan), regional inequality has actually increased in many other developing countries. The latter has particularly been the case in countries experiencing fast overall real GDP growth (e.g., China, Indonesia, and Thailand), but also in other developing countries as well (e.g., Brazil, Philippines, Sri Lanka, and Vietnam).16 In many cases, the divergence in regional incomes has occurred despite targeted regional policies, including infrastructure and other public investments in the “poorer” regions.

51. In contrast, developed countries generally experienced convergence in per capita incomes. There was a large reduction, for example, in inequality of regional per capita incomes in the United States in the past century (1880–1990).17 In addition, several studies have found convergence in per capita incomes among European Union (EU) countries,18 as well as across regions in most individual EU countries since the end of World War II, particularly among regions of Spain and Portugal.19 However, there have been cases in the EU where regional income inequalities seem to have increased (e.g., Greece).20

52. It seems that the theories based on decreasing returns to scale provide an adequate description of the main forces underlying the convergence process in developed countries. Evidence from U.S. cities reported by Ciccone, Peri, and Almond (1999), for example, overwhelmingly supports the idea that, if externalities are at all present, they are too weak to overcome the usual effects of decreasing returns. While the tendency for divergence in developing countries seemingly supports theories based on increasing returns to scale, it may well be that the observed divergence reflects the fact that factors of production are either unable or unwilling to move, due to impediments and/or age and ethnicity considerations (Shankar and Shah, 2001).

53. The speed of convergence, in countries and regions where it has taken place, has been relatively slow. In India, for example, the estimated speed of regional income convergence between 1961 and 1991 amounted to about 1.5 percent per year of per capita income (World Bank, 1996). The speed of convergence among EU countries since World War II has been similar, about 2 percent of per capita income a year (Barro and Sala-i-Martin, 1991), as poorer countries have grown faster than wealthier ones, and the dispersion of per capita incomes has fallen (Pritchett, 1996). The speed of convergence has been highest in the United States (Barro and Sala-i-Martin, 1991) and Japan, although not much higher than elsewhere—about 2.5 percent and 3 percent per year, respectively.

Policies that have helped promote regional convergence

54. The experience of countries, where convergence occurred, demonstrates that free factor mobility, integrated domestic markets, and strong regional institutions contribute to convergence in regional incomes. However, evidence on the positive impact on income convergence of financial transfers to the regions, decentralization, and special economic zones is mixed.

Enhancing factor mobility

55. The convergence in regional incomes in developed countries has generally been associated with policies of open trade across regions and removal of barriers to factor mobility. The United States provides one of the best examples of convergence in regional incomes, primarily because of enhanced mobility of labor and capital. The high mobility of the factors of production, together with an improved business environment in the poorer Southern regions, led gradually to the displacement of low-wage labor and its migration North, while increased mobility of capital led to its move toward the South. Both factors have contributed to convergence.

56. Regional economic developments in the EU also highlight the positive impact of high labor and capital mobility in achieving regional economic convergence. Boldrin and Canova (2000), for example, show that the regional convergence rate in Europe was the highest in the period between the initial post-war period until about mid-1970s, which was characterized by high mobility of both labor and capital. During this period, regional policies were largely absent in the EU area. The authors note that the EU’s convergence rate has slowed down since then, despite gradually increasing Structural and Cohesion Funds allocated by the European Commission to poor regions, as the mobility of labor was reduced considerably due to increasingly restrictive labor policies. These policies included restrictions on hiring and firing of workers, as well as the adoption of binding national minimum wage levels. In the absence of substantial labor mobility across regions, and given the rigid wage structures that have increasingly characterized national labor markets in Europe since the mid-1970s, long-term unemployment in the poorer regions has increased.

Enhancing decentralization

57. International experience suggests that fiscal decentralization offers no panacea. Evidence indicates that successful development strategies are compatible with various balances of power between the center and the regions, which are often related to the size of the country, its history, as well as ethnic and political factors. While fiscal decentralization is positively correlated with per capita income growth across countries (Oates, 1999), the direction of causality is not clear.

58. Decentralization can strengthen the delivery of local public services with consequent efficiency gains by providing greater local accountability for results and interjurisdictional competition, and allowing local governments to take into account the specific local conditions and preferences. However, experience indicates that in terms of progress in convergence some decentralized states (e.g., Brazil) have done worse than centralized ones (e.g., South Korea), and some centralized states (e.g., Thailand) have had a worse record than decentralized ones (e.g., United States). Implementation capacity at the level of regional governments and the strength of regional institutions appear to be essential preconditions for successful decentralization.

Increasing capital and current transfers

Investing in infrastructure

59. Improving infrastructure seems to have been critical to help achieve convergence in per capita incomes in most developed countries and some developing countries. Studies by Barro and Sala-i-Martin (1991 and 1992), for example, suggest that if there had been no differences in infrastructure endowments in the EU, the speed of convergence in per capita incomes among EU countries would have been faster. In addition, studies on the economic performance of Chinese (Fleisher and Chen, 1997; Mody and Wang, 1997; and Demurger, 2000) and Spanish provinces (De la Fuente, 2001) also suggest that differences in transport infrastructure and telecommunication facilities account for a significant part of the observed variation in growth performance of provinces. Furthermore, evidence from the EU area indicates that there is crowding-in effects from public investments.21

60. However, the benefits from investments to poorer regions for regional income convergence have proven elusive in many countries. In Brazil, for example, regional development policy in the 1970s was marked by massive state-led investments in protected capital-goods industries, subsidies, and incentives for the private sector and infrastructure development in the poorer regions (north and northeast). However, these efforts were largely unsuccessful in reducing the income and social gaps—primarily because many of these projects were ill-conceived and did not benefit the local, agriculture-based economy (World Bank, 1996). In fact, it seems that such infrastructure investments, which aimed at integrating the national economy and lowering business costs in peripheral regions, had perverse effects—companies in and around the richer Sao Paolo achieved greater economies of scale by reaching distant markets, contributing to increased divergence in regional per capita incomes.

61. Similarly, in Italy, the large-scale investments in basic infrastructure over the last 40 years targeting the “poorer” South (Mezzogiorno) have had little wider regional impacts. The literature suggests that the root problems of regional development of the Mezzogiorno may have more to do with overall institutional weaknesses in this region—for example, rent-seeking behavior (itself engendered in part by the fiscal subsidies and transfer programs), inability of the state to adequately protect and enforce private property rights, the same minimum wage as the rest of the country, and relatively underdeveloped financial institutions in poorer regions (World Bank, 1996).

62. These experiences suggest that public investments in infrastructure are effective only in the presence of efficient institutions and when driven by private sector demand (World Bank, 1996), bolstering convergence effects of competitive domestic markets. These experiences also indicate that investing in small but essential infrastructure works, such as rural and urban water supply, or a simple network of roads, is more effective in promoting income convergence than large-scale projects.

Investing in health and education

63. While investing in infrastructure is beneficial under appropriate conditions, investing in health and education is key to supporting balanced regional development. As noted in Demurger, Sachs, and others (2002), once market economy institutions are in place, technological advancement is the fundamental engine of sustainable development. The objective is to create sufficient local scientific capacity to accelerate the diffusion of new technologies from the richer to poorer regions, a process that requires a qualified labor force.

Using income-equalizing transfers to regions

64. Fiscal income-equalizing transfers from the center to poor regions have undoubtedly helped to maintain a higher standard of living and per capita incomes in poorer regions. However, such transfers have not contributed to real dynamism in the poor regions (in the sense of having led to self-sustained growth). Indeed, depressed regions have often tended to become more reliant on such central assistance over time. In some cases, such reliance may have spawned other distortions (i.e., rent-seeking in Southern Italy), and the richer regions usually have become more intolerant of such resource transfers.

Establishing special economic zones

65. International experience indicates that the costs associated with special economic zones (SEZs) are often too high (IMF, FAD Guidance Note, 2003). The main costs of SEZs to the government arise from tax breaks. Firms established in free zones often obtain concessions from host countries—including tax holidays, reduced income tax rates, exemptions from domestic indirect taxes, and the like. There is ample evidence that such privileges are not among the main determinants of the decision of foreign firms to invest; for these firms, it is more important to be subject to a stable and fair tax regime. A study by McKinsey Global Institute (2003), for example, finds that official targeted incentives rarely have a positive effect, and have often resulted in inefficiency and waste of resources. In fact, successful free trade zones, such as Santa Cruz in India, grant very limited privileges. Enforcement costs, to prevent producers/consumers outside the zone from taking advantage of the zone’s privileges, are the second source of direct costs and tend to increase over time. The most serious costs associated with free zones come from tax leakages from less-than-fully secured zones—when the enforcement efforts are inadequate. No enclave is fully hermetic, and significant tax revenue can be lost through abuse. Over time, pressure increases for privileges to be extended to other regions as well, and more companies take advantage of the tax concessions (e.g., for local investors, by teaming up with foreign partners; for existing foreign investors, by dismantling and rebuilding their corporate structures), thus increasing the costs.

E. Lessons for Azerbaijan

66. Many elements of the regional economic development program adopted recently by the Azerbaijan authorities are consistent with international best practices. However, the key to achieving dynamic and sustainable economic growth in the regions will be the effective implementation of the measures incorporated in the regional economic development program. In particular, for a successful implementation of the strategy the authorities will need to do the following:

  • First and foremost, create a level-playing field across Azerbaijan by removing barriers to labor and capital mobility, developing the financial sector, reducing red tape, and tackling governance weaknesses;
  • Improve planning and institutional implementation capacity of the central government to ensure that resources earmarked for poor regions are used effectively, including by approving rolling medium-term investment strategies for regional development in the context of the annual budget preparation process and medium-term expenditure framework;
  • Direct public spending and investment in poorer regions toward health, education, and demand-driven basic infrastructure, such as a simple network of regional roads, urban and rural power, gas and water supply, and irrigation facilities;
  • Increase regional fiscal autonomy gradually, in recognition of limited institutional capacity in the regions;
  • As international experience indicates, the convergence in regional incomes is likely to be slow. Thus, in the medium term, the rapid labor migration toward Baku is likely to continue because of the persistence in regional income disparities. To address this challenge, the government needs to consider improving infrastructure in Baku. This will help avoid the costs associated with congestion, such as slums, overburdened sewage systems, and an increasing incidence of poverty.
  • Carefully consider the SEZ policy, given that any benefits from SEZs are likely to be short-lived, and clearly offset by associated costs, which tend to increase over time. Should such zones be established, the authorities will need to consider (i) limiting the number of zones (one or two maximum) (ii) physically enclosing the zones; (iii) limiting tax privileges to trade taxes; (iv) maintaining a stable and fair tax regime with no special privileges; and (v) allowing the private sector be responsible for investment and administration in the zones.

Prepared by Niko Hobdari.


Following the war with Armenia, the Nakhchican Autonomous Republic, with a population of about 0.3 million people, is separated from the main territory of Azerbaijan by Armenia.


Local (municipal) taxes include (i) taxes on land and other property of physical persons; (ii) taxes on the use of construction materials of local importance; and (iii) taxes on profits of enterprises and institutions that are in the ownership of municipalities.


The basic model is that of Solow (1956) and Swan (1956).


As noted in De la Fuente (2002), some studies have found that a one euro increase in public investments increases private investments by as much as twenty cents.


    BarroR. and S.Sala-i-Martin1991Convergence across States and Regions,Brookings Papers on Economic Activity Vol. 1 (Washington: Brookings Institution).

    • Search Google Scholar
    • Export Citation

    BarroR. and S.Sala-i-Martin1992Convergence,Journal of Political Economy Vol. 100 (2) pp. 223251.

    BoldrinMichele and FabioCanova2000Inequality and Convergence: Reconsidering European Regional Policies,” (unpublished; Minneapolis: University of Minnesota and Barcelona: Universitat Pompeu Fabra).

    • Search Google Scholar
    • Export Citation

    CaselliFrancesco and John WilburColeman1999How Regions Converge,Discussion Paper No. 2191 (United Kingdom: Center for Economic Policy Research).

    • Search Google Scholar
    • Export Citation

    CicconeA.G.Peri and D.Almond1999Capital, Wages and Growth: Theory and Evidence” (unpublished; Barcelona: Universitat Pompeu Fabra and Berkeley, Calif.: University of California, Berkeley).

    • Search Google Scholar
    • Export Citation

    De la FuenteAngel2001On the Sources of Convergence: A Close Look at the Spanish Regions,European Economic Review Vol. 46 (3) pp. 569599.

    • Search Google Scholar
    • Export Citation

    De la FuenteAngel2002The Effect of Structural Fund Spending on the Spanish Regions: An Assessment of the 1994-99 Objective,Discussion Paper No. 3673 (United Kingdom: Center for Economic Policy Research).

    • Search Google Scholar
    • Export Citation

    DemurgerS.2000Infrastructure Development and Economic Growth: An Explanation for Regional Disparities in China?Journal of Comparative Economics Vol. 29 (1) pp. 95117.

    • Search Google Scholar
    • Export Citation

    DemurgerS.J.D.SachsW.T.WooS.BaoG.Chang and A.Mellinger2002Geography, Economic Policy, and Regional Development in China,NBER Working Paper 8897 (Cambridge, Mass.: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    FleisherB. and J.Chen1997The Coast-Noncoast Income Gap, Productivity and Regional Economic Policy in China,Journal of Comparative Economics25 (2) pp. 220236.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund2003Special Economic Zones—Costs and Benefits”,FAD Guidance Note (unpublished; Washington).

    KrugmanP.1991Increasing Returns and Economic Geography,Journal of Political Economy99 (3) pp. 483499.

    KrugmanP. and A.Venables1995Globalization and the Inequality of Nations,Quarterly Journal of Economics110 (4) pp. 857880.

    • Search Google Scholar
    • Export Citation

    MartinPhilippe1997Can Regional Policies Affect Growth and Geography in Europe” (unpublished; Geneva: Graduate Institute of International Studies).

    • Search Google Scholar
    • Export Citation

    McKinsey Global Institute2003available on Internet at:

    ModyA. and WangF.1997Explaining Industrial Growth in Coastal China: Economic Reforms…and What Else?The World Bank Economic Review11 (2)).

    • Search Google Scholar
    • Export Citation

    OatesW.1999An Essay on Fiscal Federalism,Journal of Economic Literature37 (3) pp. 11201149.

    PrinchettL.1997Divergence, Big Time,Journal of Economic Perspectives11 (3) pp. 317.

    ShankarR. and A.Shah2001Bridging the Economic Divide within Nations: A Scorecard on the Performance of Regional Development Policies in Reducing Regional Income Disparities” (unpublished; Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    SolowR.1956A Contribution to the Theory of Economic Growth,Quarterly Journal of Economics70 (1) pp. 6594.

    SwanT.1956Economic Growth and Capital Accumulation,Economic Record XXXII pp. 334361.

    World Bank1996“The Changing Fortunes of Sub-National Regions” (unpublished; Washington: World Bank.

Other Resources Citing This Publication