Advertisement

Dutch Disease in Post-Soviet Oil Exporting Countries: Impact of Real Appreciation on De-industrialization

  • Kwansik YunEmail author
Original Article
  • 112 Downloads

Abstract

With the recent economic slowdown, there has been a growing concern that de-industrialization might result from resource abundance in the post-Soviet region. By using a comparative panel analysis of three countries of the post-Soviet region and 14 countries of selected oil exporters, this paper investigates whether real appreciation for the oil exporters contributed to de-industrialization in the region. We fail to find any evidence to support the claim that real appreciation caused de-industrialization in the region.

Keywords

Dutch disease Oil price De-industrialization Exchange rate 

JEL Classification

P28 P51 O14 

1 Introduction

Whether the abundance of natural resources is a blessing for the economy or not has been the interest of many scholars over the year. Some claimed that resource abundance is a curse, while others argued that it could be a blessing. Some of the post-Soviet oil exporting countries are not exceptions to this debate.

In 1982, Corden and Neary proposed the model of Dutch disease, which was again discussed by Corden (1984) and many other scholars. At its core, the theory of Dutch disease suggests that the resource dependency of resource-rich countries can harm the economy through the de-industrialization in the long run. Ever since the proposal of Dutch disease, there have been researches supporting the theory in different countries (Corden and Neary 1982; Corden 1984).

According to the theory of Dutch disease, resource dependency causes the de-industrialization mainly through two different channels; ‘the resource movement effect’ and ‘the spending effect.’ ‘The resource movement effect,’ also known as ‘direct de-industrialization,’ refers to the direct movement of labor factor from manufacturing sectors to the oil sectors. ‘The spending effect,’ known as ‘indirect de-industrialization’ refers to the indirect movement of labor factor from manufacturing sectors to the non-tradable sector through real appreciation (Corden and Neary 1982; Corden 1984).

The post-Soviet countries are not exceptions. A large number of researches tried to determine whether the post-Soviet resource-rich countries had been vulnerable to Dutch disease, but no consensus was reached due to the lack of data and other distinct features of the region.

Despite a lack of consensus on the existence of Dutch disease, there seemed to be an agreement on the correlation between the resource boom and the real appreciation of exchange rate. Oomes and Kalcheva (2007) stated that there exists a correlation between oil price and the real exchange rate in Russia. Also, Mironov and Petronevich (2015) found the positive correlations between oil price and real effective exchange. However, not many types of research were conducted on de-industrialization due to lack of data.

Besides, a few existing kinds of literature on the topic have different views. Hasanov (2013) claimed relative de-industrialization of the manufacturing sector in Azerbaijan. Mironov and Petronevich (2015) and E’gert (2012) argued that there exists de-industrialization in Russia. Oomes and Kalcheva (2007) also agreed with the existence of de-industrialization in Russia but argued that it could be due to other factors. Dobrynskaya and Turkisch (2010) and Katsuya (2017) claimed that there are no signs of de-industrialization in Russia.

Therefore, instead of focusing on the matter, which has already reached consensus, this paper focuses on de-industrialization of the manufacturing sector, in particular, the correlations between de-industrialization and real appreciation. In addition, the model that reflects the whole process of the theory of Dutch Disease was implemented in order to draw the whole picture. Regarding the region, this paper analyzes the presence of Dutch disease in post-Soviet states, especially in three major oil exporting countries: Russia, Kazakhstan, and Azerbaijan.

The remainder of this study is organized as follows. Section 2 provides a literature review. Section 3 describes the empirical model and data. Section 4 presents and discusses the main results. Section 5 concludes the study.

2 Literature Review

2.1 The Model of Dutch Disease

The model of Dutch disease was first proposed by Corden and Neary (1982). Dutch disease model suggests that a strong dependency on natural resources can harm the economy’s potential in the long run through the process of de-industrialization. De-industrialization refers to the decline of the manufacturing industry. In the original model of Dutch disease, several assumptions were made.

One of the major assumptions of the Dutch disease model is a small and open economy. Small and open economy suggest that the price of tradable goods will be subjected to the world price but would not have enough power to affect the world price. Dutch disease model also assumes that the economy consists of three different sectors: two tradable sectors and the non-tradable sector. Two tradable sectors refer to the manufacturing sector and the booming sector. The non-tradable sector refers to the service sector (Corden and Neary 1982).

All factor movement, including the labor movement, is perfectly mobile within the economy, allowing the wage to be equalized. However, factor movement is immobile internationally. Furthermore, all goods are used for final consumption only. The price increase is called as ‘boom.’ Therefore, ‘resource boom’ indicates the increase in natural resource price. Resource boom affects economy mainly through two channels; ‘the resource movement effect’ and ‘the spending effect’ (Corden and Neary 1982, P830).

When experiencing a boom in the resource sector, the marginal productivity of labor in the booming sector improves. Increased marginal productivity drives up the wage, naturally attracting more labor. Under the assumption that labor is mobile within the country, an increase in labor demand would lead to a shift of labor from the oil sector. This is called ‘the resource movement effect.’ As briefly mentioned above, ‘the resource movement effect’ happens directly without the appreciation of the real exchange rate and the non-tradable sector. Therefore, it is also called as ‘direct de-industrialization’ (Corden 1984, P361).

Also, the resource boom leads to increased income due to marginal productivity gain. Higher income would yield an increased demand for goods and services. Since the price of manufacturing products is determined in the international market, this increased demand raises the price of the non-tradable sector. As it is assumed that the economy only consists of three sectors: a non-tradable sector and two tradable sectors, and increased price of non-tradable goods indicates a real appreciation (Corden 1984, P361).

The real appreciation and increased price of non-tradable goods trigger the increase in demand for labor within non-tradable sector. Such increase in demand for labor will lead to the additional labor movement from the manufacturing sector to the non-tradable sector. This is ‘the spending effect.’ Since ‘the spending effect’ happens through real appreciation and includes non-tradable sector within the process, it is also called as ‘indirect de-industrialization.’

In short, ‘the resource movement effect’ directly causes de-industrialization through labor movement from the manufacturing sector to the booming sector. ‘The spending effect’ indirectly causes de-industrialization through real appreciation and labor movement from the manufacturing sector to the booming sector. Through both channels, labor migrates toward non-tradable sectors and oil sector from the manufacturing sector as shown in Fig. 1 (Corden 1984, P361).
Fig. 1

The model of Dutch disease

Ever since the proposal of the model, many resource-rich countries were diagnosed to suffer from Dutch Disease. In particular, many oil exporters were subjected to the symptom. Netherland was one of the most well-known examples. After the discovery of natural gas in the 1960s, Netherland enjoyed the oil revenue in the short run. However, starting from the 1970s, the country suffered severely due to the negative effect on the manufacturing sector (Corden and Neary 1982).

As shown in Table 1, many oil exporting countries are known to have a weak manufacturing sector. Besides, the share of manufacturing sector among these countries tends to decrease even further. According to the theory of Dutch disease, such a trend can be the result of resource dependency and potentially harm the economy through the de-industrialization in the long run.
Table 1

Share of manufacturing output in GDP by countries of selected oil exporters

Country

2012

2013

2014

2015

2016

Algeria

3.71

3.85

4.05

4.52

4.57

Angola

4.21

4.61

4.83

5.22

4.88

Brunei Darussalam

16.62

16.50

16.13

14.53

11.46

Colombia

12.22

11.82

11.48

11.40

11.60

Gabon

4.95

5.15

2.84

2.81

 

Islamic Republic of Iran

14.19

13.61

13.85

12.45

12.05

Kuwait

5.33

5.27

4.9

5.84

5.34

Libya

3.65

3.65

3.65

3.65

3.65

Nigeria

7.70

8.93

9.64

9.43

8.68

Oman

9.98

10.13

9.1

9.2

9.48

Qatar

10.31

9.99

9.87

9.33

8.66

Saudi Arabia

9.79

9.93

10.8

12.42

12.57

Trinidad and Tobago

18.88

16.05

15.97

15.32

15.14

Venezuela (Bolivarian Rep. of)

11.77

11.51

12.07

  

The data were obtained from World Bank. Unit is %. Display until the second decimal place

Research shows that not only the manufacturing sector, but also other tradable sectors, can be the victim of resource dependency as well. In 2014, the study on the effect of oil rents on agriculture in the Middle East and North African countries was done by Apergis (2014). Through his research, he proved that not only the manufacturing sector but also other tradable sectors could be suffered due to the resource boom.

However, not every country that was blessed with the abundant natural resource were suffered from the symptom. With appropriate policy practices, many countries have avoided the resource curse. Figure 2 shows that among selected oil exporters, some countries were able to maintain their manufacturing sector while others show significant de-industrialization.
Fig. 2

Share of manufacturing output in GDP by countries of selected oil exporters (unit: %). Average share of manufacturing between 2012 and 2016.

Source: World Bank and UNSTATS

Also, among countries that were subjected to Dutch disease, there exist ones that were able to overcome the symptom of Dutch disease as well. Netherland is one particular case. As mentioned above, Netherland suffered from de-industrialization after the discovery of natural gas in the 1960s (Corden 1984). However, Netherland was able to overcome the Dutch disease with the proper policy management later on.

Regarding the proper policy direction toward batting Dutch disease, there seemed to be consensus on the utilization of oil revenue on developing other industries. Apergis (2014) stated that exchange rate control as a response to Dutch disease is inefficient and highly distortionary. Ploeg (2013) also argued that the fixed exchange rate could worsen the symptoms of Dutch disease. He suggested the use of a flexible exchange rate in order to avoid Dutch Disease. Pomfret (2012) suggested that the utilization of sovereign-wealth was one of the key policies that helped Azerbaijan and Kazakhstan to avoid Dutch disease.

2.2 Dutch Disease in Post-Soviet Oil Exporting Countries

Oil exporters in post-Soviet region, particularly Russia, Kazakhstan, and Azerbaijan, are one of the major players in the oil market. In 2017, Russia, Kazakhstan, and Azerbaijan were ranked as 3rd, 16th, and 25th largest producer of total petroleum and other liquids, respectively. Russia produced over 11,210 thousand barrels per day, Kazakhstan produced over 1879 thousand barrel per day, Azerbaijan produced over 799 thousand barrels per day (U.S Energy Information Administration 2018).

As resource-rich countries showed growing dependency on natural resources, there has been a concern on the possibility of the long-term adverse impact of resource dependency in the region. For instance, in 2017, oil export accounted for 72.3% of the entire export in Azerbaijan (World Bank 2018).

In addition to such high dependency on natural resources, these countries were not able to perform better regarding economic growth, compared to other members of post-Soviet states. Figure 3 shows the GDP growth rate of Russia, Kazakhstan, Azerbaijan and other former members of Soviet Union. Besides Azerbaijan, where the GDP growth rate is extremely volatile, there does not seem to be a significant difference between major oil exporters and other countries.
Fig. 3

GDP growth of post-Soviet oil exporting countries (unit: %).

Source: World Bank

As mentioned earlier, there has been a continuous debate on the existence of Dutch disease in these countries. Many scholars have suggested that there exist the symptoms of Dutch Disease in the region, while others took the different stance regarding the issue. In an effort to identify the existence of Dutch disease, many scholars have focused on capturing the impact of indirect de-industrialization channel.

Hasanov (2013) claimed the existence of relative de-industrialization of the manufacturing sector in Azerbaijan, mainly due to ‘the spending effect.’ Kutan (2005) argued that there exist the signs of Dutch disease in Kazakhstan, mainly due to real appreciation resulted from capital inflows. However, he also mentioned that as a transitioning economy, their currency was significantly undervalued after their independence. He also mentioned that it is not necessarily an immediate concern and there is no need for policy action to counter the symptom. E’gert (2012) also argued that Central and South-West Asia shows the symptoms of Dutch Disease. Through his study, he argues that though the presence of resource movement and spending effect has not established yet, with a time lag, increased oil price leads to real appreciation.

Regarding Russia, Dülger (2013) argued that Russia shows signs of Dutch disease, in particular, the real appreciation caused by the increased price of oil and relative de-industrialization. However, in his study, he suggested that de-industrialization is likely to result from the eradication of soviet disease, which refers to the underdevelopment of the service sector. Observing imports from EU 25 countries, Barisitz and Ollus (2007) claimed that at least regarding trade, there exist symptoms of Dutch disease.

On the contrary, other scholars have argued that resource-rich countries in the region are currently not subjected to symptoms of Dutch disease. Katsuya (2017) claimed that though, real appreciation of exchange rate exists, there seemed to be no signs of Dutch disease since manufacturing output continues to rise. Dobrynskaya and Turkisch (2010) also claimed that the appreciation of ruble exists, but the manufacturing sector is rather increasing.

Mironov and Petronevich (2015, P108) took a neutral stance regarding the existence of Dutch disease in Russia. Through their research, they suggested that Russia shows several signs of Dutch disease. However, the manufacturing sector still shows positive growth, which can be possible due to ‘the eradication of soviet disease.’ Pomfret (2012) stated that Central Asian countries, including Kazakhstan and Azerbaijan, are not suffering from the severe resource curse. Tabata (2013) argued that Dutch disease presents in Russia; however, symptoms are offset by significant differences between internal and external oil price.

As shown above, a consensus has been made in the existence of a positive correlation between real exchange rate appreciation and an increase in oil price, which can be observed from Fig. 4. Not only scholars who agreed on the existence of Dutch disease in the region, but also scholars who do not agree to Dutch disease admitted the existence of real exchange rate appreciation.
Fig. 4

Real effective exchange rate of post-Soviet oil exporting countries and global oil price (1997 = 100, annual).

Source: IMF and World Bank

2.3 De-industrialization

However, regarding the de-industrialization of the manufacturing sector, there have been debates. In Azerbaijan, Hasanov (2013) claimed that ‘the relative de-industrialization’ rather than ‘the absolute de-industrialization’ has taken place. Furthermore, he suggested that direct de-industrialization is not significant and the increase in real wage was conspicuous. E’gert (2012) argued that with a time lag, increased price of oil leads to the real and nominal appreciation of currency in Central and South-West Asia. As evidence, he presented the recent decline of manufacturing share over GDP along with economic development and the increasing importance of the mining sector in oil exporters of the region. Furthermore, Oomes and Kalcheva (2007) found the relative de-industrialization in the region by showing a slowdown in manufacturing growth. However, they noted that there is no direct evidence that shows a clear relationship between real appreciation and de-industrialization.

To the contrary, Katsuya (2017) suggested that the manufacturing sector continues to grow positively even with an appreciation of the real effective exchange rate in Russia. Thus, de-industrialization does not exist. Mironov and Petronevich (2015) also claimed that the manufacturing sector shows positive growth even with an appreciation of the real effective exchange rate.

As shown above, even those who suggested the de-industrialization admits that the link between oil boom and real appreciation of the currency. However, the link between de-industrialization and real appreciation is not clear in the region. Regarding the correlations between real appreciation and the de-industrialization, Hua (2007) proved the significant negative correlations of the real effective exchange rate on manufacturing employment in China. Hua suggested that real appreciation leads to the decline of manufacturing employment through three different channels; ‘the technological channel,’ ‘the export volume channel,’ and ‘the efficiency channel’ (Hua 2007, P339). Among three channels, two channels reflect the increased productivity in the manufacturing sector; ‘the technological channel’ and ‘the efficiency channel.’ ‘The export volume channel’ reflects a decrease of manufacturing export.

According to ‘the technological channel,’ the appreciation of the real exchange rate encourages the change of factors from workers to imported inputs, increasing labor productivity (Hua 2007, P340). Replaced human resources migrate to other sectors, as suggested in the model of Dutch disease. ‘The efficiency channel’ measures the impact of the real exchange rate appreciation through the improved efficiency of labor (Hua 2007, P342). Hua claimed that a real appreciation triggers efficiency improvement by increased real wage and international competition. ‘The export volume channel’ captures the impact of the real exchange rate appreciation through export (Hua 2007, P341). Worsened terms of trade decrease the export volume, having a negative impact on the manufacturing employment (Hua 2007, P341).

Through ‘the technological channel’ and ‘the efficiency channel,’ real appreciation triggers the productivity gains while causing de-industrialization. Through ‘the export volume channel,’ real appreciation negatively affects the manufacturing employment due to worsening terms of trade and decline of export, which will cause overall decline of manufacturing sector. Through three channels, Hua focused on the decline of manufacturing employment (Hua 2007, P339).

2.4 Contribution

As mentioned above, the vast majority of studies conducted on the region focused on the existence of a correlation between real appreciation and the oil boom. There have been debates on the existence of de-industrialization. Furthermore, the link between the real appreciation and de-industrialization remained ambiguous in the region. In order to identify the existence of Dutch disease and de-industrialization in the region, this paper will be using two model.

The first model focuses on identifying the correlation between real appreciation and de-industrialization. Such a link has not been clearly identified in the literature due to lack of data in the region. In an effort to identify the link, I have followed the model developed by Hua (2007). Modifications were made in order to generalize the model, due to lack of data.

The second model focuses on the whole picture of Dutch disease. On the basis of the first model, which identifies the indirect de-industrialization, this paper attempts to examine the traditional model of Dutch disease, following the work of Mironov and Petronevich (2015). Modifications were made in order to clarify the cause of labor migration from the manufacturing sector to the service sector.

3 Estimation Model and Data

3.1 Overview of Estimation Models

As mentioned earlier, Dutch disease can be explained through two channels; ‘direct de-industrialization’ and ‘indirect de-industrialization.’ Indirect de-industrialization includes the process of real appreciation. Since there have been many studies confirming the real appreciation caused by the oil boom, the first model solely focuses on the existence of de-industrialization caused by the real appreciation of the currency.

On the basis of the first model, which identifies the part of ‘indirect de-industrialization,’ I intended to capture the whole process of Dutch disease, using the second model. Following the process of both ‘the resource movement effect’ and ‘the spending effect,’ the second model identifies the impact of oil price on manufacturing output. With two models, this paper attempted to identify what was left from existing literature as well as covers the big picture of Dutch disease.

3.2 Estimation Model 1: Real Appreciation and Manufacturing Employment

Two different channels were used in order to explain the impact of real appreciation on manufacturing employment; ‘productivity channel’ and ‘terms of trade channel.’ Instead of using two different explanatory channels for the increased productivity, the paper attempted to capture the productivity gain in the manufacturing sector by using labor productivity variable.

Productivity channel refers to the increase in productivity exerted by the appreciation of the real exchange rate. As suggested by Fu and Balasubranmanyram (2005) and Hua (2007), the real appreciation triggers the efficiency gain in the manufacturing sector, since only the competitive suppliers can survive in the market. The real appreciation also encourages the change of factors from workers to imported inputs, increasing labor productivity (Hua 2007). Furthermore, replaced human resources migrates to other sectors, as suggested in the model of Dutch disease. Therefore, real appreciation causes de-industrialization, while causing productivity gain.

Terms of trade channel refers to the adverse impact on the trade resulted from worsened terms of trade. The appreciation of the real exchange rate worsens terms of trade. Due to unfavorable terms of trade, the manufacturing sector loses the price competitiveness in the world market, resulting in a decrease in manufacturing export. Furthermore, the increase in real wage exerts a higher demand for imported goods. Along with the appreciation of the real exchange rate, the demand for imported goods increases even further, leading to a decrease in labor demand in the manufacturing sector (Mironov and Petronevich 2015).

The dependent variable in this model is the share of manufacturing employment within the labor market. The real effective exchange rate was added as an independent variable to identify the correlation between real appreciation and de-industrialization. Price of oil was considered as an independent variable to identify whether there is a significant link between the increase in oil price and de-industrialization. In order to capture the significance of each channel, labor productivity, manufacturing output, manufacturing export, and real FDI stock are considered. Labor productivity variable, which is expressed as output per employed person, captures the productivity channel. As suggested by Hua (2007) terms of trade channel are expressed by three variables; manufacturing output, manufacturing export, and inward real FDI stock. Both manufacturing output and export reflect the performance of the manufacturing sectors. Real FDI stock was considered in order to capture the export of foreign-funded firms.

In a formula, this can be expressed as:
$${\text{LM}} = \alpha + \beta_{11} {\text{dlPOIL}} + \beta_{12} {\text{dlREER}}_{\text{it}} + \beta_{2} {\text{LP}}_{\text{it}} + \beta_{31} {\text{YM}}_{\text{it}} + \beta_{32} {\text{XM}}_{\text{it}} + \, \beta_{33} {\text{FDIS}}_{\text{it}} + \, \varepsilon_{\text{it}}$$
LM refers to the share of manufacturing employment in the labor market. REER refers to the real effective exchange rate. POIL refers to the price of oil. LP refers to labor productivity. YM refers to the share of manufacturing output in GDP. XM refers to the share of manufacturing export in GDP. FDIS refers to the foreign direct investment stock inflow. εit represents the error term. Time fixed effect was not considered.

3.3 Estimation Model 2: Oil Price and Manufacturing Output

Since the de-industrialization regarding the employment was considered through the first model, the second model measures the impact on the output of the manufacturing sector. Following the study of Mironov and Petronevich (2015), I attempted not only to identify the existence of Dutch disease but also draw the whole picture, regarding through which components de-industrialization took place.

Mironov and Petronevich (2015) used the employment in manufacturing and mining sectors as explanatory variables in order to measure ‘the resource movement effect.’ The disposable income growth and the employment of the service sector were used in order to capture ‘the spending effect.’ Since their study was conducted on Russia, in addition to variables reflecting different channels of Dutch disease, they controlled the employment of state-owned companies and the financial crisis in 2007 (Mironov and Petronevich 2015). Also, ‘capital accumulation’ was considered in order to reflect the transition from planned economy to the market economy (Mironov and Petronevich 2015, P104).

Several modifications were made on the existing model in order to generalize the model and better reflect the mechanism of Dutch disease. Since panel countries of both FSU oil exporters and the selected oil exporters were used, employment of state-owned companies and capital accumulation period were not considered. The financial crisis in 2007 was not considered as well. Price of oil was added as independent variable to identify the correlation between de-industrialization and the oil boom. Population growth variable was employed along with other explanatory variables suggested by Mironov and Petronevich (2015) to ensure that the manufacturing output growth was not due to the growth of labor force within the country.

Following the Balassa–Samuelson effect, the real effective exchange rate was considered in order to capture the process of ‘the spending effect.’ Balassa–Samuelson effect suggests that productivity growth in the booming sector increases real wage in the sector. Since the increase real wage spills over the economy, an increase in real wage in the sector triggers the appreciation of the real exchange rate (Asea and Corden 1994; Kutan 2005). The appreciation of the real exchange rate can cause additional labor movement toward the non-tradable sector, possibly resulting in the decline of the manufacturing sector (Corden 1984).

Manufacturing output was implemented as a dependent variable. Price of oil was used as an independent variable. Employment of manufacturing, mining, and service sector along with population growth was implemented to capture ‘the resource movement effect.’ GNI per capita and real effective exchange rate were used to capture ‘the spending effect.’

In a formula, this can be expressed as:
$$\begin{aligned} Y\_{\text{Man}} & = \alpha + \beta_{1} {\text{dlPOIL}} + \, \beta_{21} L\_{\text{serv}}_{\text{it}} + \, \beta_{22} L\_{ \hbox{min} }_{\text{it}} \\ & \quad + \, \beta_{23} L\_{\text{man}}_{\text{it}} + \, \beta_{24} {\text{POP}} + \, \beta_{31} {\text{GNIC}}_{\text{it}} + \, \beta_{32} {\text{dlREER}}_{\text{it}} + \, \varepsilon_{\text{it}} \\ \end{aligned}$$
Y_Man refers to the share of manufacturing output in GDP. POIL refers to the price of oil, L_serv refers to the employment of the service sector. L_min refers to the employment of the mining sector. L_man refers to the employment of the manufacturing sector. GNIC refers to GNI per capita growth. REER refers to the real effective exchange rate. εit represents the error term. As in the first model, the time fixed effect was not considered.

3.4 Data

Regarding the selection of each panel, the panel countries of post-Soviet oil exporting countries and the selected oil exporters were used. Russia, Kazakhstan, and Azerbaijan were used as the panel countries for post-Soviet oil exporters, since they are the top oil exporters in the region while meeting the criteria suggested by UNCTAD.1 The panel of selected exporters was chosen based on the UNCTAD criteria, which is based on the average share of fuel exports between 2013 and 2018. The share of fuel export should be greater than 50% of their total exports and greater than 0.1% of world total fuel exports. Though UNCTAD selected 22 countries in 2018 based on their criteria, only 14 countries were used in this paper due to lack of data. Table 2 presents the list of selected exporters of petroleum, based on UNCTAD criteria.
Table 2

The selected exporter of petroleum based on UNCTAD Criteria

Country code

List of countries

2370

Selected exporters of petroleum

012

Algeria

024

Angola

031

Azerbaijan

096

Brunei Darussalam

148

Chad

170

Colombia

178

Congo

226

Equatorial Guinea

266

Gabon

364

Iran (Islamic Republic of)

368

Iraq

398

Kazakhstan

414

Kuwait

434

Libya

566

Nigeria

512

Oman

634

Qatar

643

Russian Federation

682

Saudi Arabia

780

Trinidad and Tobago

795

Turkmenistan

810

Union of Soviet Specialist Republics

862

Venezuela (Bolivarian Rep. of)

Chad Congo, Equatorial Guinea, Iraq, Turkmenistan, are excluded due to lack of data. Union of Soviet Specialist Republics are excluded. Russian Federation, Kazakhstan, and Azerbaijan are classified as post-Soviet oil exporting countries panel

Regarding the sample period, it begins from 1992 since the Union of Soviet Specialists Republics was dissolved in December 1991. The latest observations used are observations from 2017. Regarding the data classification for each sector, all dataset used in this paper were based on ISIC classification, besides the selection of the panel countries. UNCTAD classification for selected exporters of petroleum, which is based on SITC classification was used in panel selection.

Tables 3 and 4 provide the descriptive statistics for the respective panel data set. Regarding the oil price, the global price of Brent crude was used in all countries, instead of reflecting the local price of oil. Data on the global price of Brent crude was obtained from IMF. Unit of the oil price is US dollar per barrel. Real effective exchange rate index variable is based on the consumer price index. 1997 was selected as the base year within all countries, besides Brunei Darussalam. In Brunei Darussalam, 2005 was selected as a base year due to lack of data. Besides Angola and Brunei Darussalam, where data were obtained from ISDB and UNCTAD respectively, Data on the real effective exchange rate index were collected from IMF.
Table 3

Descriptive statistics for selected oil exporters

Variables

Obs.

Mean

SD

Min

Max

Share of manufacturing employment (LM)

364

8.54

4.50

1.70

19.50

Real effective exchange rate (REER)

352

108.32

107.57

22.61

1402.30

Price of oil (POIL)

350

49.72

33.74

12.72

111.96

Output per employed person (LP)

364

39,693.57

28,196.66

4184.00

110,427.00

Manufacturing output (YM)

350

10.29

5.21

1.95

37.51

Manufacturing export (XM)

289

237.32

774.66

0.00

5625.60

Real FDI stock (FDIS)

360

24.25

23.93

− 6.31

115.37

Share of manufacturing output in GDP (Y_man)

350

10.29

5.21

1.95

37.51

Employment in service (L_serv)

363

4341.48

5457.87

83.00

28,395.00

Employment in mining (L_min)

364

76.49

60.52

4.00

280.00

Employment in manufacturing (L_man)

363

877.31

1203.48

5.00

5155.00

Population growth (POP)

360

2.49

2.13

− 0.04

16.33

GNI per capita growth rate (GNIC)

351

12,280.51

15,082.04

170.00

83,699.74

All data are displayed until the second decimal place. Real effective exchange rate is based on consumer price index Global price of oil is expressed as US dollar per barrel. Output per employed person is in US dollar. Manufacturing output, manufacturing export, and real FDI stock are in US million dollar. Employment in each sector are in thousand people. GNI per capita is in growth rate is the percentage changes of GNI per capita

Table 4

Descriptive statistics for post-Soviet oil exporting countries

Variables

Obs.

Mean

SD

Min

Max

Share of manufacturing employment (LM)

78

9.79

5.71

3.70

20.40

Real effective exchange rate (REER)

75

94.56

25.23

7.39

138.50

Price of oil (POIL)

75

49.72

33.92

12.72

111.96

Output per employed person (LP)

78

13,500.14

6015.88

3105.00

23,199.00

Manufacturing output (YM)

75

12.92

5.34

4.22

23.52

Manufacturing export (XM)

62

18.25

26.10

0.01

86.52

Real FDI stock (FDIS)

73

36.80

26.82

0.04

118.09

Share of manufacturing output in GDP (Y_man)

75

12.92

5.34

4.22

23.52

Employment in service (L_serv)

78

15,957.35

18,831.91

1249.00

47,626.00

Employment in mining (L_min)

78

530.15

595.79

23.00

1566.00

Employment in manufacturing (L_min)

78

4244.60

5578.51

126.00

14,726.00

Population growth (POP)

78

0.47

0.92

− 1.75

2.64

GNI per capita growth rate (GNIC)

76

4651.97

3981.53

400.00

15,200.00

All data are displayed until the second decimal place. Real effective exchange rate is based on consumer price index Global price of oil is expressed as US dollar per barrel. Output per employed person is in US dollar. Manufacturing output, manufacturing export, and real FDI stock are in US million dollar. Employment in each sector are in thousand people. GNI per capita is in growth rate is the percentage changes of GNI per capita

Data for employment in each sector and labor productivity were obtained from the International Labor Organization. Data for variables reflecting manufacturing output and export, population growth and GNI per capita growth were obtained from World Bank. However, manufacturing output in Angola, Trinidad and Tobago, and Russia was collected from UNSTATS. The Data regarding Real FDI stock is available in UNCTAD.

4 Results

4.1 Empirical Results

Table 5 shows that within the panel of selected oil exporters, there does not exist significant correlations between real effective exchange rate and the share of manufacturing employment in the labor market. Furthermore, the price of oil does not have a significant impact on the manufacturing employment. Therefore, in the panel of selected oil exporters, the real appreciation does not seem to cause de-industrialization, particularly regarding employment share in the labor market. One interesting fact that can be drawn from the result is that there does not seem to exist a significant correlation between the manufacturing employment and manufacturing output along with manufacturing export. This suggests that the positive performance of the manufacturing sector does not lead to the growth of manufacturing employment share in labor market.
Table 5

Summary of regression analysis on real appreciation and manufacturing employment

Panel countries

Selected exporters of petroleum

Post-Soviet oil exporting countries

LM

Robust

Robust

Coefficients

SE

P value

Coefficients

SE

P value

dlREER

− 0.0143178

0.561554

0.799

− 0.022235

0.05861

0.706

dlPOIL

− 0.1117809

0.3489471

0.749

0.1773371

0.3425577

0.607

LP

0.00000999

0.000016

0.532

0.0000515

0.0000348

0.145

YM

− 0.0707939

0.479462

0.141

0.3812376***

0.0730052

0.000

XM

− 0.2061034

0.1984442

0.300

123.9979**

61.79556

0.050

FDIS

0.018946***

0.8123834

0.001

0.142981***

0.0032657

0.000

Constant

8.597088

0.8123934

0.000

4.27224

1.330008

0.002

 

Number of observations = 271

Number of observations = 62

 

R-squared = 0.8937

R-squared = 0.9793

 

Adjusted R-squared = 0.8857

Adjusted R-squared = 0.9762

Dependent variable: employment of manufacturing sector

*, **, *** Significance at 10%, 5%, and 1% respectively

Regarding the panel of post-Soviet oil exporting countries, Table 5 also suggests that there does not exist significant correlations between real appreciation of the currency and manufacturing employment. Unlike in the panel of selected oil exporters, positive correlations between the share of manufacturing employment in the labor market and manufacturing output as well as manufacturing export and real FDI stock. Such correlation suggests that within post-Soviet oil exporting countries, the positive performance of the manufacturing sector leads to the employment growth. Regarding the existence of the link between real appreciation and de-industrialization, as it was in the panel of selected oil exporters, panel countries of post-Soviet oil exporters do not show the correlation between appreciation of real effective exchange rate and de-industrialization.

Table 6 suggests that there exist significant negative correlations between the price of oil and the share of manufacturing output in GDP within the panel of selected oil exporters. Such a result indicates that the oil boom leads to the decline of manufacturing output, suggesting the existence of Dutch disease within the panel countries of selected oil exporters. However, the existence of both spending channel and the resource movement channel are not confirmed, since the impact of mining employment growth and real appreciation on manufacturing output was not confirmed. Furthermore, the significant negative correlations between service employment and manufacturing output can possibly be the result of changes in economic structure.
Table 6

Summary of regression analysis on oil price and manufacturing output

Panel countries

Selected oil exporters

Post-Soviet oil exporting countries

Y_man

Robust

Robust

Coefficients

SE

P value

Coefficients

SE

P value

dlPOIL

− 1.034398**

0.4106246

0.012

− 0.43126

0.8586232

0.617

L_serv

− 0.0007367***

0.0001852

0.000

− 0.0005124***

0.0001902

0.009

L_min

− 0.0053139

0.0047521

0.264

0.0096422**

0.0041695

0.024

L_man

− 0.000317

0.0006747

0.639

− 0.0000662

0.0003885

0.865

POP

0.1729557**

0.0781067

0.028

0.994549**

0.4559059

0.033

GNIC

0.0000201

0.0000173

0.247

− 0.000551***

0.000151

0.001

dlREER

− 0.0549915

0.0871772

0.529

− 0.0987581

0.0983454

0.319

constant

13.50075

0.608139

0.000

18.09978

2.43607

0.000

 

Number of observations = 319

Number of observations = 72

 

R-squared = 0.8427

R-squared = 0.8675

 

Adjusted R-squared = 0.8322

Adjusted R-squared = 0.8483

Dependent variable: employment of manufacturing sector

*, **, *** Significance at 10%, 5%, and 1% respectively

Regarding the panel of post-Soviet oil exporters, there does not exist a significant correlation between the price of oil and the manufacturing output. However, significant negative correlations between service employment and manufacturing output are observed along with significant negative correlations between GNI per capita growth and manufacturing output. Besides, the significant positive correlations between the employment of the mining sector and the output of the manufacturing sector were observed.

Such positive correlations between manufacturing output and mining employment are opposite to the model of Dutch disease. Furthermore, the result suggests that the decline of the manufacturing sector is neither due to the oil price increase nor real appreciation. The decline of manufacturing seemed to have resulted from the gradual growth of the service sector within the economy along with the economic development and increase in income.

Figure 5 also suggest that within all three countries, the noticeable growth of mining sector employment is not observed. Also, the manufacturing sector in both Kazakhstan and Azerbaijan does not show significant decline within the given period. In the case of Russia, the rapid decline of manufacturing took place in the earlier period of development. However, the manufacturing employment was stabilized, since 2009. Such stabilization seemed to be the result of government effort toward structural reform in order to lower oil dependency.
Fig. 5

Sectoral employment in Azerbaijan, Kazakhstan, and Russian Federation

In short, the panel of selected oil exporters showed the signs of Dutch disease. However, the existence of both ‘the resource movement effect’ and ‘the spending effect’ seemed to be ambiguous. Regarding the channel of De-industrialization within the panel, further studies are required.

Contrary to the panel of selected oil exporters, the panel of post-Soviet oil exporters did not show the symptoms of Dutch disease. Relative de-industrialization regarding the share of manufacturing output in GDP was observed. However, such relative de-industrialization could potentially be influenced by the rapid growth of the service sector along with economic growth as shown in Fig. 5.

4.2 Implications

Even though post-Soviet oil exporting countries do not show the signs of Dutch disease along with the de-industrialization caused by real appreciation of the currency, relative de-industrialization presents in the region. Compared to OECD average, the share of manufacturing output is significantly low, particularly in the case of Azerbaijan. Therefore, the region needs to utilize their oil revenue appropriately in order to nurture the manufacturing industry.

Also, in the case of Kazakhstan and Azerbaijan, over 50% of the export revenue is from the export of oil. Especially, the share of oil export in Azerbaijan recorded 72.30% in 2017. Therefore, to lower the vulnerability related to the oil price, the Export portfolio needs to be diversified as well.

As shown in Fig. 4, Russia has a relatively balanced economic structure, compared to Kazakhstan and Azerbaijan. The share of manufacturing employment is still low, but it has been stabilized. Also Figs. 6 and 7 shows that the share of manufacturing GDP is still low, but somewhat close to the OECD average. Export revenue portfolio is relatively diverse as well. However, in the case of Azerbaijan, the country is facing a great need for structural reform. As shown in Fig. 8, the country not only shows the high dependency on oil revenue but also shows the significant volatility regarding economic growth, following the volatility of oil price.
Fig. 6

OECD share of manufacturing in GDP. Average share of manufacturing between 2012 and 2016 (unit: %).

Source: OECD

Fig. 7

Share of manufacturing GDP in post-Soviet oil exporting countries. Average share of manufacturing between 2012 and 2016 (unit: %).

Source: World Bank

Fig. 8

Global price of oil and GDP growth of Azerbaijan.

Source: IMF and World Bank

However, a positive trend was observed as well. Though it is not significant, in all three countries, the recent trend of growing manufacturing share is observed as shown in Fig. 7. For the stability of the economy, such growth of the manufacturing sector needs to be continued in the long run.

5 Conclusion

This paper examined the existence of de-industrialization caused by Dutch disease and impact of real appreciation on de-industrialization in post-Soviet oil exporters; Russia, Kazakhstan, and Azerbaijan. The empirical evidence indicates that the real appreciation does not lead to de-industrialization in post-Soviet oil exporting countries. Furthermore, the oil exporters of post-Soviet states are not vulnerable to Dutch disease. Even though the relative de-industrialization has taken place in the region, it is not likely to be the result of Dutch disease. Combining with descriptive evidence, the empirical results suggested that the relative de-industrialization is rather a result of rapid service sector growth, following the economic growth.

Regarding each country, Azerbaijan showed the most significant de-industrialization among three countries. Considering the significant dependency on oil revenue, the country needs an urgent political measure on structural reform. Both Kazakhstan and Russian Federation showed relatively less dependency on the oil revenue. Besides, both countries recently showed the growing trend of manufacturing sector regarding the share of output. Regarding the share of employment, manufacturing sector also seemed to be stabilized. Such a trend suggests that even though relative de-industrialization took place in the region, both countries no longer shows the symptom of de-industrialization.

There exist several limitations of the study. The first limitation is a lack of data. Due to lack of data, many countries were excluded from the panel. Even after elimination of panel countries, the data set is still not complete. Secondly, for the convenience of panel studies, the global price of Brent Crude was used as a reference for oil price in all countries. Due to the different prices of each oil, this can cause potential issues, not fully reflecting the oil price of each region. The third limitation is regarding the first model on the impact of real appreciation on de-industrialization. In this model, instead of labor productivity in the manufacturing sector, the average labor productivity was used due to lack of data. Lastly, further studies are needed on the matter of correlation between de-industrialization and service sector growth.

Footnotes

  1. 1.

    Developing and transition economies of which from 2013 to 2018 the average share of fuels exports (SITC 3) was

    • greater than 50% of their total exports (SITC 0-9) and

    • greater than 0.1% of the world total fuels exports (SITC3).

References

  1. Apergis, N. 2014. Dutch Disease Effect of Oil Rents on Agriculture Value Added in Middle East and North African (MENA) Countries. Energy Economics 45: 488–489.CrossRefGoogle Scholar
  2. Asea, P., and W.M. Corden. 1994. The Balassa–Samuelson Model: An Overview*. International Economics 2: 191–200.Google Scholar
  3. Barisitz, S., and S.E. Ollus. 2007. The Russian Nonfuel Sector: Signs of the Dutch Disease? Evidence from EU-25 Import Competition. Focus on European Economic Integration 1: 157–164.Google Scholar
  4. Corden, W.M. 1984. Booming Sector and Dutch Disease Economics: Survey and Consolidation. Oxford Economic Papers 36: 359–380.CrossRefGoogle Scholar
  5. Corden, W.M., and J.P. Neary. 1982. Booming Sector and De-Industrialization in a Small Open Economy. The Economic Journal 92: 826–842.CrossRefGoogle Scholar
  6. Dobrynskaya, V., and E. Turkisch. 2010. Economic Diversification and Dutch Disease in Russia. Post-Communist Economies 22: 298–300.CrossRefGoogle Scholar
  7. Dülger, F. 2013. Is Russia Suffering from Dutch Disease? Cointegration with structural break. Resources Policy 38: 610–611.CrossRefGoogle Scholar
  8. E’gert, B. 2012. Dutch Disease in the Post-Soviet Countries of Central and South-West Asia: How Contagious is it? Journal of Asian economics 23: 580–583.Google Scholar
  9. Fu, X., and V. Balasubranmanyram. 2005. Exports, Foreign Direct Investment and Employment: The Case of China. The World Economy 28: 607–625.CrossRefGoogle Scholar
  10. Hasanov, F. 2013. Dutch Disease and the Azerbaijan Economy. Communist and Post-Communist study 46: 474–476.CrossRefGoogle Scholar
  11. Hua, P. 2007. Real Exchange Rate and Manufacturing Employment in China. China Economic Review 18: 338–349.CrossRefGoogle Scholar
  12. Katsuya, I. 2017. Dutch Disease and Russia. International Economics 151: 66–70.CrossRefGoogle Scholar
  13. Kutan, A. 2005. Explaining the Real Exchange Rate in Kazakhstan, 1996–2003: Is Kazakhstan Vulnerable to the Dutch Disease? Economic Systems 29: 243–254.CrossRefGoogle Scholar
  14. Mironov, V., and A. Petronevich. 2015. Discovering the Signs of Dutch Disease in Russia. Resources policy 46: 100–110.CrossRefGoogle Scholar
  15. Oomes, N., & Kalcheva, K. (2007). Diagnosing Dutch disease: Does Russia have the symptoms? IMF Working Paper, 16–22.Google Scholar
  16. Ploeg, F. 2013. Absorbing a Windfall of Foreign Exchange: Dutch Disease Dynamics. Journal of Development Economics 103: 239–240.Google Scholar
  17. Pomfret, R. 2012. Resource Management and Transition in Central Asia, Azerbaijan and Mongolia. Journal of Asian Economics 23: 150–155.CrossRefGoogle Scholar
  18. Tabata, S. 2013. Observations on Russian Exposure to the Dutch Disease. Eurasian Geography and Economics 53: 237–242.Google Scholar
  19. U.S Energy Information Administration. 2018. https://www.eia.gov/beta/international/.
  20. World Bank. 2018. https://wits.worldbank.org.

Copyright information

© Asiatic Research Institute 2019

Authors and Affiliations

  1. 1.Graduate School of International StudiesKorea UniversitySeoulRepublic of Korea

Personalised recommendations