University of Amsterdam
Dutch disease in oil-exporting countries
The effects of oil exports on different sectors of the domestic economy
Bachelor Thesis Economics and Business
Author: Daniël Heijtel
Supervisor: Ron van Maurik
Table of contents
Introduction 3
Theoretical framework 7
• The Dutch Disease 7
• Positive effects of a boom in the oil sector on different sectors 11
Formulation of hypotheses 13 Data introduction 14 Data analysis 16 Conclusion 21 Discussion 23 Bibliography 26 Appendix 27 2
Introduction
Natural resources are one of the most important aspects of economics. No matter how advanced the global economy may become, natural resources will always be essential for the production of goods. Therefore, the possession of natural resources is of great importance for an economy. Knowing the importance of natural resources, one might argue that the possession of natural resources is a blessing for the economy of a country. However, a quick glance at the scientific economic literature reveals that there have been many countries whose economies actually suffered from the discovery of natural resources. This apparent paradox is known as the Natural Resource Curse.
In his paper ‘The Natural Resource Curse: A Survey’, Frankel (2010) gives an overview of several aspects of the Natural Resource Curse. It has been observed for several decades that the possession of natural resources like oil, gas and minerals does not necessarily confer economic success. Although many African countries are rich in natural resources, their peoples experience low level of incomes and a low quality of living, whereas some East-Asian economies have achieved Western-level standards of living with virtually no exportable natural resources. In general, the relationship between exports of primary products as a fraction of GDP and economic growth is slightly negative (Frankel, 2010, p. 3). Frankel states that this relationship is not very strong, masking almost as many resource successes as failures. However, this relationship certainly suggests no positive correlation between natural resource wealth and economic growth.
A well-known publication on the Natural Resource curse is the paper ‘Natural Resource Abundance and Economic Growth’ by Sachs and Warner (1995). In this paper, Sachs and Warner do not only suggest that there is no positive correlation between natural resources and economic growth, but even state that ‘One of the surprising features of economic life is that resource-poor economies often vastly outperform resource-rich economies in economic growth’ (Sachs and Warner, 1995, p. 3). According to these authors, this phenomenon can be observed throughout economic history. Sachs and Warner indicate that this negative association between resource abundance and growth poses a conceptual puzzle, since natural resources raise the wealth and the purchasing power over imports, so that resource abundance might be expected to raise an economy’s investment and growth rates as well (Sachs and Warner, 1995, pp. 3-4). However, after analyzing 97 developing countries over the period of 1971 to 1989, the authors find evidence for a statistically significant negative relation between natural resource intensity and subsequent growth, even after controlling for a large number of additional variables that other studies have claimed to be important in explaining cross-country growth. These additional variables include initial GDP, trade policy,
investment rates, terms of trade volatility, inequality, and the effectiveness of the bureaucracy. Sachs
and Warner also find that trade policy matters enormously for growth (Sachs and Warner, 1995, pp. 21-22).
However, there has been a lot of critique on the conclusions of Sachs and Warner. These scholars state that the low GDP growth in resource-abundant countries is actually caused by other factors than natural resource abundance itself. For example, Manzano and Rigobon (2001) state that the conclusions of the Sachs and Warner mentioned above are subject to econometric problems. Instead of a curse on natural resources, the bad economic performance of resource-abundant countries is due to the fact that these countries decided to take advantage of high commodity prices on the 70’s to use them as implicit collateral and found themselves on a debt overhang when commodity prices fell in the 80’s (Manzano and Rigobon, 2001, pp. 25-26).
Additionally, there are also papers indicating that the Natural Resource Curse simply does not exist. In the paper ‘The Resource Curse Revisited and Revised: A Tale of Paradoxes and Red Herrings’, Brunnschweiler and Bulte argue that there exists a discrepancy between the theory behind the Resource Curse and the empirical work used to support it. According to these authors, abundant resource rents are a crucial element in the theory, but the bulk of the empirics is based on resource dependency instead. Brunnschweiler and Bulte indicate that resource dependency is not a proper exogenous variable in regression analysis, and in treating it as an endogenous variable, it is no longer significant in growth regressions. In contrast, following their analysis these authors claim that greater resource abundance leads to better institutions and more rapid growth (Brunnschweiler and Bulte, 2007, p. 22). Brunnschweiler and Bulte conclude that ‘’the empirically significant relationship between institutional quality and resource dependence reflects that countries with poor institutions are unlikely to develop non-primary production sectors to reduce their dependence on resource exports. If so, the causality would be from institutions to dependence, and not the other way around. It would be inappropriate to talk about the “curse of resources” then. Instead, growth regressions in the resource curse literature may be viewed as a reminder of the important direct and indirect impacts of institutions on economic outcomes’’ (Brunnschweiler and Bulte, 2007, p. 23).
The papers mentioned above make clear that the scientific literature has not yet reached consensus on the consequences of natural resource abundance. However, knowing the economic consequences of natural resource exploitation is of vital importance for economic policy-making in resource abundant countries. This thesis will therefore try to shed some light on the economic consequences of resource exploitation. Frankel (2010, p.34) states that there are at least five
channels whereby natural resources might possibly have negative effects on economic performance. These channels are:
1. the high commodity price volatility of primary goods
2. the crowding out of the manufacturing sector by specializing in natural resources
3. civil war caused by the struggle for resources 4. poor institutions
5. the Dutch Disease
Unfortunately it is not feasible to evaluate all these channels in this thesis. Since this thesis will contribute more to the scientific literature when it thoroughly discusses a smaller part of the
resource curse rather than briefly discussing the entire theory, only a few of the channels mentioned by Frankel will be dealt with. The institutions channel and the civil war channel are more sociological than economic of nature, which makes these channels less interesting for this thesis. The high commodity price volatility is also not very interesting, since it has already been proved that oil and gas prices in particular, but also mineral and agriculture prices in general are far more volatile than prices of most manufactured products or services (Frankel, 2010, p.10). Therefore this thesis will only deal with the Dutch Disease and the crowding out of the manufacturing sector. These channels are highly interconnected, since the crowding out of the manufacturing sector also is a part of the effects of the Dutch Disease. This makes it relatively easy to thoroughly discuss and evaluate both channels in one thesis.
Nevertheless, even when focusing solely on these two channels, it is still very hard to draw interesting conclusions about these channels in general. To make the analysis more feasible, this thesis will therefore focus on the effects of the Dutch Disease and the crowding out of the manufacturing sector in a particular situation. This situation is the extraction of oil by major oil-exporting countries. The reason for choosing this case is twofold. Firstly, oil production is a well-known example of natural resource exploitation. There are relatively many countries whose economies rely for a significant extent on oil exports. Also there is more data available on the amount of oil exports than on the extraction of other natural resources. This makes the effects of oil exports by major oil-exporting countries relatively easy to analyze. Secondly, the literature on the Dutch disease indicates quite clearly that there indeed will be negative effects of natural resource extraction on economic performance of different economic sectors. However, there are reasons to expect that these effects may not apply to oil-exporting countries. These reasons are discussed extensively in the theoretical framework, so they will not be elaborated on in this section. This makes oil-exporting countries more interesting to analyze than other natural resource exploiting countries.
To summarize the previous section, what this thesis will deal with is the Dutch Disease in oil-exporting countries. As indicated by the literature mentioned above, the oil-oil-exporting countries might experience Dutch Disease effects from oil exports. The research question therefore reads as follows: does the Dutch Disease occur in major oil-exporting countries after an increase in oil exports?
To answer this research question, the next section is devoted to outlining the Dutch Disease. Afterwards, the positive effects of a boom in the oil sector on different sectors are described. Following this section, the hypotheses that are used to test the research question are formulated. Subsequently, the data that is used to test these hypotheses is introduced. Thereafter this data is used to test the formulated hypothesis. After testing these hypotheses, the results of the tests are summarized in the conclusion. Finally the limitations of this thesis are discussed in the last section.
Theoretical framework
The Dutch Disease
In the introduction it has been stated that the Dutch Disease is an important phenomenon in explaining the natural resource curse. This section takes a closer look at the Dutch Disease. In the subsequent part, the economic mechanisms behind the Dutch Disease and the effects of the Dutch Disease are described. Afterwards it is explained why these effects could be harmful for an economy.
The term Dutch Disease refers to the adverse effects on the Dutch manufacturing sector of the natural gas discoveries in the Netherlands in the nineteen sixties. After exploiting the newly found gas fields, The Dutch manufacturing sector suffered from the subsequent appreciation of the real exchange rate (Corden, 1984, p.359). To understand how an increase in endowments in one sector can lead to a fall in output in a different sector, it is very useful to first take a look at the Heckscher-Ohlin model. The Heckscher-Ohlin model assumes an economy with two countries producing two goods with two factors of production, namely labour and capital. The model further assumes that both goods require labour and capital, which are both mobile between sectors in the long run (Krugman, Obstfeld and Melitz, 2012, p.111). In general, the choices of the amount of labour and capital used in the production of a good depend on the factor prices for labour and capital. Furthermore, the model allows the possibility of substituting capital for labour and vice versa in production. However, the opportunity cost in terms of good of producing one more unit of good B rises as the economy produces more of good B and less of A (Krugman, Obstfeld and Melitz, 2012, p.112-113). Finally, the model assumes that good A requires relatively more labour compared to capital than good B. In other words, the production of good A is labour-intensive, while the production of good B is capital intensive (Krugman, Obstfeld and Melitz, 2012, p.115).
If this economy produces both good A and good B, competition among producers will ensure that the price of each good equals its cost of production. This cost of production depends on the prices of labour and capital: if wages rise, then the price of any good whose production requires labour will also rise. However, the importance of the price of labour to the cost of producing a good depends on the amount of labour required in the production of that good. Therefore, there is a one-to-one relationship between the ratio of the wage rate to the rental rate of capital, and the ratio of the price of good A of that of good B (Krugman, Obstfeld and Melitz, 2012, p.116). This implies that if the relative price of the labour-intensive good A were to rise, the ratio of the wage rate to the capital rental rate would also rise. Because labour is then relatively more expensive, the ratios of labour to capital employed in the production of good A and good B would drop (Krugman, Obstfeld and Melitz, 2012, p.117).
Now let us suppose that the economy is an oil-rich country that produces cloth as well. Natural resources like oil are generally considered to be capital-intensive products, since mining projects and oil installations can cost billions of dollars to build (Sadorsky, 2001, p. 17-18). Therefore oil would represent the previously mentioned good B, and cloth, as an example of a labour-intensive product, good A. It has been shown that a given relative price of oil is associated with a given wage-rental ratio, which determines the ratio of labour to capital employed in the cloth and oil sector. If we suppose that the capital stock suddenly grows, the economy’s aggregate labour to capital ratio decreases. However, at given relative price of cloth, we saw that the ratios of labour to capital employed in both sectors remain constant. The question then is how the additional capital stock is employed by the economy. Since the labour-capital ratio in the oil sector is lower than in the cloth sector, the economy can increase the employment of capital to labour by allocating more capital and labour to the production of oil, which is capital-intensive. As labour and capital move from the cloth sector to the oil sector, the economy produces more oil and less cloth (Krugman, Obstfeld and Melitz, 2012, p.118). Another way to see the results of an increase in capital is in terms of how capital and labour affect the economy’s production possibilities. After an increase in capital, the production possibility frontier shifts out. This enables the economy to produce more of both cloth and oil. However, the shift is strongly biased towards oil. When oil is highly capital-intensive and cloth production is highly labour-intensive, the expansion is so strongly biased towards oil production that at unchanged relative prices, the movement of production factors causes an actual fall in the output of cloth. This effect, where an increase in one factor of production leads to a more than proportional increase in the output of the good that uses that factor intensively, and an absolute decline in the output of the other good, is known as the Rybczynski theorem (Krugman, Obstfeld and Melitz, 2012, p.118-119).
The Rybczynski theorem could offer an explanation for the economic downfall of certain economic sectors after an increase in the exploitation of natural resources. However, the model makes use of some assumptions that may not hold in reality. For instance, it is very doubtful that all capital can move freely between sectors. Therefore it would be wise to take a look at a different model that could explain the Dutch Disease, but does not assume full capital mobility. Probably the most interesting model would be the Core Model, developed by Corden and Neary in 1982. This model uses booming and lagging sectors to describe the effects of a boom in a certain part of the economy on the economic performance of other economic sectors. The Core Model provides some insights that are essential for understanding the mechanisms behind the Dutch Disease.
This Core Model, as explained by Corden (1984), assumes an economy consisting of three sectors, namely a booming sector, a lagging sector, and a non-tradable sector. The booming and the lagging sectors produce goods that are tradable at world prices. Output in each sector is produced by
a factor-specific and labour, which is mobile between all three sectors and moves between sectors so as to equalize wages in all sectors. Further it is assumed that all factor prices are flexible and that all factors are internationally immobile (Corden, 1984, p.360).
A boom in the booming sector has the initial effect of raising incomes of the factors initially employed in the booming sector. The outcomes of this boom for the rest of the economy are split up in two effects, namely the Spending Effect and the Resource Allocation Effect. The Spending Effect indicates that a part of the additional income in the booming sector, whether directly by factor owners or indirectly via taxes and government spending is spent on goods produced by the non-tradable sector. This additional demand increases the prices of non-non-tradable goods. However, the prices of goods of the booming and lagging sectors are set internationally, so these prices cannot change. This causes a real appreciation of non-tradable goods, drawing away resources from the booming and the lagging sector, as well as shifting away demand from non-tradable goods to the booming and the lagging sector (Corden, 1984, p.360).
The Resource Allocation Effect assumes that the marginal product of labour in the booming sector rises as a result of the boom. Then, at a constant wage in terms of tradable goods, the demand for labour in the booming sector rises, which induces a movement of labour out of the lagging and the non-tradable sector. According to Corden, this effect has two parts. Firstly, the movement of labour out of the lagging sector and into the booming sector lowers output in the lagging sector. This effect is called direct de-industrialization. Secondly, there is a movement out of the non-tradable sector into the booming sector at a constant real exchange rate. The resource allocation effect reduces supply in the non-tradable sector, and thus brings about additional real appreciation. Since the price of the lagging sector is still set internationally and does not change, this causes additional labour to move out of the lagging sector into the non-tradable sector. This effect reinforces the shifting away of resources from the lagging sector caused by the Spending Effect. These two effects, shifting away labour from the lagging sector into the non-tradable sector, are called indirect
deindustrialization (Corden, 1984, pp. 360-361). Overall, the boom will attract more resources to the booming sector and thus increase output in this sector, while drawing away resources from the lagging sector and thus decrease output in this sector. The effects on the non-tradable sector are more ambiguous, since the amount of resources used by the non-tradable sector could go either go up or down; the Spending Effect would predict an increase in output, whereas the Resource
Allocation Effect would predict a decrease in output.
What the Rybczynski theorem and the Core model tell us, is that we can expect at least two negative effects from an increase in oil exploitation on different economic sectors. Firstly, the shifting of resources away from the lagging sector to the oil sector will lead to a decrease in output in the lagging sector. Secondly, we can expect an increase in the real exchange rate, which makes the
lagging sector less competitive. This second effect obviously has some negative consequences, but we can ask ourselves why the reallocation of resources is such a bad thing. After all, one of the most important insights of international economics is that a trade between two countries can benefit both countries if each country exports the goods in which it has a comparative advantage (Krugman, Obstfeld and Melitz, 2012, p.56). Thus, if a country has a comparative advantage in the production of oil, according to the theory of comparative advantage it should specialize in oil production. This means that the shifting of resources away from a lagging sector to a booming sector is a wise thing to do. However, there are several reasons that can explain why the reallocation of resources could turn out to have negative economic consequences, especially in the long run.
The first reason has already been mentioned in the introduction, namely the price volatility of primary goods. The world market prices of oil are more volatile than those of any other mineral and agricultural commodities, which in their turn are far more volatile than prices of most
manufactured products. It has been suggested in several studies that it is precisely the volatility of natural resource prices that is bad for economic growth (Frankel, 2010, p. 10). Therefore, if a country reallocates its production factors from the manufacturing sector to the oil sector, the price volatility of its output increases, which could lead to lower economic growth.
The second reason is that natural resources like oil will inevitably be depleted after a certain period. When this happens, the country again has to reallocate its production factors. However, sometimes an economic sector cannot recover as quickly as it has left. It is easily imaginable that a lot of knowledge gets lost when an economic sector disappears or contracts. Therefore these economic sectors could not be as efficient after the depletion of the natural resources as they were before the production factors were shifted to the oil sector. It could also be that the lagging sector is characterized by learning by doing, while the booming oil sector is not (Frankel, 2010, p. 14). That would imply that the lagging sector will become more and more efficient in the long term, while the efficiency in the oil sector stays the same. In that case, diversification out of the lagging sector will lead to lower output in the long term then if the production factors would have stayed in the lagging sector. On top of that, the shifts of movable production factors may incur needless transaction costs like frictional employment, incomplete utilization of the capital stock and incomplete occupancy of housing. These deadweight costs evidently have negative consequences on economic growth (Frankel, 2010, p.11).
Considering all these arguments, it seems clear that a boom in the oil sector usually causes a decline in output of other economic sectors. Furthermore, it could even be that such a boom could have negative consequences for the economy as a whole. However, there are also reasons to believe that this may not be the case. These reasons are discussed in the following section.
Positive effects of a boom in the oil sector on different sectors
Despite the reallocation of production factors and the appreciation of the real exchange rate, a boom in the oil sector does not necessarily mean that the output of other sectors falls. There are good reasons to believe that such a sector can actually benefit from a boom in the oil sector. In this section several of these reasons are discussed.
First of all, the resource curse and the Dutch Disease are well-known economic phenomena among policy makers. Therefore there already have been many attempts to sterilize the effects of a boom in the natural resources sector on different sectors. The resource allocation effect seems to be more or less inevitable, but there are definitely government policies that can reduce the spending effect. For example, many governments have created institutions insuring that export earnings are put aside during the boom time into a saving fund. Examples of such funds are commodity funds and sovereign wealth funds. These funds make sure that the export revenues are not immediately spent, and thus prevent the real exchange rate from rising too much. Additionally, the revenues in such a fund can be used later to make long-term investments that can structurally improve the economy or to sustain the standards of living after the natural resources are depleted. Besides a savings fund, a government can also decide to run a countercyclical fiscal policy. Such a policy does not only prevent an excessive appreciation of the real exchange rate, but can also moderate the effects of an
economic downturn later. However, the problem with saving funds and countercyclical policies is that there often is a lot of pressure on politicians to spend the revenues immediately for political purposes. Therefore these policies are often not implemented in the right way (Frankel, 2010, pp. 30-32). Nevertheless, if these policies are correctly implemented, the effects of a boom in the oil sector on other economic sectors can partially be sterilized.
The second reason is somewhat related to the saving funds mentioned above. If a
government uses the export earnings to make long term investments in for example infrastructure, education or healthcare, this could boost the competitiveness of all economic sectors. If the marginal productivity increases enough, it is possible that this increase in productivity can offset the decline in production factors in a lagging sector. This is another reason that a boom in the oil sector does not necessarily imply a fall in output in other sectors.
Thirdly, it has been stated before that oil is a capital intensive industry. Therefore, a boom in the oil sector requires relatively little additional labour. If we assume that installed capital is largely immobile between sectors, which seems like a reasonable assumption since cloth factories cannot suddenly become oil refineries, this means that the resource allocation effect will be quite small. However, we can go even further by assuming that even labour is not mobile in the short run. Even this assumption does not seem to be very unreasonable, since it takes some time for a labourer to
acquire the skills that are necessary to work in the oil industry. This assumption leads to a model that does not allow any participation by the booming sector in domestic factor markets. In effect, it is an enclave. There is then only a spending effect, there is no direct de-industrialisation, and the key mechanism of resource reallocation is the real appreciation (Corden, 1984, p. 362).
Finally, it is important to realize that oil is very price inelastic. In the article ‘Price elasticity for demand for crude oil: estimates for 23 countries’, it is stated that the US Federal Energy Office estimated that the long run price elasticity of demand for oil in consuming countries ranges between -0.2 and -0.6, while the short-run elasticity’s are expected to be even lower (Cooper, 2003: 3). This means that a slight increase in the supply of oil leads to a large price decrease. Since oil is used in many production processes throughout the economy, it is easily imaginable that a boom in the oil sector leads to lower costs in different economic sectors. These lower costs could boost the competitiveness of these non-booming sectors and thus lead to higher revenues and output. Following this argumentation, a boom in the oil sector could even induce a boom in the
manufacturing sector as well. It has to be noted that so far we have assumed that the domestic oil prices are the same as the world market price for oil. In this case, the reasoning above is only relevant if the increase in oil supply of a county is so large that it significantly affects the world market prices. However, many oil-exporting countries have detached domestic oil prices from the world market price (Corden, 1984, p. 368). If this is the case, it could be that the government of an oil-exporting country decides to use the increased oil supply to lower domestic oil prices instead of exporting it at the world market price.
Formulation of hypotheses
Following the predictions in the theoretical framework, an increase in oil exports in major oil-exporting countries is expected to have a negative effect on other economic sectors in two ways. Firstly, the reallocation of resources will reduce the availability of production factors in other sectors. Secondly, the additional revenues brought in by the oil sector will cause an appreciation of the real exchange rate, which will reduce demand for other economic sectors. However, we have also seen that the government can reduce these effects with certain policies and that the revenues of the oil industry can be used to invest in the competitiveness of lagging sectors. Additionally, the effects of a boom in the oil industry may not be as large as it would be in other sectors, since the production factors may not be as mobile as the described theories assume. Finally, a boom in the oil sector can lead to a decrease in production costs of different sectors, which could actually increase their output.
In the introduction, the following research question was formulated: does the Dutch Disease occur in major oil-exporting countries after an increase in oil exports? We have seen that if an increase in oil exports indeed leads to Dutch Disease, the negative effects of a boom in the oil sector are expected to outweigh the positive effects. In this case, an increase in oil exports would lead to a decrease in output in other economic sectors. However, if the negative effects do not outweigh the positive effects, an increase in oil exports should not lead to a decrease in output in other economic sectors. Therefore, we can test the research question using the following hypotheses:
H0: an increase in crude oil exports in a major oil exporting country does not lead to economic contraction in the agricultural, services and manufacturing sector of that country.
H1: an increase in crude oil exports in a major oil exporting country leads to economic contraction in the agricultural, services and manufacturing sector of that country.
If the null hypothesis is to be rejected in favor of the alternative hypothesis, this means that a major oil exporting country indeed experiences Dutch Disease after an increase in oil exports. However, if the null hypothesis is not rejected, there is no evidence to assume that such a country indeed experiences Dutch Disease.
In the formulation of hypothesis, the exports of oil has been changed to the exports of crude oil, and the agricultural, services and manufacturing sector are used as potentially lagging sectors. The reason for these choices, as well as an introduction of the quantitative methods used to test the hypotheses, are given in the next section.
Data introduction
The research question of this paper indicates that the analysis in this thesis will look at ‘major oil-exporting countries’. Although the reason for looking at oil oil-exporting countries has been previously outlined, one may wonder why only major oil-exporting countries are taken into account. The reason for this is that the effects described in the theoretical framework are only expected to materialize if the oil industry in a country is large enough to have a significant effect on the rest of a country’s economy. Therefore, in the following analysis only oil-exporting countries are taken into account whose GDP relies for a significant amount on the oil industry. To be precise, all countries in the analysis have a GDP that consists for at least 5% of oil rents. It is probably safe to assume that oil rents of more than 5% have a significant impact on the economy, and that oil rents of less than 1% do not have a significant impact on the economy. Since there are only very few countries that have oil rents as a percentage of GDP between 1 and 5 percent, this cut-off point of 5% separates
countries whose oil industry does have a significant influence on the economy from countries whose oil industry does not have a significant influence on the economy. However, it can be seen later that the analysis concerns the oil exports and the performance of different economic sectors for these countries between 2001 and 2010. Therefore, only countries whose GDP has consisted for at least 5% on oil rents on average between 2001 and 2010 are taken into account in this analysis. The calculation of this average is based on the World Development Indicators dataset of the World Bank, which calculates oil rents as the difference between the value of crude oil production at world prices minus the total costs of production. The list of these countries can be found in appendix 1.
The increase in crude oil exports is calculated based on the database of the U.S. Energy Information Administration (EIA). This database describes the exports of crude oil for all countries per year in thousands of barrels per day. The database has information dating back to 1984 until 2010. However, this analysis is only looking at the time period 2000-2010, as there are many missing data point before 2000 and some of the analyzed countries have only started exporting crude oil between 1990 and 2000. These countries exhibited very high growth rates in crude oil exports in this time period, while the absolute values of their oil exports was still rather small. Including this time period would therefore reduce the reliability of the analysis.
By calculating the percentage change in the export of crude oil, it is possible to determine in which years the output in the oil industry in a certain country has grown and in which year the output of the oil industry has fallen. These numbers have been calculated for all countries whose GDP has consisted for at least 5% on oil rents on average between 2001 and 2010 and are described in the EIA database. The countries Bahrain, Bolivia and Uzbekistan were deleted since the database provides insufficient information for these countries. In the end, the analysis includes Algeria, Angola,
Azerbaijan, Brunei, Cameroon, Chad, Colombia, Congo-Brazzaville, Ecuador, Egypt, Equatorial Guinea, Gabon, Iran, Iraq, Kazakhstan, Kuwait, Libya, Malaysia, Mexico, Nigeria, Norway, Oman, Papua New Guinea, Qatar, Russia, Saudi Arabia, Sudan (including South Sudan), Syria, Trinidad and Tobago, Turkmenistan, United Arab Emirates, Venezuela, Vietnam and Yemen. An overview of the amount of crude oil exports in thousand barrels per day for all these countries between 2001 and 2010 can be found in appendix 2. It has to be noted that the export of refined petroleum products is not included in this thesis. The exclusion of these products from the calculation may cause problems if countries have made sudden changes in the amount of oil they refine before exportation. However, since some countries import crude oil in order to refine it and sell the refined products, including these refined products as a measure of natural resource extraction would reduce the reliability of the analysis much more.
Finally, to test the effect of a change in the crude oil exports on other economic sector, the database of the World Bank is used to calculate the change in the total value of the agricultural, services and manufacturing sector. In their database, the economy of a country has been divided in four sectors, namely agriculture, services, industry excluding manufacturing and manufacturing, according to the International Standard Industrial Classification (ISIC), revision 3. The oil industry is included in the industry sector. By looking at the change in the total value of these sectors, it is possible to determine in which year the value of these sectors has grown and in which year the output has fallen. Since the entire economy is included in one of these sectors, they are an ideal measure of the effects of an increase in oil exports on other sectors of the economy. Therefore these sectors were described in the hypotheses as potentially lagging sectors. The change in the total value of the agriculture, services and manufacturing sector for all countries between 2001 and 2010 can be found in appendices 3, 4 and 5 respectively.
The variable GDP growth has also been included in the dataset. This variable has not been discussed in the theoretical framework as an explanatory variable. Nevertheless, it sounds very intuitive that an increase in GDP will cause both an increase in crude oil exports and an increase in the productivity of certain economic sectors. To control for this possibility, the GDP growth has been included as a control variable.
Apart from the variables mentioned above, there are many important variables that have been omitted for several reasons. In the discussion, a selection of these omitted variables and the impact of omitting these variables on the reliability of the regression analysis are discussed.
Regression analysis
Now that the hypotheses are formulated and the dataset to test these hypotheses has been created, it is now possible to test the hypotheses. To do this, the statistical program Stata is used. Since the dataset consists of 34 cases over 10 points in time, a panel data analysis is the right tool to analyze the data. The main tool for regression analysis for panel data is a fixed effects regression. This regression is an extension of multiple regression that exploits panel data to control for variables that differ across entities but are constant over time (Stock and Watson, 2007, p. 349). The equation of this regression is as follows: Yit=β1X1,it+…+ βkXk,it +αi+uit, where i= 1,…,n and t=1,….t (i representing the entity and t representing the time period) where X1,it is the value of the first regressor for entity I in time period t, X2,it is the value of the second regressor, and so forth, and α1,….., αn are entity-specific intercepts (Stock and Watson, 2007, p. 359). Alternatively, with panel data it is also possible to do a regression with time fixed effects. Time fixed effects can control for variables that are constant across countries but evolve over time (Stock and Watson, 2007, p. 361). In that case the possibility to control for variables that are constant over time but differ across countries would be lost. Since we are interested in variables that vary over time and countries have unique time-invariant
characteristics that influence the variables, the fixed effects model would be the most appropriate tool to analyze the dataset. Therefore the following analysis is based on a fixed effects regression.
After entering the dataset as panel data into Stata, it is possible to summarize the variables. The descriptive statistics of the variables can be found in table 1.
Variable Obs Mean Standard
Deviation
Min Max
Country ID 340 17.5 9.825168 1 34
Year 340 2005.5 2.876515 2001 2010
Crude oil exports (% change) 336 2.687189 21.72687 -62.37709 149.2308
Agriculture, value added (% change) 256 3.899137 9.717916 -34.2 76.81411
Services, value added (% change) 256 6.958533 8.504255 -20.69086 101.6303
Manufacturing, value added (% change) 250 10.77742 6.424854 -4.148838 30.37808
GDP growth 337 5.975188 7.547692 -33.10084 63.37988
Table 1. Descriptive statistics of the variables.
The descriptive variables of the variable countryID confirm that we have included 34 countries in the dataset in 10 different points of time. Since these countries have been assigned a number from 1 to 34 in alphabetical order, the mean and the standard deviation of this variable are meaningless. This is the same for the variable year. We can see from the minimum, maximum and number of
observations that we indeed have 340 observations from 2001 to 2010. Again the mean and standard deviation of this variable are meaningless.
We can see from the number of observations of the variable crude oil exports (% change) that there are four missing values. The average increase in crude oil exports for all included countries between 2001 and 2010 was 2.7%, with a standard deviation of 21.7%. The largest decrease in crude oil exports in one year was 62.4%, whereas the largest increase was 149%.
The descriptive statistics of the variable agriculture, value added (% change) indicate that average growth in the agricultural sector between 2001 and 2010 for all included countries was 3.9%, with a standard deviation of 9.7%. The number of observations is only 256, which means that there are 86 missing values for this variable. This relatively large number of missing values is due to the fact that not all values for all countries were included in the Worldbank database.
The variable services, value added (% change) has a mean of 7.0%. This means that the average growth in the services sector between 2001 and 2010 for all included countries was 7%. The standard deviation of this growth is 8.5%. The number of observations indicates that this variable has 86 missing values. The missing values are all from the same country and year as the missing values of the variable agriculture, value added (% change).
The descriptive statistics of the variable manufacturing, value added (% change) indicate that average growth in the manufacturing sector between 2001 and 2010 for all included countries was 10.8%, with a standard deviation of 6.4%. Given the number of observations, this variable has 90 missing values. These missing values are not necessarily the same as the missing values from the services and manufacturing variables.
Finally, the variable GDP growth indicates that the average economic growth between 2001 and 2010 for all included countries was 6.0%, with a standard deviation of 7.5%. The number of observations indicates that this variable has only 3 missing values.
Although these statistics may be quite interesting, they do not say anything about the effect of an increase in crude oil exports in the manufacturing, services and manufacturing sector. To find out this effect, we can use a fixed effects regression analysis. In the regression analysis, a significance level of 0.05 is used. That means that the chance of concluding that a variable is significant while there is no relationship in reality is less than 5%. However, given the research question we are only interested in a potential negative relationship between oil exports and growth in different economic sectors. We will therefore run a one-sided test, whereas Stata automatically runs a two-sided test. In order to interpret the significance of variables, we should therefore divide the observed values of P>|t| by two. If this number is below 0.05, we can conclude that a variable is significant.
From the theoretical framework we have reasons to expect that an increase in crude oil exports influences the growth in different economic sectors. Additionally, since production factors
may not be very mobile in the short run, it is also wise to include the change in crude oil exports of previous year as a variable in our regression analysis. Having these variables as explanatory variables, we should also include control variables. There is one factor that is expected to have a major
influence on both oil exports and growth in different economic sectors, namely GDP growth.
However, since GDP growth partly consists of the growth in a particular sector, including GDP growth in the same period would lead to large endogeneity problems. In this analysis, we will therefore use the growth in GDP of the previous period as a control variable, as well as the growth in the services and manufacturing sector when testing for the agricultural sector, the growth in the manufacturing and agricultural sector when testing for the services sector, and the growth in the services and agricultural sector when testing for the manufacturing sector. The reason behind this is that we can expect that growth in one sector leads to higher revenues in this sector. These revenues will partly be spend in other sectors. In this way, growth in one sector can lead to growth in another sector.
Admittedly, this relationship goes in both ways, so these control variables still entail some endogeneity. Nevertheless, since omitting these variables could lead to false conclusions when analyzing the effects of a change in crude oil exports on different economic sectors, including these variables probably leads to the most reliable conclusions.
Now that the testing method and all variables have been outlined, it is time to run the first regression analysis. First we will test the effect of an increase in crude oil exports on the agricultural sector. The results can be found in table 2.
Agriculture Coefficient Standard Error T P>|t| [95% Conf. Interval] Oil -.061997 .0389867 -1.59 0.114 -.1389511 .014957 R2 within: 0.05 Oil L. .0590848 .0366182 1.61 0.108 -.0131941 .131364 R2 between: 0.08 GDP L. -.1611223 .1546413 -1.04 0.299 -.4663614 .144117 R2 overall: 0.00 Services -.01566 .1531771 -0.10 0.919 -.3180089 .286689 Nr. Obs: 205 Manufacturing .6164155 .4520183 1.36 0.174 -.2758018 1.508633 Constant -2.409473 5.189297 -0.46 0.643 -12.65238 7.833431
Table 2. fixed effects regression with agriculture, value added (% change) as dependent variable and crude oil exports (% change), crude oil
exports (% change) of the previous year, GDP growth of the previous year, services, value added (% change) and manufacturing, value added (% change) as independent variables. Please note that in a fixed effects model, the R2 within has to be used to estimate the
proportion of total variation explained by the model.
As can be seen from the table, the coefficient of oil is negative. This would imply than an increase in crude oil exports leads to a decrease in the change in the total value of the agricultural sector. However, as the P>|t|/2 is higher than 0.05, this relationship is not significant. The coefficient of the
change in crude oil exports of the previous period is positive. This result is quite surprising, as the economic theory predicts a negative relationship. However, the P>|t| is 0.108. This means that this relationship is also not significant. We can therefore conclude that the change in crude oil exports does not significantly influence the growth in the agricultural sector.
Secondly, we will test the effect of an increase in crude oil exports on the services sector. The results can be found in table 3.
Services Coefficient Standard Error T P>|t| [95% Conf. Interval] Oil .0074735 .0195403 0.38 0.703 -.0310961 .046043 R2 within: 0.14 Oil L. .0345751 .0181747 1.90 0.059 -.001299 .0704492 R2 between: 0.23 GDP L. .3051755 .0736289 4.14 0.000 .159843 .450508 R2 overall: 0.24 Agriculture -.0038802 .0379535 -0.10 0.919 -.0787947 .0710344 Nr. Obs: 205 Manufacturing .0260021 .2262056 0.11 0.909 -.4204944 .4724986 Constant 4.930156 2.557212 1.93 0.056 -.117402 9.977715
Table 3. fixed effects regression with services, value added (% change) as dependent variable and crude oil exports (% change), crude oil
exports (% change) of the previous year, GDP growth of the previous year, agriculture, value added (% change) and manufacturing, value added (% change) as independent variables. Please note that in a fixed effects model, the R2 within has to be used to estimate the
proportion of total variation explained by the model.
This table contains some remarkable results. The coefficient of Oil is positive; this implies that, contrary to the hypothesis, an increase in crude oil exports leads to an increase in the services sector. However, since the value of P>|t| is very high, this result is far from significant. The coefficient of the change in crude oil exports of the previous year, on the other hand, is positive and significant. The P>|t|/2 is lower than 0.05, so we can conclude that an increase in crude oil exports leads to an increase in the change of the total value of the services sector in the next year. This results seems contrary to the theory on Dutch Disease. However, the Core Model can provide an answer for this odd result. The model assumes that there is a non-exportable sector whose output will increase after a boom in the booming sector, leading to an appreciation in the real exchange rate. In our dataset, the economy is divided in four sectors, namely the industry, agriculture, services and manufacturing sector. Services are usually non-exportable products, while all the other sectors mostly contain products that are easily exportable. Therefore this result can be explained by assuming that the services sector represents the non-exportable sector in the Core Model.
Thirdly, we will test the effect of an increase in crude oil exports on the manufacturing sector. The results can be found in table 4.
Manufacturing Coefficient Standard Error T P>|t| [95% Conf. Interval] Oil -.0047869 .006579 -0.73 0.468 -.017773 .0081992 R2 within: 0.03 Oil L. -.0085594 .0061557 -1.39 0.166 -.0207098 .003591 R2 between: 0.01 GDP L. -.0074454 .0260215 -0.29 0.775 -.0588081 .0439173 R2 overall: 0.00 Agriculture .0173525 .0127246 1.36 0.174 -.007764 .0424691 Nr. Obs: 205 Services .0029542 .0257001 0.11 0.909 -.047774 .0536824 Constant 11.05094 .2213503 49.93 0.000 10.61402 11.48785
Table 4. fixed effects regression with manufacturing, value added (% change) as dependent variable and crude oil exports (% change), crude
oil exports (% change) of the previous year, GDP growth of the previous year, agriculture, value added (% change) and services, value added (% change) as independent variables. Please note that in a fixed effects model, the R2 within has to be used to estimate the proportion of
total variation explained by the model.
From this table we can see that, as predicted by the economic theory, the coefficients of the change in crude oil exports in the same year and in the previous year are negative. However, the values of P>|t| indicate that this relationship is not significant. This means that there is not enough statistical evidence to conclude that a change in oil exports negatively influences the growth in the
manufacturing sector.
Now that all the required tests have been done, the results of these tests can be used to reject either the null or the alternative hypothesis. Afterward we can use this result to answer the research question. Please note that all commands entered in Stata and the output produced by Stata that are described above are included in appendix 6.
Conclusion
This thesis has looked at the effects of an increase in oil exports in major oil-exporting countries on different sectors of the domestic economy. By looking at these effects, it has been tried to shed some light on the question whether the Dutch Disease occurs in major oil-exporting countries after an increase in oil exports. The review of the scientific literature has shown that there are two main mechanisms by which oil exports can negatively influence other economic sectors. Firstly, the reallocation of resources will reduce the availability of production factors in other sectors. Secondly, the additional revenues brought in by the oil sector will cause an appreciation of the real exchange rate, which will reduce demand for other economic sectors. However, we have also seen that there are several reasons to believe that these effects will either not materialize in oil-exporting countries, or will be significantly reduced by the positive effects of a boom in the oil sector on other economic sectors. To see if the negative effects of a boom in the oil sector on other economic sectors
outweigh the positive effects on other sectors, the following hypotheses have been formulated:
H0: an increase in crude oil exports in a major oil exporting country does not lead to economic contraction in the agricultural, services and manufacturing sector of that country.
H1: an increase in crude oil exports in a major oil exporting country leads to economic contraction in the agricultural, services and manufacturing sector of that country.
In the regression analysis we have seen that there is insufficient statistical evidence to conclude that the agricultural, the services and the manufacturing sector experience economic contraction after an increase in oil exports. This means that the null hypothesis cannot be rejected.
The research question of the thesis reads: does the Dutch Disease occur in major oil-exporting countries after an increase in oil exports? Based on the analysis of oil exports and economic performance of 34 major oil-exporting countries over 10 years, we can conclude that an increase in oil exports does not lead to Dutch Disease.
However, that is not to say that these countries can rapidly increase their oil exports without having to fear for negative economic effects. The conclusion of this thesis is based on an analysis that did not control for changes in economic policy. Therefore it is possible that an increase in oil exports does lead to economic contraction in other economic sectors, but that these effects are normally sterilized by good economic policies. Also this analysis did not specifically look at excessive growth in oil exports, but at changes in oil exports in general. We therefore cannot rule out the possibility that Dutch Disease does occur when the growth in oil exports is very high. However, under normal growth
conditions and with a properly functioning government, it is safe to say that in major oil-exporting countries an increase in oil exports does not lead to Dutch Disease.
Discussion
Although it has been tried to make this thesis as reliable as possible, it is inevitable to make use of some assumptions and simplifications that may not totally hold in reality. This section is devoted to the discussion of some of these assumptions and simplifications.
The first of these assumptions has to do with the selection of countries in the analysis. As indicated in the data introduction, the effects of an increase in oil exports on other economic sectors are only expected to materialize if the oil industry in a country is large enough to have a significant effect on the rest of a country’s economy. Therefore it has been decided to only include countries whose GDP consisted for at least 5% of oil rents in the period 2000-2010. However, the selection of this cut-off point is rather arbitrarily. It is possible that some countries have an economy that depends much more on oil rents than other countries, while the oil rents as a percentage of GDP actually is lower. However, for a lack of reliable quantitative information on oil dependence of a particular country in a particular year, this cut-off point of 5% has been selected. This 5% is based on the assumption that oil rents of more than 5% have a significant impact on the economy, and that oil rents of less than 1% do not have a significant impact on the economy. Since there are only very few countries that have oil rents as a percentage of GDP between 1 and 5 percent, this cut-off point of 5% separates countries whose oil industry does have a significant influence on the economy from countries whose oil industry does not have a significant influence on the economy.
Another simplification that should be mentioned is the fact that only crude oil exports have been taken into account. The reason for this is that the research question is aimed at natural
resource abundant countries that export resources that they themselves have extracted. Since there is no public database on oil exports that makes a difference between self-extracted and originally imported oil, we would also have to take into account countries that import crude oil, then refine it and subsequently export these refined products if we were to include refined petroleum products in our analysis. In that case the theories mentioned in the theoretical framework are not expected to hold anymore. It has therefore been decided to only take into account crude oil exports. However, this choice may cause problems if countries have made sudden changes in the amount of oil they refine before exportation.
The next potential problem that has to be discussed is endogeneity. In the regression analysis the growth in sector A is used as a control variable for the growth in sector B. Since growth in sector A leads to higher revenues in sector A, it is likely that this growth also leads to a growth in sector B, since these revenues will partially be spend in that sector. However, this relationship also goes the other way around. Therefore it could be that some of the growth in the dependent variable is attributed to the control variable, while the dependent variable is actually causing the change in the
control variable instead. This endogeneity problem could potentially bias the results of the regression analysis. However, since the growth in other sectors is a very important driver of economic growth in the sector concerned, this variable should not be excluded from the analysis.
The last problem that should be discussed is omitted variables. Of course there are numerous drivers behind economic growth in a certain sector. However, including all these variables is not feasible in this thesis. Nevertheless, one may notice that the number of included control variables is very small. The reason for this is that is has proved to be very difficult to successfully quantify the relevant control variables for all countries over the relevant time period. For example, the economic growth for the main trading partners is a very important factor in exports and growth of different economic sectors. However, it is not feasible to determine the most important trading partners for every country in every year and then calculate the weighted average growth in these countries. Another important control variable that was mentioned in the theoretical framework is technological growth. This is one of the main drivers behind growth in a particular sector, but quantifying this growth for every country has not been possible, especially since the countries concerned are not always developed countries. Usually these less developed countries provide less statistical information about their economy and are less often the subject of a reliable study or report. The same reasoning applies to the consumption preferences in these countries, which could greatly influence the growth in an economic sector. Then there were also control variables that were quantifiable and could thus be included, but were not expected to have a measurable effect in this analysis. Examples of such control variables are education and population growth. It is beyond doubt that these factors influence the growth in an economic sector, but these effects take very long to materialize. Since the immediate effects of these changes are not expected to influence some economic sectors more than others, including these variables would have been useless.
Then there is another control variable that was both relevant and easily measurable, namely foreign direct investments. However, this control variable has also not been included in the analysis. The reason for this is that only the net inflows of investments would have been a good control variable, as the effects of an increase in inflows can easily be offset by an increase in outflows. However, since this thesis is concerned with the effect of changes in one factor on changes in another sector, we would have to work with the percentage change in net inflows of investments. Many of the included countries have net inflows that vary around 0. Therefore the percentage change in inflows in these countries is often excessive, while the absolute change is rather small. Growth rates in net FDI of over 100% are often observed, while some growth rates even exceed 10000%. Including these growth rates would therefore greatly bias the results of the analysis.
Nonetheless, omitting all these control variables does bias the results of the analysis. It would therefore be very interesting to test the hypotheses of this thesis again with a regression
analysis that includes these omitted variables. As indicated, it requires much work to successfully quantify these variables, so this is beyond the scope of this thesis. However, including these variables is the only way to make a more reliable statement about the occurrence of the Dutch Disease in major oil-exporting countries after an increase in oil exports. Therefore this subject would be a good starting point for further research on the Dutch Disease in major oil-exporting countries.
Bibliography
Brunnschweiler, Christa and Bulte, Erwin (2007). The Resource Curse Revisited and Revised: A Tale of Paradoxes and Red Herrings. Journal of Environmental Economics and Management, Vol. 55, Issue 3, May 2008, pp. 248-264.
Cooper, J. (2003) Price elasticity for demand for crude oil: estimates for 23 countries. OPEC Review, Vol. 27, Issue 1, pp. 1-8.
Corden, W. M. (1984). Booming Sector and Dutch Disease Economics: Survey and Consolidation. Oxford Economic Papers, New Series, Vol. 36, No. 3, pp. 359-380.
Frankel, Jeffrey A. (2010). The Natural Resource Curse: A Survey. National Bureau of Economic Research, NBER Working Paper Series, Working Paper 15836.
Krugman, Paul R., Obstfeld, Maurice and Melitz, Marc J. (2012) International Economics. Theory & Policy. Ninth Edition, Pearson Education, Essex.
Manzano, Osmel and Rigobon, Roberto (2001). Resource Curse or Debt Overhang? National Bureau of Economic Research, NBER Working Paper Series, Working Paper 8390.
Sachs, Jeffrey D. and Warner, Andrew M. (1995). Natural Resource Abundance and Economic Growth. National Bureau of Economic Research, NBER Working Paper Series, Working Paper 5398.
Sadorsky, Perry (2001) Risk factors in stock returns of Canadian oil and gas companies. Energy Economics, Vol. 21, Issue 1, pp. 17-28.
Stock, James H. and Watson, Mark W. (2007) Introduction to Econometrics. Second Edition, Pearson Education.
Appendix
Appendix 1. Oil rents as percentage of GDP
Source: World Bank Database (http://data.worldbank.org/indicator/NY.GDP.PETR.RT.ZS)
Oil rents (% of GDP) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Average Algeria 11.13607 12.28073 14.63142 16.98301 21.42789 23.10564 22.14524 23.08087 15.6179 16.99776 17.74065 Angola 50.85354 44.70741 47.52137 54.16869 70.63797 65.60649 59.96503 66.84189 38.26968 45.72037 54.42924 Azerbaijan 35.05896 33.66379 34.73765 40.21977 55.51502 61.8005 55.76653 53.07226 37.78423 42.16194 44.97807 Bahrain 14.69785 14.16263 14.9545 17.03314 20.2204 20.75126 19.5659 22.60078 14.68 17.14142 17.58079 Bolivia 2.970723 3.16457 4.037613 5.974562 8.78885 8.539163 8.278529 8.552347 4.05716 4.976648 5.934016 Brunei Darussalam 25.12023 25.48238 27.5095 29.63903 37.50191 37.97591 34.44358 39.38771 26.73849 30.25852 31.40573 Cameroon 7.908275 7.465644 6.724556 7.489037 9.504173 10.83234 9.93414 11.55551 6.297818 7.166125 8.487761 Chad .. .. 7.991144 47.86587 47.67904 45.63603 40.48151 41.20637 24.84249 29.30954 35.6265 Colombia 4.377657 4.282372 4.965748 5.374687 6.251912 6.876518 5.933773 7.716215 5.197874 6.582892 5.755965 Congo, Rep. 63.12992 57.4843 54.11887 57.67773 70.12655 75.70786 62.84458 65.00268 53.55856 64.14433 62.37954 Ecuador 11.95241 10.08682 11.38027 17.31026 22.57619 24.45384 23.54948 26.52984 14.91538 17.52706 18.02815 Egypt, Arab Rep. 5.023462 5.672713 7.35054 9.765892 12.05552 12.07134 11.23194 13.09074 6.426384 7.42377 9.01123 Equatorial Guinea 78.35678 76.98812 76.02296 75.48384 69.64439 68.27109 58.67842 66.14852 50.78753 50.24861 67.06303 Gabon 40.37902 38.39006 39.94151 45.53364 54.1128 53.21162 49.26979 48.68222 37.16649 42.82366 44.95108 Iran, Islamic Rep. 23.89306 23.74245 26.46864 29.74164 38.32613 40.28699 34.99251 37.8932 21.41105 24.12824 30.08839 Iraq .. .. .. 67.92459 66.42057 64.02631 53.8598 56.14322 40.58335 41.79745 55.82219 Kazakhstan 26.45018 28.62685 29.57593 33.04882 37.51929 34.13873 29.74761 33.82122 24.72225 26.32866 30.39795 Kuwait 43.7311 36.64951 41.56466 48.80201 58.07896 57.38189 54.54038 60.67308 42.60186 50.1032 49.41267 Libya 32.67017 47.49147 50.50464 54.69734 64.46733 63.22079 55.71007 57.7899 42.28938 .. 52.09346 Malaysia 5.027849 5.008077 5.781958 7.09082 8.502904 8.639748 7.902956 9.160345 5.866702 6.174371 6.915573 Mexico 3.05787 3.111627 4.120207 5.30161 6.794967 7.227272 6.940683 8.239874 5.318117 5.949847 5.606207 Nigeria 37.40815 24.3997 28.94638 32.64213 37.59949 33.24794 29.54707 31.12611 22.42154 25.02476 30.23633 Norway 12.48044 10.90922 10.83374 11.96147 13.68218 13.67761 12.53737 14.35661 9.216979 9.954259 11.96099 Oman 34.12441 32.49472 33.14819 37.8878 44.03679 42.68126 39.69651 40.31963 30.70235 35.53592 37.06276 Papua New Guinea 13.35704 11.50499 12.08729 14.09529 16.40677 18.13651 18.40978 .. .. .. 14.85681 Qatar 26.32902 25.4749 24.20467 27.69817 28.98871 27.38952 22.37417 22.69588 13.49616 13.62134 23.22725 Russian Federation 13.78684 13.58568 14.67375 16.16581 18.94773 18.10314 15.48424 16.79524 12.61643 13.78682 15.39457 Saudi Arabia 33.7405 30.33254 37.15982 43.57594 51.75135 53.62335 51.33454 58.87597 38.28397 40.40504 43.9083 Sudan 11.48689 12.08981 13.27028 16.82276 20.00446 20.66128 24.28513 27.22928 16.69161 17.38102 17.99225 Syrian Arab Republic 19.90798 21.43566 24.29187 21.09323 24.86281 24.37684 21.16264 .. .. .. 22.44729 Trinidad and Tobago 9.110074 10.49012 10.38802 11.29796 15.58035 16.35248 12.91554 13.00674 10.08578 11.38775 12.06148 Turkmenistan 29.34876 27.87348 27.73571 33.43421 39.27268 32.81831 31.99873 31.36845 16.35254 20.27506 29.04779 United Arab Emirates 14.54289 13.05891 15.39688 18.3138 22.05244 23.4368 21.89353 25.08857 15.75592 19.0063 18.85461 Uzbekistan 8.537311 10.05226 12.57466 14.89455 12.61833 12.85407 9.790725 10.55972 4.745854 4.400944 10.10284 Venezuela, RB 19.06029 23.69359 27.55376 35.07226 40.88787 39.1868 30.00796 30.78437 17.05182 18.53574 28.18345 Vietnam 6.968315 6.563475 6.920548 9.763755 11.35642 10.83667 9.757632 9.850348 5.917109 6.568126 8.450239 Yemen, Rep. 30.91685 29.05515 30.84953 33.27077 39.71942 38.8103 32.80436 32.35149 20.0587 19.31412 30.71507 Note: Bahrain, Bolivia and Uzbekistan were deleted as the US EIA Database did not provide enough information on these countries
Appendix 2. Amount of crude oil exports in thousand barrels per day
Source: US Energy Information Administration
(http://www.eia.gov/cfapps/ipdbproject/iedindex3.cfm?tid=5&pid=57&aid=4&cid=all,&syid=2000&e yid=2010&unit=TBPD)
Crude oil exports, thousand of 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Algeria 809.4618 442.4971 875.9059 1154.619 1279.2 903.317 882.693 864.96 783.58 1175.128 1096.927 Angola 694.5338 697.5936 853.7489 860.0947 1010.716 1219.755 1392.808 1658.86 1849.38 1908.881 1928.107 Azerbaijan 108.0096 150 170.3965 176.7824 177.2988 287.095 498.44 695.6 737.16 871.0813 907.5395 Brunei 189.4135 185 178.7517 183.1556 192.5886 197.83 207.468 170.46 163.6 142.0522 147.9241 Cameroon 55 107.6195 101.5792 97.84846 91.67546 83.4817 87.926 78.52 76.52 63.89661 55.68499 Chad 0 0 0 0 170 176.7 157.92 143 126 119.7278 125.7278 Colombia 384.079 295.9562 278.8728 229.9453 213.388 236.789 226.808 233.62 266.94 371.0315 483.8403 Congo (Brazzaville) 272.2222 245 240.3462 235.8276 224.3118 235.283 265.989 213.393 229.22 260 287.5685 Ecuador 240.8971 250.9334 235.15 265.2624 379.073 377.907 394.495 355.796 364.838 333.9723 341.7273 Egypt 220 138.1007 140.2078 99.90706 37.58794 57.4752 52.334 43.4 89.5 86.19612 85 Equatorial Guinea 166 181.44 212.56 206 371.7 375.44 362.87 345.37 331.118 346.12 319.12 Gabon 300 256.8123 234.7234 226.4191 220.8868 248.8469 221.2551 227.44 219.16 226.9923 225.2711 Iran 2309.127 2229 2093.6 2296.3 2555.7 2497.36 2539.65 2617.63 2475.16 2296.621 2377.195 Iraq 2071.664 1850 1494.6 911.935 1600 1381.15 1479.82 1617.92 1766.93 1902.536 1913.97 Kazakhstan 521 586.5239 784.9792 869.7648 1037.871 892.694 903.156 929.926 980.918 1366.387 1406.183 Kuwait 1316.807 1220.736 1138 1249.2 1479.1 1689.74 1759.521 1645.49 1785.38 1495.173 1395.026 Libya 1110 1050 983.6 1126.6 1218.6 1322.312 1389.47 1314.64 1335.56 1404.995 1378.396 Malaysia 400 390.3558 392.1352 426.14 420.2245 369.8662 352.762 327.78 289.88 254.4605 244.9758 Mexico 1843.426 1882.54 1913.09 2114.423 2117.969 2021.561 2001.692 1808.295 1505.373 1311.759 1460.27 Nigeria 2069.177 2034.1 1893.2 2163.5 2176.1 2260.33 2190.28 2120.22 1931.94 2115.367 2340.587 Norway 2894.839 2978.752 2775.38 2731.019 2681.854 2338.632 2175.584 1980.713 1674.238 1771.667 1601.572 Oman 905 840 838.9041 763.0137 720.2186 756.697 673.43 640.622 623.911 665.5693 705.1279
Papua New Guinea 68.5 66.5 55 48.77958 44.57787 39 24.2 19 13 32.4 28.4
Qatar 680 680 570.8 753.98 867.8 933 982.4963 623.94 684.36 1041.13 1106.123
Russia 3150 3300 3953.426 4520.399 5211.14 5222.327 5106.261 5171.58 5120 4890.627 4887.826
Saudi Arabia 6444 6256.5 5984.6 6872.8 7143 7215.77 7036.1 6968.74 7298.7 6250.406 6844.096
Sudan and South Sudan 142 142 178.92 196.98 262.3907 227.993 252.493 374.18 367.82 370.7 389.08
Syria 305.7851 367.3219 355.1859 294.0373 211.0537 209.277 162.244 153.388 156.814 134.3208 152.4125
Trinidad and Tobago 52.45183 50.22658 68.47123 86.5765 80.95507 75.0472 74.5652 61.22 55.62 93.60474 75.33696
Turkmenistan 29.39303 55.01732 55.21381 58.35765 58.78607 50.463 30.1233 38.1562 48.0656 47.1577 35.36827
United Arab Emirates 1870 1743 1674 1848 2172 2106.6 2324.35 2289.44 2338.72 2180.886 2142.1
Venezuela 2094.297 1947.087 1621.9 1535 1587.4 2418.48 2348.97 2224.72 1860.76 1594.082 1645.031
Vietnam 315 356.75 322.0809 342.0421 399.2106 360.817 345.414 300.64 275.04 309.4692 214.7735
Yemen 340.2422 353.0493 357.9833 350.2754 306.1951 284.745 243.992 214.12 197.8 199.7205 175.1762
Appendix 3. Percentage change in the total added value of the agricultural sector
Source: World Bank Database (http://data.worldbank.org/indicator/NV.AGR.TOTL.ZS)
Agriculture, value added (% change) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Algeria 13.27494 -1.2 19.5 5.4 2.3 8.1 2.5 -3.8 21.1 4.9 Angola 18 12.1 12.1 14.1 17 9.8 26.6824 1.711408 27.74666 5.888461 Azerbaijan 11.1 6.4 5.6 5 7.5 0.9 4 6.1 3.5 -3.21808 Brunei Darussalam 5.845335 5.180359 11.30361 11.98977 1.307225 -9.86381 -4.46761 3.741231 5.709992 -5.86141 Cameroon 4.500002 3.726127 3.669024 3.502846 2.557036 3.010406 .. .. .. .. Chad .. .. .. .. .. 0.636836 -1.17983 -4.92855 -3.05756 16.13473 Congo, Rep. 8.022599 8.577406 6.262042 4.805077 4.411765 5.385253 5.000157 5.579554 -3.18641 6.256173 Ecuador 4.965135 1.766688 6.607248 2.385051 7.058366 4.308752 3.419526 1.667103 1.715204 0.739268
Egypt, Arab Rep. 3.70817 3.600498 3.480101 2.759423 3.256159 3.248658 3.680369 3.34757 3.170502 3.474481
Equatorial Guinea .. .. .. .. .. .. .. .. .. ..
Gabon 2.984039 -4.78437 1.4862 1.185495 3.308063 2.134757 5.300197 -0.19819 2.950565 -7.75753
Iran, Islamic Rep. -2.26375 11.35501 7.113891 2.171959 9.319345 4.670984 6.169653 .. .. ..
Iraq 1.198518 16.98105 -29.126 17.4402 31.35477 4.315105 -27.699 -13.1862 3.386475 17.21342 Kazakhstan 17.1 3.2 2.2 -0.1 7.1 6 8.9 -6.2 13.2 -11.6 Kuwait 18.57923 17.28111 8.840864 .. .. .. .. .. .. .. Libya .. .. .. .. .. .. .. .. .. .. Malaysia -0.17294 2.866575 6.030949 4.674998 2.593833 5.835857 1.37799 3.83498 0.053961 2.39698 Mexico 2.613601 -0.39624 2.885098 3.735492 -4.78004 6.920239 2.238058 1.315218 -2.50004 0.755952 Nigeria 3.880849 55.18264 6.982713 6.292025 7.055755 7.403043 7.191591 6.266296 5.879197 5.822678 Norway -2.69096 8.987486 -1.59387 11.35817 -0.07327 1.455545 4.838826 5.415187 -1.52377 7.846328 Oman 3.388135 0.455911 0.681841 6 .. .. .. .. .. ..
Papua New Guinea -4.74491 -4.12871 4.962708 4.6085 .. .. .. .. .. ..
Qatar .. .. .. .. .. .. .. .. .. ..
Russian Federation 11.39933 2.905362 -1.79091 1.036177 0.338185 2.719886 1.317164 6.391701 1.645905 -12.0079
Saudi Arabia 0.567207 1.283619 0.814729 3.589018 1.058576 1.011099 1.947918 1.327646 1.031262 -0.99361
Sudan 4.613434 2.972598 -0.22493 -2.50127 2.237293 3.196632 -1.37731 3.348108 -0.75239 -1.16994
Syrian Arab Republic 10.41698 7.815395 .. .. .. .. .. .. .. ..
Trinidad and Tobago 8.677563 8.710576 -15.2604 -34.2 -5.4 -10.1 21.8 7.6 -32.4292 76.81411
Turkmenistan 23 0.095 0.099 19.3 20.3 24 .. .. .. ..
United Arab Emirates .. 3.210507 -3.11064 -1.97008 -4.01935 -8.63707 -1.84615 -10.8961 -0.40034 -10.5238
Venezuela, RB 2.037019 -0.80008 -1.30891 4.379339 9.83519 1.034372 2.58822 3.485358 1.000007 0.919749
Vietnam 3.171575 4.186935 3.713448 4.45799 4.189322 3.798011 3.955248 4.692286 1.909818 3.291204
Yemen, Rep. .. .. .. .. .. .. .. .. .. ..
Colombia 1.767643 4.550784 3.090086 2.977318 2.812989 2.370714 3.91049 -0.37562 -0.65352 0.191564
