Does Russia suffer from the Dutch Disease?

Bachelor Thesis

Does Russia Suffer from the Dutch Disease?

Final version

13 – 03 – 2014

Pieter Speksnijder

6132936

Faculty of Economics & Business

BSc. Economics & Business

Track: Economics

2

CONTENT

CONTENT 2

1. INTRODUCTION 3

2. LITERATURE REVIEW 7

2.1 Literature related to the Dutch Disease. 7

2.2 The Dutch Disease model. 9

2.2.1 Equilibrium 10

2.2.2 The effects of a boom 11

2.3 Empirical research on the Dutch Disease in Russia 14

3. METHODOLOGY 17

3.1 Appreciation of the real exchange rate. 17

3.2 A decline in manufacturing output and a rise in service output 18

3.3 A rise in overall wages. 19

3.4 Transition state. 20

4. RESULTS 22

4.1 Appreciation of the real exchange rate. 22

4.2.1 A rise in the share of services in GDP relative to manufacturing. 23 4.2.2 Decline of manufacturing employment and increase in service employment. 25

4.3 Increase in overall wages. 26

4.4 Transition state of the Russian economy. 27

5. CONCLUSION 30

6. DISCUSSION 32

APPENDIX A 34

LIST OF TABLES AND FIGURES 36

Figures 36

Tables 36

3

1. INTRODUCTION

Since the collapse of the Soviet Union in 1991 the Russian economy has gone through turbulent times in the transition of a central planned communist economy to a free market economy. Problems with implementing new fiscal structures, declining productivity, artificially high exchange rates between the ruble and foreign currencies and a chronic fiscal deficit in the 1990s led to a financial crisis in 1998 in which the Russian government defaulted on its domestic debt, devaluated the ruble and postponed payments to foreign creditors.

In the years after the crisis of 1998 the Russian economy recovered and started to grow, from 2000 to 2008 the Russian economy averaged a 7% annual real GDP growth, as shown in figure 1.1.

Real wages more than doubled, the average salary eight folded from $80 to $640, unemployment fell and the percentage of people living below the poverty line decreased from 40% in 1999 to 13% in 20131_{. The financial crisis of 2008 and the global economic downturn} that followed it did not leave Russia untouched, GDP decreased in 2009 but the economy is recovering with GDP growing again, although slower, from 2010 and with inflation at all-time low levels around 4%, economic prospects are looking good for Russia.

Much of this economic growth however was caused by the exports of natural resources. Russia holds the 8th largest oil reserves, the second largest gas reserves in the world and with 10.37 billion barrels of oil per day Russia has overtaken Saudi Arabia as the world largest oil producer in 2012. Furthermore the energy sector accounts for around two-thirds of Russian exports, 30% of GDP and about 50% of the fiscal budget revenue (White, 2011).

At first this does not seem like a problem, however abundance in natural resources has a Figure 1.1: Real GDP growth in % of total GDP of Russia from 2000-2012.

The yearly % change in GDP for Russia from 2000 to 2012. Source: IMF world outlook database.

10,0 5,1 4,7 7,3 7,2 6,4 8,2 8,5 5,2 -7,8 4,5 4,3 3,4 -10,0 -5,0 0,0 5,0 10,0 15,0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

4 proven negative correlation with economic performance. Sachs and Warner (1995) conducted a large cross country empirical analysis of 97 developing countries between 1970 and 1989 and found a robust negative relationship between real GDP growth and the ratio of resource exports to GDP, this results was repeated by Oomes and Kalcheva (2007) with an extended data period of 1970 until 2000.

In economics this negative relationship is often referred to as the natural resource curse and several explanations for it exist. One explanation for the natural resource curse is that income generated by natural resources tends to be volatile because the elasticities of supply for natural resources are usually small (Oomes and Kalcheva, 2007). This volatility is negatively correlated with growth (Ramey and Ramey, 1995) and investment (Aizenman and Marion, 1999). An explanation for this is that in the presence of financial market imperfections, volatility leads to lower investments and lower welfare as a result of higher costs of capital. (Hausmann and Roberto, 2003).

Another explanation is that the large rents that can be obtained from the natural resources incentivize governments, institutions and private agents to show all kinds of self-enriching behaviour that is socially undesirable, and cause lower growth. The reasoning behind this is that the possible wealth generated by the natural resources is so high that it gives incentives for governments and private agents to try to capture a large part of the existing wealth, for instance through corruption or conflict, instead of engaging in activities that produce more wealth. Mauro (1995), Leite and Weidmann (1999) and Gylfason (2004) empirically found that natural resource dependence is positively related to corruption and Collier and Hoeffler (2004) found a strong positive effect of the share of natural resources in GDP on the probability of civil conflict in a country. Yang (2009) contradicts this with an empirical study in which he concludes that it is government policy, rather than weak institutions, that causes the emergence of the resource curse. Although weak institutions contribute to the adverse effect of resource abundance on growth, it is government policy that plays the decisive role.

This thesis however focusses on a third explanation for the natural resource curse namely the Dutch Disease. The term Dutch Disease was mentioned for the first time in the journal The Economist in an article that tried to explain the decline of the Dutch manufacturing sector as a consequence of the natural gas exports in the late 1970s (The Economist, 1977). In short the Dutch Disease is known as the negative effect of the export of natural resource on the exports of traditional export sectors. The theory behind this is that the real exchange rate appreciates through the exports of natural resources, which causes the traditional non-natural resource sectors to become less competitive internationally and face less demand which could lead to lower exports of traditional export sectors.

5 Russia seems like a perfect candidate for the Dutch Disease with exports of natural resources accounting for roughly two-thirds of total exports. Opinions however are divided on the question if Russia suffers from the Dutch Disease. On one side are Ahrend et al. (2007) and Beck et al. (2007) who conclude that the observed symptoms on the surface might look like Dutch Disease symptoms but in fact are caused by other factors than natural resource abundance like the transition of the Russian Economy or rebounds after the 1998 crisis. On the other side we find Barisitz and Ollus (2007) and Algieri (2011) who find confirming evidence of the Dutch Disease symptoms of de-industrialisation and appreciation of the exchange rate and somewhere in the middle are Oomes and Kalcheva (2007), Westin (2005) and Roland (2005) who conclude that there are some indications of the Dutch Disease but do not go as far as concluding that Russia suffers from a full-blown Dutch Disease. This thesis aims to add another argument to the discussion by analysing if Russia suffers from the Dutch Disease, with a longer data period (1996-2013) than the previous studies and by simultaneously analysing the transition state of the Russian economy. This leads to the main question of this thesis: To what extent does Russia suffer from the Dutch Disease?

This question is answered by analysing if Russia suffers from all four symptoms of the Dutch Disease: 1. An appreciation of the real exchange rate, 2. An increase in service sector output, 3. A decrease in traditional export sector output and 4. An overall increase in wages. The derivation of these symptoms is described in the next chapter. Furthermore the transition state of the Russian economy is analysed in order to account for the argument that the symptoms of the Dutch Disease are in fact caused by the transition of the economy.

The Dutch Disease however might not actually be Dutch, as Cordon stated “It could be argued that the true Dutch Disease in the Netherlands was not the adverse effects on manufacturing of real appreciation but rather the use of Booming Sector revenues for social service levels which are not sustainable, but which have been political difficult to reduce.” (Cordon W. M., 1984). Furthermore Barker (1981) and Kremers (1985) conclude that it is difficult to ascribe the decline in the Dutch manufacturing sector to the natural gas discoveries alone. Most of Western European countries were facing the same problems, regardless if they were net energy exporters or importers1_{. Although the Dutch Disease might not be Dutch, the} term is widely used in economic literature to describe the previously mentioned phenomenon, and thus it will remain the terminology in this thesis.

In the next chapter relevant literature for this thesis will be discussed as well as a theoretical model of the Dutch Disease. Chapter 3 describes the research methodology that is used to answer the main question as well as the technique of analysing the transition state of the economy, chapter 4 discusses the results of that research methodology. Furthermore a

6 conclusion is presented in the fifth chapter and in the final chapter limitations to this thesis and opportunities for further research are discussed.

7

2. LITERATURE REVIEW

This chapter discusses the literature that is relevant for this thesis and consists of three parts. The first part presents an overview of the literature in the field of the Dutch Disease, the second part discusses the theoretical model of the Dutch Disease by Cordon and Neary (1982) and in part 3 empirical literature is discussed that focussed on the question whether Russia suffered from the Dutch Disease.

2.1 Literature related to the Dutch Disease

Gregory (1976) was one of the first to describe the negative effect of mineral exports on traditional export sectors, which was later named the Dutch Disease. He uses a two sector partial equilibrium model to analyse the impact of mineral development on the Australian Economy. The main result of his model is that a discovery of minerals leads to an increase in exports, which leads to a surplus on the current account. In order to restore equilibrium on the current account either the domestic inflation has to rise or the exchange rate has to appreciate. The consequence of this is that traditional export and import-competing sectors are squeezed.

Snape (1977) extended the work of Gregory (1976) by adjusting the model so that it would fit general equilibrium theory while maintaining Gregory’s general assumptions about the effects of a mineral discovery. His results provide the following extensions or modifications to his work. 1) Even though in general non-mineral goods production is expected to decline, it is possible that for some goods production may rise, 2) a social gain is still possible even if outputs of other goods remains unchanged, 3) the price of non-tradable goods is expected to rise but it is not clear if the production of those goods will increase or decrease. Although the Gregory and Snape laid the base for the Dutch Disease literature, the term Dutch Disease is never mentioned in their articles.

The extensions made by Snape laid the foundation for the most influential model in Dutch Disease theory by Cordon and Neary (1982), which will be discussed in part two of this literature review. Cordon (1984) extended their model two years later to fill some theoretical gaps, these however are not relevant for this thesis since the 4 symptoms of the Dutch Disease that followed from their original model remained unchanged.

Van Wijnbergen (1984) added another argument to the discussion when he argued that a temporary decline in the manufacturing sector caused by a boom in the energy sector could lead to a permanently lower national income when most economic growth is caused by Learning by Doing induced technological progress. The Learning by Doing effect was introduced by Arrow (1962) and it refers to the capability of workers to raise their productivity through

8 repeating the same type of action. Oomes and Kalcheva (2007) state that Learning by Doing effects are common in manufacturing sectors and manufacturing sectors tend to be more competitive and innovative than other sectors. Furthermore they argue that the technological progress from the manufacturing sector is known to spill over to other sectors and thus a temporary decline in manufacturing could thus lead to less of these technological spill overs and permanently lower economic growth.

Krugman (1987) analysed the Dutch Disease by treating the income generated from a resource boom as a pure transfer payment from abroad in a model of trade in which comparative advantage changes over time as a consequence of Learning by Doing. The effects of the transfer in his model depend on the duration and the size of the transfer; if the transfer is small domestic wages will increase but it would not change the pattern of specialization. A larger transfer however will raise domestic wage by so much that the comparative advantages for some sectors changes through increasing labour costs and as a result of that some sectors will move abroad. For the longer run implications the duration of the transfer payment is important, if the duration is short, wages will return to their old positions after the transfer ends and the sectors that moved abroad will return. But for a large transfer of long duration all of the sectors that moved abroad will remain abroad even when the transfer ends and this causes the home country’s exports and relative wages reduced permanently.

Torvik (2001) extended the literature on modelling the Dutch Disease with Learning by Doing effects with a new model in which both the traded and non-traded goods sector can contribute to learning and the model assumes that there are learning spill-overs between the two sectors. In this model it is showed that a foreign exchange gift results in long run real exchange rate depreciation because the relative steady-state productivity between the traded and non-traded goods sector shifts in favour of the non-traded sector. Stijns (2005) argues that Dutch Disease effects might be present in countries that are rich in land, oil and gas. Natural resource abundance however is not a significant determinant of economic growth. Furthermore he argues that what matters most is what countries do with their natural resources. Finally Matsen and Torvik (2005) state that some Dutch Disease is always optimal because it is part of an optimal growth path, in the sense that a positive fraction of the resource wealth should be consumed in each period. Lower growth in resource rich economies is thus part of an optimal growth path and not a sign of a malfunctioning economy.

Early empirical research on the effects of oil windfalls is Gelb (1988) who conducted an comparative analysis of six developing resource rich economies during the period of 1973 and 1979. He concludes that much of the potential windfalls from oil revenues have been lost and that some countries are actually worse off despite the extra revenues. This work was followed by the work of Sachs and Warner (1995) mentioned in the introduction and the work of Davis

9 (1995) who found that developing mineral economies as a group do not perform worse in the long run than non-mineral developing economies in a study of 97 countries, which somewhat contradicts the results of Sachs and Warner (1995). Later Sachs and Warner (1999) found more evidence in favour of their earlier work in an attempt to analyse whether natural resource booms work as catalysts for economic development. They analysed 7 Latin-American countries and found that natural resource booms rather slow economic development down than increase it.

Several studies have tested whether individual countries suffer from the Dutch Disease. In a comparative study between the government policies on oil booms of Indonesia and Mexico, Usui (1997) concludes that Mexico clearly suffers from the Dutch Disease but that the Indonesian government reduced the effects of the natural resource curse through balanced spending of the revenues generated by it. Larsen (2004) shows how Norway has escaped the Dutch Disease through deliberate economic policy, strong institutions and a good juridical system. Beine et al. (2012) did an extensive econometric analysis on the basis of which they concluded that the Dutch Disease is a problem for the Canadian economy and that it challenges current economic policy. Some other country studies diagnosing the Dutch Disease are Pegg (2013) on Botswana, Hjort (2006) on Botswana, Indonesia and Norway and Mainguy (2011) on Mali.

2.2 The Dutch Disease model by Cordon and Neary (1982)

This section describes the theoretical model on the effects of a boom in one specific sector introduced by Cordon and Neary in 1982. The model assumes a small open economy producing two goods that are traded internationally at exogenously given world prices: ‘energy’ (XE) and ‘manufacturing’ (XM) and a third good ‘services’ (XS) which is traded domestically at a price that equalizes domestic demand and supply. The model is a purely real model, monetary effects are ignored, all prices determined are relative and expressed in terms of given prices or traded goods. It is assumed that national output and expenditure are always in equality so that overall trade is balanced. This does not mean that trade in one specific factor has to be balanced; it is possible that there is a trade surplus for one factor but then there has to be a trade deficit for the other factor, so that overall trade is balanced. Cordon and Neary (1982) provide no clear reason for this assumption in their original paper but it is assumed that this assumption is made to simplify the analysis.

The boom is modelled as a once and for all Hicks-neutral improvement in technology, a change is considered Hicks-neutral if the change does not affect the balance of capital and labour in production. This means that the relative share of capital and labour used in production does not change; both have been multiplied with the same factor. Although a natural resource

10 discovery is not exactly the same as an improvement in technology, the differences raised by the special case of natural resource discovery are not necessarily crucial. (Cordon & Neary, 1982)

Furthermore it is assumed that output in each sector is produced by a single factor specific to that sector, this can be interpreted as capital, and by labour, which is perfectly mobile and moves flexibly between sectors as to equalise its wage in all three employments and that the wage rate is perfectly flexible to ensure full employment. Furthermore there are no distortions in factor and commodity markets, this excludes the possibility of immiserising growth. Immiserising growth is a theoretical situation described by Bhagwati (1958) and it refers to the rare situation in which economic growth could leave a country worse off than before the growth. This can only happen when the country is able to influence world prices and when the growth is heavily export biased. If the growth then expands exports to a point where world prices are lowered so much that it worsens the terms of trade significantly this could lead to a lower real income.

Finally it is assumed that the terms of trade are fixed so that the relative price of energy and manufactures does not change. However the real exchange rate can change as it is defined as the relative price of non-traded goods to traded goods, a rise in the price of services (non-traded good) would correspond to a real appreciation.

2.2.1 Equilibrium

The starting point is the pre-boom equilibrium, which in figures 2.1 and 2.2 correspond to the points A and a. Figure 2.1 describes the labour market with the wage rate measured in manufactured goods on the vertical axis and the total labour supply on the horizontal axis OSOT.

Figure 2.1: Labour market and the effect of a boom.

The labour market with wage on the vertical axis and labour input on the horizontal axis. Ls, Lt and Lm show demand

for services, traded goods and manufactured goods respectively. Source: (Cordon & Neary, 1982).

11 Labour input in services is measured from the left by the distance with OS and labour input in the two traded goods is measured from the right by the distance with OT. It is assumed that the demand for labour in each sector is a decreasing function of the wage rate relative to the output in that specific sector. Thus LS represents the labour demand schedule for the service sector. Similarly LM corresponds to the labour demand schedule for the manufacturing sector which is laterally added to the labour demand schedule for the energy sector to obtain LT, the total labour demand schedule for the two traded goods combined. The initial equilibrium for the labour market is at point A at the intersection of LS and LT and thus the initial wage rate is w0.

The price of services is determined endogenous and illustrated by figure 2.2, with traded goods on the vertical axis and services on the horizontal axis. It is allowed to aggregate energy and output in a single composite good XT because the terms of trade are fixed. The production possibilities curve is represented by TS and equilibrium is at point a where the production possibilities curve is tangential to the highest attainable indifference curve I0, which represents

demand. The price of services, and thus the real exchange rate since PT is fixed, is given by the slope of the common tangent to the two curves in point a.

2.2.2 The effects of a boom

In order to consider the effects of a boom in a structural manner a distinction has to be made between the two effects that are the consequence of the boom, namely the spending effect and the resource movement effect. The boom causes marginal products for the mobile factor employed there to rise, this draws out resources from the other sectors and gives rise to various

Figure 2.2: Commodity market and the effect of a boom.

The commodity market, with services measured in goods on the horizontal axis and traded goods on the vertical axis. The indifference curves represent demand.

12 adjustment mechanisms in the economy, one of those being the real exchange rate, this is in short the resource movement effect. However if the booming sector does not use many resources the major effect of the boom could become the spending effect, which is the extra spending on services resulting from the higher real income caused by the boom. This extra spending on services raises their price and thus the real exchange rate, which in turn leads to further adjustments. Next both effects will be discussed in depth.

The resource movement effect

We will use a two-stage approach to discuss the resource movement effect, at first the real exchange rate will be held constant and then it is allowed to vary to restore equilibrium in the market for services. The boom causes the energy sector’s labour demand schedule to shift upward proportional to the technological improvement. This causes the aggregate labour demand LT to shift to L’T and thus there is a new equilibrium at point B, illustrated in figure 2.1. This causes the wage rate to shift from w0 to w1 and causes labour to shift from services and manufacturing to the energy sector. Employment in the manufacturing sector falls from OTM to OTM’ and this effect thus gives rise to direct de-industrialisation.

Figure 2.2 illustrates that the production possibilities curve changes to T’S because the maximum output of traded goods is raised from OT to OT’ as a consequence of the boom and the production changes from point a to point such as b, where it intersects with the highest attainable indifference curve. The movement of labour out of the service sector leads to a fall in service output and thus point b lies to the left of point a.

The second stage of the resource movement effect is to no longer hold the real exchange rate constant. Furthermore we want to exclude the spending effect, so we assume that the income-elasticity of demand for services is zero. This implies that the income-consumption curve is a vertical line through point a intersecting T’S at point j in figure 2.2. At this point the initial real exchange rate (the price of services) leads to an excess demand for services, since point j lies to the right of point b. In order to restore equilibrium the price of services must rise, switching demand away from services and thus restore some of the output lost because of the resource movement effect. However the effect cannot be reversed and thus the equilibrium in figure 4 will lie somewhere between points b and j on T’S.

The spending effect

In order to consider the spending effect on its own we have to assume that the energy sector uses no labour. Hence the boom will have no effect on the labour market because LT and LM would coincide to one curve. In the market for commodities the production possibilities curve is moved vertically upwards, with point b lying vertical above point a. Assuming that the

13 demand for services rises with income (it is assumed that services are a normal good) the demand for services at the initial exchange rate moves along an income-consumption curve as On which intersects T’S at point c. Again there is an excess demand for services and therefore the price of services must rise, which corresponds to an appreciation of the real exchange rate. The equilibrium will lie between the points j and c on the curve T’S and thus output will rise compared to the initial situation, since any point between j and c is further away from O than point a.

Both effects combined

When both effects are combined the outcome in the market for commodities is ambiguous. Both effects contribute to a real appreciation, in the final equilibrium point g in figure 2.2 there is a higher price for services than in the initial equilibrium point a. However the spending effect raises the output of services while the resource movement effect lowers the output of services and it is unclear which effect will dominate. Figure 2.2 illustrates the case where the spending effect dominates and thus point g lies to the left of point j, which is horizontally above a and thus represents the initial output of services.

For the labour market the effect is not ambiguous. Because the price of services has

risen the labour demand schedule for services shifts upward from LS to L’S and the new equilibrium lies in point G, illustrated in figure 2.3. In the new equilibrium the wage has shifted upwards to w2 causing manufacturing employment to fall even further to M’’, this effect is called indirect de-industrialisation. The boom thus gives rise to direct en indirect de-industrialisation, the direct effect is caused by the resource movement effect alone and is represented by the

Figure 2.3: Resource movement and Spending effect combined.

The effects on the labour market when both effects are combined. Source: (Cordon & Neary, 1982).

14 distance MM’ and the indirect effect is the consequence of the real appreciation of the exchange rate which is caused by the resource movement effect as well as the spending effect.

When both effects are combined it is thus uncertain whether the output of services will rise or fall. But this uncertainty does not hold for the other consequences of the boom, the real exchange will rise, manufacturing output will fall and the wage rate rises for all sectors. Table 2.1 summarizes all the consequences of the boom per effect and for both effects combined.

2.3 Empirical research on the Dutch Disease in Russia

The vast literature strand that empirically tests Dutch Disease symptoms in a wide range of resource rich developing and developed countries, differs from studies on the Russian economy because the Russian economy has gone, or is still going, through a transition phase from a communist economy into a market economy, therefore the economic background when analysing Russia is very different. This gave rise to discussion whether the Dutch Disease symptoms found in Russia were in fact merely transition phenomenon and not symptoms of the Dutch Disease.

To structure the empirical research on Russia a distinction is made between the 4 hypotheses that can be derived based on the analysis of the model of Cordon and Neary (1982) and which can be regarded as symptoms of the Dutch Disease:

Table 2.1: Effects of a boom.

Output Employment Wage Price

Resource movement effect

Energy sector + + + given

Manufacturing sector - - + given

Service sector - - + +

Spending effect

Energy sector - - + given

Manufacturing sector - - + given

Service sector + + + +

Combined effect

Energy sector ambiguous ambiguous + given

Manufacturing sector - - + given

Service sector ambiguous ambiguous + +

The effects of a boom for each sector for the resource movement effect, the spending effect and for both effects combined.

15 1. An appreciation of the real exchange rate.

2. A drop in manufacturing output and employment. 3. An overall rise in the wage rate.

4. A rise in service sector output and employment. Although when both effects are combined the effect on the service sector is ambiguous one can expect the spending effect to be dominant when the energy sector employs relatively few workers and labour mobility is low. Which is the case for Russia. (Oomes & Kalcheva, 2007)

There are several authors who tested one or more of these symptoms of the Dutch Disease in the resource rich Russia, with mixed results.

Oomes and Kalcheva (2007) tested all four hypotheses on Russia for the period between 1997 and 2005. Based on cointegration techniques they found that a 1% increase in the oil price leads to a 0.5% increase in the real exchange rate providing suggestive evidence for hypothesis 1. However they did not find evidence that the real exchange rate has been overvalued over the period of 1997 to 2005 and thus hypothesis 1 cannot be fully confirmed. They did find evidence, confirming hypothesis 2, of relative de-industrialisation by using OLS estimates and descriptive analysis, but for them more micro-level research is needed before this can be ascribed to the Dutch Disease alone. The same applies to hypothesis 3 and 4, that is an overall rise in the wage rate and a relative increase in service sector output and employment. Both hypotheses were confirmed using descriptive analysis, however it is hard to conclude that they are indeed the result of the Dutch Disease. All these effects could also be natural transition phenomenon or rebound effects after the 1998 crisis. With their results Oomes and Kalcheva confirm the work of Westin (2005) and Roland (2005) who found some symptoms of the Dutch Disease but find it premature to speak about full-blown Dutch Disease in Russia.

Ahrend et al. (2007) did a comparison study over the period between 2000 and 2005 between Russia and Ukraine and found no evidence of de-industrialisation in Russia. They conclude that although there are some indications that resource growth and rising commodity prices had a negative effect on the manufacturing industry, the effect does not seem to be very large. Furthermore they find the term Dutch Disease not applicable for Russia, because the term implies that Russia has suffered from the abundant resources. They state that this is not the case because Russia’s GDP per capita and household disposable incomes are much higher than in the Ukraine, mainly because of the natural resource abundance. Overall, they conclude, Russia is much better off having large natural resources even if this implies some negative effects for the manufacturing sector.

16 Beck et al. (2007) find that, confirming hypothesis 1, the ruble has appreciated significantly, in real terms, since 2000 but explain this appreciation as an adjustment mechanism to rising oil prices. Evidence found for hypothesis 2 is mixed, although the manufacturing sector has not grown significantly slower than other sectors profitability in manufacturing grew the least compared to other sectors. Hypothesis 3 is confirmed, as overall wages have increased significantly since 2000 but this might be caused by more factors than the spending effect alone, like productivity gains and recovery from the 1998 crisis. At last they find that the service sector has grown faster than the other sectors in the economy but there are two caveats to this result. The first is the fact that Russia’s service sector is not strictly non-tradable anymore because Russia also exports services nowadays and secondly the growth of the service sector may be largely caused by the transition of the Russian economy to a market economy.

Contradicting the previous mentioned papers Barisitz and Ollus (2007) find evidence of de-industrialisation. They have found a clear trend of increasing overall imports that compete with domestic production and for some sectors even outstrip it. Their results are derived from EU-25 trade data which corresponds to about half of Russia’s value-based imports and they diversify between 15 different sectors. For instance in sectors like leather products, machinery and equipment and electrical and electronic equipment imports are much higher than domestic production. Furthermore import growth exceeds production growth in all sectors except for electrical and electronic equipment, providing more evidence for de-industrialisation.

Algieri (2011) tested hypotheses 1 and 2 simultaneously in a VECM framework with a third hypothesis, namely a flourishing economic situation pushed by higher oil prices. The results suggest that there are three long-run cointegrating vectors, confirming hypotheses 1 and 2 and the third one added by Algieri. More specifically it was found that a 1% increase in oil prices lead to 0.4% appreciation of the real exchange rate (similar to the results of Oomes and Kalcheva), a 0.3% rise in GDP and 0.3% drop in the ratio of manufacturing to service output. This last result implies that a rise in the oil price causes a drop in manufacturing output and thus provides evidence for the Dutch Disease symptom of de-industrialisation. Furthermore hypothesis 3 was confirmed, since a strong overall rise in wage rates was found.

The empirical evidence for Russia thus seems mixed and further research is needed to analyse if Russia suffers from the Dutch Disease or that the observed symptoms are in fact caused by other factors than the Dutch Disease. Further research on the Dutch Disease in post-soviet economies is Rosenberg and Saavelainen (1998), Singh and Laurila (1999) and Hasanov (2013) on Azerbaijan and Kutan and Wyzan (2005) and Egert and Leonard (2006) on Kazakhstan.

17

3. METHODOLOGY

In this chapter the 4 symptoms of the Dutch Disease that followed from the model of Cordon and Neary (1982), are formulated as hypotheses and it is described how these hypotheses will be tested. The first paragraph discusses symptom 1, the second and third symptom are treated simultaneously in paragraph 2 and the third paragraph discusses symptom 4. Furthermore in paragraph 4 it is described how the transition state of the Russian economy is analyzed.

3.1 Appreciation of the real exchange rate

The first symptom of the Dutch Disease is an appreciation of the real exchange rate as a consequence of a boom in the oil sector. To test this the real effective exchange rate (REER) has to be used since the REER reflects the purchasing power of a currency relative to another currency at current exchange rates and prices. This basically means that the real exchange rate is the nominal exchange rate that takes the price differences among the countries into account; this yields the first hypothesis of this thesis:

1. The real effective exchange rate appreciates as a consequence of a boom in the oil industry.

To test this hypothesis the REER of Russia is plotted against the Urals oil price, which is a reference oil brand that is commonly used as a basis to price the exports of Russian natural resources. To fully prove that higher oil prices cause appreciations in the REER a regression analysis has to be made with the REER as dependent variable and the Urals oil price as independent variable while controlling for other factors that could influence the REER like government consumption, net international reserves or the production differential. However the econometrics of such a regression analysis become quite difficult, and is beyond the scope of this thesis, because of simultaneous causality between the dependent and independent variables.

The REER is the weighted average of the real exchange rates of the currency of one country to the currencies of its main trading partners and it is calculated in two steps. First the real exchange rates of the national currency to the currencies of its main trading partners have to be calculated. Real exchange rates are usually converted to indices and are calculated in the following way. The real exchange rate of, for instance, the currency of Russia to that of the currency The United States is:

18

= 100 * (E

/E

) * (CPI

/CPI

)

Where RE$t is the real exchange rate of the ruble to the dollar, that is the ruble divided by the

dollar, for Russia at time t,

E$t is the nominal exchange rate of the ruble to the dollar at time t,

E$0 is the nominal exchange rate of the ruble to the dollar in the base year 0 and

CPIUS and CPIRU are the consumer price indices for period t relative to the base year 0, for

respectively the United States and Russia.

The second step in calculating the REER is to calculate the weighted average of the real exchange rates of the national currency to that of the currencies of its main trading partners. Following the previous example the weight of country i in the overall trade of Russia is calculated according to the following formula:

∑

∑

With being the weight of country i in the overall trade of Russia with ∑ , Mi being the imports to Russia from country i,

Xi being the exports from Russia to country i,

∑

∑

When the weights for all main n trading partners have been calculated the REER can be calculated as follows:

With REit being the real exchange rate of the ruble to the currency of country i and wi being the

weight of country i in the overall trade of Russia.

3.2 A decline in manufacturing output and a rise in service output

The second and third symptom of the Dutch Disease are discussed simultaneously in this paragraph and are respectively a decline in output and employment of the manufacturing sector and a rise in output and employment of the service sector. For comparative reasons the symptoms are discussed simultaneously while differentiating between the effects on output and employment.

19 An absolute decline in manufacturing output, de-industrialization, however is unlikely to expect because the Russian economy has grown rapidly since the beginning of the data period. Similarly it is very likely that the service sector has grown in absolute sense since 1996. Furthermore the predictions of the Dutch Disease model are ceteris paribus, meaning that when all other determining factors remain equal a boom in the oil sector causes an absolute decrease in manufacturing output. However it is very likely that other factors determining manufacturing output have offset the possible decrease in manufacturing output caused by a boom in the oil sector, like technological improvements, increases in domestic demand through wealth increases.

Therefore symptoms two and three are changed to a slower than average growth of the manufacturing sector and a higher than average growth of the service sector, implying that the share of services in GDP relative to the share of manufacturing will rise, which is called relative de-industrialization. An absolute decline in manufacturing employment or absolute increase in service employment however are not unlikely but in order to correct for economy wide declines or raises in employment they have to be compared with total employment. This yields the second and third hypothesis of this thesis:

2. The share of service sector output in GDP rises relatively to the share of manufacturing sector in GDP as a result of a boom in the Oil sector.

3. Manufacturing employment decreases relative to total employment and service employment rises relative to total employment as a result of a boom in the Oil sector.

To test the second hypotheses total Russian output is divided in three sectors: fuel, non-fuel and construction and services. To construct the necessary data firstly the data provided by the Russian authorities on the size of the fuel and non-fuel sectors had to be merged because Russian Federal State Statistics Service used a new industrial classification scheme after 2005. Secondly the fuel sector was divided in construction and services and the remaining non-fuel part.

3.3 A rise in overall wages

The fourth symptom of the Dutch Disease is an overall increase in the wage rate. Since all variables in the model of the Dutch Disease are purely real one’s the real wage rate is used to analyze the fourth symptom. This yields the fourth hypothesis of this thesis.

20 The real wage rate index is commonly used when analyzing wages since it reflects the changes in purchasing power of an individual. It is the nominal wage corrected for changes in consumer prices and it is calculated as follows:

The real wage rate was calculated on the basis of data on the nominal wage rate provided by Russia’s federal state statistics service and data of the consumer price index which was taken from the database of the International Monetary Fund. Unfortunately the Russian federal state statistics service only provided data of the nominal wage rate in 2000, 2005 and 2007 until 2012, thus a year on year analysis is not possible the period of 2000 to 2007. Furthermore no data is provided for the period before 2000.

3.4 Transition state

To correct for the transition effects on the Russian economy the EBRD transition report is used to assess the current state of transition in Russia from a central planned economy to a free market economy. The EBRD is the European Bank for Reconstruction and Development and their mission is to foster transition to open and democratic market economies from central Europe to central Asia and in southern and eastern Mediterranean countries. Each year the EBRD publishes a transition report in which it updates the current understanding of transition economics with new insights in the field and in which it assesses the progress of transition on a number of indicators for each country in transition.

Since the EDRB began quantifying transition indicators in 1994 the amount of indicators has grown to 16 indicators in four categories: Corporates, Energy, Infrastructure and Financial Institutions in 2013. Each indicator is given a value between 1 and 4+, with 4+ being the highest score, where a score of 1 represents a central planned economy and a score of 4+ a fully free market economy. The scores are based on publically available data or on observable market structures and institutions. Based on the scores the EBRD defines the transition gaps for each sector of each country, with the scores of 1, 2, 3, and 4 representing large, medium, small and negligible transition gaps respectively. The transition gap is defined as the difference between the current state of the economy and the free market economy.

The EBRD transition indicators are used widely in academic research when the link between transition and other factors is studied. For instance studies that research the link

21 between transition and economic growth are Berg et al. (1991), Havrylyshyn and van Rooden (2003), Falcetti et al. (2002), Falcetti et al. (2006) and Eicher and Schreiber (2010).

22

4. RESULTS

This chapter describes the results from testing the four hypotheses that were drafted in chapter 3 and the transition state of the Russian economy. The first three paragraphs discuss hypothesis 1, 2 and 3 simultaneously and 4 respectively, and finally paragraph four treats the transition.

4.1 Appreciation of the real exchange rate

The following graph shows a plot of the REER and the Urals oil price between 1994 and 2013. From the graph it seems that the REER and the Oil price follow a similar path and that changes in the oil price are followed by smaller changes in the REER. This best illustrated at the

beginning of the financial crisis at the end of 2008 when the oil dropped sharply, followed by a less severe but also sharp drop of the REER. This is further strengthened by the high correlation of 0,81 between the two variables. The period mid-1995 to 1998 seems an exception to this relationship when the REER rose sharply while the oil price stayed relatively stable. But this was probably caused by the fixed exchange rate regime and the very high inflation in Russia, which ended in a financial crisis in 1998 after which the REER dropped back to values of the period before 1995.

The plot seems in line with earlier results of Spatafora and Satvrev (2003), Oomes and Kalcheva (2007) and Algieri (2011) who found long run elasticities of the oil price on the REER of 0,372; 0,49 and 0,31 respectively, this means that a 1% increase in the oil price leads to

Figure 4.1: The REER and Urals oil price.

The Urals oil price and the REER for Russia between 1994 and 2013. Calculation of the REER was done by Nienke Oomes and her colleagues, special thanks to them for contributing this data.

Source: Bank for International settlements & Bloomberg 0 20 40 60 80 100 120 140 19 94-01 19 95-01 19 96-01 19 97 -01 19 98 -01 19 99-01 20 00-01 20 01-01 20 02-01 20 03-01 20 04-01 20 05-01 20 06-01 20 07-01 20 08-01 20 09-01 20 10-01 20 11-01 20 12-01 20 13-01 REER Oil price

23 roughly a 0,4% increase in the REER. Although their estimations of the effect differ somewhat the standard errors of the coefficients are roughly the same, implying that their results are not statistically different. The differences in the strength of the effect are probably caused by the relative short time span of the analysis. This result combined with the previous research provide confirming evidence for the first symptom of the Dutch Disease that the real effective exchange rate is dependent of changes in the oil price.

4.2.1 A rise in the share of services in GDP relative to manufacturing

Figure 4.2 shows the growth path of the oil, services and manufacturing sector with 1996 as base year and the Urals oil price. The figure shows that output of services and manufacturing followed a similar path until the end of 2004, in the beginning of 2005 however the service sector started outgrowing the manufacturing sector, at the same time the oil price

started rising. Output in both sectors dropped during the financial crisis at the end of 2008 and started rising again in the beginning of 2010. The data represented in figure 4.2 shows that during times of high oil price output of services grew faster than manufacturing output providing suggestive evidence for the second and third symptom of the Dutch Disease.

Further evidence for these symptoms is represented in figure 4.3, which shows the growth rates per year per sector. A growth rate of 0,05 represents a 5% growth relative to the previous year. The figure shows that the service sector outgrew manufacturing each year starting from 2000 except for 2003, 2009 and 2010 and especially during the times of high oil prices from 2004 until 2007. Naturally there are more factors than oil prices influencing the output per sector and it is not unlikely that the exceptions in the years 2003, 2009 and 2010 were caused by non-oil shocks. The decrease in services in 2009 for instance could be caused by

Figure 4.2: Growth index services, oil and manufacturing (1996=100) and the oil price.

The growth indexes of the services, oil and manufacturing sector and the Urals oil price. Source: Russian Federal State Statistics Service

0 50 100 150 200 250 1Q9 6 4Q9 6 3Q9 7 2Q9 8 1Q9 9 4Q9 9 3Q0 0 2Q0 1 1Q0 2 4Q0 2 3Q0 3 2Q0 4 1Q0 5 4Q0 5 3Q0 6 2Q0 7 1Q0 8 4Q0 8 3Q0 9 2Q1 0 1Q1 1 4Q1 1 3Q1 2 2Q1 3

24 the freezing of the financial service sector, which is part of the service sector, during the financial crisis. Similarly the winter Olympics being held in Sotsji in February 2014 could be a factor that stimulated the manufacturing sector in 2010 and 2011. However when following this line of reasoning it could also be that the relative increase of the service sector was caused by other factors than the Dutch Disease. De-industrialization for instance is also found in other economies that are not abundant in natural resources such as the United States, because when wealth increased demand shifts away from goods towards service Oomes and Kalcheva (2007). Further econometric analysis is needed to study if there is a causal relationship between oil

prices and the output of services and manufacturing, however the data illustrated in figure 4.2 and 4.3 point in this direction.

The financial crisis, or the very sharp rise in oil prices in the years preceding the crisis, did not seem to have any effect on the growth path of the oil sector. This is not surprising since demand for oil is known to be quite inelastic, short run price elasticities have been estimated at -0,034 (Hughes, Knittel, & Sperling, 2008), -0,05 (Cooper, 2003) and -0,07 (Dahn & Sterner, 1991), with long run price elasticities at roughly three or four times the size of the short run elasticities. In the short run the price elasticity of supply is also inelastic, mainly due to large lead times between the discovery of an oil field and the moment the oil actually arrives at a refinery Hamilton (2008).

Figure 4.3: % growth rates per sector per year.

The annual % growth for the service, oil and manufacturing sector. Source: Russian Federal State Statistics Service

-0,15 -0,1 -0,05 0 0,05 0,1 0,15 0,2 Services Oil Manufacturing

25

4.2.2 Decline of manufacturing employment and increase in service employment

Figure 4.4 shows manufacturing employment and total employment in thousands of persons, with the values for manufacturing displayed on the left vertical axis and the values for total employment on the right vertical axis.

Manufacturing employment drops sharply over the period of 2005 until 2008, while total employment kept rising. This shows that manufacturing employment decreased in absolute terms as well as in relative during the period of high oil prices.

However manufacturing output kept rising throughout the period of 2005 until 2008 as shown in figure 4.1, implying that productivity in the manufacturing sector has risen. This could be explained in two ways, either productivity increased which allowed employers to employ less workers and still increase output. Or employers were faced with declining profits, through rising real effective exchange rate, and were forced to increase productivity and employ less workers in order to stay competitive. In line with the Dutch Disease hypothesis the second explanation seems more plausible, however confirming evidence for this is yet to be found. Unfortunately the Russian Federal State Statistics Service does not report on employment specific in the service sector and no other data for this was found. Therefore it is uncertain if employment in the service has increased or not. Thus only half of the hypotheses can be answered in the affirmative.

Figure 4.4: Manufacturing and total employment.

Manufacturing and total employment in thousands of persons. The left horizontal axis represents manufacturing employment and the right axis represent total employment.

Source: Russian Federal Statistics Service data on industrial output.

63000 64000 65000 66000 67000 68000 69000 70000 71000 72000 9500 10000 10500 11000 11500 12000 12500 13000 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Manufacturing Employment Total employment

26

4.3 Increase in overall wages

The changes in the real wage rate are shown in figure 4.5. The bars represent the annual percentage change in the real wage rate in the year mentioned secondly relative to year mentioned first on the horizontal axis. For instance the real wage rate in 2008 was 11,47%

higher that the real wage rate in 2007. For multiple years the bars represent the compounded average annual growth rate, for instance the wage rate grew on average 14,01% each year from 2000 to 2005. In the years 2000 to 2008 the real wage rate grew with more than 10% each year, then there was a slight dip during the crisis year of 2009, but after that the real wage growth started to recover and almost reached 10% again in 2012. During the period of high oil prices 2000 until 2008 and in 2012 when the oil price started to recover, the wage rate thus increased faster than during the period of lower oil prices from 2009 until 2011.

Further evidence for the fourth hypothesis of this thesis is found in figure 4.6, which shows the annual percentage change in real wage rate divided by the change in GDP. From 2000 the real wage rate outgrew GDP each year except for 2011, this means that the wage rate has grown relative to GDP. This further strengthens the hypothesis of increasing wages, since a rising GDP implies that wages will rise too. Thus for a relative rise in the wage rate it has to outgrow GDP which has been the case for Russia as illustrated in figure 4.6.

Figure 4.5: % Change in real wage rate.

Annual % change in real wage rate of Russia. Source: Russian Federal State Statistics Service and IMF 14,01% 15,28% 11,47% -3,46% 5,21% 2,85% 9,24% -0,05 0,00 0,05 0,10 0,15 0,20 2000-05 2005-07 2007-08 2008-09 2009-10 2010-2011 2011-12 % change in real wages

27 Although this result confirms hypothesis 4 it is uncertain whether this rapid increase in the real wage rate was caused by the Dutch Disease alone. For instance other factors that might have caused the sharp rise in the period of 2000 until 2007 could be a recovery from the 1998 crisis, productivity increases or the transition from a central planned economy to a free market economy. Furthermore the negative and slower growth from 2009 to 2011 could have been caused by other factors than low oil prices, for instance by the worldwide slowdown after the financial crisis of 2008.

Figure 4.6: % change in Wage/GDP.

The % change in the annual growth in the real wage divided by the annual growth of GDP. Source: Own calculations based on data from the Russian Federal State Statistics Service. 7,46% 6,45% 5,96% 4,71% 0,68% -1,39% 5,65% -0,02 -0,01 0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 2000-05 2005-07 2007-08 2008-09 2009-10 2010-11 2011-12 % Change

28

4.4 Transition state of the Russian economy

The transition indicator scores of the EBRD for Russia of 2013 are illustrated in table 4.1. A value of 4,5 represents a score of 4+, which is the maximum score and it represents a

transition gap that is negligible, similarly a value of 1 represents a transition gap that is large. Scores of 2 and 3 represent transition gaps that are medium sized and small respectively. From the table it is clear that the Russian economy has not yet finished transitioning from a central planned economy to a market economy since no sector has obtained the maximum score and only 4 sectors are close to the maximum score. Furthermore there are still 3 sectors that still score a low value of 2.

It thus seems that the Russian economy remains in transition phase and that more structural reforms are needed to reach a free market economy. However the transition indicator scores for Russia have not changed much in the last 10 years as illustrated by table 4.2. Table 4.2 shows the change in transition indicator scores for Russia from 2002 until 2012. It shows 6 indicators instead of the 16 indicators of table 4.1, this is because the EBRD extended their indicator system from 6 to 16 indicators in 2010. Thus in order to compare the scores of before and after the extension of indicators, the indicators of after 2010 were merged by the EBRD so that they fitted the division of indicators of before 2010. In table 4.2 a minus sign represents a drop of a score by 0,33 with respect to the previous year, an equal sign shows that the score remained the same as in the previous year and plus sign shows an increase in score of 0,33 with two plus signs indicating an increase of 0,66 in indicator score.

In half of the indicators there has been no change at all in the scores over the last ten years and the indicators large scale privatization and competition policy have only had one 0,33

Table 4.1: EBRD transition indicators Russia in 2013

Transition indicator scores for Russia in 2013. Source: EBRD transition report 2013 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5

29 change in their scores. The only indicator that seems to have changed is trade & forex, however it also remained equal from 2004 until 2011.

The Russian economy may thus be far away from a free market economy, the transition however remained quite stable over the last 10 years. Which leads to the question if the transition from a central planned towards a free market economy still was an important factor influencing the Russian economy in 21th century? It seems unlikely that while the transition remained relatively unchanged it still caused large fluctuations in the real exchange rate of the ruble, relative de-industrialization and a rise of the real wage rate.

However this argument does not rule out the transition argument entirely, it could be argued that it is not the progress of transition which causes the previously mentioned effects but that it is the economic environment of an economy that is not finished transitioning to a fully free market economy which is the decisive factor. And although there has been no progress in the transition, the economic environment is still that of a transition economy. But still it remains questionable if a factor that remained fairly constant could be causing large fluctuations in other important macro-economic factors.

Table 4.2 Change in transition indicator score

Indicator 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Large scale privatization = = = - = = = = = = = Small scale privatization = = = = = = = = = = = Governance & enterprise = = = = = = = = = = = Price liberalization = = = = = = = = = = = Trade & Forex + + = = = = = = = = ++

Competition

Policy = = = = = = = = = + =

Change in transition indicator score for Russia in the period 2002 to 2012. Source: EBRD transition reports 2002-2012

30

5. CONCLUSION

In this chapter a conclusion is drawn and an answer is formulated on the main research question of this thesis: To what extent does Russia suffer from the Dutch Disease? This question was answered by analyzing if Russia suffers from the 4 symptoms that are widely accepted as symptoms the Dutch Disease.

Supporting evidence for the first symptom, an appreciation of the real exchange rate as a consequence of an oil boom, is found in the high correlation and similar growth path between the Urals oil price and the real effective exchange rate of Russia, which confirmed earlier studies that found a similar relationship; Spatafora and Stavrev (2003), Oomes and Kalcheva (2007) and Algieri (2011). Some supporting evidence was found for the second and third symptoms, an increase in service sector output and employment and a decrease in manufacturing sector output and employment. The service sector has outgrown the manufacturing sector since 2000 and especially during times of high oil prices, however no absolute decrease in manufacturing output was found. Manufacturing employment however did absolutely decrease but unfortunately no data of service sector employment was found to compare it with. Furthermore evidence in favor of the fourth hypothesis of an overall increase in the wage rate was found, since the real wage rate has increased rapidly since 2000, except during the crisis year of 2009. Similar results on the increase of the service sector relative to the manufacturing sector and increases in the real wage were found by previous studies; Beck et al. (2007), Oomes and Kalcheva (2007) and Algieri (2011). However all analysis in this thesis is descriptive therefore more econometric analysis is needed to prove any causal relationships.

Although it looks like Russia is suffering from all symptoms it is unclear whether these observed symptoms are caused by the Dutch Disease alone. The argument that these symptoms are merely transition phenomenon presented by Ahrend et al. (2007) and Beck et al. (2007) is weakened somewhat through analyzing the EBRD transition reports of the last 10 years. Although the Russian economy is far from a free market economy the EBRD transition indicators have not changed much over the last 10 years, which shows that the transition has slowed down over the last years. This evidence might not entirely rule out the argument that the symptoms are caused by transition but it seems less likely that while the transition remained fairly stable it still caused large fluctuations in the exchange rate, real wage and output of the service and manufacturing sector.

On the basis of this thesis and the results of earlier studies it is thus concluded that the observed symptoms of the Dutch Disease still remain present to some extent in Russia with a longer data period than previous studies, this data period for this thesis was 1996 to 2013, whereas previous studies covered periods up to 2008. Furthermore the transition argument is

31 weakened a little bit, providing more evidence in favor of the Dutch Disease as the most plausible hypothesis for the resource curse in Russia. The answer to the research question of this thesis is thus that Russia shows much of the symptoms affiliated with the Dutch Disease, however it remains to be concluded if the observed symptoms are caused by the Dutch Disease alone.

But other explanations than the transitioning from a communist economy to a free market economy exist to explain the observed symptoms in Russia. The increase of the service sector relative to the manufacturing sector is commonly found in economies in which wealth increases, whether they are abundant in natural resources or not, because when wealth increases demand shifts away from goods towards services Oomes and Kalcheva (2007). More research is needed to compare the growth rates of de-industrialization in Russia with the growth rates found in other countries, which faced de-industrialization. Furthermore the sharp growth of the service sector could be partially explained by the strong growth of the financial service sector (Deutsche Bank, 2007).

What can be concluded is that the Russian economy is highly dependent on the revenues of oil and thus vulnerable for commodity-price volatility. Several authorities such as the World Bank, the IMF and the EBRD report this as one of the weaknesses of the Russian economy as well as the scientific studies on Russia Barisitz and Ollus (2007), Oomes and Kalcheva (2007) and Algieri (2011). It could then be argued that the true Dutch Disease in Russia is the lack of diversity in the economy and the high dependency on the revenues generated by the exports of natural resources and that the true challenge for Russia in the next decades is to reform the economy in such a way that it is stable, diverse and less dependent on the price of oil.

In the next chapter some policies will be discussed on how Russia could overcome this challenge as well as the limitations of this thesis and opportunities for further research.

32

6. DISCUSSION

This chapter will discuss the limitations of this thesis, opportunities for further research and policy measures that Russia could take to diversify and strengthen their economy.

The first limitation to this thesis is that only descriptive analysis is used to answer the hypotheses, therefore no causal relationships have been established. The reason for that is that simple regression analysis would give biased results because of simultaneous causality between the variables modelled. To overcome this cointegration techniques could be used, these techniques however are beyond the scope of this bachelor’s thesis. Furthermore other authors have proved a causal relationship between the real exchange rate and oil prices; Oomes and Kalcheva (2007) and Algieri (2011), therefore it was safe to assume that this causal relationship existed. However more micro level econometric is needed to analyze the effect on an oil boom on the different sectors.

Secondly data on service sector employment missed in order to completely analyze the effect of the oil boom on the service sector. Thirdly the data period is still relatively short and only one significant boom occurred in the oil industry in the period 2005-2008, and it is unwise to draw conclusion on the basis of a single event. Furthermore there were large gaps in the data for the real wage rate making a precise analysis impossible. The data did not go back further than 2000 and the years 2001 to 2004 and 2006 were missing.

Possible further studies could thus include an econometric analysis on the micro level comparing the effect of a rise in oil prices on the output of the service sector with the effect on manufacturing sector output. Furthermore the relationship between the oil price and the real exchange rate could be examined further than current studies have done for instance by doing a asymmetric analysis to differentiate between upward en downward effects. Other approaches could be to examine if there exists a causal relationship between the oil price and total employment and average real wage as well as differentiating analysis on the effect per sector.

Although there might thus be limitations to this thesis and more research is needed to fully conclude that the observed problems in the Russian economy are caused by the Dutch Disease, it is fairly evident that the Russian is economy highly dependent on oil revenues and that policy makers in Russia face the challenge of diversifying the Russian economy.

One of the problems limiting the growth of Russian firms are the large differences in the business environment between regions and the regional implementation of federal reforms. In order to stimulate economic growth and diversification these differences have to be reduced and the federal government should ensure that their policies are implemented correctly (EBRD, 2012). Another problem is the poor quality of management in the Russian economy, in order to diversify the productivity in non-resource firms has to rise, strong management however is

33 essential in achieving this. The EBRD suggests policies that strengthen competition, specialist management training, allowing entry of multinationals and improvement in the capital market. Furthermore the EBRD argues that stronger protection for intellectual property rights, more finance availability for research sectors and more private incentives to finance R&D are needed in order for Russia to catch up with the rest of the developed world in terms of innovation. Finally the Russian government should reduce its dependence on the oil revenues for the fiscal balance in order to maintain stability in the long run.