Resource curse in the 21st century

(1)

Resource curse in the 21st_century.

Abstract:

This paper is based on the ongoing research of Sachs and Warner’s resource curse. These papers examined the relation between resource abundance and average growth in the 21st

century. Because of resource depletion the resource abundant countries need to adapt to other kind of activities to be profitable in the future. In a cross sectional regression of 73 countries the resource curse in the 21st_{century wasn’t present. However more research}

should be done over more years to proof that indeed the resource curse effect declined.

Ivo Postma 11056509

Bachelor thesis (6013B0345Y) 26-06-18

Supervisor: dhr. C.W. (Kees) Haasnoot

(2)

Statement of Originality

This document is written by Student Ivo Postma who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this

document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

discovered that resource abundance countries had a slower GDP growth per capita from 1970 to 1990. Old Dutch disease models from Cordon and Neary or van Wijnbergen, give an explanation for this phenomenon . The resource curse can be linked to the way countries deal with their gains. A bad example of dealing with gains from resource export is

Venezuela. In Venezuela profits from the oil resources where managed badly by consuming it all without saving or investing. At the moment the price of oil decreased those countries ended up in a crisis (Monaldi, 2015). The fear of big shocks on the oil price market that affect a country severely, makes that these oil resource driven countries tend to move away from their oil and more diversify their economy. Studies have shown that countries like Botswana, Canada, Australia and Norway are less vulnerable against the curse or even seemed to have escaped the curse. Most of these studies gave as explanation for their escape the way these countries handle the gains from their resources. Mehlum et al. (2006) categories institutions in grabber friendly and producer friendly. The paper of Torvic (2008) gives an overview why some countries are more affected by the curse then others. What kind of political policies should countries apply to manage their resources correctly and diminish the resource curse or eliminate it completely? Since the paper of Sacks and Warner was

published a few things have changed in the 21st_{century. Some of the main resources will}

run out (Shafee and Topal, 2008). Also, countries have seen the example of Venezuela that showed how they shouldn’t handle their resource rents. This raises the question how

countries should manage their gains from their resources to avoid a crisis like Venezuela. Because in such case they can’t rely on their resources anymore. In this thesis the resource curse in the 21st_{century will be examined by using a regression based on the regression in}

Sachs and Warner’s papers. Do we still see a negative effect of resource abundance on growth in the 21st_{century? Have countries adapted a more efficient way of managing their}

resource gains by diversifying their economy to be relevant in the post-oil future? And what kind of policies should countries use to resist the resource curse?

(5)

2. Literature review 2.1 Overview

The topic of the resource curse has quite some literature. This literature review starts with the main papers of Sachs and Warner and some relevant papers that investigate more in debt the resource curse. Also some papers that oppose to the works of Sachs and Warner will be mentioned, followed by an explanation of the resource curse using three Dutch

disease models. At last the changes that happen during the transition from the 20th_{to the}

21st_{century will be described and it will be explained how these changes affect the Dutch}

disease models.

2.2 Resource curse

In 1995 Sachs and Warner published their first paper about the resource curse. In their studies they showed evidence of the resource curse. They discovered that high resource abundant countries tended to grow slower than countries who are low in their resource abundance (Sachs and Warner, 1995). The time period of the GDP growth per capita of their data was from 1970 to 1990. The resource abundance was taken from 1970 and was measured by all the resource exports divided by the GDP. In their cross sectional regression, they used the control variables openness, share of export, growth before 1980, gross domestic investment and rule of law index. These variables were all taken from the year 1970. Although they found evidence for a resource curse, they couldn’t give a explanation to the problem. One of the reasons was that their data were not complete. In their research they excluded oil rich countries like Bahrain, Qatar and Saudi-Arabia. In a later paper they added omitted geographical and climate variables (Sachs and Warner, 2001). These variables however seemed to have no significant impact on the growth. Also, the oil rich countries that were excluded in the first paper were included in the paper of 2001.

Not all research that has been done points into an existence of a resource curse like the papers of Sachs and Warner. Davis (1995) observed in his research no correlation of slow growth in mineral rich economies. In his research he tested if 22 developing mineral economies underperformed compared to other developing economies with the basis year of 1970. No evidence was found for underperformance by the mineral economies. This would suggest that mineral resource economies aren’t really affected by the resource curse. However, the paper of Torvik (2009) contradicted these result of Davis (1995). With a regression that only used oil and minerals as measurement for resource abundance he showed that the resource curse effect with only using minerals and oil had a stronger effect than the resource measurement used in Sachs and Warner (2001) in which all kinds of

(6)

resource exports where included in the measurement. Davis couldn’t find a direct

explanation why mineral economies didn’t suffer from the resource curse. He argued that the mineral economies compared to other developing countries are coincidental less dictatorial (Davis, 1995).

2.3 Dutch disease models

The Dutch disease models gives an explanation for the mechanism of the resource curse ( Sachs and Warner, 2001). The Dutch disease was first described by the Economist in 1977. The Dutch disease was a development in the Netherlands in 1959 when a large gas reserve was discovered. After the discovery of this gas reserve the competitiveness and growth in the Netherlands declined. This effect was quite unexpected for most people (Kiev, 2014). There are quite a few different models that explain the Dutch disease. Most of these Dutch disease models are based on the booming sector model of Cordon and Neary (1982). In this model they explain what happen to a small open economy during a booming of the traded goods sector. What follows after the boom is the de-industrialization in which the non-booming traded goods sector declines and with it the production and employment. After the decline of production follows a decline in the balance of trade of the non-booming traded goods sector and a fall in the real return on factors of the non-booming traded goods sectors.

A later model of the Dutch disease is the model of Matsuyama (1992). Matsuyama describes a Dutch disease model consisting of two sectors, the Agriculture- and the manufacturing sector. The difference between the two is that the manufacturing sector is more based on the learning by doing factor. This means that the more effort is put into the sector the return will increase over time. The learning by doing factor is an important factor for growth in the paper of Matsuyama. The model of Matsuyama showed that in a closed economy the agricultural sector had a positive relation with growth. However, in a small open economy the model had a negative relation with growth because the labour and capital shift from the manufacturing sector to the agricultural sector and thus reduce the learning by doing factor (Matasuyama, 2009). This would only count for resources that are labour and capital intensive like agricultural resources. For resources like oil this would be less off a problem because it doesn’t require a lot of labour and capital (Sachs and Warner, 1997). Following the model of Matasuyama the resource curse will still be relevant in the 21th century if countries don’t diversify between their manufacturing and agricultural sector.

(7)

overall than less educated people. Sachs and Warner (1997) points out that the

manufacturing sector will lead to a more complex labor division which results in a higher standard of living unlike the natural resource sector.

Another model of the Dutch disease is the model of van Wijnbergen (1984). In his model he showed how a high but temporarily oil revenue increases the consumption of the non-traded goods. The high demand for non-traded goods increases the production of the non-traded goods and decreases the production of the non-oil traded goods. The non-oil traded goods sector is like the manufacturing sector in (Matasuyama, 2009) characterized by learning by doing. When the traded goods sector shrinks, so will the Learning by Doing effects. With a decreasing in the Learning by Doing effects the country gets a comparative disadvantage in the production of the non-oil traded goods. The paper of van Wijnbergen (1984) argues that countries should subsidies the non-oil traded goods in periods of high resource gains if the non-oil traded goods show signs of Learning by Doing effects. Van Wijnbergen also argues that the extra subsidy on the non-oil traded goods isn’t necessary if the effect of high oil revenue on consumption decreases. By acquiring foreign assets when the revenue of oil is high, the cash flow of the foreign assets compensates for the lower oil revenue after the boom and thereby makes the consumption smoother. Countries who did this unintentional are countries like Saudi Arabia, etc. (van Wijnbergen, 1984).

2.4 Countries not affected by the resource curse

Although Sacks and Warner showed evidence for the resource curse not all countries suffer from it (Mehlum et al, 2006). Acemoglu et al. ( 2001) described in a working paper how the high in diamonds resource country Botswana witnessed one of the biggest growth in the last 35 years of the 20th century. A plausible cause of the growth in the paper of

Acemoglu was dedicated to the good institutions in Botswana, in which they had secure property rights and they had the opportunity to invest (Acemoglu et al, 2001). Sachs and Warner would argue in their paper of 2001 that resource rich countries are high price economics and therefore miss-out on export-led growth, unless it is the export of the resource which explains the growth in Botswana. More countries that don’t seem to suffer from the Dutch disease are: Canada, Australia and Norway ( Mehlum et al, 2006).

The paper of Mehlum et al. (2006) gives like Acemoglu as the cause of the resource abundance growers and losers their countries openness and respective institutions. The paper of Acenmoglu described good institutions as institutions of private property and bad institutions as extractive institutions. The paper of Mehlum et al. (2006) describes two other kinds of institutions namely grabber friendly and producer friendly institutions. Grabber

(8)

friendly institutions aim more to rent seeking rather than productive activities. Their model and regression validate their belief that institutions which are producer friendly can make the effect of resource abundance on growth rather positive than negative. Later work from Torvik (2009) showed in a regression that the quality of institutions had an impact on the resource curse. Van der Ploeg (2011) continues on the work of Mehlum et al. (2006) and Torvik (2009). Van der Ploeg states in his concluding remark that resource rich economies should invest their resource wealth in reproducible assets like human capital and infrastructure. In his research he found that especially resource rich countries with high corruption and bad quality institutions had a bad growth record. The institutions could be the same as in Mehlum et al. (2006) either be rent grabbing friendly or production friendly . In rent grabbing friendly institutions a fixed number of entrepreneurs will use their knowledge to seek rent from their resources, while they abandon the production sector. The production sector could, like in the Dutch disease models, use those entrepreneurs to boost the production which will benefit the whole economy. Van der Ploeg suggested that well developed countries should invest their gains from their resources in a sovereign wealth fund while less developed countries should use their gains to pay off their debt and to lower the interest rate so that countries can increase their capital accumulation and increase their economic development. Atkinson (2003) found evidence that resource abundant countries with good qualities institutions had greater rates of investments. Also, the resource rich countries that suffer from the resource curse appear to be the countries who had a low genuine savings rate. Van der Ploeg continues to argue that the saving pattern of a country is affected by politicians who may be voted out of office next period.

Countries with lots of resources are more easily infected with corruption. In Kolstad and Soreide (2009) the reason for this phenomenon is due to the large rents that are on some of the resources because of scarcity. These rents give agents more incentives to act corrupt to acquire some parts of these rents. Kostad and Soreide argues further that resource exploitation and extraction requires complex contractual and financial

arrangements. It is difficult to make standardized procedures for these arrangements. This makes it easier to exploit corruption. In most of the developing resource rich countries the institutional legacy is really poor resulting in limited control of corruption. Like van der Ploeg (2011) corruption is a catalyst for the resource curse. Besides rent seeking/grabbing does the paper of Kolstad and Soreide also mention that patronage lead to government payoff supporters to stay in power, which will lead to a misallocation of public funds.

(9)

2.5 21st_century

Most of the research that has been done about the resource curse was from a time period of 1980 or older. Sachs and Warner (1997) used a time period from 1970 to 1990 in which the resource abundance was measured in 1970. A lot of papers that has been

published after the paper of Sachs and Warner were based on approximately the same time period. Mehlum et al. (2006) used the period 1965 to 1990, Atkinson (2003) used a period from 1980 to 1995. Most of the exported resources are crude oil, coal and gas which are the main sources of energy in the world (Shafiee and Topal, 2008). In the recent years the effect of these energy sources on the environment has been in debate. The paper of Sinton and Friedley has stated that China in 1994 started to reduce their coal consumption. One of the main reasons is that China was the second largest gas polluter in the world which has an impact on the environment (Stinton and Fridley, 2000). Countries also have to consider that these resources are running empty. The paper of Shafee and Topal (2008) quote that the crude oil and gas resources are approximately depleted in 2042. The only fossil fuel that is left after 2042 is coal which has a lot of environmental problems (Shafee and Topal, 2008). Countries whose main income come from oil and gas exportation have to come with other options for the future to succeed as a country. The biggest oil contributor Saudi Arabia is already planning for the future when their oil reserve will be depleted. The paper of Moser et al. (2015) describes a mega city that Saudi Arabia is constructing. The city is designed for a population of 2 million people and their goal is to provide 1 million jobs. The city comes with 6 different zones: education, industrial, business, resort, sea port and residential areas and is located near the red sea. There are more examples of oil rich countries that draw their oil economy to these master planned cities, Lusail City in Qatar and Masdar in Abu Dhabi (Moser et al, 2015). Saudi Arabia is trying to diversify their economy since 1975. The paper of Albassam (2014) showed how Saudi Arabia is making the economy more diversified by introducing big projects each five year. In his research on the oil sector he discovered that Saudi- Arabia is making progress into making their economy diversified but it isn’t there yet. The main economic driven factor in Saudi-Arabia still remains oil.

A possible other reason for oil rich economies to switch to a master planned city or other ways to diversify and to turn the focus away from the oil-reserves is to be independent from oil shocks. Countries like Saudi Arabia , UAE and Kuwait have a large share of their GDP contributed to oil revenue. The paper of Basher et al (2018) investigated if the stock market returns are affected by oil shocks. Eight oil abundance countries were examined: Canada, Kuwait, Mexico, Norway, Russia, Saudi Arabia, UEA and the UK. With the use of a two-state Markov-switching model the impact of the oil shock was examined on the stock market. The results showed that from the eight countries only Mexico wasn’t affected by the oil shock while the other countries did (Basher et al, 2018). These two reasons cause

(10)

resource dependent countries to prepare for the future of their country. This results in that the gains from their resources are invested in programs which will supply a steady economy in the future. The case of Venezuela showed how intense a shock in the oil price can lead to a severe economic crisis. Venezuela under the control of Chávez wasted its revenue from their oil on domestic consumption while it builds up its debt and abandoned the increase in their productive investment (Monaldi, 2015).

When the oil rich countries reduce their dependence on oil, the Dutch disease model described by van Wijnbergen (1984) becomes less of a problem. A boom in the oil-sector would be less of an effect on the consumption of non-traded goods. This means that the non-oil traded sector will shrink less hence the learning by doing effect would be less damaged. When we apply the shrinkage of the oil economy in the 21st_{century to the Dutch}

disease model of Cordon and Neary (1982) we see that the booming traded sector declines and thereby the booming traded sector won’t decline. The balance of trade for the non-booming traded goods sector will then be untouched.

The papers of Mehlum et al (2006) and van der Ploeg (2011) will also conclude that the resource curse would be diminishing with the developments of these countries. If these oil driven economies are using various plans to diversify their economy the institutions would become producer friendly instead of rent-grabbing friendly. This will result in a flow of

(11)

3. Methodology and hypothesis 3.1 The regression

The aim of this paper is to examine the resource curse in the 21st_{century. I wish to}

do this by comparing countries growth with their resource abundance. This paper not only aims to show if there is a resource curse in the 21st_{century, but also wants to investigate if}

the resource curse declined over time. The data in this investigation are from the year 2000. The regression is a cross section regression with 73 countries. To compare the 20th_century

with the 21st_{century I also take data from 1980. The cross-sectional regression from 1980}

contains 51 countries. The regression used for the year 2000 will be as followed: Average GDP growth per capita2000-2017= βresource abundance2000+ βinitial GDP2000 + βgrowth before2000 + βopenness2000 + βgovernment consumption2000+ βCPI2000 + βCPI2000* βresource abundance2000

H0: βresource abundance2000 = 0 H1: βresource abundance2000 < 0

The regression from the year 1980 is exactly the same with only a different year. 3.2 The variables

The growth in the 2000 regression is the average GDP growth per capita from 2000 till 2017 and in the 20st century from 1980 till 1999. This measurement is used because in the literature the main measurement for growth was GDP growth per capita. Other

measurements for growth that also could be used is the Inclusive Wealth Index. This measurement includes natural-, human,- social- and manufactured capital (Thiry and Roman, 2014) . This measurement is interesting because it also pays attention to natural capital. This means that the effect of resource abundance on the Inclusive Wealth Index would be less . So, the more resource abundant a country is the more decline in natural capital. This paper however aims to be a continuation on previous papers so the best way to compare it with other papers is to use the same measurement. The resource abundance for the regression is measured in the export of the resources: fuel ( consists of: mineral fuels, lubricants and related materials), raw agricultural material ( consists of: crude materials like farm products, wood, sand and gravel) and Ores and metal (consists of: crude fertilizer, minerals nes, metalliferous ores, scrap and non-ferrous metals). These exports of resources are divided by their respective GDP to get the resource abundance. This measurement would take almost full coverage of the resource abundance. It will absorb a big part of the

(12)

Algeria Australia Austria Bahamas, The Bangladesh Belgium Botswana Brazil Burkina Faso Cameroon Canada Chile China Colombia Costa Rica Cote d'Ivoire Denmark Ecuador Fiji Finland France Gabon Ghana Greece Honduras Hong Kong SAR, ChinaIceland India

Iran, Islamic Rep. Ireland Italy Jamaica Japan Kenya Korea, Rep. Luxembourg Madagascar Malawi Malaysia Mexico Morocco Netherlands New Zealand Niger Nigeria Norway Pakistan Panama

Papua New Guinea Peru Philippines PortugalSenegal Singapore South Africa Spain Sri Lanka Sudan Suriname Swaziland Sweden Thailand Togo Turkey Uganda United Kingdom United States Uruguay

United Arab Emirates Saudi Arabia Venezuela, RB Qatar Russian Federation 0 2 4 6 8 G D P gr ow ht p er c ap ita fr om 2 00 0 til l 2 01 7 0 20 40 60 Resource abundance in 2000

Source: World Bank

resource exports since fuel and minerals are one of the biggest exported products in US$. Torvik (2009) would argue that measuring the resource abundance by the resource export divided by GDP would lead to a wrong interpretation of resource abundance. This would mean as Torvik explains that a poor country is more resource abundant than a rich country with the same resource export. If the curse exists in the 21st_{century than the β in the}

regression on the resource abundance should be lower than 0. If that’s the case than

resource abundance has a negative influence on average growth. Some variables are added in the regression to make the regression more robust. The initial GDP is taken into account because the countries converge to a certain point in their growth. Countries with an initial lower GDP per capita are expected to grow faster than countries with an already high GDP per capita (Yanikkaya, 2003). Graph 1 shows how China is one of the biggest growers in the 21st_{century. However, China’s initial GDP per capita in 2000 is almost 30 times smaller than}

a developed nation. We see the same with the developing countries Sudan and Ghana which economies are almost 100 times smaller than the developed countries when looking at the GDP per capita (world bank, 2018).

(13)

Algeria Australia Austria Bahamas, The Bangladesh BelgiumBotswana Brazil Burkina Faso Cameroon Canada Chile China Colombia Costa Rica Cote d'Ivoire Denmark Ecuador Fiji Finland France Gabon Ghana Greece Honduras

Hong Kong SAR, China Iceland

India Iran, Islamic Rep.

Ireland Italy Jamaica Japan Kenya Korea, Rep. Luxembourg MadagascarMalawi Malaysia Mexico Morocco Netherlands New Zealand Niger Nigeria Norway Pakistan Panama Papua New Guinea

Peru Philippines

Portugal

Senegal Singapore

South Africa Spain

Sri Lanka Sudan

Suriname

SwazilandSweden Thailand Togo UgandaTurkey

United Kingdom United States Uruguay

United Arab Emirates Saudi Arabia Venezuela, RB Qatar Russian Federation 0 2 4 6 8 G D P gr ow ht p er c ap ita fr om 2 00 0 til l 2 01 7 -5 0 5

GDP growht per capita from 1980 till 1999

Source: World Bank

Yearly average growth of GDP per capita from 1980 to 1999 is also considered in the regression. A country that is already growing is likely to grow more. However, Sacks and Warner (1997) found that the GDP growth per capita a decade before their data period did not have much influence on the resource curse effect. When we take a look to the scatter plot in graph 2, there is no clear relation. We see that China was also a grower before 2000. Sudan, Saudi-Arabia, Niger and Ghana did not grow before 2000 but start growing in the upcoming period. It is also important to mention that the Russian Federation was one of the biggest shrinkers. In a closer look at the data of the Russian Federation, it states that from 1990 till 1998 the Russian federation went through a severe decline in growth with the biggest peak in 1992 of -14,53 percent growth due to the collapse of the Soviet Union in 1990-1991. The growth after 2000 could be the recovery of the Russian Federation.

Graph 2 ( GDP growth per capita from 1980 until 2000 against growth per capita from 2000 until 2017)

Another variable that is included in the regression is government consumption in the year 2000. The paper of Atkinson (2003) showed evidence that resource abundant countries suffer from a low growth when the government consumption is high. The variable trade openness is also included in the regression. In the papers of Sachs and Warner (1997) and Mehlum et al (2006) the relation between openness and growth was significant positive. The openness in the regression of this paper is measured by import plus export divided by their countries respective GDP in the year 2000. The paper of Yanikkaya (2003) discovered that

(14)

Algeria Australia Austria Bahamas, The Bangladesh Belgium Botswana Brazil Burkina Faso Cameroon Canada Chile China Colombia Costa Rica Cote d'Ivoire Denmark Ecuador Fiji Finland France Gabon Ghana Greece Honduras

Hong Kong SAR, China Iceland

India

Iran, Islamic Rep.

Ireland Italy Jamaica Japan Kenya Korea, Rep. Luxembourg Madagascar Malawi Malaysia Mexico Morocco Netherlands New Zealand Niger Nigeria Norway Pakistan Panama Papua New Guinea

Peru _Philippines

Portugal

Senegal Singapore

South AfricaSpain Sri Lanka Sudan Suriname Swaziland Sweden Thailand Togo Turkey Uganda United Kingdom United States Uruguay

United Arab Emirates Saudi Arabia Venezuela, RB Qatar Russian Federation 0 2 4 6 8 G D P gr ow ht p er c ap ita fr om 2 00 0 til l 2 01 7 0 100 200 300 400 Openness in 2000

Source: World Bank

previous research used this type of measurement to measure the openness of a country. These papers showed that openness had a positive influence on growth. Another way to look at openness is to look at trade barriers. In the results of Yanikkaya’s paper he found that trade barriers are positively associated with growth contradicting previous research in his literature review in which trade barriers have a negative influence on growth. Looking at the data for the regression we don’t see the positive relation between openness and growth described by the works of Sachs and Warner and Mehlum et al.

Graph 3 (Openness in 2000 against GDP growth per capita from 2000 until 2017)

3.3 Corruption variable

Another important variable measured in this paper for the regression is the Corruption Perception Index (CPI) of the year 2000. It states in the literature review that corruption had a negative influence on growth (van der Ploeg, 2011). The CPI is a surveys-based measurement in which business people, risk analyst and the general public grade their countries corruption. The grading is on a scale of 1 tot 10 in which 10 means clean and 1 means highly corrupt. This measurement is more of a perception of corruption than the actual corruption (transparency international). Unlike the paper of van der Ploeg, which stated that corruption has a negative influence on growth, we see in the data that countries

(15)

Algeria Australia Austria Bahamas, The Bangladesh Belgium Botswana Brazil Burkina Faso Cameroon Canada Chile China Colombia Costa Rica Cote d'Ivoire Denmark Ecuador Fiji Finland France Gabon Ghana Greece Honduras

Hong Kong SAR, ChinaIceland India

Iran, Islamic Rep.

Ireland Italy Jamaica Japan Kenya Korea, Rep. Luxembourg Madagascar Malawi Malaysia Mexico Morocco Netherlands New Zealand Niger Nigeria Norway Pakistan Panama

Papua New Guinea Peru

Philippines

Portugal

Senegal Singapore

South Africa Spain

Sri Lanka Sudan Suriname Swaziland Sweden Thailand Togo Turkey Uganda United Kingdom United States Uruguay

United Arab Emirates Saudi Arabia Venezuela, RB Qatar Russian Federation 0 2 4 6 8 G D P gr ow ht p er c ap ita fr om 2 00 0 til l 2 01 7 2 4 6 8 10 CPI in 2000

CPI goes from 1 (highly corrup) to 10 (no corruption)

countries in terms of GDP growth per capita. Additionally, I want to examine the interaction between corruption and the resource curse with an interactive variable between the resource abundance and the CPI.

Graph 4 (Coruption Perception Index against GDP growth per capita from 2000 till 2017)

3.4 Time variable

Eventually this paper investigates if the resource curse declined in the 21st_century. To look at the difference between the 20th_{and the 21}st_{century, an extra regression is made} with the data from both 1980 and 2000. In this regression a variable “Time*Resource abundance” is introduced. This is an interactive variable that is generated by the resource abundance multiplied by time. This will help the regression to prove if time is relevant for growth. If the resource curse is reducing over time, the “Time*Resource abundance” variable is bigger than 0. The last regression will look like:

Average GDP growth per capita = βresource abundance1980 or 200+ βinitial GDP1980 or 2000+ βgrowth before1980 or 2000 + βopenness1980 or 2000+ βgovernment consumption1980 or 2000 +

(16)

With time = 1 if resource abundance in 1980 and time = 2 if resource abundance in 2000 H0: βTime*resource abundance = 0

H1: βTime*resource abundance > 0

With an extra regression, I plan to examine if there is a difference in resource abundance when including only the fuel or non-fuel resource exports. According to the literature, not all resources behave the same way. Furthermore, there is the fact that the oil price in the year 2000 was lower than that of 1980. This could influence the resource abundance of a country. When looking specifically at the non-fuel exports, we could exclude the price difference. The resource curse of 2007 will also be examined as an extra check. In 2007 the price of oil was around the same as that of 1980 (macrotrends). However, 2007 has fewer periods to

(17)

4. Results

4.1 Resource curse

Table 1 shows the results from the regression on the data from the years 1980 and 2000 combined. This table shows a negative non-significant correlation between resource abundance and growth in the second and third regression. A negative β of resource abundance can be interpreted as an existence of the resource curse. However, it is not certain in this regression because the β is insignificant. The variables with the most influence in the regressions are the “initial GDP per capita” and the “Growth before”. Prior studies show a negative correlation between “initial GDP per capita” and growth and a positive correlation between “growth before” and growth. The results of this study confirm these findings. “Openness” and “Government consumption” seem to have little influence in the regression. Two of the main variables this thesis focuses on are “time” and “corruption”. 4.2 Results on the time variable

When examining the resource curse of 2000 (table 2), the effect of resource

abundance on growth is positive. The result from the 20th_{century curse (table 3) are similar}

to the results from Sachs and Warner (1995, 1997, 2001). They display negative results for the resource abundance. The difference between 1980 and 2000 gives the impression that the effect of the resource curse is declining over time. The time variable on its own is significantly positive in this regression. This would suggest that the average growth in the 21st_{century from 2000 to 2017 was bigger than the average growth from 1980 to 2000. The}

interactive variable between time and resource abundance gives a positive insignificant result. This is further evidence that the resource curse is diminishing over time. The prices of the specific export products are excluded in the regression. Using crude oil as an example, the price was lower in the year 2000 than in 1980 (macrotrends). As a result of this the export of crude oil in 2000 was lower than 1980 and therefore the resource abundance was lower. In tables 4 and 5 the measurement for resources abundance is split between fuel abundance and non-fuel abundance. Table 4 shows that when looking exclusively at fuel as resource export the resource curse is present between the two years, but when looking at the interactive variable between time and resource abundance it is declining over time. The regressions in table 5 show that the non-fuel resource abundance acts like the opposite of fuel resource abundance. Non-fuel resource abundance has a positive relation with growth. The effect of the resource curse with the non-fuel resource measurement is declining over time. This effect is mentioned in the Dutch disease model by van Wijnbergen (1982). When the oil sector discovered high oil revenue due to the high prices in 1980, the other export sectors declined. Looking at the regressions of 2007 (table 6), where the price of oil was

(18)

around the same as in 1980, the β of resource abundance is still negative but insignificant and much smaller than in 1980. However, it cannot be said that the resource curse in the 21st _{century has declined. When combining the 2007 data and 1980 data in one regression}

the interactive time and resource abundance appears to be positive but insignificant. 4.3 Results on the corruption variable

The last variable this thesis is interested in is that of corruption. The interactive variable between CPI and the resource abundance shows a negative β. This means that unlike the literature, the data shows that corruption has a positive effect on growth in

resource abundant countries. The regression in 2007 shows similar results. However, these βs are insignificant, and consequently conclusions cannot be drawn from these results.

4.4 Tabels

Table 1 (Resource curse 1980 and 2000)

(1) (2) (3)

Growth GDP per capita Growth GDP per capita Growth GDP per capita Resource abundance 0.0194 * _-0.0413 _-0.0430 (2.22) (-1.16) (-0.77) Initial GDP per capita /10000 -0. 262 * _{-0. 514}*** _{-0. 684}*** (-2.55) (-4.47) (-4.57) Consumption 0.00280 -0.00685 (0.13) (-0.32) Growth before 0.164*** _0.157*** (3.69) (3.53) Openness 0.00200 0.00105 (0.98) (0.49) Time 1.107** _1.202*** (3.25) (3.45) Resource abundance * Time 0.0335 0.0377 (1.71) (1.59) Corruption Index 0.128 (1.54) Corruption Index * Resource abundance -0.000180 (-0.04) _cons 1.510*** _-0.639 _-1.137 (9.61) (-0.95) (-1.43)

(19)

Table 2 (Resource curse 2000)

(1) (2) (3)

Growth GDP per capita

2000 Growth GDP per capita 2000 Growth GDP per capita 2000 Resource abundance 2000 0.0235 * _0.0240* _0.0235 (2.29) (2.21) (1.91) Initial GDP/10000 -0. 435*** _{-0. 456}** _{-0. 443}* (-3.91) (-3.14) (-2.14) Consumption 2000 -0.00883 -0.00828 (-0.31) (-0.28) Openness 2000 -0.000388 -0.000326 (-0.15) (-0.12) Corruption index -0.00935 racorrupt -0.00147 (-0.20) _cons 1.880*** _2.000*** _2.028** (8.97) (3.78) (3.24) N 73 73 73 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

Table 3 (Resource curse 1980)

(1) (2) (3)

1980 Growth GDP per capita 1980 Growth GDP per capita 1980 Resource abundance 1980 0.00389 -0.0371 * _-0.0317 (0.27) (-2.15) (-1.85) Initial GDP 1980/10000 0. 579 0. 171 -0. 629 (1.91) (0.45) (-1.10) Consumption 1980 -0.0251 -0.0138 (-0.78) (-0.43) Openness 1980 0.00959 ** _0.00711* (3.05) (2.13) Corruption index 0.185 (1.84) Curruption index* Resource abundance 1980 -0.00631 (-1.05) _cons 0.895*** _0.494 _-0.294 (3.77) (0.99) (-0.45) N 51 51 51 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(20)

Table 4 ( Resource curse 1980 and 2000 with Fuel abundance)

(1) (2) (3)

Growth GDP per capita Growth GDP per capita Growth GDP per capita Fuel resource abundance 0.0299 ** _-0.0944* _-0.122 (3.11) (-2.01) (-1.89) Initial GDP per capital/10000 -0. 272 ** _{-0. 553}*** _{-0. 739}*** (-2.69) (-4.95) (-5.23) Time 1.124*** _1.244*** (3.69) (4.07) Fuel resource abundance * Time 0.0668 ** _0.0787** (2.66) (2.76) Corruption index 0.127 (1.82) Fuel resource abundance * Corruption index 0.00320 (0.63) _cons 1.523*** _-0.868 _-1.352 (10.52) (-1.39) (-1.95) N 124 124 124 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

Table 5 ( Resource curse 1980 and 2000 with non-fuel abundance)

(1) (2) (3)

Growth GDP per capita Growth GDP per capita Growth GDP per capita Non fuel resource

abundance 0.00162 0.0475 0.0258 (0.06) (0.59) (0.21) Initial GDP per capita/10000 -0. 257 * _{-0. 486}*** _{-0. 558}*** (-2.42) (-4.09) (-3.59) Time 1.483*** _1.527*** (3.97) (3.97) Non-fuel resource abundance * Time -0.0239 -0.0215 (-0.50) (-0.43) Corruption index 0.0469 (0.55) Non-fuel resource abundance * Corruption index 0.00291 (0.22) _cons 1.668*** _-0.993 _-1.134 (9.85) (-1.34) (-1.33)

(21)

(1) (2) (3)

Growth GDP per capita Growth GDP per capita Growth GDP per capita Resource abundance 2007 -0.00309 -0.00533 -0.00298 (-0.86) (-1.32) (-0.34) Initial GDP per capita/1000 -0.0904 *** _-0.0676** _-0.0783* (-4.47) (-2.69) (-2.30) Consumption -0.0295* _-0.0309* (-2.26) (-2.27) Growth before -0.0241 -0.0199 (-0.36) (-0.29) Openness 2007 0.000301 0.000281 (0.37) (0.32) Corruption 0.0201 (0.48) Resource abundance 2007 * corruption -0.000455 (-0.22) _cons 0.488*** _0.909*** _0.842** (6.73) (4.14) (3.14) N 72 72 72 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(22)

5. Conclusion and Discussion

This thesis was mainly based on the resource curse of the 21st_{century. Prior studies}

have shown the existence of the resource curse. Factors such as bad quality institutions and corruption have previously shown to amplify the curse. A country’s spending is also a factor in the resource curse. When looking at the resource economy today we see that significantly oil rich countries are diversifying their economy to safeguard their position when their main resources are depleted in the future. Following the Dutch disease models the process of diversifying the economy and stepping away from oil should lower the growth decline and thereby the resource curse. In a cross-sectional data analysis of the 21st_{century using the}

year 2000 as a base year, there was no presence of a resource curse. The resource abundance even showed a positive correlation with growth. The combined regression showed how the resource curse was declining over time. This result could be due to the diversification discussed in the literature review. Another reason could be that countries reinvest their gains from resources, making the countries more resistant against the curse (van der Ploeg, 2011). Further research should be done to prove this point with attention paid to the prices of resources in a specific year. In the year 2000 the oil price was low compared with the oil price in 1980. The result from the regression from the resource curse in the year 2007 disprove the resource curse decline in the 21st_{century. In further research}

this could be examined by measuring it over more years as well as including omitted

variables to make the regression more robust. The resource curse is especially present if the price of the resources are high. The regression of the fuel resource abundance showed that the curse was high when the price of oil was high. The effect of high prices on the resource curse is related to the boom in the Dutch disease model from Cordon and Neary (1982). In the regressions the CPI in combination with resource abundance has almost no effect on the growth of a country, unlike the research done by van der Ploeg (2011). This can be due to the fact that less developed countries are high in both corruption and growth.

(23)

6. References

Acemoglu, D., Johnson, S., Robinson, J.A., 2001. An African success story: Botswana. Working Paper series, pp. 01-37.

Albassam, B.A., 2015. Economic diversification in Saudi Arabia: Myth or reality? Resour. Policy 44, pp.112–117.

Basher, S.A., Haug, A.A., Sadorsky, P., 2018. The impact of oil-market shocks on stock returns in major oil-exporting countries. Journal of International Money and finance. 86, pp. 264-280.

Corden, W.M., Neary, J.P., 1982. Booming sector and de-industrialization in a small open economy. Econ. J. 92, pp. 825–848.

Davis, G.A., 1995. Learning to love the Dutch disease: evidence from the mineral economies, World Dev. 23, pp. 1765–1779.

Géraldine Thiry, Philippe Roman. The Inclusive Wealth Index. A Sustainability Indicator, Really? FMSH-WP-2014-71. 2014.

Kiev, C.W., 2014. What Dutch disease is, and why it's bad. The Economist.

https://www.economist.com/blogs/economist-explains/2014/11/economist-explains-2 Consulted: 08-04-2018

Kolstad, I., Søreide, T., 2009. Corruption in natural resource management: implications for

policy makers. Resources Policy 34 (4), pp. 214-226

Krugman, P., 1987. The narrow moving band, the Dutch disease, and the competitive consequences of Mrs. Thatcher: notes on trade in the presence of dynamic scale. Journal of development economics. Volume 27 pp. 41-52

http://www.macrotrends.net/1369/crude-oil-price-history-chart consulted: 15-06-2018

Matsuyama, K. "Agricultural Productivity, Comparative Advantage, and Economic Growth." Journal of Economic Theory. 1992. 58, pp. 317-334.

(24)

Monaldi, F., 2015. the Impact of the decline in oil prices on the economics, politics and oil industry of Venezuela. SIPA Center on Global Energy Policy.

Mehlum, H., Moene, K., Torvik, R., 2006. Institutions and the resource curse. Econ. J. 116, pp. 1–20.

Moser, S., Swain, M., Alkhabbaz, M.H., 2015. King Abdullah Economic City: Engineering Saudi Arabia’s post-oil future. Elsevier Cities 45, pp. 71–80.

van der Ploeg, F., 2011. Natural Resources: Curse or Blessing? Journal of Economic Literature. 49-2, pp. 366-420

Rajan, R.G., Subramanian, A., 2011. Aid, Dutch disease, and manufacturing growth. J. Dev. Econ. 94, pp. 106–118.

Sachs, J.D., Warner, A.M., 1995. Natural Resource Abundance and Economic Growth. Technical Report 5398, National Bureau of Economic Research.

Sachs, J.D., Warner, A.M., 1997. Natural Resource Abundance and Economic Growth. Technical Report, Center for International Development and Harvard Institute for International Development.

Sachs, J.D., Warner, A.M., 2001. The curse of natural resources. Eur. Econ. Rev. 45, pp. 827–838

Shafiee, S., Topal, E., 2009. When will fossil fuel reserves be diminished? Elsevier Energy Policy. 37, pp. 181-189.

Sinton, J.E., Fridley, D.G., 2000. What goes up: recent trends in China’s energy consumption. Elsevier Energy Policy. 28, pp. 181-189.

Spilimbergo, A., Londoño, J.L., Székely, M., 1999. Income distribution, factor endowments, and trade openness. Journal of Development Economies. Vol. 59, pp. 77-101

(25)

consulted: 25-05-2018

Torvik, R., 2009. Why do some resource-abundant countries succeed while others do not? Oxford Rev. Econ. Policy 25, pp. 241–256.

Van Wijnbergen, S., 1984. The ‘Dutch disease’: a disease after all? Econ. J. 94, pp. 41–55. Yanikkaya, H., 2002. Trade openness and economic growth: a cross-country empirical investigation. Journal of Development Economics 72, pp. 57-89

(26)

7. Appendices 7.1 List of countries

Tabel 7 (countries used for the regressions)

Algeria Fiji Madagascar Spain

Australia Finland Malawi Sri Lanka

Austria France Malaysia Sudan

Bahamas, The Gabon Mexico Suriname

Bangladesh Ghana Morocco Swaziland

Belgium Greece Netherlands Sweden

Bolivia Honduras New Zealand Thailand

Botswana Hong Kong SAR, China Niger Togo

Brazil Iceland Nigeria Trinidad and Tobago

Burkina Faso India Norway Turkey

Cameroon Iran, Islamic Rep. Pakistan Uganda

Canada Ireland Panama United Kingdom

Chile Israel Papua New Guinea United States

China Italy Peru Uruguay

Colombia Jamaica Philippines Qatar

Costa Rica Japan Portugal Saudi Arabia

Cote d'Ivoire Kenya Senegal Venezuela, RB

Denmark Korea, Rep. Singapore Russian Federation

(27)

7.2 Extra regressions Table 8 (Resource curse 1980)

(1) (2) (3)

1980 Growth GDP per capita 1980 Growth GDP per capita 1980 Resource abundance 1980 0.00389 -0.0371 * _-0.0317 (0.27) (-2.15) (-1.85) Initial GDP 1980/10000 0. 579 0. 171 -0. 629 (1.91) (0.45) (-1.10) Consumption 1980 -0.0251 -0.0138 (-0.78) (-0.43) Growth before 1980 0.126 ** _0.141** (2.91) (3.28) Openness 1980 0.00959 ** _0.00711* (3.05) (2.13) Corruption index 0.185 (1.84) Curruption index* Resource abundance 1980 -0.00631 (-1.05) _cons 0.895*** _0.494 _-0.294 (3.77) (0.99) (-0.45) N 51 51 51 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(28)

Tabel 9 (Resource curse 1980 with fuel abundance)

(1) (2) (3)

1980 Growth GDP per capita 1980 Growth GDP per capita 1980 Fuel Resource abundance 1980 -0.00254 -0.0655** _-0.0525 (-0.13) (-3.04) (-1.19) Initial GDP per capita/10000 0. 577 0. 166 -0. 429 (1.90) (0.45) (-0.74) Consumption -0.0260 -0.0182 (-0.85) (-0.56) Growth before 1980 0.140 ** _0.149** (3.39) (3.51) Openness 1980 0.0103*** _0.00838* (3.72) (2.09) Corruption index 0.142 (1.37) Fuel Resource abundance 1980 * corruption -0.000693 (-0.09) _cons 0.933*** _0.346 _-0.240 (4.23) (0.72) (-0.37) N 51 51 51 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(29)

Tabel 10 (resource curse 1980 with non-fuel abundance)

(1) (2) (3)

1980 Growth GDP per capita 1980 Growth GDP per capita 1980 Non-fuel Resource abundance 1980 0.0269 0.00581 0.231 ** (0.83) (0.17) (3.14) Initial GDP per capita/10000 0. 604 0. 238 -0. 782 (1.99) (0.59) (-1.47) Consumption 1980 -0.0280 -0.0316 (-0.83) (-1.05) Growth before 1980 0.133 ** _0.127** (2.87) (3.10) Openness 1980 0.00447 0.00836** (1.60) (2.81) Corruption index 0.426*** (3.88) Non-fuel Resource abundance 1980 * Corruption index -0.0459** (-3.49) _cons 0.809** _0.502 _-1.387* (3.22) (0.95) (-2.10) N 51 51 51 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(30)

Tabel 11 (resource curse 2000 with fuel abundance)

(1) (2) (3)

2000 Growth GDP per capita 2000 Growth GDP per capita 2000 Fuel resource abundance 2000 0.0333 ** _0.0375** _0.0427 (3.15) (3.21) (1.52) Initial GDP per capita/10000 -0. 436 *** _{-0. 524}*** _{-0. 599}** (-4.04) (-3.66) (-2.83) Consumption 2000 0.00566 0.00378 (0.20) (0.13) Growth before 2000 0.139 0.154 (1.06) (1.12) Openness 2000 0.0000168 -0.000255 (0.01) (-0.10) Corruption index 0.0549 (0.49) Fuel resource abundance 2000 * Corruption index -0.000600 (-0.08) _cons 1.880*** _1.718** _1.529* (9.73) (3.28) (2.29) N 73 73 73 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(31)

Tabel 12 (resource curse in 1980 and 2000 robust)

(1) (2) (3)

Growth GDP per capita Growth GDP per capita Growth GDP per capita Resource abundance 0.0194 * _-0.0413 _-0.0430 (2.22) (-1.16) (-0.77) Initial GDP per capita /10000 -0.0000262 * _-0.0000514*** _-0.0000684*** (-2.55) (-4.47) (-4.57) Consumption 0.00280 -0.00685 (0.13) (-0.32) Growth before 0.164*** _0.157*** (3.69) (3.53) Openness 0.00200 0.00105 (0.98) (0.49) Time 1.107** _1.202*** (3.25) (3.45) Resource abundance * Time 0.0335 0.0377 (1.71) (1.59) Corruption Index 0.128 (1.54) Corruption Index * Resource abundance -0.000180 (-0.04) _cons 1.510*** _-0.639 _-1.137 (9.61) (-0.95) (-1.43) N 124 124 124 t statistics in parentheses *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

(32)

Tabel 13 (resource curse in 1980 and 2007) (1) (2) Growt Growt Resource abundance 1980 and 2007 -0.00317 0.00417 (-0.51) (0.11) Initial GDP 1980 and 2007 -0.0000181 *** _-0.0000140* (-4.71) (-2.48) Consumption 1980 and 2017 -0.0418 * (-2.48) Openness 1980 and 2017 0.00172 (1.13) Corruption -0.0397 (-0.69) Resource abundance 1980 and 2007* Corruption -0.000645 (-0.20) Time 0.159 (0.75) Resource abundance 1980 and 2007 * Time -0.00760 (-0.46) _cons 0.849*** _1.339* (7.99) (2.57) N 124 124

Resource curse in the 21st century

Table of Contents