!
Do Changes in Oil Prices and
Green Policies Affect Norwegian
Oil Investments?
!
Master Thesis
MSc International Business and Management
Jon Vikse
!
University of Groningen
Faculty of Economics and Business
Abstract
A drop in oil prices could have very negative effects oil exporting countries. Norway has built up a strong economy from their oil resources in the North Sea, but could be volatile from a drop in oil prices. I investigate what effects a drop in oil prices can have for a very important driver for the Norwegian economy, namely the oil investments. At the same time, I take a look at what effects «green» policies can have for the oil investments.
I use time series data on oil investments in Norway from 1986 to 2013 to analyze this. I find a positive relationship in oil production investments from an increase in oil prices, but a negative effect for oil exploration investments. I find that the oil investments are not effected by carbon taxes, but increase in alternative energy investments, have a negative effect on oil investments.
!
Preface
1. Introduction 5
1.1 Motivation 5
1.2 Research Question 5
2. Theory 6
2.1 The Norwegian Oil Industry 6
2.2 Challenges for Norway 7
2.2.1 US Shale Oil Revolution 9
2.2.2 Dependency on China 10
2.3 The Oil Industry’s Value Chain 11
2.4 Oil Price Setting 12
2.5 Historical Oil Prices 14
2.6 Oil Investment Behavior 16
2.6.1 Characteristics of Oil Industry Investments 17
2.7 Oil Investments in Norway 19
2.8 Alternative Energy 20 3. Existing Literature 21 3.1 Alternative Energy 23 4. Methodology 25 4.1 Data 25 4.2 Descriptive Statistics 26 4.3 Methodology 26 4.4 OLS Properties 27
4.5 OLS Model Specification 28
6. Results 30
6.1 Data tests 30
6.2 OLS regressions 33
7. Conclusion 37
7.1 Limitations and future research 38
References 40
Appendix 43
!
1. Introduction
1.1 Motivation
Norway is one of the world’s wealthiest countries due to being blessed with a vast amount of oil in the Norwegian sector of the North Sea. This has put the country in a special situation, as it has become dependent on continuous oil revenues. A source of a heated debate in the country, is just how serious the oil dependency is. What consequences will a big drop in oil prices have for the Norwegian economy? This not a very unrealistic scenario, as a US shale revolution, or a drop in demand from China could cause such a drop. It is a very interesting issue to research, in a country which is unique in its dependency on the oil industry. Norway is also one of the more «green» countries in the world, and the government have introduced a set of policies to decrease the emission of greenfield gases. How do these policies effect the oil investments? I want to contribute to the discussion about resource dependency, while at the same time look at the environment policies, by doing an empirical research on the effects of changes in oil price, and environment policies for Norwegian oil investments.
!
1.2 Research Question
The relationship between oil prices and the Norwegian economy, and possible outcomes from a drop in oil prices has been researched in great detail in the past (BI 2013, Bjørnland and Thorsrud 2013, and Holden 2013). However, to my knowledge there are no studies which have researched the effect oil prices have on oil investments in the Norwegian oil sector. However this type of study has been conducted on oil investments from oil companies in other regions (Osmundsen 2006, Favero et al. 1992, Mohn and Misund 2009). The findings of these previous papers do not draw a clear picture of how oil investments reacts to changes in oil prices. By using oil price as an explanatory variable for oil investments I aim to determine how well the oil price explain the investment behavior in the Norwegian oil sector.
!
first time these effects are researched. I control for exchange rates, and steel prices in my study. Drawing on these insights, the research question is as follows:
!
Research question: Does change in oil prices and green energy policies have an effect on Norwegian oil
investments?
!
This thesis proceeds as follows. Next chapter is about the Norwegian oil industry, where I explain the importance it has for Norway and how the supply chain is built up. I continue by explaining how the oil price is set, current theories on oil investments behavior, and alternative energy focus in Norway. Chapter three covers the previous literature relevant for my thesis, and I develop my hypotheses based on the content presented. Chapter four and five presents the data and methodology used, and chapter six presents the results. Finally, I conclude, as well as limits and future research, in chapter seven.
!
2. Theory
In this section I present the oil industry, both for the world market, and specifically for the Norwegian market. I elaborate on how important the oil industry is for the Norwegian economy, and some challenges related to Norway’s dependency on the oil industry. The second part of this section is about the oil price, specifically what determines the price, and about the history of Brent oil price. In the third part, I present theory on investments behavior in the oil industry, and historical data for oil investments in Norway. The final part covers the alternative energy strategy in Norway.
!
2.1 The Norwegian Oil Industry
for Norway’s rise to a wealthy nation. Today Norway is ranked as one of the countries with the best level of welfare in the world, and it is currently ranked as number four worldwide in GDP per capita. (HDI index 2014).
!
Today the petroleum sector is very important for the overall economy of Norway. In 2013 it counted for 23 percent of all value added (Bjørnland and Thorsrud, 2013), and Norway is today the world’s fifth largest exporter of oil, and third largest gas exporter (Index Mundi). In 2013 the investments in the sector counted for 30 percent (NOK 175 billion ) of overall real 1
investments in the country (NDP 2014).
Figure 1: Macroeconomic indicators for the Norwegian petroleum sector 2013. Source: NPD 2014
!
As we can see from Figure 1, the petroleum sector also counts for 21,5 percent of GDP, 29 percent of state revenues and 48,9 percent of total exports. This describes well just how crucial the sector is for Norway.
!
2.2 Challenges for Norway
Research has shown that historically, many countries that are heavily dependent on natural resources had lower economic growth than other countries (Holden, 2013). Norway on the other side, have avoided such a resource curse. Countries that suffer from a resource curse are usually countries with bad institutions, and struggle with from bad infrastructure, no property rights, and corruption. Again, this does not fit the description of Norway, who actually were among the richest nations in Europe even before the oil discoveries. In figure 1 you can see
100NOK = 12,1 EUR
1
12 • FACTS 2014
Current petroleum activities
The petroleum industry is Norway’s largest industry measured in value creation, State revenues and export value. Since production started on the Norwegian continental shelf in the early 1970s, the industry has contributed approx. NOK 11 000 billion to the Norwegian GDP, measured in 2013 NOK. The industry has thus been highly important for the Norwegian economy and the financing of the Norwegian welfare state. The State’s tax revenues are currently transferred to the Government Pension Fund – Global, which was valued at more than NOK 5000 billion as of 1st of January 2013. However, only 44 percent of the projected recoverable resources on the Norwegian continental shelf have been produced.
Significance to the Norwegian economy
The Norwegian petroleum management system is based on the principle that exploration, development and operations must gene-rate the greatest possible values for society, and that the revenues shall benefit the State and thus the Norwegian society as a whole. The petroleum resources are highly valuable. This is the primary reason why the State claims a large share of the value creation through taxes, fees and the State’s Direct Financial Interest (SDFI). Tax revenues in 2012 totalled approx. NOK 232 billion. The State also receives substantial income from direct ownership in fields through the SDFI scheme. The State’s total net cash flow from petroleum activities in 2012 totalled NOK 401 billion, measured in 2013 NOK. Total revenue from the sector amounted to about 29 per cent of the State’s total revenues.
The State’s revenues from the petroleum activities are transferred to a special fund, the Government Pension Fund – Global. The expec-ted returns from the fund can be spent over the fiscal budget. At the end of 2013, the Fund was valued to NOK 5038 billion. This cor-responds to about one million kroner per Norwegian citizen.
Figure 1.3 The net government cash flow from petroleum activities
(Source: Norwegian Public Accounts)
Figure 1.4 The net government cash flow from petroleum activities
in 2012 (bill. 2013-NOK) (Source: Norwegian Public Accounts)
232.1 4.1 151.1 14.1 401.4 Taxes
Environmental taxes and area fees SDFI Statoil dividend Total: -100 0 100 200 300 400 500 1971 1976 1981 1986 1991 1996 2001 2006 2011
Bill. NOK 2013 value
Statoil dividend Royalty and area fee SDFI
Environmental taxes Taxes
State net cash flow
Figure 1.2 Macro-economic indicators for the petroleum sector 2013 (Source: Statistics Norway, Ministry of Finance)
The petroleum sector’s
share of GDP The petroleum sector’s share of state revenues The petroleum sector’s share of total investment The petroleum sector’s share of total exports
21.5 %
29.1 % 30.7 %
the GDP per capita history from before the oil was discovered until today for Norway, Sweden and EU.
Figure 2 . GDP per capita in USD. Adjusted for current prices and PPP. Source OECD. 2
!
From the figure we can see how Norway has grown rapidly economically, measure by GDP per capita, compared to a similar country like Sweden. It is clear that Norway has done very well because of their oil income, and that the country has used the revenues well. Mideksa (2013) studies this more in detail and concludes that 20 percent of the increase in GDP per capita for Norway is due to the oil revenues.
!
Despite of these promising facts, Norway also has its challenges. An immediate threat, more so than the fact that the oil is an exhaustible resource, is the possibility of a drop in the price of oil. Recent research done by BI Norwegian Business School (2013) unveil that more than 40 percent of the changes in the economic activity in Norway is directly or indirectly related to the oil sector, oil investments, everyday activity or variations in the oil price. Moreover the research predicts that a 50 percent drop in oil prices will reduce the Oslo Stock Exchange by 30-40 percent and GDP in Norway will drop by two to three percent the first year.
EU include Belgium, France, Italy, Luxembourg, Netherlands, Germany, Denmark, Ireland, United Kingdom,
2
Greece, Spain, Portugal, Austria, Sweden, and Finland
!
An important part of the Norwegian economy and oil sector is the oil investments. The investments in the petroleum sector counts for 29 percent of the total level of investment in the country, and for 6 per cent of the total GDP. A large decrease in these investment will have large macro economical consequences, and a 50 per cent decrease will decrease GDP for that year by three per cent (Mohn 2007). A decrease in investments could also have negative impacts on the many oil supplying companies in the region, which is already struggling from harsh competition from cheaper suppliers from South East Asia. Many oil related companies, especially shipyards, have in fact been forced to suspend workers because of large contracts being given to Asian countries, instead of been kept in Norway (tu.no 2013). A decrease in investments would inevitably mean less contracts handed out by the oil companies, and thus fewer contracts to local suppliers, leaving many workers unemployed. The last decade we have seen a steady and rapid increase in both oil prices and oil investments, indicating a possible correlation, where drop in oil prices will result in lower investments.
!
The question is therefor, can the oil price can drop that much? As history shows, it has done so before, and although the oil price has been increasing steadily for many years, there are threats that can cause big drops in the near future. The next two paragraphs is devoted to two example of such possible scenarios, namely the US shale oil revolution, and a drop in demand from China.
!
2.2.1 US Shale Oil Revolution
The shale oil revolution in the US can have a major impact on oil prices (PWC 2013). In a report from February last year, PwC estimate that shale oil can displace 35-40 percent of waterborne oil imports to the US. As the oil demand from the US is such a large part of the world’s total demand, a drop in demand from the country would cause a considerable drop in oil prices. As China might start extracting shale oil themselves, it is definite that a shale oil revolution could cause lower oil prices. PwC (2013) puts it simply, «Increased shale oil production could lead to oil prices that are significantly lower than projected in current forecasts». Using data from the US Energy Information Administration, PwC predict a likely scenario where the oil price will fall to around $83 per barrel in 2035 (from $110 today).
2.2.2 Dependency on China
Another possible scenario that can lead to significantly lower oil prices in the future is the possibility of a drop in demand from China. Research done by Aastveit et al. (2012) show that about 30-40 percent of the changes in oil prices the last ten years is due to changes in demand from China and other Southeast Asian emerging economies. There are several concerns about China’s extreme growth the last decade, which now might be over or even worse, turn into a recession. As OECD (2014) points out in their latest report from May 2014, that China must implement drastic improvements on their insufficient financial system, in order to avoid a financial crisis. A report done by Nomura Securities (2014) claim that the biggest threat for China is their property market, which is claimed to be overinvested, and that a market correction could lead to a housing crisis, similar to the one we saw in the US in 2008. If the Chinese economy would enter a recession, this would cause a drop in demand for oil from China, which again would lead to lower oil prices, negatively impacting the Norwegian economy.
Figure 3: Chinese consumption and oil prices. Source BP.
!
The figure above illustrate the importance Chinese consumption of oil had for the rise in the oil price. Both consumption from China and the oil price have more than tripled since 2002.
!
!
!
0 2750 5500 8250 11000 0 30 60 90 120 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 20112.3 The Oil Industry’s Value Chain
The oil industry has a distinguishable value chain with three major sectors. Upstream is exploration and search for oil under the surface, as well as oil production. Companies in this sector is referred to as E&P (exploration and production) companies. Midstream is processing and raw refinement of the oil being pumped up by the E&P companies, as well as transportation and storage. The downstream sector is final refinement, and distributing the oil products to the end-user.
Figure 4: Oil industry value chain. Source: Own illustration
!
An E&P company needs to get permission from the Norwegian government before it can start exploration for oil in a specific field in the North Sea. If a company gets approved it will get a consensus for that specific oil field (NPD 2014). If the E&P company decides to start drilling for oil, it will have to rent an oil rig at a daily rate. The day rates of drilling rigs are determined on the world drilling rig market, and fluctuations are correlated with the oil prices (Sørensen and Skjerve, 2010). Current day rates for rigs which are used in the North Sea varies between $150 000, and $500 000 (RigZone 2014), in other words, the decision start drilling for oil is a substantial investment. After drilling is done, platforms are connected to the newly drilled holes, and oil production can start. Oil companies build specific oil platforms for each oil field that they start production on. Building an oil platform is also a major investment, and is today mostly done by Asian shipyards, specifically in South Korea or Singapore.
!
Midstream Downstream
Upstream
Pipelines from the oil fields to land connects the oil to the midstream part of the value chain. Refineries buy crude oil from the platforms, and process it for further distribution. A large part of the crude oil is turned into gasoline, and the rest is used for jet fuel, natural gas, motor oil, among other things. The refineries then sell their products to the downstream sector, which transport and sell it to the end users (Kahn Academy, 2012). Worth mentioning is that this value chain is often vertically integrated (one company owns all parts of the value chain), for example Shell, who pump up the oil, refine the crude oil and sell it to the end user via Shell gas stations.
!
2.4 Oil Price Setting
The oil prices are set on the world oil market and is in the long run determined by basic macroeconomic demand and supply, as well as the market sentiment (Kosakowski, 2011). Concerning oil prices, it is normal to regard the short run as a few months to even a year, and the long run as a year and longer. The reason why the short run is such a long period is because the demand and supply changes slowly. If the oil prices would go up, it would take a lot of time for the consumers to change their behavior and purchase less oil products. For example if gas prices would go up, you would not swap your gas heavy car for an electric vehicle immediately, but you might do so if gas prices would increase over a long period. Demand is therefore quite inelastic in the short run.
!
Since oil production includes high fixed costs, and because the lead time from exploration of an oilfield to actual production is so long, supply is to a lesser extent affected by changes in price in the short run. In other words, the supply is very inelastic for price changes in the short run (Hamilton, 2008). In the long run however, it can be affected, as it is possible to change the number of extraction fields or the capacity at existing fields (Akram & Holter, 1996). The supply can also be affected by regulations on the extraction rates set by politicians.
!
supply of oil to consumers, and their decisions influence the oil price (OPEC.org). Their production quota, together with world oil reserves determines the supply side in the long run (Amadeo, 2012), and the demand is given by worldwide demand for crude oil, and basic microeconomic models give us the price equilibrium.
!
In the long run, and increase in price could for example be caused by increasing demand from China, which we have seen the last 10-15 years. Ceterius paribus, increasing demand give us a higher oil price and vice versa. As a result of increased price, oil companies will work towards more supply by investing in new oil fields and increased exploration.
!
The price of oil can fluctuate significantly in a short time, and cause a price shock (a large increase or decrease in the oil price, within a short period of time). The reason for this is that oil is purchased on forward contracts, that is, people purchase a contract which secure a price for oil purchased at a specific date in the future. The prices of the contracts are set through speculations, specifically, how sellers and buyers believe that the price of oil will evolve in the future. If people believe the price will increase in the future, price will increase and vice versa. The next section provides example of such price shocks, as the history of oil prices is presented.
2.5 Historical Oil Prices
Figure 5: Brent oil price 1970-2012 in USD per barrel. Source: British Petroleum
!
The figure above show the Brent oil price since the start of the Norwegian oil era till today. Brent oil is a benchmark price for a mix of 15 different varieties of crude oil used in the North Sea, and it is the price used in the Norwegian oil sector, as well as two thirds of the world trade. Therefor, I use Brent oil price in my study. Next, the timeline of Brent oil price from 1970 till today is explained in detail.
!
1970 - 1986. As seen in the graph, the oil prices increased rapidly from the start of the
1970’s, until it came to a halt, and started falling in the 1980’s caused by a serious surplus of crude oil and fall in demand after the 1970’s Energy Crisis. This caused the price to start dropping, which culminated in a collapse in oil prices in 1986, when the prices fell by nearly 50 percent. The crisis affected oil importing countries positively, but the countries who were large exporters of oil had a bad time following the crisis. Countries like Mexico, Nigeria, and Venezuela came close to bankruptcy (wtrg.com, 2011).
D ol la rs pe r ba rr el 0 30 60 90 120 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012
!
1986 - 1999. After the collapse in 1986, the price stabilized. Iraq’s invasion of Kuwait and
low production caused the prices to rise again at the start of the 1990’s. The Gulf War then caused the prices to fall again, before the strong US growth and emerging Asian economies again made the prices rise. The Asian crisis in 1997 resulted in lower demand from Asia, which combined with increased production from OPEC caused the prices to plummet.
!
1999 - 2008. In 1999 the prices began to recover again, before we saw a new drop after 9/11.
It did not last long before the prices started increasing again. In the graph you can see a dotted line where the big increase started in 2002. Increased demand from emerging countries in Asia, especially China, has the main factor for the large increase (BI 2013). In midst of the Financial Crisis in 2008, there was a huge drop in demand and price, but as the economy has begun to recover again, demand and prices have recovered within a couple of years.
Figure 6: Monthly brent oil price per barrel in USD. Source: BP
!
2008 - 2014. If we take a closer look at the price shock in 2008, we can see the forces at work
in the short run. Figure 6 show the monthly price history from 2008 until today. We see a
US D P er B ar re l 0 35 70 105 140
feb-2008 oct-2008 june-2009 feb-2010 oct-2010 june-2011 feb-2012 oct-2012 june-2013 feb-2014
clear drop after the fall of Lehman Brothers, the big turning point for the worse during the financial crisis, in september 2008. The price fell from $113 per barrel in August 2008 to $40 per barrel in December the same year. This is a very clear example of the forces of futures contracts that determine the price in the short run. Naturally, people’s view for the future became negative after the crisis hit, it was assumed that consumption of oil would drop and that the price would fall. Therefor contracts were sold for less than half the price in December 2008, compared to August the same year. If we turn the clock one year after the fall of Lehman Brothers, we can see that the price had risen to $70 per barrel in the long run. $70 is still a long way from where the price was before the crisis, but this is because of the actual drop in demand after the crisis, and we get a new equilibrium in the long run.
!
2.6 Oil Investment Behavior
In order to research what effect oil prices have on oil investments we need to know more about what factors that influence an oil company’s decision to invest or not. In this section I will briefly present the economic theory of investment behavior, then explain some specific characteristics for investments in the oil industry, and finally present the history of oil investments in Norway.
!
therefor also increase the willingness to invest with increased risk. In other words, it is uncertain what the outcome of increased risk is.
!
2.6.1 Characteristics of Oil Industry Investments
For oil companies, the investments are fully irreversible, and the initial costs are sunk. The uncertain of future rewards depend on the level of risk attached to revenue, political risk, reservoir, development, and production (Bøhren and Ekern 1985). The uncertainty concerning revenue is the future oil prices, which historically have been volatile. For the oil industry, revenue is by far the most important source of uncertainty. These types of investments are large in size, often in the billion dollar numbers, which makes the future revenues of these investments the most important.
!
Exclusive for investments in the oil industry is the concept of reservoirs and the exploration activities, as well as imperfect competition. Oil companies mainly have two different substantial investment decisions to choose between if they want to expand production. These two are exploration of new fields, or expand production in an existing field. Exploration of new fields are definitely the most risky investment, but often have much higher potential for future revenue. If an oil company decides to explore a new oil field, there is a big uncertainty regarding how much oil and gas the oil reservoirs will contain. Having access to large reservoirs is extremely important for oil companies, and a big part of their long term strategies.
!
Oil investments are typically divided into three separate phases, exploration, development and extraction (Favero and Peseran, 1991). Recall from figure 4, we can say that the exploration investments are upstream activities, development investments are mid- and downstream activities, and extraction investments are mostly upstream. Investments in a new oil field in the North Sea have very long time horizons. The average time from discovery to cash payback is on average about 15 years.
Figure 7 illustrates the investment decisions an oil company will go through, from exploration to extraction.
!
!
Figure 7: Oil investment decisions Source: Based on Smit and Trigeorgis (2004)
!
Exploration investments have three distinct activities, scouting, concession, and exploration. After the scouting of an oil field is done, the oil company decide whether or not to apply for a concession to start exploration drilling in that specific field. After a concession has been granted from the Norwegian government, the drilling can start (NDP 2014).
!
The first drilling will unveil if there is oil or if it is a dry hole. If the latter is the case, the oil company will abandon further investment in the oil field. The development phase starts with appraisal drilling, which will unveil if the reserves are large enough to start production, or to abandon. If a decision to start production is made, large investments in further development is made. This is investments in permanent platform(s), infrastructure and pipelines to the refineries. When to start production will be a crucial decision here, but if the field is abandoned, there is small chances for it to open later. Once the production has started, there will be continuous investments decision on how large the production will be, at what times. The costs of extraction, and operating expenditure is included in the extraction investments, also called production investments. In my analysis I will use two dependent variables,
!
exploration investments, and production investments. Exploration investments covers the exploration activities explained in this section, and production investments covers the development and extraction activities. Recall from the value chain activities, production investments covers activities from the whole value chain, and exploration covers the upstream activities.
!
2.7 Oil Investments in Norway
Figure 8 (left) and 9 (right): Norwegian exploration investments and total petroleum investments. Source: Statistics Norway
!
Since the production on Norwegian continental shelf started in 1971, 91 oil fields have been developed (NDP 2014).
Figure 10: Spudded exploration wells on the Norwegian Continental Shelf 1970-2013. Source: The Norwegian Petroleum Directorate
!
0 10000 20000 30000 40000 1985 1990 1995 2000 2005 2010 Exploration Investments 0 55000 110000 165000 220000 1985 1990 1995 2000 2005 2010 Total petroleum investments34 • FACTS 2014
To ensure that the resources on the Norwegian shelf are efficiently exploited, the process from when a discovery is made until it is developed and production starts, is both long and meticulous. At the same time, the industry must continuously adapt to new infor-mation, new technology, changing requirements and market condi-tions. This chapter briefly describes the cycle from when a field is first discovered until the field is decommissioned. The most impor-tant market conditions are also described, and the chapter is con-cluded with climate and environmental considerations.
Exploration activity
In order to extract the petroleum resources located on the Norwegian continental shelf, the resources must be proven through exploration. Exploration policy is therefore an important part of Norway’s long-term resource management and the Government wants to give the companies access to attractive exploration acreage.
The Norwegian Parliament (Storting) has opened most of the North Sea, the Norwegian Sea and the southern Barents Sea for petroleum activities. The Norwegian Petroleum Directorate’s estimate of undis-covered resources in areas on the shelf, is approx. 3 billion Sm3 of
recoverable oil equivalents. Undiscovered resources are distributed as follows between the different ocean areas: 28 per cent in the North Sea, 29 per cent in the Norwegian Sea and 43 per cent in the Barents Sea (see Figure 4.2).
A period of low exploration activity was followed by a surge in 2006, see Figure 4.1. A new record was set in 2009 with 65 spudded explo-ration wells. Fifty-nine exploexplo-ration wells were spudded in 2013, resulting in 20 discoveries. Recent years have also seen major disco-veries such as Johan Sverdrup in the North Sea and Johan Castberg in the Barents Sea.
Exploration policy in mature and frontier areas
Licensing system
The Norwegian licensing system consists of two types of equal licensing rounds. The first type is the numbered licensing round which comprise less mature parts of the shelf. These rounds have been used since 1965, and have been held every second year in recent years. Numbered licensing rounds start with inviting the oil companies to nominate blocks they want to be announced. Based on this and the authorities’ own assessments, a proposed announce-ment is submitted for public consultation. The Governannounce-ment finally announces the round.
The second type is the Awards in Pre-defined Areas (APA), which was introduced for mature parts of the shelf in 2003. This system entails the establishment of pre-defined exploration areas compri-sing all of the mature acreage on the shelf. Companies can apply for acreage within this defined area. As new areas mature, the areas will be expanded, but not reduced. A regular, fixed cycle is planned for licensing rounds in mature areas. So far, eleven annual rounds have been held (APA 2003–2013).
Applicants in the licensing rounds can apply as joint ventures or individually. Impartial, objective, non-discriminatory and pre-announced criteria form the basis for award of production licences. Based on the applications received, the Ministry of Petroleum and Energy awards production licences to a group of companies. The Ministry designates an operator for the joint venture, to be respon-sible for the operational activities authorised under the licence. The production licence applies for an initial period (exploration period) that can last up to ten years.
Mature areas
Petroleum activities on the Norwegian continental shelf started in the North Sea and have gradually moved north, based on the prin-ciple of stepwise exploration. This means that large parts of the North Sea are now considered mature from an exploration perspec-tive. The same applies to the Halten Bank and the area around the Ormen Lange field in the Norwegian Sea, as well as the area sur-rounding Snøhvit and Goliat in the Barents Sea.
Mature areas are characterised by known geology and well-develo-ped or planned infrastructure. Discoveries in these areas are likely, but new, large discoveries are less likely. It is important to prove and recover the area’s resources before the area infrastructure is shut down. If this cannot be done, profitable resources could potentially be left behind if the discoveries are too small to justify independent infrastructure development. Additional resources from the area sur-rounding a producing or planned field may also increase the field’s
0 10 20 30 40 50 60 70 1970 1975 1980 1985 1990 1995 2000 2005 2010
Number of exploration wells spudded
Appraisal wells Wildcat wells
Figure 4.1 Spudded exploration wells on the Norwegian continental
Figure 10 presents the number of active exploration wells on the Norwegian continental shelf. There was a lot of activity in the 1980’s, as well as in the recent six years, when exploration activities has peaked. Figure 9 show the total investments in the Norwegian petroleum sector from 1985 until today. We see that there is a almost constant increase in investments, with with some fluctuations. As with oil prices, investments have increased rapidly since around 2002 until today. Figure 8 show only the exploration investments, which was fairly stable from the first observations in 1985, before it tripled in the span of two years in 2005. Note that the oil exploration investments naturally correlates well with the number of exploration wells.
!
By now, everyone in the world seems to agree that the global warming is caused by humans, and that the main driver is CO2 emissions, and fossil fuels. Actions have been made to reduce emissions, which has strong impacts the oil industry. In the following section I will explain some of these actions.
!
2.8 Alternative Energy
The global warming effect from carbon dioxide (CO2) emissions is an issue that the whole world is aware of today. The petroleum operations in Norway account for 31 percent of the country’s total CO2 emissions (Norway’s Ministry of Petroleum and Energy 2007). The offshore platforms create emission through combustion of gas in turbines, flaring of gas, and diesel consumption. Taxes on CO2 emission in the Norwegian offshore petroleum sector was introduced in 1991 in order to increase government revenues from the industry, and reduce CO2 emissions. However, the improved technology to reduce emissions, did not reduce overall emissions as oil demand increased (see appendix 10). Oil companies is met with pressure to reduce production, not only from a global warming perspective, but also because oil is a non-renewable resource. Although it is currently predicted that there will still be large oil reserves on the Norwegian continental shelf for the coming decades, the oil will eventually run out, and we will need other energy sources.
!
there are also environmental issues concerning this source. Because of this, wind energy and solar energy seems like the best alternatives, with almost no environmental damage. In Norway almost all electricity is produced from hydroelectricity, but wind energy production has been increasing rapidly the last decade (see appendix 5).
!
In order to reduce emissions, politicians must solve the problem by aiming at reducing demand for fossil fuel products. Popular ways to deal with this is either through tax increase on fossil fuel consumption or by developing alternative energy sources. In my analysis I will focus on both emission taxes, and alternative energy sources.
!
3. Existing Literature
Harold Hotelling’s paper on extraction path for exhaustible resources from 1931 is to this day a much cited paper on oil investment theory. He introduced a standard model of resource extraction, which states that the price of natural resources must grow at the rate of the scarcity rate, and that the value of investments will consist of the shadow value of actual and new reserves.
!
Guerra (2007) builds on Hotelling (1931), and studies the long run relationship between oil prices and aggregate oil investments. He uses the aggregate number of oil rigs as a proxy of oil investment, and analyze the bidirectional relationship between oil prices and oil investment. Guerra (2007) considers the relation between oil prices and oil investments as endogenous, and analyze the interrelationship. By analyzing both oil investment response the shocks in the oil price, and vice versa, he finds significant results. Aggregate oil investment respond positively to shocks in oil price (big increase in a short time). For the opposite relationship, results suggest that the oil price response to shocks in oil investment is barely significant. A possible explanation for this could be that more oil investments do not necessarily lead to greater production, and that most investments are being directed to less productive parts of the oil fields, which limits the increase in supply.
!
the effect of oil price volatility. They conclude that there is a highly significant relationship between oil price volatility and investments.
!
Osmundsen et al. (2005) investigate the investment levels from oil companies in the world, and the oil price. They find contrasting evidence to the general economic theory and Guerra (2007), and states that oil exploration and oil field development should respond positively to increasing oil prices. When demand for oil and oil price goes up, the oil companies make more money, and a healthy cash flow should increase spendings in order to feed the new demand. Contrary to this theory, Osmundsen et al. find that exploration spending failed to respond to oil prices since the oil prices started to increase around year 2000 (Appendix 1). These results could be explained by the increasing focus towards short-term profitability in the oil industry. Since there is a long lead time for exploration projects, investments in exploration will not be profitable in the short run, but first after a few years. Osmundsen et al. imply that this has caused a shift in management attention to cost-cutting and value-maximization of existing reserves. In other words, their findings suggest that oil companies focus more on increase production on oil fields where they are currently operating in, instead of exploring for resources in new fields.
!
On the other hand, Osmundsen et al. also suggest that the equilibrium with low-capacity/ high-price is not sustainable, so we should expect to see more investments in exploration from the date this research was performed (2005), till today. Based on the previous literature presented, the following hypotheses are formulated.
!
H1: An increase in oil prices will result in an increase in oil production investments. H2: An increase in oil prices will result in an increase in oil exploration investments.
!
Since such a substantial part of the investments are paid in USD, for example drilling rigs, and platforms, it is necessary to control for exchange rates, as it can lead to a substantial change in costs for an investment. It is expected that an increase in USD compared to NOK will have a negative effect on investments, and vice versa.
!
A large part of oil investments in the North Sea is construction and rent of oil rigs and platforms. Steel is the larges input in rig construction, and an increase in the price of steel will increase the price of construction of oil rigs. An increase in the price of steel is therefor expected to influence oil investments, and will be a good control variable in my study.
!
H1b: An increase in oil prices will result in an increase in oil production investments, after controlling for exchange rates and steel prices
H2b: An increase in oil prices will result in an increase in oil exploration investments, after controlling for exchange rates and steel prices
!
!
3.1 Alternative Energy
In order to reduce the greenhouse effect, governments have introduced tax regimes on greenhouse gas emissions. Norway has relatively high emission taxes, and introduced a high carbon tax already in 1991. According to Bruvoll and Larsen (2002), the Norwegian oil production is heavily influenced by political decision, however they claim that the oil production will not be effected by emission taxes. Instead of lowering production, the oil companies have reacted to the carbon taxes by improving the technology and equipment on oil platforms, in order to decrease their emissions. They researched how the carbon taxes introduced in 1991 had influenced CO2 emissions, in the period from 1991 to 1999. Their results indicate that the taxes contributed to a reduction of 2.3 percent, compared to what the emissions would have been without the carbon taxes. This is a small reduction, considering that the carbon taxes have been on average about 20 percent (even higher for the petroleum sector). Their findings indicate that there is a marginal effect on emissions from carbon taxes in Norway. Thus I have the following hypothesis.
H3: The introduction of carbon taxes will not have a considerable effect on oil production investments
!
Sinn (2007) introduced the concept of the «green policy paradox». The concept describes that policies to reduce CO2 emissions reduce the discounted value of oil in the future more than in the present, oil producers will have an incentive to predict the the price cuts, and extract earlier. This is directly linked to investments behavior theory, as the policies, according to Sinn, will only increase price uncertainty.
!
H4: There is a positive relationship between the introduction of carbon taxes and oil production investments
!
On the contrary, Hoel (2010), tested Sinn’s green paradox, and found that if the carbon taxes are sufficiently high, the emissions will go down. The carbon taxes that were introduced for petroleum activities in 1991, was record high at the time (Sumner et al. 2009). Because of this, I hypothesize that in the short run after the introduction of carbon taxes, there is a drop in oil production investments.
!
H4b: There is a negative relationship between the introduction of carbon taxes and oil production investments
!
As a consequence of the environmental concerns, and the fact that oil is exhaustible, there have been rapid exploration of new sustainable sources of energy. These sources, which were discussed in the last chapter, have the potential of meeting world demand many times (Herzog et al. 2001). Large technological investments in renewable energy sources has cut costs substantially the last decades (see appendix 11), as well as increased the productivity. The huge potential of the renewable market creates a big opportunity for energy companies to invest in these new energy sources, and Norway has already invested in alternative energy, and newly announced that they will use a part of the oil pension fund to invest in alternative energy (Reuters, 2014). Most of the investments already made have been placed into wind energy. In my analysis I will use wind energy production in Norway as a proxy for investments in alternative energy in the country.
!
4. Methodology
4.1 Data
My data have been retrieved from multiple sources. All time series are quarterly. I have sorted the data in Excel and used STATA for regression analysis.
!
All my dependent and independent variables are from 1986 until 2013, with no missing data. I want to go this far back in time, in order to get as many observations as possible, and my dependent variables are only available from 1986.
!
Dependent variables
My dependent variables are exploration investments, and production investments. These are all gathered from Statistics Norway.
!
The investment data from Statistics Norway is included in the Norwegian annual national accounts. The data covers accrued and estimated investments in the oil and gas sector. The investments are divided into investments for exploration, field development, fields on stream, onshore activities, and pipeline transport. The exploration investments covers the activity from when the production license of the specific oil field is given to an oil company, until the exploration program is finished. The development investments and pipeline transport are included in the production investments. The data on investments is given in millions of NOK.
!
Independent variables
Brent oil prices are gathered from British Petroleum’s online database. The prices are not adjusted and given for the price at the time. They are not adjusted because my dependent variables are not adjusted.
!
contains all electricity produced from wind energy in Norway, which has been measured in gigawatt-hours (GWh) from the generators.
!
Control variables
Exchange rates for one USD to NOK were gathered from fxtop.com, which is a website specifically designed to provide currency converters, and historical exchange rates. The data was gathered in monthly average NOK, which I converted to quarterly average rates.
The prices of steel was gathered from the Index Mundi database. The data was given in USD per metric ton, in monthly average data, which I converted to quarterly average data.
!
4.2 Descriptive Statistics
The following table show descriptive statistics for variables that I use in the analysis. It contains an overview of number of observations, minimum values, maximum values, means, and standard deviation.
Table 1: Descriptive statistics of the variables used in my thesis
!
4.3 Methodology
In this thesis I will use econometric methodology in order to answer my research question. I will use multiple regression analysis in order to test the proposed relationships. The purpose of regression analysis is to explain how one or more independent variable(s) (y) can explain the changes in the dependent variable (x) (Wooldridge, 2009). Since I have more than two
N Min Max Mean Std. Dev.
variables in my analysis, I will use multiple regression. The multiple regression will allow me to explicitly control for other variables that affect my dependent variable.
!
!
This is the basic equation for a multiple regression model, where y is the dependent variable,
β0 is the intercept, β1 is the parameter associated with x1, β2 is the parameter associated with x2 and so on. The factors explaining y, which is not included in our independent variables is collectively contained in the error term, u.
!
More specifically, an ordinary least squares (OLS) method is used in my analysis. OLS method is used to estimate the unknown parameters (β0 andβ1), in a linear regression model (Schmidt, 2005 p103-107). Since we do not know the real value of the parameters, an estimation from an OLS regression is our best guess. The method minimize the distance from a fitted line to the observed residuals.
!
4.4 OLS Properties
When I perform a OLS regression analysis with time series data there are some properties that needs to be fulfilled. These are for stationarity, normality, multicollinearity, heteroscedasticity, and autocorrelation.
!
In order to test my variables in an OLS regression, my time series data needs to be stable. Each variable must have a variance and an expected value that do not depend on time. Values that are unstable tend to act random, and be impossible to predict. With stable variables, a shock in values (for example a price shock in oil prices) will affect the values after the shock, whereas an unstable variable would not. Unstable values will therefor be meaningless in a time series analysis. It is possible to test if my variables have the properties that are required for a time series analysis, which I will do in the next chapter following section. Next I will explain the properties, and explain the consequences, if they are not fulfilled.
• Stationarity
The first variable test I will perform is to see if they are stable enough. If they fulfill the requirements of stability I can do regressions on the time series. A time series is stable if the mean and variance is stable over time, and if the mean and variance change, it is non-stationary. If a time series is non-stationary, this can be due to a trend in the data, or it has a random trend with a unit root, or a combination of both. (Schmidt, 2005, page 329-331)
• Normality
The next assumption we need is the assumption of normality. The errors of u need to be independent from x and independently and identically distributed as normal (Wooldridge, 2009).
• Multicollinearity
Multicollinearity occurs when there is a high correlation between two or more predictor variables in a multiple regression. With multicollinearity the coefficients estimated from the regression will be unreliable and unstable. (Schmidt, 2005, page 181-183)
• Heteroscedasticity
Heteroscedasticity occurs when the error term does not have the same variance given any values of the explanatory variables. Heteroscedasticty causes the standard errors of OLS regression models to be wrong, which means that the hypothesis tests is unreliable (Schmidt, 2005, page 243-245).
• Autocorrelation
The error terms are statistically independent. Correlation of a time series with lagged values is called autocorrelation. That is, the error term correlates with the error term in the next period, which is a very common problem in time series data. For example, it is natural that the change in the price of oil in one quarter has an effect on the price in the next one.
!
4.5 OLS Model Specification
In order to test my hypotheses, I will use the following three models.
!
!
Model 2, Oil Exploration Investments:
!
I will also run the models in two different time periods, to see if it changes the results Recall from oil exploration, and production investments (figure 8 and 9), there is a clear separation which can be made around year 2005, when the investments increased massively. Because of this it will be interesting to run the models twice, with, and without a dummy variable to separate two time periods.
!
Model 3, Alternative Energy Effect: 3a:
3b:
3c:
!
Model 3a resembles model 1, but with an added variable for change in carbon taxes. In this model I will see if there is any changes from the results in model 1 when I add a dummy variable to separate the time when the carbon taxes were introduced in 1991. Model 3a is used to test hypothesis 4 and 4b. In model 3b, and 3c wind energy production is added as an independent variable. This variable will be used as a proxy for alternative energy production in Norway. This way it is possible to check if increase in alternative energy investments has any effect on either exploration or production investments, which accordingly tests hypothesis 5.
6. Results
6.1 Data tests
Stationarity
When I look at the graphs of my dependent variables (appendix 2 and 3), it is clear that there is a positive trend for both exploration and production investments. In order to perform a regression analysis with time series data, I need to remove the trend as it can give wrongly correlated results because of biased coefficients.
In order to solve the stationarity problem I will use first-difference estimators for my variables. Appendix 7 and 9 displays a histogram of my variables after using first-difference, and we can clearly see that the problem of stationarity is solved.
!
Normality
Appendix 6 to 9 show the normal distribution of my variables, before and after first-difference. It is clear that the problem I had with normality is consequently solved.
!
Multicollinearity
In order to test my data for multicollinearity I checked the variance inflation factor (VIF) for each model. In the following table I have presented the variance inflation factors for each model:
!
Model 1 and 2:
Table 2: VIF results for model 1 and 2
!
!
!
Variable VIF 1/VIF
Oil Price 2.64 0.378
Steel Price 4.09 0.244
!
Model 3:
Table 3: VIF results for model 3
!
We can see that for model 1 and 2, multicollinarity is not a problem, as the VIF values are well below five for all variables. In my third model however, the VIF scores are much higher. The wind energy production variable is even above 10, so I can confirm that there is strong multicollinarity between the variables. It is clear that it is the oil price and wind energy variable which is causing the multicollinarity, but I can not remove them, because they are vital for the model. This is a limitation in my study.
!
Heteroscedasticity
In order to test my data for heteroscedasticity I will use the Breusch-Pagan test, which will test the null hypothesis that the variance of the residuals is homogenous. Following is the results of the Breusch Pagan test for my models.
!
Table 4: Breusch-Pagan test results
!
From the results we can see that model 1 and 3 which has production investments as the independent variable has a low chi2 value and a p value which indicates that we can safely
Variable VIF 1/VIF
Oil Price 8.85 0.112
Steel Price 4.77 0.209
Exchange Rate 2.09 0.479
Wind Energy Production 11.87 0.084
Model Chi2 P value
Model 1 4.84 0.0278
Model 2 31.25 0.0000
Model 3a 1.48 0.2243
confirm the null hypothesis of homoscedasticity. However, there is seems to be a problem with heteroscedasticity in the models where exploration investments is the independent variable.
!
I solve the problem of hetroscedasticity in model 2 and 3b, by adjusting my estimated standard errors. I adjust the standards errors, by using White’s standard errors.
!
Autocorrelation
In order to test my models for autocorrelation I performed a Durbin Watson test for each model. This test is used to detect if there is autocorrelation in the residuals.
Following is the result of the Durbin-Watson tests:
Table 5: Durbin-Watson test results
!
The Durbin-Watson test gives a score from 0 to 4, where a score of 2 indicates that there is no autocorrelation. A rule of thumb is that a score under 1 is cause for concern regarding autocorrelation. From the results we can see that there is no problem with autocorrelation in my models.
!
As all assumption for my models are now fulfilled, except the multicollinearity issues in model 3, and I can safely proceed to the OLS regression models.
!
!
!
!
!
Before first order difference After first order difference
Model 1 0.40 2.76
Model 2 0.329 2.78
Model 3a 0.87 2.73
!
6.2 OLS regressions
Following is the results of the OLS regressions performed in STATA.
!
Oil Production Investments
As explained chapter 2.4, there will be a lag from a change in prices, to a change in investments in the oil industry. In order to find the optimal number of lags to add to the model, I used a stepwise approach. I looked at the outcomes of the model, and found that the optimal lag for oil prices was 24. In other words, the effect of a change in oil prices on oil investments after six years.
!
The results of model 1 is presented in the following table.
Table 6: Results of model 1
* Significant at the 5% level. ** Significant at the 1% level
!
From the results we can see that the model is indeed significant for oil price as an explanatory variable for oil production investments, also when controlling for the steel price and USD to NOK exchange rate. The results imply that a one dollar increase in oil prices will increase the oil production investments with 54.85 million NOK, six years later. The adjusted R2 value
implicate that the oil price can explain for 6.5 percent of the change in oil production investments.
!
Oil Prod.
Inv. Oil Price Steel Price Exchange Rate F R R
The second result presented in the table is the same model with an added dummy variable to divide between the periods before and after 2005. We now see that the significance is even stronger, and that the coefficient increase from 54.85 to 83.56. Also, the R2 adjusted value
increased, implying that the model can explain 9.3 percent of the changes for investments.
!
The lag time of six years is surprisingly long, as it expected that oil companies react to oil prices quicker.
!
Oil Exploration Investments
In order to find the optimal lag for model 2, I proceed the same way as I did with model 1. The optimal lag for oil prices in model two is 14 (three and a half year).
!
The following table presents the results from model two, oil exploration investments.
Table 6: Results from model 2
* Significant at the 5 percent level. ** Significant at the 10 percent level *** Significant at the 1 percent level
!
We can see that the model is only significant after adding control variables. The results are a bit surprising though. The oil price is significant at a five percent level, and the coefficient is actually negative. The results imply that a one dollar increase in the oil price will result in a decrease of 8,36 million NOK, and 15,67 million with the period dummy. We can therefor reject hypothesis 2a and 2b, an increase in oil prices do not have a positive effect on oil exploration investments.
Oil Expl.
Inv. Oil Price Steel Price Exchange Rate F R R
!
!
!
Alternative Energy Effect
In order to test hypothesis 3 and 4 I have added a dummy variable for when the carbon taxes were introduced in 1991. In order to test these hypotheses I will have to use a short lag time, because it is only five years from the start of the observation to the policy change in 1991. Therefor I have used a lag of one year in this model. I have performed the regressions with and without the carbon tax dummy, to see if there was any effect.
!
Table 7: Results from model 3a
!
From the results we can clearly see that the model is insignificant, which rejects both hypothesis three and four. The F-test for the model is insignificant, both with the carbon tax dummy and without it. None of the independent variables are significant, and there is minimal change in the coefficients after adding the carbon tax dummy. In other words, the introduction of carbon taxes in 1991 did not have any effect on oil production investments in Norway, which supports hypothesis 3, and reject hypothesis 4.
!
In order to look at the effect of alternative energy on oil investments, I have added wind energy production as an explanatory variable. I have used a lag of one year on both oil prices and wind energy, when oil production investments is the dependent variable. With oil exploration investments there is no lag in the independent variables.
!
Oil Price Steel
Price Exchange Rate Carbon Tax F R adjustedR
!
Table 8: Results from model 3b and 3c
* Significant at the 5 percent level. ** Significant at the 1 percent level.
!
We can see that the model is significant for both production and exploration investments. Interestingly oil prices are highly significant now that we added wind energy production as an explanatory variable, but keep in mind, that when the wind energy variable is added, there is a problem with multicollinearity, so I would advice to read the results from this model with some precaution.
!
Wind energy is also significant, and the model implies that an increase in one GWh produced wind energy in Norway, decreases oil production investments by 14.17 million NOK, and oil exploration investments by 4.87 million NOK. These surprising results could come from the fact that wind energy production has been increasing a lot more than the investments have (see appendix 2, 3, and 5), and I therefor read these results with some precaution. I also read the highly significant results for the effect of oil price changes here with some skepticism, because they have become more significant after adding wind energy production, which is not a very good explanatory variable for oil investments. However, the results from this model confirms hypothesis 5, investments in alternative energy has a negative effect on oil investments.
!
To summarize, the results from the OLS regression models support hypothesis 1a, 1c, 3 and 5. Hypothesis 2a, 2b, and 4 are rejected.
!
Oil Prices Wind
Energy PriceSteel Exchange Rate F R adjustedR
7. Conclusion
In this study I have analyzed the effect of changes in oil prices, and green policies on oil investments in Norway. This effect is important to study because oil investments are so crucial for the Norwegian economy, and drop in future oil prices can lead to a severe drop in oil investments. Based on theory on investment behavior in the oil industry, and previous studies done on oil investments reaction to change in oil price on the world market, I hypothesized that an increase in oil prices will effect in an increase in oil investments. I have studied the effect on investments in oil exploration, and oil production separately. To test the effect of green policy effects, I analyze the effect of the introduction of carbon taxes in 1991, and investments in alternative energy.
!
The results implies that an increase in oil prices will lead to substantial increase in oil production investments. However these results are only significant with a very high lag value for the oil price variable. I have found that there is a strong effect from change in oil prices on oil production investments after six years. This is somewhat puzzling, as existing literature would imply a more immediate effect.
!
The results show that changes in the oil price have a highly significant, but negative effect oil exploration investments after three and a half years. This surprising finding could be explained by Osmundsen et al. (2005), who finds that oil exploration investments failed to respond to the large increase in oil prices from 2001-2005.
!
decrease in oil investments. By using wind energy production as a proxy for investments in alternative energy, I found significant results supporting the hypothesis.
!
To answer my research question, «Does change in oil prices and green energy policies have an effect on Norwegian oil investments?». Yes, oil prices are a good explanatory factor for changes in oil investments, and a decrease in oil prices will lead to a substantial decrease in oil investments. Results on green policies show that investments in alternative energy has a negative effect on oil production investments, and that the introduction of carbon taxes had no significant effect on oil investments.
!
For the ongoing discussion on how dependent Norway is on the oil industry, and what will happen if there is a big drop in oil prices. My study shows that a drop in oil prices will have a severe negative effect on oil production investments, which will result in lower growth in GDP for Norway. However, the results show that there is a strong lagged effect, and that oil production investments do not react until six years after. This would imply that we should expect a continuous high growth in oil investments in Norway, since the oil prices have been growing rapidly for the last six years. The results for exploration investments are the opposite, but these investments are substantially smaller compared to production investments, and would not have huge impact on the Norwegian economy in the short run. In the long run however, decrease in exploration investments today, means there will be less production investments in the future.
!
7.1 Limitations and future research
My thesis has some important limitations. First of all I do not have enough control variables in my analysis. Oil investments are influenced by many factors, and I only have two control variables, due to lack of available data for my time period. Important factors like oil industry wage levels, oil rig rental costs and oil platform construction costs, would make my models much more robust, than they are now.
It is also a limitation that the production investments covers such a large part of the value chain, and data more separated into different parts of the value chain would improve the analysis.
!
Also my analysis on green policies could benefit from more indicators of taxes and regulations for the industry. Another limitation is the multicollinearity problems I had when analyzing the effect of wind energy.
