|
|
|
|
CIAO DATE: 3/99
The Political Economy of Energy Taxes, OECD 1973-1995: The Role of Political Institutions, Public Opinion and Ideology (A Pre-Test) *
Claremont Graduate University
School for Politics and Economics
International Studies Association
40th Annual Convention
Washington, D.C.
February 16–20, 1999
DISCLAIMER: None of the results reported in this work are considered to be final findings. The author reserves the right to change interpretation of the preliminary results upon the acquisition of the complete data-set.
Please, do not cite this work without permission. Thank you.
Abstract
The paper presents a theoretical cross-national study of energy taxation, concentrating on the heavy fuel oil tax. It theoretically investigates the effects that public opinion, institutional corporatism and left-wing ideology may have on the cross-national variance in manufacturing energy taxes, controlling for the plausible influence of budget deficits, energy import-dependency and deindustrialization. It is hypothesized that in more corporatist nations public opinion supportive of energy conservation, in combination with the Left-wing ideology of governing legislative coalition, will lead to higher energy taxes. Deindustrialization, proxied by the declining employment and output value in/of energy-intensive industries is believed to be responsible for a certain share of energy tax variance in the OECD countries. Finally, it is argued that energy import-dependency brings affects national manufacturing energy taxes.
The research offers a number of pre-tests which examine the significance of economic predictors: deindustrialization, budget deficits and energy import-dependency. Due to the fact that this study represents an initial testing stage of a much broader project, not all the data necessary for empirical analysis are available at this time. The preliminary test demonstrated statistically-significant aggregate model fit for the above economic predictors. While a number of OLS regressions generated expected and significant results for several nations, the tests for France, Japan and Germany did not confirm the expectations. The inconsistency of results is explained by extremely small number of observations per nation (with the maximum of 22 per each variable) and by a large number of missing observations in some of the predictors. Individual analysis of deindustrialization effects yields promising results: for the countries where the data permitted this type of testing, deindustrialization exhibited statistically significant relationship with the level of heavy fuel oil tax. Several correlation tests applied to the effects of public opinion remain inconclusive due to the limited data availability and inconsistency of results. The same can be reported with respect to the effects of corporatism. In sum, the lack of necessary data does not permit conclusive analysis but allows to demonstrate some preliminary trends and exhibit the truly unique data and theoretical arguments.
Introduction: Research Issue
The paper attempts to explain decisions of governments to levy energy taxes or increase the level of an existing energy tax. It represents the pre-test stage of my dissertation work on political economy of energy taxation. The spatial domain of this study includes the OECD economies within the time frame of 1973-1995. Energy taxes considered include various forms of national consumption and excise taxes 1 on oil derivatives, 2 coal and natural gas in manufacturing and transportation excluding customs duties and taxes on sales of electricity.
The research attempts to develop a theory to answer the following questions:
To provide a tentative and preliminary answer to the above questions, I employ statistical modeling to test a number of empirically operationalizable hypotheses inducted with the help of the public choice framework.
Specifically, this research outlines the theory of energy taxation as a function of political influence of domestic energy-intensive manufacturers (versus unorganized random energy consumers), the effects of public opinion, political corporatism, and ideological preferences of governing coalitions (among other predictors of energy efficiency), and the reaction of responsible decision-makers to the political ‘inputs’ of the above factors. 3 In this paper the reported empirical model pre-tests 4 the economic conditions leading to the cross-national variance in energy tax which is expressed as a function of several well-researched economic variables: energy intensity of manufacturing, budget deficits, deindustrialization 5 and energy import-dependency. Later I argue that energy taxes are also stipulated by the political values of decision-makers, policy-making routine in each expected country, and by the social setting in which decision-making takes place (public opinion). In my theory energy taxes are also influenced by the government’s perception of expected costs and benefits imposed on a very heterogeneous energy consumer groups by an introduction of an energy tax. The addition of political variables to the theoretical analysis and the subsequent use of the case study allow me to sense the way in which selected social factors and political values of decision-makers augment the purely economic justifications for institutionalization or increase of energy taxes.
Why study the motivations of decision-makers and institutional structures that lead to cross-national variance in energy taxes? — The main argument for researching energy tax policies rests on the fact that energy taxes may induce the reduction of the wasteful energy consumption. In its turn, the reduction of wasteful energy consumption is important for a number of reasons.
he first argument: currently the world is facing the challenge of increasing global energy demand 6 in the environment of decreasing fossil fuels’ reserves. As developing countries industrialize 7 and their standard of living increases rapidly, their commercial energy intensity tends to rise. In the first place, this global growth of energy demand is accelerating the depletion rate of non-renewable energy resources. Though it is argued that substantial reserves of carbon fuels are available to secure stable economic growth throughout the XXI century, serious disagreements exist about the exact time horizon when fossil fuels may be exhausted. 8 It is known, however, that global energy intensity changes point to the direction of sky-rocketing fossil fuels’ consumption:
Primary Energy, GDP and Energy Intensity Changes (1981-1991) (Per Cent/Year)
| Regions of the world | _E/E | _GDP/GDP | _I/I |
|---|---|---|---|
| South Asia | 6.5 | 5.2 | +1.3 |
| East Asia | 7.7 | 6.6 | +1.1 |
| Latin America | 2.9 | 1.8 | +1.1 |
| Africa | 4.1 | 2.7 | +1.4 |
| OECD | 1.4 | 3.7 | -2.3 |
where _E/E are the changes (increase) in primary energy 9 consumption; _GDP/GDP represents the rate of economic growth; and _I/I demotes change in energy intensity of production.
Source: Energy in Developing Countries— A Sectoral Analysis, OECD/IEA, Paris (1994).
Next, projected energy consumption growth threatens to increase carbon-dioxide (CO2) emissions that are directly responsible for the global warming effect which, in its turn, is expected to adversely affect economic growth in a large group of nations. 10 A great deal of recent studies on climate change, headed by the IPCC research team, 11 demonstrate that even a marginal increase in global temperature may have detrimental effects on agriculture, forestry, human health, and the well-being of coastal communities. Recent statistics show that current level of carbon dioxide emissions accounts for 55% of already observed global warming 12
Third, the expected growth of global energy demand coincides with the relative decline in the number of world oil-exporters. 13 This phenomenon is widely discussed in the literature on international regimes under the label of rising vulnerability of domestic producers due to the decreasing security of international oil supply (Keohane and Nye, (1977)). Dependence on fossil fuel exporters for the OECD members was partially reduced in 1974 when the International Energy Agency (IEA) was created. 14 However, as most of the major fossil-fuel exporters — such as Russia, Saudi Arabia, or Iran — are not subjects to any IEA’s agreements, the argument of increased energy-supply vulnerability resulting from highly concentrated pool of oil exporters remains valid. 15
Finally, implicit in much of the recent literature in political economy is the belief that utilization of energy resources is determined, primarily, by economic factors, such as world price of fossil fuels, structural changes in the economy, degree of dependency on imported fuel, etc. 16 While intense academic debates occurred around the issue of behavioral changes of consumers in response to increased energy prices, 17 social scientists have not invested adequate rigor in examining the political and social determinants of various energy policies, including energy taxes. This study attempts to bridge this gap. Taking into consideration that a great volume of economic literature has already scrutinized the topic of efficient energy use by taxing energy consumption, my paper attempts to control for empirically tested economic predictors of energy taxes and innovates in examining energy tax policies from political, social, and institutional standpoints.
The central theme of my research is to show how selected non-economic factors may influence energy tax leading to more efficient national energy consumption. Given all the problems associated with wasteful energy use (which are illustrated above), I view this inquiry as a very timely and innovative study.
The existing literature: What we already know
Historically, the issue of energy saving did not attract scholastic attention until the first oil price shock of 1973. The industrialized countries, particularly Japan and most countries in Western Europe, reacted quickly to this first oil shock and initiated efforts to use less energy in 1974 or shortly thereafter. As a result, the majority of social science literature examined the issue of energy saving in the light of conventional neo-classical economics. Namely, a great deal of research concentrated on the effects of world price fluctuation on initiation of energy saving policies. 18 The same earlier clusters of classical economic literature pointed out the important role of energy taxes in reducing energy consumption through manipulating consumers’ behavior. 19 In this context, energy saving policies appear to be a direct response to perceived scarcity of a vital resource.
As economists progressed in their understanding of mechanisms driving energy policies, they came to the conclusion that average national income is a great determinant of whether energy-saving practices are going to take roots within a particular state. 20 Empirical research in energy economics can also be credited for discovering that energy-saving practices are directly related to the sectoral structure of national economies. 21 Finally, success in energy conservation was found to be related to the national trading position in the energy sector: a net importer of fossil fuels is expected to be more prudent in its energy policies 22 (though the empirical evidence in support of this argument is not highly conclusive). From the above factors, average national income and sectoral structure of national economies appear to influence levels and targets of energy taxes. 23
The 1980s brought with them the security of understanding that the world is not likely to run out of fossil fuels in the next 40 (and beyond) years. However, academic attention shifted to the problem of global warming and the associated world-wide pollution threats. The renewed attention to the issue of the efficient energy use as well as to the problem of the effects of energy and/or carbon taxes resulted in a great volume of econometric modeling. The models frequently provided widely divergent results as they were based on different assumptions related to energy-saving technologies, discount rates, energy demand elasticity, etc. 24 With this new wave of research on energy tax and savings the problem has acquired a great depth of rigor characteristic to deductive analysis while becoming even more distant from examining the political underpinnings of energy taxation and resulting conservation.
Energy taxes have been elaborately analyzed with respect to their effects on income distribution, economic growth, achieved reduction of carbon emissions, and competitiveness of the taxed industry. Numerous tax policy analyses have been carried out with the help of General Equilibrium and climate impact modeling: Boero et. al. (1991), Nordhaus (1990, 1991), Hoeller et. al. (1991), Cline (1992: Ch. 4), Barker et. al. (1995) — to name just a few. 25 The results of these studies illustrate that energy taxes entail a marginal decrease in the rate of economic growth, 26 which leads to the suggestion that energy taxes should be levied in a revenue-neutral fashion (by recycling them back to offset increased costs of production) (Boero et. al. 1991:93, Nordhaus (1993)). 27 Further, energy taxes are seen as one of the most effective methods to induce more efficient fossil fuel consumption (Griffin and Steele (1980); Schipper et. al. (1992); Poterba (1991), Pearce (1992), Barker (1993)) 28 with one important reservation: nations with presently higher energy taxes will achieve lower energy conservation results after new energy taxes or tax increases because of the decreased marginal effectiveness of an additional tax increment. 29 The most recent research addresses the phenomenon of carbon leakage due to energy taxes in developed economies: 30 Amano (1997), 31 Polidano (1997); however, the current stand on the issue is that carbon leakage is the issue of the medium term importance — not the immediate — concern and the dimensions of this problem cannot be presently determined (Amano (1997)). Finally, the literature is inconsistent in its findings about the effects of energy taxes on disposable income: depending on the assumptions about the terms of trade, effects of energy taxes on income can be either positive (when taxes are partially transferred to foreign consumers through exports) 32 , or negative (when the balance of trade is negative or when trade balance is omitted from the model). 33 The above are the illustrations of the studies examining possible impacts of an energy tax.
A recent IMF research 34 examined the reasons for energy taxation, concentrating on petroleum products. The study concluded that energy taxes can be raised, among other reasons, to achieve higher energy supply security (in addition to plausibly improving the distribution of income, raising revenue with low administrative costs, and — for the energy exporters — charging the export opportunity price for domestic sales of petroleum products in order to ensure a more efficient use of resource).
To summarize the results of the existing economic literature, energy-saving policies are initiated by governments in industrialized economies as a response to economic price-signals (response to the condition of scarcity). Energy taxes, with the exception of the recent concerns about carbon leakage, are the leading policy to reducing energy wastes. The most important empirically tested over the recent two decades propositions can be presented as follows:
A variety of non-economic frameworks encompass numerous descriptive research related to various government energy policies and to energy taxes in particular. As one example, organizational theory within the field of public administration examined how the structure of national policy-implementing agencies affects energy taxes and savings. 36 The field of public policy has produced numerous case studies of energy taxes and energy efficiency policies within the policy process/policy analysis framework. 37 However, both academic and policy-oriented research in public policy and public administration remains limited to detailed descriptive studies which do not provide opportunity for theoretical generalization about specific predictors of energy taxes and/or other successful energy conservation policies.
A useful insight in the relationship between energy-saving, taxes, and efficiency of government institutions was given in the 1996 International Energy Agency’s study of governmental role in energy conservation. 38 On comparative basis the study examined the relationship between various energy policies for several nations, and compared the components of initiated energy-saving programs. Until the present date, this report remains the most comprehensive analysis of cross-national policies and the source which elucidated some national institutional characteristics of energy-saving, including characteristics of national tax policies.
Two attempts to bridge politics and economics with respect to energy use across nations were undertaken by Nigel Lucas (1985) 39 and John Clark (1990) 40 *. Regardless of the great volume of analyzed material and attention to details, both studies do not suggest a coherent theory of what drives energy policies or how the policies could be classified with respect of achieving the set-up goals. This research deficiency impairs our ability to understand the factors which would explain the variance between energy taxes and energy use practices within OECD and beyond it.
The effects of political parties and institutions on climate control policies 41 were theorized in the most recent cluster of literature on energy and global warming: Kawashima 42 (1997), O’Riordan and Jordan (1996), Haigh (1996), and a number of other works in O’Riordan and Jagel: Politics of Climate Change. 43 This cluster of research did not examine the relationship between energy taxes and institutions/partisan orientations per se.
None of the numerous existing studies of energy taxes 44 directly examines political determinants of differences in the tax levels and the resulting energy savings. One common argument — that energy taxes are used as a mean of redistributing income between various social groups of energy users 45 — is widely quoted in the literature; however, it was not addressed from the standpoint of politics which would be the logical starting ground for exploring the reason why (and how) governments would use energy taxes for the redistributive purposes.
To sum up, despite the overwhelming number of normative, empirical, and descriptive studies related to energy pricing, taxation, subsidies, renewable energy R&D and fiscal incentives — all of which would lead to higher energy savings — the literature has rarely addressed the fact that these economic policies are the output of politically-induced manipulations of governments, where governmental actions are themselves functions of numerous political pressures of interest groups, institutional structure, social preferences of voters, and so forth. Once again, my research attempts to demonstrate this connection.
Theoretical Assumptions
The foundation of my research is the assumption that higher energy taxes indeed lead to higher energy efficiency. 46 I stipulate that rational actors (voters and decision-makers) are the unit of analysis in this study, where the key research questions are: what combination of economic factors (both domestic and international) will induce governments to push for higher/additional energy taxes to achieve efficient energy use? (what is the ‘formula’ of economic predictors of a higher energy tax?); to what extent do selected political factors influence energy tax policies that result in reduced energy intensity in manufacturing and transportation?
In constructing the theoretical assumptions and the subsequent model, I rely on several arguments of the public choice framework. First, I assume that governmental officials act as rational politicians: they strive to maximize the likelihood of their reelection which is a function of strong economic performance and voters’ political support for the existing policies (including energy taxes). Thus, energy tax policies will be raised only when taxes entail such a change in consumers’ behavior that the net perceived political and economic effects of the tax are positive. In other words, the plausible reduction of energy consumption due to the higher after-tax energy prices should not create politically-visible adverse economic effects; to the contrary — it should generate strong economic incentives for technological innovation, fuel substitution and overall energy conservation in the environment of voters’ consent. Further, being rational decision-makers, government officials would not advance an energy tax proposal that would hurt a politically powerful interest group: i.e. a well-organized financially-strong small interest group. 47
Second, following Schnider and Frey(1988), and Frey (1997) I further assume that political actors will consider and value the position of voters with respect to energy tax issues (especially at the times closer to the elections), but they will also pursue their own ideological preferences:
A government receives utility from carrying out its ideological program. A government formed by a left-wing party (or parties) may, for example, be assumed to increase the budget...., while a right-wing government wants to decrease it. The most important constraint to the government is political: it may stay in power only if it is reelected.....Government is also restricted in its activity by administrative and legal constraints...
F. Schneider and B. Frey, p. 240 48
Finally, since governments are restrained in their activity by administrative and legal constraints, I recall the argument of the state-centered public choice literature which stipulates that certain political institutions appear critical for understanding variance in policies and policy outcomes across nations (Ikenberry, Lake, Mastanduno (1988)). 49 Thus, I will address the effects of institutions which appear to be the most conducive to higher energy taxes. Specifically, I assume that political institutions that facilitate economic and political trade-off in decision-making, and which allow for the limited penetration of public pressure should be facilitating adoption of higher and more stable, non-fluctuating energy tax rates.
In sum, I attempt to merge a state-centered and interest-group centered public choice approach. 50 Thus, I induct a politico-economic model of energy taxation which, in the first place, controls for several specific economic factors which are likely to generate high variance in the levels of taxation, and, secondly, adds political variables to control for the effects of interest groups pressure, public opinion, ideological preferences of governing elites, and state decision-making practices/structures.
The logic of my theoretical induction is based on several additional assumptions consistent with the broad public choice literature which could be summarized as follows.
Energy taxes are raised for a broad range of reasons, including relatively low elasticity of energy demand, the desire of governments to close budget deficits, attempts to conserve energy due to the mentioned security externality. In addition, certain shifts in the structure of domestic markets may facilitate introduction of higher energy taxes: such as, the natural process of deindustrialization would lead to relatively higher energy taxes for the nations with low percentage of employment and output originating in energy-intensive sectors. I stipulate that all the listed reasons constitute valid arguments that have to be accounted for in the construction of my theory and empirical testing. Further, energy taxes are not homogenous for all energy consumers. The difference in energy tax rates, which the governments levy (with public consent) is stipulated by the political power of affected domestic consumers, which is the function of the degree of their political cohesiveness (how well-organized a given interest group is), the ability of energy consumer groups to get access to the top governmental decision-makers and the relative economic wealth of the affected groups. In this light, better-organized energy consumers with larger economic wealth and — through their better organization and wealth — a better access to top governmental officials will be subject to lower energy taxes or tax exemptions. For the purpose of my research I make a key simplification: the wealth of energy consumer groups is dependent on the share of GDP output such groups generate. I further simplify the ‘reality’ and add that in operationalizing the political power of interest groups, I will be concerned only with the energy-intensive industry interest groups and with their ability to exert pressure on governments for tax exemption or lower overall tax rates.
Nevertheless, any policy a government considers is subject to benefits and costs assumed by various interests beyond the cluster of energy-intensive producers. Policies with concentrated benefits and dispersed costs will have the most success in adoption and implementation (Olson); in other worlds, concentrated benefits satisfy political preferences of well-organized small interests. On the contrary, policies with widely dispersed benefits and concentrated costs will have hard time getting approval of any national Legislature because of expected strong political opposition.
Further, I assume that historical structure of political institutions 51 influences the access that various interest groups have to political bargaining process happening within each nation. This bargaining ultimately determines the level and shape of energy taxes because I consider democratic states where positions and stands of powerful interest groups matter. The historical institutional structures are relatively stable over time (absent revolutions or other major regime changes 52 ), and I consider them fixed 53 for the purposes of my analysis. I assume that institutions that reduce the bargaining costs of interest groups over the specifics of energy taxes, incorporate policy stands of all well-organized energy consumers and which overcome free-riding during tax policy implementation would be conducive to higher energy taxes and higher resulting energy efficiency.
Next, the extent to which voters’ pressure, exerted through public opinion, is influential with respect to the energy tax policies is also pre-determined by the aforementioned institutional structures. Specifically, governments with relatively closed (to public) decision-making practices are less likely to be influenced by public opinion, even if it is strongly opposing an increase in the existing energy tax. On the contrary, governments that are highly exposed to public criticism by the virtue of their openness to interest group pressure will be more receptive to public opinion. Should more open institutional structures coincide with strong opposition to energy taxes and low historical levels of taxation, any economic reasoning for imposing a higher energy tax will be suppressed by public opposition.
Finally, as I underscored on the previous pages, in democratic states policy-makers (governments) are interested in winning the support of the majority of voters with respect to any policy for the purpose of securing future (re-)elections. Thus, they will consider the opinion of a median voter on energy tax issues, especially close to the time of elections. 54 However, in between-elections, the opinion of voters may be less influential with respect to energy tax policy (especially so since energy tax debates are highly technical and in between-elections time they may not be broadly discussed in the public domain). In between elections, ideological preferences and partisan affiliations of decision-makers in national legislatures begin to play more important role than voter opinion.
The above assumptions call for an integrated politico-economic model of decision-making about the level and type of energy taxation that follows:
The Model 55
After two energy-supply crises governments are increasingly concerned about the security of energy supplies. 56 As governments are interested in sustaining stable economic growth, which depends on undisrupted supplies of fossil fuels, they attempt to reduce energy consumption waste and substitute for less energy-intensive technologies. It is also recognized that reduction of fossil fuel consumption would lead to decreased air pollution associated with combustion of fossil fuels. The most universal way to affect energy consumption is an energy tax because taxes adjust energy consumer behavior.
Following the above assumptions, higher taxes are more likely in nations that depend on energy imports and attempt to sustain high rates of economic growth. For nations with relatively lower rates of economic growth, adoption (or increase) of an energy tax is unacceptable economically and difficult politically. 57
Next, the availability of alternative fuels in the economy would affect the level of taxes. For example, if a nation under consideration has an abundant fossil fuel resource, such as coal, it may tax heavy fuel oil or/and light fuel oil to stimulate the consumption of coal thus reducing the national ‘energy bill’ and stimulating employment in the coal-mining sector.
Furthermore, in the countries that has been experiencing historically-significant shifts of manufacturing employment and production into the service sector, energy taxes may be higher due to the natural effect of deindustrialization. Especially so since after the oil shocks of the 1970s (1973 and 1979), the deindustrialization in the fossil fuels importing countries should be accelerating because increased prices of oil speeded up shifts into more capital intensive sectors.
58
Energy tax levels for the manufacturing sector
59
are expected to be lower relative to the energy taxes in transportation sector due to the political opposition of concentrated energy-intensive industries affected by the manufacturing tax. Their interests are going to be of paramount significance for decision-makers because politicians are always concerned about maintaining high employment levels in the unionized energy-intensive manufacturers as well as supporting energy-intensive producers in export-competitiveness.
The economic, geographical and historical reasons listed above are insufficient to explain levels and types of energy taxes. Given an economic incentive to impose the tax and given no serious political opposition to this tax from a domestic energy-intensive industry (because, for example, this industry may be made tax-exempt at the initial stages of tax implementation), three additional social factors would affect the level and type of energy taxes across the OECD nations:
Left-wing partisan ideology of governing coalitions (which I define later) may play a positive role in adopting energy taxes because left-wing political parties are usually more likely to support tax increase and environmental sentiments as they are trying to court the support of New Social Movement groups. 60 They also advocate increases in taxes for publicly-funded social services. Countries with high accommodative or integrative capacity in decision-making (corporatist countries) are likely to have more consistent (politically non-contested) energy-saving goals and associated higher energy taxes (Janicke and Weidner (1997), p. 13; Badaracco (1985); Ricken (1995); Janicke (1992)). This assumption would hold only when the interests accruing net economic costs from an energy tax/energy saving policy are given an incentive to cooperate (for example, through tax exemption in other areas or/and other tax recycling).
A combination of left-wing governing ideology and political corporatism would facilitate higher energy taxes even in the absence of public opinion concerned about air pollution and resource depletion. In the nations of a more pluralist-type decision-making, where the policy process is more open to public scrutiny, left-wing ideology of the governing coalition 61 and public opinion supportive of energy conservation may facilitate higher energy taxes. Relatively more closed corporatist decision-making culture combined with the right-wing ideology will not be conducive to the adoption or/and marginal increase of energy taxes. Corporatist states with left-wing governments will be more likely to institute higher energy taxes.
In sum, the model works as represented schematically in Appendix B. 62 The following section represents a set of hypotheses to be tested.
Hypotheses
A nation with high energy imports, especially oil imports (relative to exports) and a high past rate of economic growth lacks security of energy supply which mobilizes governments for actions to guarantee undisrupted supplies. Development of non-carbon fuels and renewable energy infrastructure may be considered. This is exactly what has been done to mitigate energy dependency since the beginning of the 1990s; however, as mentioned before, the commercial costs of using renewable energy sources relative to the conventional fossil fuels are high. This fact would be the first reason prompting governments to tax the conventional fossil fuels: to create incentives encouraging energy users to shift to various alternative fuels thus commercializing their large-scale consumption.
For a nation experiencing a high rate of economic growth the security of stable energy supplies becomes a priority policy. 63 In the times of economic recessions governments would not be concerned about energy saving because political priorities shift into keeping up with inflation and reducing unemployment: as energy is a cheaper production input relative to labor, during economic recessions a rational decision-maker would not push for energy taxes because he/se would lose votes.
When securing the undisrupted energy flow becomes a policy priority, governments have an urgent call to reduce energy waste per unit of production (or per unit of any function/operation, such as driving) in order to limit the dependence on imports. This argument especially applies to the nations with energy-intensive industries because — over the long-run — the dependence of domestic manufacturers on imported fuels leads to higher costs of manufacturing (additional transportation costs) and to the increasing transaction costs of lobbying and negotiating in international regimes to secure undisrupted fossil fuels supplies.
Politics are introduced in the picture exactly at this point. As a rational decision-maker, one would want to initiate a policy which would: provide a universal economic incentive to shift away from energy-intensive practices into new energy-saving technologies; reduce the high costs of this transition to your energy-intensive producers since a nation with a high ratio of energy-intensive industry employs a large number of voters affecting election outcomes; generate revenues for subsidizing investments 64 in energy-saving technology; and not disrupt the existing political alliances and voter distribution. The last point is especially important since political incumbents always want to remain in power while their challengers attempt to "replace" incumbents by suggesting policies only marginally different from the existing ones (radical policies would not lead to large number of votes supporting a policy, according to the public choice assumptions: recall the median voter theorem).
Hypothesis I
The next question to be answered is to which extent energy taxes are raised to reduce energy waste, and to which extent are they the function of the need to close budget deficits, as suggested in numerous literature. As follows from the aforementioned assumptions, it is not possible to single out the direct independent causal relationship between energy taxes and energy efficiency because any energy tax can be introduced for reasons other than attempts to reduce energy use. Further, ability of governments to collect taxes for funding public services is not homogeneous across nations: for example, income taxes are notorious for being avoided in the nations of Southern Europe (for example, Greece). An energy tax — be it an excise tax on heavy fuel oil for manufacturing or a gasoline tax — can hardly be avoided by energy consumers for a number of reasons (the major one being the low demand elasticity for energy). Thus, it is a good tool for rescuing national budgets.
Policy-makers are assumed to have a clear understanding of this property of energy taxes. For this reason, politicians have an incentive to place higher energy taxes in countries facing chronically weak tax collection records and high running deficits. This, the next hypothesis will attempt to control for this very obvious factor affecting energy taxation. 66
Hypothesis II:
In the presence of persistent budget deficits (up to 2 years time-lag before the energy tax introduction or/and increase in value), energy tax may be raised for the fiscal reason of closing the budget deficit.
Next I return to the nexus between deindustrialization and energy taxes. Economic deindustrialization is not equivalent to raising unemployment; it is rather a more complex issue of shifts in employment patterns stipulated by changing productivity of various sectors within economies. Thus, since deindustrialization is represented by the decreasing employment in manufacturing 67 (specifically, in energy-intensive manufacturing for the purposes of my paper), energy taxes should be higher for the nations with more sweeping deindustrialization because in those nations energy taxes affect a smaller proportion of active voters. Energy taxes may be higher in those nations also due to a smaller proportion of national economies being adversely affected by high manufacturing energy tax relative to the service industries, where consumption of energy occurs at a lower rate.
Hypothesis III:
Nations with higher rate or/and absolute measure of deindustrialization will have higher energy taxes
However, deindustrialization does not affect all types of energy taxes in the identical manner: national differences in fossil fuel endowment would also influence the level and type of energy tax. Specifically, if a nation under consideration has an abundant fossil fuel resource — say, coal—, it may tax the imported heavy fuel oil or/and light fuel oil to encourage the consumption of domestic coal thus reducing the national energy import bill and retaining current employment levels in the coal-mining sector. From all of the above it follows that one should consider the interactive effect of energy import dependency and deindustrialization in my model. 68
I reiterate that the model assumes that policy-makers carefully consider how introduction/adjustment of an energy tax will affect their chance of being reelected to a public office (where being reelected yields a higher utility relative to being ‘beaten’ by an opponent). Elected officials responsible for energy policy-making will also analyze the degree of cohesion and organization of the voters directly and adversely affected by energy taxes. Less numerous voters do not exert significant political pressure unless they are highly organized and have direct access to decision-makers. In the absence of strong unionization and/or well-organized producer interest groups 69 specifically opposing energy taxes, politicians have easier time pushing for a higher tax on energy because a higher energy tax may mean a lower income tax and resulting political credit in the basket of a decision-maker who proposed this “tax-exchange”. In the nations with high degree of deindustrialization both unionization and organized 70 power of manufacturers are lower relative to the nations with low degree of deindustrialization. A lower number of energy-intensive manufacturers will also be conducive to higher energy tax because the energy-intensive interests would be pressuring the decision-makers for tax exemptions or/and lower levels of taxes. It is expected that deindustrializaing nations with a lower portion of their national products generated in energy-intensive sectors will have higher energy taxes, where the remaining group of energy-intensive producers will be tax-exempt.
Since the ability of voters to mobilize in a representative democracy matters to the decision-makers, politicians will carefully consider the affiliation of voters with unions, where unions represent institutions of collective bargaining enhancing the bargaining power of voter groups (here voters become politically active). As in a deindustrializing economy with a low share of energy-intensive manufacturing (the one with the predominant number of voters employed outside of manufacturing) most jobs are not unionized, the bargaining leverage of manufacturing interests regarding the preferred level of energy taxes will be reduced. Under these circumstances, governments have political freedom to raise higher energy taxes.
Hypothesis IV :
Nations with a low energy-intensive share of GDP output (relative to the average output in the OECD sample) and with a low proportion of energy-intensive unionization in major export-oriented sectors 71 will have higher manufacturing energy taxes
Since my public choice framework is applied to the cluster of democratic economically developed countries, I have to control for the latent power of voters expressed through public opinion. Public opinion with respect to energy policies and regulations is expected to have some effect on the type, level and even existence of energy taxes since voters’ opinion has the potential to affect both election outcomes and party platforms (which, in their turn, affect tax policies). Within the public choice framework public opinion is an influential determinant of a policy outcome since an opinion of a median voter matters. 72
Public opinion supportive of energy conservation programs will be conducive to the adoption of energy efficiency policies (with the exception of high taxes on transportation fuels ). However, environmentally-supportive public opinion by itself may not be sufficiently influential for the adoption of energy-saving policies in the scenario where national decision-makers are relatively insulated from the public pressures (relatively closed corporatist decision-making systems), or in the periods between parliamentary and executive elections in both pluralist and corporatist economies (given good performance of the economy). Considering this argument, it remains difficult to determine the direct and unique effects of public opinion on energy taxes, especially for the case of examining energy tax and energy efficiency in manufacturing: public opinion is likely to be less influential in affecting policy choices for energy taxes in manufacturing because public opinion cannot stand up against better-organized political pressure of industrial energy consumers. However, the extent to which public opinion can be influential in manufacturing taxes can be determined only empirically.
The stipulated influence of institutional decision-making structures will be examined in a separate hypothesis and will be controlled for in the empirical specification. Considering the separate effects of pre-election and between-election timing, I come to the following hypothesis:
Hypothesis V:
Two additional hypotheses are related to non-economic factors influencing energy taxes. I next consider the effects of left-wing partisan ideology of the governing Parliamentary/Congressional coalition and political corporatism.
There seems to be a marked connection between the left-wing ideology of national governments and their propensity to levy or increase existing energy taxes; left-wing governments appear to be more supportive of energy conservation programs relative to the ‘right-wings’.
I define a left-wing governing coalition as an alliance of Social-Democrats (or Socialists) with the welfare-state supporting Liberal parties, such as the U.S. Democratic Party. 74 Historically, within parliamentary democracies, Social Democrats 75 have also been in alliance with the Greens to broaden their declining electorate (Eatwell & Wright (1993)). What similarities between ideological platforms make these political coalitions plausibe? — First, social-democrats stand for mixed economy based on the notion that governmental correction of market imperfections is the main solution for economic crises. This ideological platform is supported by the ‘reformed’ welfare-state liberals. As Green parties stand for the conservation of non-renewable fuel resources as well as for the reduction of air pollution 76 which, given the current structure of production, can be achieved only by the means of careful economic planning and market intervention, political coalition of the Socialists, welfare-liberals and the Greens becomes rationally justified 77
In sum, left-wing governments are closer connected to environmental interests relative to the right-wing groups because the majority of environmental interests advocate controlling economic growth for the sake of preservation of natural resources, and left-wing governing coalitions campaign on the platform of constrained, regulated markets. 78 Thus, there is a clear proximity of political preferences and ideological orientation between left-wing governments and environmental interests.
However, energy taxes (and other price-increasing solutions for energy problems) are found to have negative welfare effects on the low-income energy consumers since these consumers spend a higher portion of their income on energy consumption. Left-wing governments have traditionally avoided advocating policies leading to adverse welfare effects of the low-income groups. 79 Thus, during persistent economic recessions even left-wing governing coalitions will not support energy taxes or any policy leading to energy price increases, at least in transportation, because of regressive distributive effects of such a policy. 80
Finally, as mentioned before, the public choice literature argues that ideological preferences of governments become especially pronounced in between elections when political actors are free from the pressures to accommodate the diverse voters’ preferences. Taking into the account all the above, I come to hypothesis VI:
Hypothesis VI :
Finally, I consider the effect of accommodative or consensual decision-making practices within the governments. Accommodative structures (by which I imply corporatism) may facilitate initiation of a new energy tax and/or tax increase. 81 For example, corporatism is expected to have positive effects on introduction of energy taxes given that some incentive to cooperate is shown to producers in the form of tax recycling into reduction of other taxes, funding of energy-related R&D or other similar policies. Corporatist arrangements usually provide a context for more effective policy adaptation because, in the first place, the power of national peak associations over local units facilitates the pursuit of national rather then particularistic interests and, secondly, corporatism presumes compromice between diverse policy positions which leads to stability of policy processes. Given the risks related to introducing new policies in the energy sector (which result from the lack of practical knowledge in this field), corporatist systems may prove to be more flexible and responsive to changes in energy-related knowledge (Lyle Scruggs (1998)).
Most importantly, corporatism implies accommodating practice of decision-making not only on the part of energy consumers (including manufacturers), but also on the part of the government agents introducing a policy. Thus, within an accommodative structure of decision-making political actors will seek opportunities to create tangible incentives for energy consumers which would induce consumer’s acceptance of an energy tax. At the same time, in corporatist environment a broad spectrum of energy consumers will work with the elected officials to relate back the information about the actual economic and political effects of the tax policy. It is important to stress that accommodative process implies compromising on both sides of the negotiation table.
Further, accommodative decision-making is usually characterized by a small number of highly professional actors involved in negotiating any policy. The process of political bargaining remains representative of industry, labor and other directly affected interests, but limited in its openness to the general public. 82 Small size of the bargaining groups reduces side payments necessary to be made in larger groups in the process of achieving political accommodation. 83 Thus, accommodative process of decision-making reaches stable, uncontested policy decisions which are not likely to be challenged after they were adopted.
Finally, corporatist decision-making structures imply centralization of policy process which reduces the free-rider problem 84 : one manufacturer can accept an energy tax if all manufacturers in his industry do, and there is a mutual consent about the acceptance of a certain tax level. Hence, my last hypothesis:
Hypothesis VII:
Accommodative/Corporatist structure of decision-making leads to the provision of tangible side-payments to the actors adversely affected by an energy tax. For this reason, higher and more comprehensive energy taxes (including the Carbon Tax) are to be expected in the countries with corporatist structure of decision-making.
Empirical Specification and Results of Several Pre-Tests
As shown in the introduction to this research, spatial domain of my study includes the OECD countries; temporal domain encompasses 1973-1995.
Yip = Energy tax in manufacturing (taxes on heavy fuel oil 85 and natural gas, excluding VAT: including various excise taxes and CO2 tax for some nations in the 1990s); 86
X1 — net importer of energy (import substitution for coal, oil derivatives and natural gas);
X2 — public opinion (percent and weighted frequencies of OECD respondents supporting energy conservation policies) 87
X3 — national economic growth rate (cumulative, adjusted for inflation: in constant 1990 prices);
X4 — energy intensity of manufacturing (ratio obtained by dividing the GDP generated by several 88 energy-intensive sectors to the overall GDP);
X5 — Relative Degree of Unionization (number of energy-intensive manufacturing unions and membership numbers for the key unions);
X6 — deindustrialization (number of workers for several energy-intensive industries and economic output + value added in energy intensive sectors);
X7 — corporatism 89 (with variance among nations, no variance across the time);
X8 — ideology of governing coalitions (ordinal variable, coded from original information: 1 relating to the Left-wing Coalition up to 5 standing for Right-wing Coalition);
X9 — budget deficits (difference between annual expenditures and revenues);
X10 — dummy for the proximity of legislature/executive elections (1 = one year before the election time; 0 = otherwise).
Yip = Energy tax in manufacturing (taxes on heavy fuel oil 90 and natural gas, excluding VAT: including various excise taxes and CO2 tax for some nations in the 1990s); 91
X1 — net importer of energy (import substitution for coal, oil derivatives and natural gas);
X3 — national economic growth rate (cumulative, adjusted for inflation: in constant 1990 prices) 92 ;
X6 — deindustrialization (number of employees for several energy-intensive industries and economic output + value added in energy intensive sectors);
X7 — corporatism 93 (with variance among nations, no variance across the time);
X9 — budget deficits (difference between annual expenditures and revenues);
The hypothesized relationship between energy tax and the predictors will be found by a set of pooled-data regressions: each hypothesis is tested by assessing the significance of the respective beta-coefficients. At this stage of research it is necessary to stress that for each nation under consideration my data remains incomplete: for example, the data on heavy fuel oil taxes does not report the implied maximum up to 23 observations for any single nation (the time-span of the study is 1973-1995) 94 . Furthermore, a large group of nations have clusters on missing observations on both independent and dependent variables, which brings a serious bias in the observed pre-tests, making them highly inefficient. While a serious effort is being made to solve the problem with missing observations and reduce the resulting regression bias, it remains important to demonstrate some pre-test findings which would facilitate the future research. The pre-test is reported only for the nations which did not have missing observations in the dependent variable. 95
LINEAR SPECIFICATION TO TEST HYPOTHESES I—VII
+ ß 1X1 +ß 2X3 + ß 3X6 + ß 4X7 + ß 5X9 + 
I
To the equivalent degree, both in the aggregate and country-by-country runs I expect to observe a negative relationship between energy tax and deindustrialization with a set of positive signs for the remaining predictors.
On this stage of empirical analysis I do not employ non-linear specifications. However, as one may see on the example of Sweden given below, a series of data plotting and curve-estimation tests suggested the use of both linear and logarithmic specifications as plausible alternatives: 96
Curve fitting estimation between Swedish iron production (a proxy for deindustrialization, total value of output, national currency) and Swedish heavy fuel oil tax, with 10 missing observations (out of 25)
| Dependent Mth | Rsq | d.f. | F | Sigf | Upper bound |
b0 | b1 |
| SWEDTAXH LIN | .623 | 12 | 19.86 | .001 | -629.21 | .0462 | |
| SWEDTAXH LOG | .658 | 12 | 23.04 | .000 | -12700 | 1309.51 | |
| SWEDTAXH CUB | .683 | 11 | 11.85 | .002 | -1589.7 | .0993 |
In this example, the log functional form exhibits clear significance (five additional functional forms were tested for each country and model).
The graphic representation of the relationship between iron manufacturing output and the level of heavy fuel oil tax (for Sweden):
The above curve does not graph the Heavy Fuel Tax for Sweden; it simply captures the goodness of fit between the key indicator of deindustrialization (iron manufacturing) and the heavy fuel oil tax (the dependent variable).
As indicated above, the same procedure was performed for each single nation in the sample (data permitting) with respect to the available predictors. The logarithmic function appears to exhibit the best fit, followed by the linear one; the example of Italy follows:
Curve fitting for the Heavy Fuel Oil tax, Italy with respect to iron manufacturing output value (the proxy for deindustrialization):
| Dependent Mth | Rsq | d.f. | F | Sigf | Upper bound | b0 | b1 |
| ITALYIRP LIN | .516 | 12 | 12.80 | .004 | 27845.5 | .2538 | |
| ITALYIRP LOG | .761 | 12 | 38.25 | .000 | -11700 | 5143.95 | |
| ITALYIRP S | .668 | 12 | 24.18 | .000 | 10.6172 | -574.86 |
The second stage of the pre-test will attempt to examine whether the hypothesized statistical relationships hold for each individual nation in the sample. While a number of significant results may be expected for the pooled data analysis, it is doubtful that individual country runs will yield statistically-reliable output due to the small samples. However, a series of individual country pre-tests would allow me to select the candidates for more in-depth individual case-studies (not represented in this research).
The pre-test of the aggregate model yields the following results: 97
Number of obs = 125
F( 4, 120) = 9.46
Prob > F = 0.0000
R-squared = 0.2397
Adj R-squared = 0.2143
Dependent variable: Heavy Fuel Oil tax (converted into millions USD) = Adjtaxhf
| adjtaxhf | Coef. | Std. Err. | t | P>|t| | Beta |
| emplche | .1110832 | .0364964 | 3.044 | 0.003 | .3206667 |
| ironemp | -.0685233 | .0158679 | -4.318 | 0.000 | -.4592598 |
| impsubo | -.2716674 | .0789775 | -3.440 | 0.001 | -.2890039 |
| year 98 | 2.945361 | .7309888 | 4.029 | 0.000 | .3296496 |
| _cons | -5799.948 | 1447.188 | -4.008 | 0.000 |
Though the above model exhibits low values of statistics responsible for the overall model fit (with adjusted R squared of only 24% and F-statistics equal to 9.46), it is important to remember that half of the theoretically significant variables were not entered in the equation due to the temporary lack of data. What is surprising is that the findings for the overall model fit are significant even when a large number of relevant variables remain omitted.
In the above analysis variables "emplche" and " ironemp" stand for employment numbers in chemical industry and employment in iron manufacturing (both energy-intensive sectors) 99 . The variable representing import substitution (to capture the effects of import dependency) exhibits a wrong sign though appears to be significant. Low individual variable coefficients signal the problem of possible muticollinearity that could be solved after all the remaining observations are added to the existing data-set. At this stage of analysis it remains impossible to assess the validity of Hypotheses I (which addresses the effects of import substitution) and Hypothesis III (which captures the influence of deindustrialization) until the data-base is expanded.
The aggregate data OLS also tested the relationship between corporatism and heavy fuel oil taxes across the nations (bivariate regression):
Adjtaxhf = dependent variable, regressed on corporatism (index, no variance over time)
Number of obs = 163
F( 1, 161) = 21.97
Prob > F = 0.0000
R-squared = 0.1201
Adj R-squared = 0.1146
| adjtaxhf | Coef. | Std. Err. | t | P>|t| |
| corporat | 7.991736 | 1.704964 | 4.687 | 0.000 |
| _cons | 13.22028 | 1.742626 | 7.586 | 0.000 |
The significant relationship between corporatism and the heavy fuel energy tax disappears when the corporatist variable is added to the overall model containing all hypothesized and available for testing predictors. This inconsistency of results is easily explainable. First, and most importantly, the measure of corporatism used for this analysis 100 covers the period 1960-1980 while the available data on heavy fuel oil tax begins in 1979. Thus, each country contributes only 2 observations to the overall pooled data analysis (making the total of corporatist observations equal to 36) while the deindustrialization and import substitution contribute of about 240 observations to the analysis each. 101 Secondly, the corporatist variable does not contain variance across the time which is characteristic to all the remaining predictors. For this reason, no comparable computation of variance is possible for both time-varying and constant independent variables. As suggested before, in the future analysis the effects of corporatism can be tested through a separate ANOVA model. Still, it remains important to note that in a bivariate pooled model there appears to be a statistically-significant relationship between the corporatist index and oil taxes.
Do the corporatist nations indeed have higher energy taxes, as is suggested in Hypothesis VII? By observing the following table comparing national energy taxes (converted in to US$) discriminated by the degree of corporatism, it appears that a certain positive association between the two variables does exist: 102
| year | France | FrancCorp | Germany | GermCorp | Italy | ItalCorpor | Netherlan | NethCorp |
| 1979 | 0.188058 | -0.725 | 8.183306 | 0.48 | 1.203514 | -0.851 | 7.527418 | 1.006 |
| 1980 | 0.189304 | -0.725 | 8.250825 | 0.48 | 1.167679 | -0.851 | 7.595573 | 1.006 |
| year | NewZealn | NewZeCor | Norway | NorwayCo | CanadaHF | CanadCor | Denmark | DenmCorp |
| 1979 | 3.554502 | -1.106 | -48.6509 | -1.106 | 0 | -1.335 | 2.926829 | 0.518 |
| 1980 | 7.288925 | -1.106 | -69.1171 | -1.106 | 0 | -1.335 | 6.729223 | 0.518 |
However, a number of nations in the sample do not fall into the above category of states where positive or/and increasing values of corporatism correspond to a higher relative value of energy tax. The comparison between heavy fuel oil taxes in Germany and Ireland given below would support an argument that more corporatist states do not necessarily have significantly higher manufacturing energy taxes. 103
| year | Germany | GermCorp | Ireland | IrelCorp |
| 1979 | 8.183306 | 0.48 | 9.414654 | -0.528 |
| 1980 | 8.250825 | 0.48 | 26.85984 | -0.528 |
Thus, the detailed testing of Hypothesis VII remains to be carried out in the future either upon acquisition of a more extended time-series corporatist measure or after the energy tax levels can be obtained for the period 1973-1979.
While the current stage of aggregate data analysis did not permit to accept or reject the significance of Hypothesis III, present data availability allowed for a more in-depth statistical analysis of the effects of deindustrialization on both light fuel oil and heavy fuel oil energy taxes. The runs were performed on the non-aggregate basis for each OECD nation, where deindustrialization was modeled as a decline in both employment and manufacturing output for pulp & paper, chemical and iron & steel manufacturing (all considered energy-intensive industries). The model can be represented by the following equation:
X tax = α + β 1 EM + β 2 EMPA + β 3 IRPRO+ β 4 CHE +β 5 EMIR +μI
Where EM= employment in industrial chemicals (thousands employed);
EMPA = employment in paper and pulp industry (thousands employed);
EMIR = employment in iron and steel industry (thousands employed);
IRPRO = production of iron and steel for any given year (in national currencies, nominal value);
CHE = production in industrial chemicals (national currency, nominal value).
Several specifications include the variable IRVA = value added in iron & steel manufacturing, in 1985 USD: the variable is included to control for the effects of inflation and to check for the robustness of the results on the IRPRO variable.
The following section reports the results of a series of multivariate country-by-country regressions: 104
For Japan (heavy fuel oil tax is the dependent variable):
F(5,12) = 25.37
Prob > F = 0.0000
R-squared = 0.9136
Adj R-squared = 0.8776
| japtaxhf | Coef. | Std. Err. | t | P>|t| |
| japanem | -20.63962 | 2.538973 | -8.129 | 0.000 |
| japemiro | 10.81648 | 1.947604 | 5.554 | 0.000 |
| japaemp | -1.538852 | 1.272913 | -1.209 | 0.250 |
| japairpr | .0359467 | .0262144 | 1.371 | 0.195 |
| japanch | -.1287826 | .0477349 | -2.698 | 0.019 |
| _cons | 1918.577 | 987.6037 | 1.943 | 0.076 |
For Japan (light fuel oil tax is the dependent variable):
F(5,11) = 23.83
Prob > F = 0.0000
R-squared = 0.9155
Adj R-squared = 0.8771
| japtaxlf | Coef. | Std. Err. | t | P>|t| |
| japanem | -26.91342 | 3.332824 | -8.075 | 0.000 |
| japemiro | 13.46003 | 2.56408 | 5.249 | 0.000 |
| japaemp | -1.771678 | 1.716238 | -1.032 | 0.324 |
| japairpr | .0504507 | .0350331 | 1.440 | 0.178 |
| japanch | -.1775188 | .0627451 | -2.829 | 0.016 |
| _cons | 2734.361 | 1351.384 | 2.023 | 0.068 |
As can be seen from the above, the Japanese models are more significant for the "numbers employed" proxy of the deindustrialization though the positive sign for employment in Japanese iron production does not support the argument that Japan in consistently deindustrializing in all energy-intensive sectors. It is important to note that in these preliminary results the extremely high R-squared statistics may be inflated due to plausible autocorrelation of the predictors over the time. This problem is to be corrected in the future aggregate analysis.
To compare the results of the Japanese OLS to the statistical runs for other nations, I report below the results for Germany.
Heavy fuel oil tax as a dependent variable :
F(4,12) = 659.17
Prob > F = 0.0000
R-squared = 0.9955
Adj R-squared = 0.9940
| gertaxhf | Coef. | Std. Err. | t | P>|t| |
| germem | .0544849 | .0270046 | 2.018 | 0.067 |
| gerempa | -.0399225 | .0016892 | -23.634 | 0.000 |
| germirpr | .0001051 | .0000253 | 4.160 | 0.001 |
| germch | -.0001018 | .0000192 | -5.306 | 0.000 |
| _cons | 21.16448 | 10.12291 | 2.091 | 0.058 |
Germany, Light Fuel oil tax as a dependent variable
F(4,12)= 74.60
Prob > F = 0.0000
R-squared = 0.9613
Adj R-squared = 0.9485
| gertaxlf | Coef. | Std. Err. | t | P>|t| |
| germem | -.3815686 | .2846295 | -1.341 | 0.205 |
| gerempa | -.1340838 | .017804 | -7.531 | 0.000 |
| germirpr | .0009988 | .0002663 | 3.751 | 0.003 |
| germch | -.0005235 | .0002022 | -2.589 | 0.024 |
| _cons | 195.3983 | 106.6957 | 1.831 | 0.092 |
As in the previous case (Japan), individual variable results appear slightly more significant for employment indicator relative to the manufacturing value one (in terms of t -statistic’s significance). Again, the appearing R-squared statistics are unreliable due to the possibility of the autocorrelation effect 105 .
The deindustrialization tests further yield significant results for Norway, USA, Netherlands, Italy and Denmark (the runs for Australia, Austria, Portugal, Finland, and the UK were not attempted due to the data problems). Thus, I conclude that the pre-test of the deindustrialization variable yields evidence that this factor is indeed a significant predictor of energy tax levels within a given nation. At this time I avoid making conclusions on the direction and significance of the effects of deindustrialization until the future tests. However, I expect that in the complete cross-sectional analysis deindustrialization would yield statistically significant results, plausibly constituting the first finding of my dissertation model. More importantly, judging by the results of several OLS regressions, the employment proxy for deindustrialization appears to be more statistically significant relative to the manufacturing output value proxy. This observation seems to tentatively support the politico-economic explanation of energy taxes relative to the purely economic one: marginally increasing energy taxes in the sectors characterized by the decreasing number of workers are the indicator of the reduction of the political clout of energy-intensive manufacturers.
After the pre-tests of deindustrialization effects have elucidated some promising trends, the paper presents, as an illustration, two country-specific tests of the overall model with the specification Yi = α + ß 1 X 1 +ß 2 X 6 + ß 3 X 9 + ß 5 X 9 + μ I , (linear specification)
Where X1 = "...importsub"— net importer of energy (import substitution for coal, oil derivatives and natural gas);
X6 = "...IRPRO"and "IRVA85"— deindustrialization (number of employees for several energy-intensive industries and economic output + value added in energy intensive sectors);
X9 = "...defmi" or "... DEFTRI" — budget deficits (difference between annual expenditures and revenues).
The results are reported only for the nations that showed the most statistically significant preliminary findings with respect to the section of the model which can be tested with the limited available data.
The OLS on the U.S. heavy fuel oil tax (dependent variable):
F( 4, 17) = 22.74
Prob > F = 0.0000
R-squared = 0.8425
Adj R-squared = 0.8055
| ustaxlf | t | P>|t| | Beta |
| usirprod | -5.239 | 0.000 | -1.465725 |
| usirva85 | 4.322 | 0.000 | .982451 |
| USdefbil | 1.175 | 0.256 | .1894429 |
| usimpsug | 1.597 | 0.129 | .22282 |
| _cons | 4.523 | 0.000 |
OLS for the Norwegian heavy fuel oil tax (dependent variable):
F( 3, 17) = 6.88
Prob > F = 0.0031
R-squared = 0.5484
Adj R-squared = 0.4688
| nortaxhf | t | P>|t| | Beta |
| norwirpr | 4.513 | 0.000 | 1.488707 |
| norirva8 | -3.552 | 0.002 | -1.114557 |
| Norwdemi | -1.844 | 0.083 | -.3302733 |
| _cons | 1.583 | 0.132 |
The results for Netherlands, Italy, U.K., Canada, and Australia have statistically significant model fit, but lower individual predictor significance. Regression tests could not be run for Austria, Australia, Finland, Switzerland, Sweden and New Zealand due to a large number of missing observations on the dependent variable.
Due to the aforementioned problems with the number of observations and possible mismeasurement of the independent variables no conclusions can be made with respect to the above pre-testing. 106 It appears that the models for the US and Norway demonstrate correct signs and are statistically significant on the individual predictors. One would expect similar results for the overall model upon the completion of the statistical analysis.
Trial tests for Germany, France and Japan yielded "no-results" regardless of the fact that the data for these nations is complete and consistent. 107 It was logically expected that France could be the outlier on the dimension of the heavy fuel oil tax because of its public policy orientation on nuclear energy production, which leads to a higher level of direct command-and control regulations versus the market-based taxing approach. At the same time, the OLS test for the French light fuel oil tax generated a highly statistically significant results for the overall model fit and two deindustrialization predictors:
Number of obs = 23
F( 4, 18) = 65.44
Prob > F = 0.0000
R-squared = 0.9357
Adj R-squared = 0.9214
| frataxlf | t | P>|t| | Beta |
| franirpr | 8.603 | 0.000 | 1.687015 |
| franirva | -6.933 | 0.000 | -.8506315 |
| Fradefbi | 1.487 | 0.154 | .1648304 |
| franimps | -0.219 | 0.829 | -.0286223 |
| _cons | 1.698 | 0.107 |
As a result, the case of the French light fuel oil tax is a clear candidate for one of the country-specific case-studies.
The cases of Germany and Japan would also require more detailed case-study investigation should the final empirical analysis lead to ‘no-results’ findings for these two nations.
The last preliminary statistical observation to be reported in my work addresses the correlation between the European public opinion on energy use and energy taxes (as in the previous pre-tests, I concentrate on the light- and heavy fuel oil taxes). As stated in the theoretical section of my research, in a democratic state public opinion is a crucial variable which may affect the level of tax or decisions to institute a new energy tax. Great problems have been encountered in collecting cross-sectional time-series data for public opinion. The available Euro-Barometer energy surveys 108 begin only 1982, where the surveys on energy conservation and use were not carried out on the annual basis. However, up to date, the Eurobarometer public opinion frequencies remain the best data available due to the consistency and diversity of questions asked. For this reason, at this stage of research, I focus my pre-tests on the members of the European Union only.
The energy-conservation frequencies reported by the Euro-Barometer cannot be used in their raw form for a number of reasons. A serious consideration should be given to weighting both frequencies and percentages of responses in several nations because the data are not homogenous in terms of their demographic characteristics. For example, if a larger number of women were polled in the U.K. relative to men, the non-weighted use of these frequencies will skew the distribution towards more conservative attitudes (because, on average, in Great Britain women tend to be more conservative in their political and social attitudes). However, a number of the EU nations do not require application of demographic weights: Belgium, Italy, Luxembourg, Ireland, Portugal and Greece. Due to the lack of time resulting in inability to apply all the necessary weights to the overall sample, I focus here on the above nations.
Among a great selection of Euro-Barometer questions related to energy conservation, I have chosen four which appear to be the closest in their meaning to the questions: " How important is it for you to save energy and to prevent its wasteful use? Would involvement of (your country’s) government be justified to prevent the waste of energy ?" These are the theoretical questions that a researcher would need to ask a respondent in order to capture his/her attitude to energy conservation. It appears that any direct question about the increase of energy taxation or introduction of such a tax would solicit the negative response from an energy consumer. Following this logic, the following Euro-Barometer questions were selected:
Since the number of observations per year for each nation is equal to one (for each respective question), and the data have been entered and sorted only for 1986-1988, regression analysis or any multivariate statistical comparison is not an option. However, by employing a simple correlation analysis I attempt to see whether there is a consistent positive relationship between the public opinion supportive of energy conservation across the non-weighted sample of nations (according to Hypothesis V, there should be a positive and statistically-significant correlation). The results are reported below, where ...TAXH = ‘heavy fuel oil tax’;... TAXL = ‘light fuel oil tax’; ...SAVE= important to save energy’; ...MONJUS= ‘spending public money on conserving the energy is justified’; ...IMPO = ‘important to minimize pollution even if it costs more’.
Correlation for Italy , heavy fuel oil tax:
| italtaxh | italsave | itaminpo | itmonjus | |
| italtaxh | 1.0000 | |||
| italsave | 0.7967 | 1.0000 | ||
| itaminpo | 0.9739 | 0.9002 | 1.0000 | |
| itmonjus | 0.6751 | 0.1122 | 0.5338 | 1.0000 |
Correlation for Italy, light fuel oil tax, with 17 observations on the dependent variable:
| italtax | italsave | itaminpo | itmonjus | |
| italtaxl | 1.0000 | |||
| italsave | 0.6497 | 1.0000 | ||
| itaminpo | 0.8669 | 0.9002 | 1.0000 | |
| itmonjus | 0.7165 | 0.1122 | 0.5338 | 1.0000 |
Correlation results for Belgium, heavy fuel oil tax:
| beltaxhf | belsaven | belminpo | bemonjus | |
| beltaxhf | 1.0000 | |||
| belsaven | -0.4383 | 1.0000 | ||
| belminpo | -0.4874 | 0.9398 | 1.0000 | |
| bemonjus | -0.3108 | 0.2168 | 0.5374 | 1.0000 |
Correlation results for Belgium, light fuel oil tax:
| beltaxlf | belsaven | belminpo | bemonjus | |
| belsaven | -0.2715 | 1.0000 | ||
| belminpo | -0.3020 | 0.9398 | 1.0000 | |
| bemonjus | -0.1925 | 0.2168 | 0.5374 | 1.0000 |
Correlation results for Ireland, Heavy fuel oil tax:
| iretaxhf | irelsave | ireminpo | iregrewa | iremojus | |
| iretaxhf | 1.0000 | ||||
| irelsave | -0.3233 | 1.0000 | |||
| ireminpo | -0.3952 | 0.8841 | 1.0000 | ||
| iregrewa | -0.3419 | 0.3912 | 0.7759 | 1.0000 | |
| iremojus | -0.2878 | 0.1776 | 0.6169 | 0.9752 | 1.0000 |
The correlation results for the remaining group of self-weighting nations are reported in Appendix...It is evident that correlation results for Belgium and Ireland are negative, contrary to the findings one would expect from the logic which lead to Hypothesis V. However, these preliminary findings cannot disprove the hypothesis itself: the correlation results are run for 1986-1988 only; out of the self-weighting sample of the European nations and correlation results for Greece, Portugal and Italy exhibit the positive sign. Further, it is important to remember: Hypothesis V deliberately states that the positive effects of public opinion will be more pronounced for the transportation energy taxes, i.e. the taxes on gasoline. 113 The correlation analysis demonstrated above applies to manufacturing energy taxes. This caveat may shed light on the inconsistency of signs related to the relationship between public opinion and energy taxes. Such a factor should be carefully considered in the future statistical work, where public opinion will be entered as a separate predictor into the aggregate pooled-data model. At this moment all I could report is that in certain demographically-homogeneous nations public opinion related to energy conservation efforts appears to be positively correlated with the levelof manufacturing energy taxes.
In sum, the statistical section of this research has demonstrated several suggestive findings. Most importantly, it appears that deindustrialization, especially when expressed through the employment proxies, is a significant predictor of manufacturing energy taxes for several nations in the sample considered. This observation allows me to specifically concentrate on the employment proxies for deindustrialization for the future statistical testing. Next, the aggregate model test demonstrated significant results on the cross-sectional bases for deindustrialization and import substitution predictors.
Though no advanced formal analysis of political predictors is possible at this stage of research, the aggregate bivariate test of the corporatist variable demonstrated that is it significantly affecting the cross-national variance of heavy fuel oil tax (which I, essentially, the main manufacturing energy tax). It appeared that public opinion is positively and highly correlated to the level of energy taxes in several selected self-weighted EU nations: Greece, Portugal, Italy. On the other hand, the negative correlation of public opinion with heavy fuel oil tax in Belgium and Ireland may prove the argument that public opinion is not a positive linear predictor of manufacturing energy taxes in the future analysis with a higher number of relevant observations.
Finally, the results of a series of country-specific trail runs suggest that for the nations with the complete and no-measurement error data the available predictors perform in the expected direction and appear to be statistically significant. Still, as reported in Appendix C, three countries — Germany, Japan and France — appear to be outliers in these preliminary runs (with respect to the heavy fuel oil tests). As a result, due to the remaining problem of the small sample size, I call even the most significant finding for the US manufacturing taxes a pre-test. I acknowledge that the small sample size may result in multicollinearity that adversely impacts individual beta-coefficients. Until this problem is solved none of the theoretical hypotheses can be either rejected or supported.
Conclusion
The research advanced a theory explaining a variance in energy tax levels for manufacturing. Though the presented theoretical arguments addressed a range of hypotheses that were not formally tested in the subsequent sections of this study, the main contribution of this paper rests exactly with its theoretical induction. This study brought together economic and political reasoning with respect to oil taxes. A number of new 114 theoretical questions was addressed: the nature and effects of deindustrialization on the level of energy taxes; the interaction between social preferences of population and the decision-making structure of political institutions charged with framing taxation policies; the political and ideological characteristics of the politicians who design and implement oil taxes; and the effects of economic endowment with fossil fuels.
The study attempted several statistical pretests of the advanced hypotheses. Due to the extremely limited availability of the necessary data, none of the findings are reported as final. However, for the limited number of observations available, the statistical fit of the aggregate model appears to be significant. This preliminary finding was accompanied by a number of successful runs testing the significance of deindustrialization as a predictor of the variance in the heavy fuel oil taxes. Deindustrialization, especially when expressed by the number of workers in energy-intensive industries, appears to be a successful predictor for the level of oil taxation for several nations of OECD: Norway, USA, Netherlands, Italy, Denmark, Japan and Germany. However, since the variable was tested outside of the aggregate model, no final conclusions could be made about its significance until the data are complete to perform the cross-sectional time-series tests.
The overall model was also tested with respect to the effects of budget deficits, deindustrialization and import substitution for Netherlands, Italy, U.K., Canada, Australia, Norway and the United States. Though no consistent results were observed across the nations, the tests showed support for Hypotheses I and III in the cases of Norway and the U.S. At the same time, the pre-tests demonstrated the inability of the model to correctly capture the universal reasons for energy tax variance. Specifically, regardless of the availability of relatively complete data for Germany, Japan and France 115 , the results of the pre-test for these nations are insignificant (with respect to all predictors). One of the reasons for this inconsistency in preliminary findings may be the low number of observations in each OLS performed for each individual nation. It is expected that upon the completion of data entry more significant and cross-nationally homogeneous results will be observed. At the same time, the possibility remains that even the most carefully developed theory cannot account for all plausible underpinnings of energy tax policies, especially when it is tested across a wide range of countries.
The paper represents the first step in the chain of empirical tests and case studies leading to a complete dissertation project. In the light of this, the contribution of this work is in its theory and in the display of the unique data most of which were never collected and tested before for the purposes of examining the political economy of energy use.
References
Baker, Terry, Ekins, Paul and Johnstone, Nick (ed.). 1995. Global Warming and Energy Demand. Routledge.
Baumol. W.J., S. Blackman, and E. Wolff. 1989. Productivity and American Leadership: The Long View. Cambridge, MA: MIT Press.
Clark, John. 1990. The Political Economy of World Energy. The University of North Carolina Press.
Cook, Earl . 1976. Man, Energy, Society. San Francisco: W.H. Freeman and Company.
Crane, George and Amawi, Abla. 1991. The Theoretical Evolution of International Political Economy. Oxford University Press.
Didsbury, H. F. (edt.) 1985. The Global Economy: Today, Tomorrow and the Transition. World Future Society.
Dalton, Russell. 1994. The Green Rainbow. Yale University Press.
Desveaux, James A. 199?. Designing Bureaucracies: Institutional Capacity and Large-Scale Problem-Solving. Stanford University Press.
Duchin, Faye and Lange, Glenn-Marie. 1994. The Future of the Environment: Ecological Economics and Technological Change. Oxford University Press.
Eatwell, Roger and Wright, Anthony. 1993. Contemporary Political Ideologies. Westview Press.
Energy in Developing Countries — A Sectoral Analysis, OECD/IEA, Paris (1994).
Ergin, Daniel et. al.. 1982. Global Insecurity: A Strategy for Energy and Economic Renewal Boston: Houghton Miffin Company.
Fuchs, V.R.. 1968. The Service Economy. New York: NBER.
Fukasaku, Yukiko. "Energy and Environment Policy Integration." Energy Policy, 1995, 23(12): 1063-1076.
Gamba, Julio; Caplin, Favid, and Mulckhuyse, John. 1986. Industrial Energy Rationalization in Developing Countries. Washington, D.C.: World Bank.
Gaskell, George and Joerges, Bernward. 199?. Public Policies and Private Actions: A Multinational Study of Local Energy Conservation Schemes. Berlin: WZB-Publications (Wissenschaftszentrum Berlin).
George, Stephen. 1996. Politics and Policy in the European Union. Oxford University Press.
Goldemberg, José. 1996. Energy, Environment and Development. Earthscan.
IEA. 1987. Energy Conservation in IEA Countries. Paris: OECD/IEA.
IEA. 1993. Taxing Energy: Why and How. Paris: OECD/IEA.
Ikenberry, John, Lake, David and Mastanduno, Michael. 1988. The State and American Foreign Economic Policy. Cornell University Press.
Jänicke, Martin and Weidner, Helmut. 1997. National Environmental Policies. Springer.
Kitschelt, Herbert et.al.. 1990. Beyond the European Left. Duke University Press.
Lee, Thomas; Ball, Ben, and Jr. Tabors, Richard. 1990. Energy Aftermath. Harvard Business School Press.
Lengelle, M. 1966. The Growing Importance of the Service Sector in Member Countries. Paris: OECD.
Lipjhart, Arend and Crepaz, Markus. "Corporatism and Consensus Democracy in Eighteen Countries: Conceptual and Empirical Linkages." British Journal of Political Science, 1991, 21(2):235-247.
Lucas, Nigel. 1985. Western European Energy Policies. Clarendon Press.
Moomaw, William and Tullis, D. Mark. 1993 (?) Charting Development Paths: A Multicountry Comparison of Carbon Dioxide Emissions. Working Paper, Fletcher School of Law and Diplomacy (Tafts).
Moroney, J.R. et.al.. "Estimating the International GNP-Energy Relation: A Further Note." The Energy Journal, 199(?), 11(1):167-174.
Myers, Norman. 1996. Ultimate Security: The Environmental Basis of Political Stability. Island Press.
Nielsson, L. 1993. “Energy Intensity in 31 Industrialized and Developing Countries, 1950-1988.” Energy, 18 (4), 309-322.
Nordhaus, William. 1994. Managing the Global Commons: the Economics of Climate Change. The MIT Press.
OECD. 1997. Sustainable Development: OECD Policy Approaches for the 21st Century.
OECD. 1996. Energy Policies of IEA Countries: 1996 Review. Paris: OECD.
Pirog, Robert and Stamos, Stephen Jr. 1987. Energy Economics: Theory and Policy. Prentice Hall.
Rodwyn, Lloyd and Sazanami, Hidehiko. 1989. Deindustrialization and Regional Economic Transformation. Boston: Unwin Hyman
Rowthorn, R. and Ramaswamy, R. 1998. Growth, Trade and Deindustrialization. IMF Working Paper # 60; Washington, D.C.;
Rowthorn, R., and Ramaswami, R. 1997 .Deindustrialization: Causes and Implications. IMF Working Paper #42; Washington, D.C.;
Rowthorn, R.E. and Wells, J.R. 1987 . Deindustrialization and Foreign Trade. Cambridge: Cambridge University Press
Saunders, Harry. "The Khazoom-Brookes Postulate and Neoclassical Growth." Energy Journal, 199., 13(4):131-148.
Schipper, Lee et.al. 1997. Indicators of Energy use and Efficiency. International Energy Agency: Paris, OECD/IEA.
Schipper Lee, Stephen Meyers et al. 1992. Energy Efficiency and Human Activity: Past Trends, Future Prospects. Cambridge University Press.
Schipper, Lee et. al. 1992. World Energy: Building a Sustainable Future. Stockholm: Stockholm Environmental Institute.
UNEP, UNDP, and World Bank. 1982-1996. World Resources: A Guide to Global Environment. Oxford University Press.
UNEP & WMO: Climate Change 1995: Economic and Social Dimensions of Climate Change. Contribution of Working Group III to the Second Assessment Report of the Intergovernmental Panel on Climate Change. 1997. Cambridge University Press.
Williamson, Peter. 1985. Varieties of Corporatism: a Conceptual Discussion. Cambridge University Press.
Endnotes
*: Prepared for presentation at the 40th annual meeting of the International Studies Association, Washington, D.C., February 16–20, 1999. Back.
Note 1: A great variety of energy taxes can exist in one nation: gasoline/local road taxes, aviation fuel taxes, general energy consumption tariffs, specific carbon dioxide charges, power-resource development taxes. etc. However, the type of taxes varies greatly from a country to a country; there is a very limited number of absolutely equivalent types of taxes across all nations considered in my sample. I chose to examine the taxes which are:
a. Most comparable across the sample of nations examined (most universal ones);
b. The taxes readily available in the data tables published by the International Energy Agency;
c. Taxes universally affecting a large section of country population to allow for generalized hypotheses about the determinants of the tax (a nation state is used as a unit of analysis in my research). Back.
Note 2: Oil derivatives are a large spectrum of oil products from high-sulfur-content heavy fuel oil (that is an input in electricity production) to gasoline (in the class of light fuel oils). Back.
Note 3: For the complete picture of the model, see Appendix B. Back.
Note 4: The term "pre-test" is used to underline that the available results of the statistical tests are not conclusive because most of the available data contain a very limited number of observations; more in-depth data gathering and testing is going to be carried out in the future. Back.
Note 5: Deindustrialization is defined as a reduced employment in manufacturing accompanied by the employment shift into the service sector. A broader definition of deindustrialization incorporates the reduction in manufacturing output as one of the key indicators of this phenomenon, but the existing literature adheres to the first definition, associated with the reduction in manufacturing employment Back.
Note 6: Especially so since China is switching from the position of net global exporter of energy resources to the position of global importer; the same could hold for Indonesia in the early 2000s — a nation which is currently a significant exporter of oil and the OPEC member. Back.
Note 7: Though a number of studies in economic growth demonstrated that increased energy inputs in production are the necessary condition for stable growth (the most prominent from these studies are Schurr et. al. (1960), and Jorgenson in Electricity Use, Prodctive Efficiency and Economic Growth (1986)), they were modeled for the conditions of closed economies, i.e. economies not dependent on energy imports. Thus, the conclusions reached in these studies do not take into consideration the current international political reality. More recent studies (José Goldenberg (1998); J. Edmonds et. al. (1994)) demonstrated that as economies mature from the industrialization phase of economic growth into the service-sector phase, it becomes economically profitable to invest in energy efficiency as this investment may free capital resources for use in other sectors of economy while leading to cost-minimizing environmental solutions (Nordhaus (1993)). Thus, in the long run, policies leading to increased energy efficiency should be seen as potentially economically profitable for developed nations. Back.
Note 8: All economic forecasts on the duration of fossil fuels supply are extremely sensitive to the assumptions about the discount rate of consumption, world fossil fuel prices and the supply capacity of global oil producers (exporters). Back.
Note 9: Primary energy resources are fossil fuels, hydropower, renewables, and nuclear power — i.e., all energy sources where energy is directly obtained from. Back.
Note 10: See Appendix A for various projections of the effect of global warming on economic growth. Back.
Note 11: IPCC = Intergovernmental Panel of Climate Change. Back.
Note 12: Lashof, D.A. and Tirpak, D.A. 1990. Policy Options for Stabilizing Global Climate. U.S. Environmental Protection Agency, Washington, D.C. Back.
Note 13: For example, as costs of domestic energy production in the United States have become prohibitively high in recent years, the nation slowed down in its energy extraction rate, reduced energy exports, and became more import-dependent on energy suppliers from Canada, Mexico and the politically-volatile Middle-East.
Since oil remains the most economically-important carbon-based fuel for the foreseeable future, undisrupted supplies of oil are critically important for stable economic growth (which helps in preventing economic recession and shocks). Back.
Note 14: The Agency requires each member-state to maintain emergency fuel reserves for cases of supply disruption. Back.
Note 15: A similar argument is developed in MacKenzie, James. 1997. Climate Protection and the National Interests. World Resource Institute. Back.
Note 16: To illustrate my argument — in a detailed study which addresses the role energy plays in economic development, José Goldenberg argues: "The evolution of the energy intensity [of production] over time reflects the combined effects of structural changes in the economy... and changes in the mix of energy sources..." (Jose Goldenberg. 1998. Energy, Environment, and Development Earthscan). Back.
Note 17: Lee Schipper et. al. (1997); Khazoom et. al. (1980); Lee Schipper and Steven Meyers (1992), and a large volume of literature in energy economics. Back.
Note 18: All college-level text-books in Environmental and Resource Economics begin analysis of energy policies from the examination of behavior of world energy prices: to take Tom Tietenberg’s Environmental and Natural Resource Economics as a classical example. Back.
Note 19: The list of the literature addressing economic effects of energy taxes is endless; the latest works include Nordhaus (1993), Nick Mabey et. al. (1997), Climate Change (1995), etc.. Back.
Note 20: Namely, it was proven hat higher income lead to more energy consumption via work-saving appliances and more leisure and travel. (James Griffin and Genry Steele. 1980. Energy Economics and Policy. N.Y., San Francisco, London: Academic Press, p. 216). Also, see: Lee Schipper et. al.: Indicators of Energy Use and Efficiency. 1997. International Energy Agency: Paris, OECD/IEA, p. 272.
At the same time, extensive empirical studies have demonstrated that nations with higher levels of economic development tend to support consistent energy-saving policies which are initiated and promoted by state governments. (Numerous country case-studies published by the International Energy Agency in the 1980s-1990s support this argument.) Back.
Note 21: In particular, in the nations with predominance of energy-intensive industries (such as steel, pulp and paper, chemicals, ferrous metals, and minerals) energy saving is not an economic and political priority. This argument is supported in: Lee Schipper et. al.: Indicators of Energy Use and Efficiency. 1997. International Energy Agency: Paris, OECD/IEA, p.p. 207-230. Back.
Note 22: See William R. Moomaw and Mark Tallis (1993?): by extensively using cross-sectional data on energy prices and energy consumption, the authors demonstrated that energy exporters did not exhibit any sensitivity to changing world energy prices after oil shocks; this conclusion was made after observing energy consumption and CO 2 emissions patterns in Mexico, Nigeria, Venezuela, and Indonesia.
In the light of the above, it is important to stress that a net energy importer (a nation) will be inclined to conserve the imported fossil fuel not only because of additional transportation and import charges that make the resource more expensive, but also due to perceived dependency on a foreign supplier. While attempting to hedge against the higher economic risks, associated with this dependency, a net energy importer becomes interested in conserving energy. Back.
Note 23: For the support of this argument, see Julio Gamba et.al. (1996), Faye Duchin and Glenn-Marie Lange (1994), Lee Schipper and Stephen Mayers (1992), Daniel Ergin et. al. (1982), Robert Pirog and Stephen Stamos (1987),
International Energy Agency (IEA) case-studies, and the IEA publications related to the 1997 Kyoto Climate Summit.
However, a number of researchers take a classical economic approach to the issue, arguing for the minimal government interference in energy markets: especially so in the areas of price adjustments through energy taxes and subsidies for R&D. For example, see Thomas Lee et. al. (1990). Back.
Note 24: The most well-known models were developed by Cline (1992), Nordhaus (1991), and a group of authors in Terry Barker, Paul Ekins and Nick Johnstone (1995): Global Warming and Energy Demand Routledge. Back.
Note 25: The mentioned studies modeled a whole set of CO 2 abatement policies: they do not solely concentrate on plausible effects of energy taxes; however, these studies represent the most well-known initial research on the issue. Back.
Note 26: For example: Aasness, Jorgen et. al. 1996. “Welfare Effects of Emission Taxes in Norway.” Energy Economics, 18: 335-346;
Yochi, Kaya and Keiichi, Yokobori. 1997. Environment, Energy and Economy. Tokyo/N.Y/Paris: UN University Press;
Whalley, J. and Wingle, R. 1990. “The International Incidence of Carbon Taxes. “ Mimeo, NBER, Cambridge, MA and Waterloo. Back.
Note 27: Some studies argue that conventional Keynesian energy demand-reduction modeling overstates the expected damage to economic growth rate (Cline, 1992: 151): under the tax-recycling scenario, negative effects to economic growth are negligible, and recycled energy taxes may have a positive employment effect (Standaert (1992), Andersen (1992), Barker et. al. (1995)). Most extensive economic studies of energy taxes and their economic effects are being undertaken by the Energy Modeling Forum at Stanford University. Back.
Note 28: All policy-oriented studies share similar opinion about the effects of taxes: see IPCC: Watson, Robert, Zinyowera, Marufu, and Moss, Richard (edt.) 1996. Technologies, Policies and Measures for Mitigating Climate Change. UNEP, WMO — this publication argues for the combination of voluntary agreements and energy taxes to achieve reduction in energy consumption.;
IEA. 1987. Energy Conservation in IEA Countries. OECD. Back.
Note 29: Skolnikoff, Eugene. 1997. Same Science, Differing Policies; the Saga of Global Climate Change. MIT Joint Program on the Science and Policy of Global Change. Cambridge: MIT. Back.
Note 30: The term carbon leakage means increased CO 2 emissions from developing countries as a result of their raising energy consumption, while the developed nations reduce energy use due to increased domestic price of fossil fuels after energy taxes (as suggested by the Kyoto protocol and previous climate control agreements, all developed (Annex I nations) economies should be subject to CO2 emissions regulations; (the Annex I nations include the OECD members with the exception of Mexico and Republic of Korea). There is an argument that decreased energy imports from the developed world would further depress the already low price of fossil fuels, allowing developing economies — which are not subject to climate control/energy taxing agreements — to consume more energy. A supporting argument states that energy taxes in developed economies would raise production costs in fossil fuel-intensive industries thus encouraging them to shift production to developing nations (over time).
This phenomenon is addressed in numerous recent literature on the effects of energy taxes and especially carbon taxes, and some models forecast the increase of carbon dioxide emissions from 2010 from the raising consumption in developing nations. For an example see Cain Polidano. 1997. The Impact of Climate Change Policies on Employment in the Coalmining Industry. ILO).
However, these models are based on the current international CO 2 abatement policy proposals which do not include developing economies and do not take into consideration:
-possible fuel substitution effects from more carbon-intensive coal and oil to less carbon-intensive natural gas — this substitution is plausible in the future due to the increasing price of oil extraction and the cheaper relative price of natural gas;
-the fact that energy importing developing nations, similar to their developed partners, have the economic incentive to conserve energy consumption due to high risks (transaction costs) resulting from dependency on foreign suppliers;
-the fact that increasing world demand for fossil fuels would increase the world price of scarce energy resources, creating an additional incentive to conserve energy and reduce CO 2 emissions.
To sum up the above arguments, the conclusions about the extent of possible carbon leakage are sensitive to the assumptions by which each forecasting model is stipulated. In particular, the closed-economy models and forecasts that do not consider technological development and fuel substitution effects appear to be overstating the damage of the CO 2 leakage. Back.
Note 31: Amano, Akihiro in Yochi Kaya and Keiichi Yokobori (edt.). 1997. Environment, Energy and Economy for Sustainability. Tokyo: UN University Press. Back.
Note 32: Aasness, Joregn et.al.(1996) Back.
Note 33: Walker, Charls et.al. 1997. Climate Change Policy, Risk Prioritization, and U.S. Economic Growth. Washington, D.C.: American Council for Capital Formation. Back.
Note 34: IMF. 1994. Taxation of Petroleum Products: Theory and Empirical Evidence. Washington, D.C.: IMF Working Paper. Back.
Note 35: This argument will not be controlled for in the empirical specification due to the difficulties with operatrionalizing the variable " before tax - after tax relative income level ": capturing differences in income and modeling the effects of energy taxes on income distribution cross-nationally is a topic for a separate dissertation. Back.
Note 36: See James A. Desveaux (199?). The author also addressed the issues of policy salience for public agencies, where salience is represented (theoretically) by budgetary commitments and by assigning key top-ranking administrative officials within a national government to oversee energy-saving policies. Back.
Note 37: The majority of country-specific case-studies combine policy analysis framework and neo-classical economics in reviewing national energy-saving policies. One example of such research is the cluster of country studies performed by the International Energy Agency (such as, Energy Policies: Poland (OECD — 1990), Energy Policies: Hungary (OECD — 1991), Energy Policies: Romania (OECD — 1993), so forth).
The most important OECD publications within this framework are the following: The Role of IEA Governments in Energy (1996), and Energy Policies of IEA Countries — 1996 Review (Compendium) Back.
Note 38: OECD. 1996. The Role of IEA Governments in Energy. The report contains detailed overview of national energy policies for all OECD countries. For the first time in social science energy-related research, this OECD-centered review provided a detailed description of national instruments of energy policy (such as: energy taxes, subsidies, tax credits, direct energy rationing and consumption standards, investments in R&D, and so forth). Back.
Note 39: Nigel Lucas. 1985. Western European Energy Policies. Oxford: Clarendon Press. Back.
Note 40: John Clark. 1990. The Political Economy of World Energy: A Twentieth Century Perspective. The University of North Carolina Press.
* Lucas provided an in-depth comparative description of European policies related to energy supply and use; the last chapter represents an attempt to explain specifics of energy policies by the differences in institutional design and in historical structures of energy markets. Clark gave detailed chronological review of national energy supply/demand patterns and corresponding energy security policies as they were being shaped by expanding interests of domestic production and strategical behavior of energy-exporters; the uniqueness of this study is in its detailed review of politics between energy suppliers and the analysis of energy issues in the developing world. Back.
Note 41: Where energy taxes are a policy aimed at reducing greenhouse gas emissions that directly affect the global climate change. Back.
Note 42: Kawashima, Yasuko. 1997. “A Comparative Analysis of the Decision-Making Process of Developed Countries Toward CO 2 Emissions Reducing Targets.” International Environmental Affairs, V. 9, n.2:95-126. Back.
Note 43: O’Riordan, Tim and J αger, Jill. 1996. Politics of Climate Change: A European Perspective. London, N.Y: Routledge. Back.
Note 44: There are about thrity detailed energy tax studies published in 1995-97 only; all of them address redistributive, macroeconomic and environmental economic effects of various energy taxes. Back.
Note 45: OECD/IEA. 1993. Taxing Energy: Why and How. Paris: OECD. Back.
Note 46: In the future empirical tests which will extend this research to the dissertation level I will employ a separate correlation analysis to demonstrate that a positive relationship exists between higher energy tax and energy efficiency (especially for transportation). This correlation is not demonstrated in this research project. Back.
Note 47: For the support of this logic see Bruno Frey in Crane, G. and Amawi, A. (1991): the author provides a detailed overview of a public choice framework. Back.
Note 48: Schneider, Friedrich and Frey, Bruno. 1988. “Politico-Economic Models of Macroeconomic Policy: A Review of the Empirical Evidence” in Willett, T.: Political Business Cycles. Duke University Press.
The literature suggests that influence of public pressures will be the most visible during the election times or during economic recessions, while more prosperous economic performance and between-election time are characterized by less visible pressure from the public, where ideological preferences of the governing elites become paramount (Ibid.; also see Zupan and Kalt (1996 ?)); some literature emphasizes that ideology exerts a greater influence on voting for general bills than it does for specific ones (Nelson and Silberberg (1987) — on the example of the trade policy for the U.S.). However, in my analysis both ideology and pressure of voters are constrained by the structure of decision-making. Thus, both in theory and in the empirical part, I do not make detailed distinctions between political business cycles due to economic performance (economic growth rate is controlled for in the model by inclusion of a respective variable for another hypothesis). Neither do I control for the proximity of the election time: my argument, instead, specifies that proximity of the election time should not matter for the countries of closed, corporatist-type decision-making. For non-corporatist economies, election cycles should exert more profound influence with respect to their effects on the governments’ ideology and/or voter pressure. If the empirical model demonstrates that it is necessary to control for the proximity of elections, the necessary variable will be included in the empirical specification. Back.
Note 49: Ikenberry, John, Lake, David and Mastanduno, Michael. 1988. “Introduction: Approaches to Explaining American Foreign Economic Policy.” In Ikenberry et.al. : The State and American Foreign Economic Policy. Cornell University Press. Back.
Note 50: On the theoretical basis, this approach is justified since recent literature encourages researchers to re-think the role of the state (including the structure of decision-making) and to incorporate it with the interest-group (societal) explanation of adopted policies — see the above chapter by Ikenberry et. al.. Back.
Note 51: By political institutions I mean the systems of political representation in national parliaments, ability of national courts to excersise judicial overview, constitutional rules related to political access to decision-making, etc.. Back.
Note 52: That were plentiful, for example, in France in XIX-XX centuries. Back.
Note 53: During the empirical analysis institutional structure variables will not vary across time, they vary only across nations. Back.
Note 54: See the literature on political business cycles: for example, Bruno Frey (1997). Back.
Note 55: See Appendix B for the complete graphical representation of the model. Back.
Note 56: See the introduction to the proposal for the justification to this argument. Back.
Note 57: In part, introduction of an energy tax during recessions is contrary to the basic macroeconomic argument that governments have to pursue expansionary policy to stabilize the growth rate. Further, in a competitive international environment a nation with a declining growth rate would be weary of raising energy taxes because it would hamper its import potential and may adversely affect the trade balance. On the other hand, positive stable economic growth allows a new tax and would even facilitate new taxes if it is necessary to reduce the import-dependency on fossil fuel. Back.
Note 58: Logically, the raise of oil prices should shift the production into labor-intensive sectors, which would be in conflict with the predictions of the deindustrialization arguments. However, I would like to refer to the existing literature on deindustrialization which argues: since the growth of labor productivity in the service sector is much slower in developed economies relative to the manufacturing sector (Rowthorn, R. and Ramaswamy, R. (1998)), the majority of labor in the deindustrialized country will be in the servicesector: this argument remains consistent with the logic of deindustrialization. Thus, skyrocketing oil prices will shift production into both capital intensive manufacturing and labor-intensive service sector. Since most of the labor is concentrated in the service sector, this structural shift is equivalent to the increased deindustrialization. This logic brings us to the direct connection: oil shocks and increased prices of oil may lead to the accelerated deindustrialization. Back.
Note 59: Energy taxes applied to transportation will not be considered in this work though they represent a part of the broader inquiry. Theoretically, transportation energy taxes are likely to be affected by a slightly different phenomena in addition to energy import dependency and economic growth rate: for example, the power of energy-intensive manufacturers is not a decisive factor in considering the reasons for cross-national tax variance. Further, transportation energy taxes are expected to be negatively affected by longer driving distances because longer driving distances imply higher demand for gasoline accompanied by larger costs of constructing public transportation infrastructure. At the same time, the political process of adopting higher energy taxes (or introducing a tax per se) in transportation may be eased by higher levels of environmental pollution, (especially in the nations with short driving distances and in the presence of public opinion concerned about environment). Back.
Note 60: Russel Dalton (1994). Back.
Note 61: Presuming that the government is not divided on the ideological bases. Back.
Note 62: In the future, a formal theoretical model will be derived. The current schematic representation in Appendix B allows to see how the variables interact and influence each other, but it remains the general graphic representation of a more precise mathematical specification. Back.
Note 63: As it is for Japan which, since the late 1960 — before the first oil shock— introduced a number of well-integrated and diverse energy saving policies in its manufacturing sector, managing to completely restructure the whole industry from energy-intensive to capital-intensive production. It is interesting to note that the Japanese chose not to set up high energy taxes since they believed that taxing energy consumption would be equivalent to taxing economic growth. This example would supports the argument that taxes are not necessarily the policy instrument one would chose for achieving higher energy savings, though they are the most universally applicable tool. Back.
Note 64: The term "subsidizing investments" in this case means providing lower-than-market-rate interest loans, distributing R&D grants and reducing the risks of technology-related investments by government-underwritten guarantees. Back.
Note 65: The existing empirical literature does not provide any guidance with respect to the most desirable time-lag for modeling the effects of economic growth. Back.
Note 66: Though the relationship between the level of energy tax and budget deficits is too obvious to be tested empirically, I include hypothesis #2 in the paper while I would like to emphasize that a great deal of economic research has previously examined this relationship. The budget deficits is purely a control variable. Back.
Note 67: In addition to the aforementioned decline in manufacturing output as a share of GDP. Back.
Note 68: Interactive effect of energy import dependency and deindustrialiation will not be tested in this paper; this task will be carried out in the future aggregate statistical analysis. Back.
Note 69: The political power of producer interest groups is extremely important for this analysis; due to the absence of data this variable is not operationalized at this stage of research. Back.
Note 70: The degree of organization of manufacturers may be very high, especially in small countries with short driving distances between enterprises and where production is concentrated in several key manufacturing branches. However, the number of manufacturing interests will be lower in the deindustrialized country relative to preponderant service-sector interests or/and environmental groups (where environmental groups are assumed to support energy taxes). Back.
Note 71: Export orientation argument matters because elected officials attempt to protect the industries generating a large proportion of export revenue. Such a precise wording of Hypothesis IV would allow me to economize on the efforts to acquire the data necessary for testing this thesis. Back.
Note 72: It is important to mention that the median voter theorem may be losing its predictive power, to an extent, when one considers nations with Proportional Representation systems of Parliaments (versus a two-party majority system in the United States). Within the institutional structure of proportional representation, the plurality of party platforms complicates predictions on what voter position may be considered "median." Nevertheless, since the preponderant number of voters tend to take non-radical positions in their policy choices, and because voter preferences can be carefully traced by public opinion polls, rational policy-makers will attempt to take moderately-centrist positions within the boundaries of their Party platforms in the PR system to the same extent as they would converge to the centrist positions in a two-party institutional system. In case politicians fail to adjust their policy proposals to meet median voter preferences, they immediately lose electoral support, as proven by the recent history of the German Green Party: after introducing a proposal to significantly raise the price of gasoline, the Party lost 2/3 of its supporters as the public showed strong opposition tot he measure. (The Economist, June 13-19, 1998). Back.
Note 73: The detalization of Hypothesis V in section B can be tested only in case-studies which are not reported in this paper. The overall test of public opinion influence on energy taxes remains outside of the scope of this study because of the current lack of public opinion data on the cross-sectional scale. Back.
Note 74: By definition, any political party of a Liberal ideology should not be in alliance with social democrats since the Liberal ideology advocates individuality, rationality and non-interference into the laissez-faire functioning of markets. In its classical form liberal ideology sees state’s function solely as a regulator and facilitator of social interactions and not an entity which can govern individual choices (which would be done by the imposition of energy consumption regulations). However, the growth of highly complex developed economies with inevitable interconnectedness of free-market operations with governmental regulations coincided with the emergence of a new ‘branch’ of classical liberalism — the so-called social liberalism (Eatwell & Wright (1993)), which maintains that in certain areas of public welfare domain intervention of governments is not only inevitable but desirable . For example, social liberals justified the arguments for the progressive taxation and pension schemes at the beginning of the XX century. The influence of Keynesian theory of mixed economy further advanced the notion that regulatory role of governments may be welcomes as long as this intervention combats unjustified privileges of particular economic groups or/and classes and helps to achieve a proper balance between the individual and social. Thus, it becomes clear that current Liberal ideology can accommodate and advance arguments supporting governmental control of economic ‘bads’, such as environmental pollution associated with the use of energy (global warming). Back.
Note 75: I do not consider parties based on the Marxist ideology, such as a Communist Party, to be a part of my hypothesis. Back.
Note 76: For the case of parsimony of my research, I omit very important distinctions between various groups within Green parties in particular and Green Social Movement in general. Specifically, I leave out the critical ideological distinction between the ecological and the general environmentalist branches of the movement. Ecologists appear to be a very radical political wing which does not mix very well with the Socialist ideology because ecologists advocate significant decentralization of governing in both political and economic spheres; consequently, they are opposing environmental governmental programs because they are perceived as inappropriate for the solution of regional community problems. In addition, ecologists believe that the notion of economic growth for the sake of achieving a larger GNP per capita is fundamentally wrong in the light of severe environmental degradation; as any public policy program is aimed at correcting market failures without re-considering the predominant economic growth paradigm, cooperation with socialist or/and liberal programs aimed at improving environmental situation is not justifiable. However, being such a radical political movement, ecologists attract less support of the electorate than a more moderate generally-based environmental branch of the movement and Green parties. For this reason, I make the critical assumption for my research: I consider only the Green ideology which admits the current economic paradigm and attempts to promote its values and policy issues without calling for the replacement of the existing political and economic structures. Thus, political cooperation between Socialist and Liberal parties and the Greens becomes possible. Back.
Note 77: Though political coalition between a Labor Party and a Green Party is not plausible because, in general, Labor parties would oppose environmental regulations since these regulations (including taxes) increase the costs of production and, under certain empirically unproved assumptions, may result in job losses. Back.
Note 78: Russel Dalton (1994). Back.
Note 79: This argument particularly apples to the Socialist and Labor Parties. Back.
Note 80: Nick Mabey et.al. (1997). Back.
Note 81: By adaptive or consensual decision-making structures I assume highly professional forum where representatives of industry, labor and governments mutually agree to pursue certain policies with shared responsibilities for implementation. Conventionally, corporatism stands to represent these structures (where the participation of labor representatives in the process of decision-making is implied). Corporatism is a contractual form of decision-making in policy arena, where state achieves domination and control over policy by securing favorable ‘contracts’ or exchanges with producer groups through bargaining (some powerful consumer groups or other interest groups may also be included in the policy negotiation process). In this environment producer group leaders, in turn, secure compliance of their members to terms of ‘contact’ by various means (Peter Williamson (1985)). "Corporatist institutional structures are so established as to generate a high degree of voluntary consent to authoritative decisions..." (Peter Williamson, p. 11). Back.
Note 82: Lijphart (1975) Back.
Note 83: Decision-making literature is consistent with this assumption: see W. Riker (1964?) on the size of the minimum winning coalition. Back.
Note 84: Lyle Scruggs (1998). Back.
Note 85: Industrial taxes are chosen for heavy fuel oil because light fuel oil is frequently used for residential heating; data are plentiful and available for natural gas, light fuel oil and coal — however, all these carbon-based fuels are also used for residential heating, and the available data do not discriminate between residential and commercial use. One exception are highly specified and separated by the usage type data from the EUROSTATA, but the EUROSTATA data do not include some Nordic countries as well as Australia, Canada, New Zealand, Turkey, and the U.S. Thus, for the reasons of data constraints I have to exclude several types of fossil fuel taxes from the analysis. Back.
Note 86: The paper does not consider transportation energy tax; it will be addressed in the larger dissertation project with the dependent variable Yit = Energy tax in transportation (taxes on leaded and unleaded premium gasoline, excluding VAT and road maintenance tax). Back.
Note 87: In a separate equation this variable will be interacted with a dummy for the presence of a Green Party and consequent effects of this interaction. For the purposes of testing this interaction, the specification will remain in its general form, as given below. Back.
Note 88: At the current stage of research the data are entered for pulp & paper, chemical and iron & steel industries. Identification of other groups of energy-intensive industries continues. Back.
Note 89: Some data on corporatism are available from Lipjhart and Crepaz (1991): Corporatism and Consensus Democracy in Eighteen Countries: Conceptual and Empirical Linkages. British Journal of Political Science, 21: 235-256.
The data discriminates countries according to the degree of their corporatist attributes on the cumulative basis for the 1960-1980s; no variance is available across the time. Since the corporatist data are not appropriate for the pooled data regression analysis, which is suggested below, for the overall model (which would include all the predictors), the effect of the corporatist variable has to be examined by using the ANOVA technique. However, at the current stage of research corporatist variable will be tested by the means of an aggregate bivariate regression on the heavy fuel oil tax. Back.
Note 90: Industrial taxes are chosen for heavy fuel oil because light fuel oil is frequently used for residential heating; data are plentiful and available for natural gas, light fuel oil and coal — however, all these carbon-based fuels are also used for residential heating, and the available data do not discriminate between residential and commercial use. One exception are highly specified and separated by the usage type data from the EUROSTATA, but the EUROSTATA data do not include some Nordic countries as well as Australia, Canada, New Zealand, Turkey, and the U.S. Thus, for the reasons of data constraints I have to exclude several types of fossil fuel taxes from the analysis. In preliminary runs, I will compare the results for the light fuel oil tax and heavy fuel oil tax; in the final analysis to follow this paper I will concentrate on the heavy fuel oil solely Back.
Note 91: The paper does not testthe transportation energy tax; it will be addressed in the larger dissertation project with the dependent variable Yit = Energy tax in transportation (taxes on leaded and unleaded premium gasoline, excluding VAT and road maintenance tax). Back.
Note 92: In the reported runs national economic growth rate was not entered in the equation because of the limited number of observations in each tested sample: though the data for economic growth are available, my theory states that economic growth should be interacted with energy import dependency to obtain reliable results. Interactive variables increase the problem of multicollinearity already present in small samples, adversely impacting statistical results. Interactive variable will be employed at a later stage of analysis where additional data would permit introduction of an interactive variable. Back.
Note 93: Some data on corporatism are available from Lipjhart and Crepaz (1991): Corporatism and Consensus Democracy in Eighteen Countries: Conceptual and Empirical Linkages. British Journal of Political Science, 21: 235-256.
The data discriminates countries according to the degree of their corporatist attributes on the cumulative basis for the 1960-1980s; no variance is available across the time. These corporatist data is appropriate for the pooled (aggregate) analysis. The effect of the corporatist variable on the country-by country bases will be examined later by using the ANOVA technique. Back.
Note 94: The list of data availability is given in Appendix B. Back.
Note 95: Missing observations on the dependent variable would lead to highly biased non-linear statistical estimates which cannot be reported as findings. Back.
Note 96: A note of caution: as energy taxes were not adjusted for the effects of inflation (due to the lack of time), some tax level fluctuation effects remain to be examined before any definite decision on the appropriate functional forms can be made. However, taking into the consideration that no hyper-inflation was observed in the majority of the OECD nations in the period of 1973-1995, I assume that my tentative conclusion on the appropriateness of the log and linear specifications should remain valid after inflation adjustments. Back.
Note 97: The aggregate model did not contain the budget deficit variable due to the lack of time to complete the currency exchange rate adjustments for each individual nation. Back.
Note 98: The variable "year" is introduced into the equation to capture the time-effects: for example, energy taxes increase over time partially due to the effects of the inflation. Since the tax levels entered into the regression were not adjusted for inflation, introduction of "year" mitigates statistical error due to various time-effects. Back.
Note 99: These are the proxies for deindustrialization in selected energy-intensive sectors. Back.
Note 100: I reiterate the corporatist data comes from Lijphart and Crepas, 1991. It constitutes an index of corporatism constructed by the authors on the basis of combining several attributes of corporatism on the scale between -1.5 to 2.0. This measure, in its turn, was derived from the previous indexes suggested by Wilensky (1981), Roland Czada (1983), so on (see Arend Lijphart and Markus Crepaz. 1991. "Corporatism and Consensus Democracy in Eighteen Countries: Conceptual and Empirical Linkages." British Journal of Political Science. V. 21, part 2, April 1991: p.p.235-247 Back.
Note 101: The problem can be corrected by entering the additional tax data for 1973-1979 or by finding a more up-to-date measure of corporatism (which is currently being done). Back.
Note 102: The first column value for each nation in the table represents a heavy fuel oil tax, converted into the US$ for the purpose of comparison (in million dollars). The second column stands for the index of corporatism. Not all countries in the sample are represented. Back.
Note 103: Exactly because of the plausible effects of the other predictors included in the regression models. Back.
Note 104: Only the most significant country results are reported. Back.
Note 105: The statistical tests for autocorrelation are to follow in the extended research. Back.
Note 106: Statistical significance of individual countries in the overall model will be tested by the inclusion of the country-specific dummy-variables in the final empirical specifications. Back.
Note 107: See Appendix C for the results of the pre-tests for France, Germany and Japan. While the pre-tests for Germany and Japan yield consistently weak results, the case of France requires an in-depth case-study because of the highly significant model fit for the Light fuel oil tax. Back.
Note 108: The public opinion data reflecting social and political attitudes for the non-European nations are available through national election surveys (Canada to serve as an example) or specialized public opinion polls on energy problems (U.S.). However, the biggest problem with aggregating the responses from these sources into the pooled data-set lies in the fact that a large number of questions asked is not comparable to the Euro-Barometer survey questions. For this reason, for the purposes of pre-test, I am concentrating solely on the public opinion frequencies reported for the European Union from 1982 until 1995. The series of correlation analyses reported below will use only the data for 1986-1988. Back.
Note 109: The statistical frequencies analysis uses option #5 only: " To save energy Back.
Note 110: Only the suggested solution # 3 was used in correlation analysis below. Back.
Note 111: Option # 1 was used in frequencies’ analysis. Back.
Note 112: The combination of option # 1 and #2 was used for computing the correlations given below. Back.
Note 113: See the section on theoretical assumptions and hypotheses to refresh the logic. Back.
Note 114: Though none of the theories considered in this study are new per se, their application to energy policies can be considered an innovation. Back.
Note 115: With the exception of highly significant model fit results for Light Fuel Oil tax in France. Back.
Email this citation