Cheon, A, and J Urpelainen. 2014. “Escaping Oil's Stranglehold: When Do States Invest in Energy Security?.” Journal of Conflict Resolution, February. doi:10.1177/0022002713520529.
The article aims to explain why states invest in energy research and innovation. The authors also declare their broader goal
... to return energy security to the core of international relations theory ...
... although the study of international politics [...] now encompasses a wide variety of concerns from human rights to environmental degradation, energy issues remain undertheorized (p. 24) It proposes a game-theoretical model to explain a country's decision to invest in a "costly energy policy such as R&D funding for new energy technologies". the model suggests that any given country A invests in energy security when (1) the international energy market is prone to supply disruptions and (2) country A already faces a dangerous external security environment. (p.6)
The model considers costs and benefits of an energy policy which enhances energy security in case of a probabilistic conflict with another country. Marginal benefits of energy policy in the model are increasing with the probability of a conflict and supply constraint (the concentration of oil production in the Middle East). This proposition is empirically tested using a data set that covers 27 IEA/OECD members between 1975 and 2007.
The dependent variable is the public R&D and the independent variables are concentration of oil production in the Middle East and country's military expenditure relative to total GDP. ECM (error correction models) to deal with time-series of dependent and independent variables. The study uses a large number of control variables such as : GDP per capita, oil price, the share of electricity generated from hydro and nuclear sources, energy intensity, executive constraints, and right-wing party prevalence. The study also considers country's "fixed effects" and adjusts for the overall decline in (total) R&D over time.
The findings confirm the propositions that energy R&D expenditures positively correlate with the concentration of oil production in Middle East and that this correlation is higher for countries with higher military expenditures. Such control variables as energy intensity, economic wealth and oil prices also tend to increase energy R&D. The results are robust when only R&D related to oil substitutes are analyzed.
The theory performs very well in explaining huge energy R&D expenditures in the US prior to the end of the Cold War and their subsequent rapid decline not reversed even with high oil prices or the war in Iraq. Likewise the theory performs well in explaining why Japan's high level of investment in energy R&D remains on the very high level independently of the Cold War and variations in the Middle East concentration of production. The authors in particular suggest that it can be explained by Japan's willingness to maintain dominance in high energy technologies.
The article demonstrates the power of an excellent statistical analysis which can provide a simple yet powerful explanation for an energy policy phenomena. There are, however, three points that could be clarified. The first is about energy security (the independent variable). The second is about energy R&D expenditures (the dependent variable). The third is about possible relations between these variables.
Independent variable: energy security - how to measure it?
To begin with, the paper would benefit from a more systematic view of energy security, its main independent variable. The paper assumes that energy (oil) security is primarily linked to a risk of a military conflict. This is because (a) oil provides fuel to the military and (b) external oil supply (imports) can be more easily disrupted in case of such a conflict.
However, security is a low risk of damage to acquired values (Baldwin 1997). Energy security is low vulnerability of vital energy systems which support such such values, be it political independence, territorial integrity, social stability or economic welfare. Thus, security in general and energy security in particular do not necessarily involve a risk of military conflict or presence or perception of military tensions.
Historically, oil security was indeed closely tied to military tensions. However, at present oil security is also tied to civilian uses of energy. Disruption of oil supplies would affect vital functions of modern society and can destabilize even societies that do not have large militaries. Societies that import all or most of their oil will be more vulnerable to such issues. Moreover, countries may experience energy security problems associated with other energy sectors, e.g. nuclear gas (as many countries in Europe). Once again, natural gas is not necessarily linked to military conflicts, but its supplies disruptions or threats of such disruptions may present a significant energy security challenge. Such problems may be sufficiently severe to influence energy R&D.
Japan may be a case in question. Japan does not have significant military, but it does have serious energy security problems. These first become apparent in the 1960s and the 1970s when the massive nuclear energy program was established to replace oil in the power sector. Energy security situation of Japan deteriorated in the last decade because of worsening relations with China and closure of nuclear power plants as a result of Fukushima. None of this was reflected in either military expenditures or the concentration of oil production in the Middle East. The authors rightly note that Japan's energy R&D is consistently very high, which does not need to be explained away by reference to its intention to retain a dominant position in the market of high energy technologies (Japan does not have such a position anyway).
It is thus sufficiently clear why energy (and oil) dependence of a particular country is not considered as a potential independent variable explaining energy R&D. For example, the declining US interest in energy R&D can be explained not only by the end of the Cold War but also by the increasing hope that its own (and Canadian) unconventional fossil resources will eventually reduce its import dependence. According to many estimates (see our study on major economies), the US is posed to become energy independent within the next 2-3 decades and this factor should be known to policy makers, but it is not taken into account by the article.
Independent variable - energy security - are military expenditures a good proxy?
Using military expenditures as a proxy for country's insecurity is an elegant move, but in some cases one can't stop wondering whether it is appropriate. According to author's table (A2 in the online Appendix) Australia (3.00%), Norway (2.79%), and the Netherlands (2.94%) are less secure than Slovakia (1.88%), Hungary (2.06%) and Ireland (1.18%), to say nothing about Japan (0.83%). This is counterintuitive because the first three countries are major energy producers as well as large and well protected economies, the last four are extremely energy dependent and vulnerable. Such difference become more apparent when one moves beyond OECD, e.g. military expenditure of extremely vulnerable Baltic countries are very small.
Dependent variable - energy R&D, but is this the best measure?
With respect to the independent variable, energy R&D expenditures would indeed reflect how a country values its energy (security). There are of course other considerations which can explain energy R&D (e.g. commercial interest already noted in case of Japan). It would be interesting to test a country's total R&D as a potential control variable. It may be that those countries which spend more on energy R&D are just more prone to invest public money in research and development in general.
More importantly, it is not clear whether this variable is best to characterize the intensity of a country's energy policy. Energy subsidies, taxes, efficiency standards and quotas (e.g. import quotas), stimulus to domestic exploration and drilling all can promote energy security at least as effectively as energy R&D.
Indeed, the authors themselves assert that public R&D expenditure does not change the distribution of [oil] production in the long run. But what are the factors that do affect such a distribution? Forces of nature? Objective evolution of technologies? Private interests of oil companies? Or perhaps state policies which are not captured in the R&D expenditures?
Relationship between the independent and dependent variables - alternative explanations
I can think of at least one explanation why energy R&D investments would correlate with military expenditures. Both are type of public sector investments which presume certain role that a country would play in the world. Small countries are likely to rely on major powers for both military protection and R&D spillovers and thus would organize their defense and their energy policies in a different way.
In summary, the paper proposes a very elegant method of explaining energy R&D in OECD countries, but it would benefit from more systematically defining its independent variable (energy security) and dependent variable (energy policy).