# Introduction he problem of climate change is of a global nature. As long as economic growth is not disentangled from an increase in greenhouse gas (GHG) emissions, the problem of climate change is likely to increase. One common argument is that the global problem encourages free-riding and reduces national incentives to contribute to climate change mitigation policies. Thus, international policy coordination is an attempt to reduce the related problems. One example of international cooperation aiming to reduce coordination problems is the Kyoto-Protocol (KP). Even though the KP was an attempt to make countries act cooperatively, strategic behavior could be observed at the ratification stage (decision to ratify or to free-ride on the agreement) as well as the implementation stage (over or underinvestment to fulfill the requirements agreed by ratification). Differences in national cost structures combined with strategic interaction between countries makes coordination difficult. A recent example was the negotiation for a follow-up agreement to the KP which took place in December 2009 in Copenhagen (e. g. Macintosh; 2010; Nicoll et al.; 2010). Despite the global nature of the problem, some governments did start to restructure their energy policies. It seems that they take the climate change problem seriously (e. g. the German government by supporting diffusion of green technologies (GTs) 1 ). Interestingly, it turns out that the same countries argue forcefully in favor of more strict environmental standards on the international platform. The fact that some countries invest relatively more than others in the abatement of climate change is somehow counterintuitive if we apply the general wisdom that free-riding of particular countries negatively affects the international competitiveness of non-freeriding-countries. Investment costs related to GTs seem to be a burden that increases the costs of energy consumption within a country. It is, therefore, an interesting question why some countries are more motivated than others in implementing policy measures that have a seemingly positive impact on the problem of global warming and promote actively high environmental standards at the international level instead of free-riding themselves. We argue that the initiative for structural change at the national level can be an outcome of international environmental agreements (IEAs) aimed at reducing problems related to climate change. However, as we also observe free-riding, not all countries are able to restructure their energy policy. Differences in political systems as well as cultural aspects might be a reason for the observed heterogeneity. In contrast to the common view, the main argument of our paper is that free-riding by some countries may encourage other countries to increase investment in abatement measures instead of reducing it. Our arguments are based on a political economy framework in combination with international trade policy. The paper is organized as follows. In section 2, we briefly discuss the costs of global climate change and the global attempt to solve the problem. In section 3, we focus on the particular German case. Different political economy explanations that help to explain the observed heterogeneity among countries follow in section 4. In section 5, we use a simple theoretical framework to explain a country's solo run to provide a global public good in climate policy. Our political economy reasoning is empirically assessed with the help of a negbin model in section 6 where we use the patent applications of German green technology firms as a proxy for their expectations about future export sales. Conclusions round off the paper. # II. # Climate Change Problem and Climate Policy There are studies trying to make predictions about the costs related to climate change (e. g. Latif; 2010; Stern; 2007). Without policy response, costs of changes in temperature are expected to increase at a level of from 5-20 percent of global annual gross domestic product (GDP). These costs can be reduced by climate policies. However, there are substantial differences between regions (cf. Hope; 2006; Mendelsohn et al.; 2000; Nordhaus and Boyer; 2003; Nordhaus and Yang; 1996; Tol; 2002). The allocation of costs has further an intertemporal dimension. Estimates came to the result that it is "cheaper" to react today than in the near future because doing nothing will increase costs (Kemfert; 2005). Another problem is related to non-cooperative behavior of particular countries and changes in relative prices. As stated by Sinn (2008), it may be the case that the abatement of industrialized countries does not affect the speed of global warming as initially intended because the reduced demand for energy by some industrialized countries simply lowers world market prices and increases the demand for energy by those countries which do not intervene to reduce energy consumption (the so-called "rebound effect"). Problems to coordinate international policies lead Lomborg (2006) to suggestions of alternatives to the option of cutting GHG emissions. It can be seen that costs related to climate change depend strongly on the policy measures implemented. Country specific costs can be reduced significantly if there is international cooperation. However, free-riding on the international level increases country specific costs of climate abatement policies. Based on these arguments, global environmental problems constitute an international prisoners' dilemma. Climate protection has the characteristics described as "tragedy of the commons" (Hardin; 1968) and countries have to cooperate to find solutions for the common pool problem (e. g. Ostrom; 1990). The Kyoto Protocol is an attempt to coordinate international policies. By signing the KP countries agreed to a reduction in the emission of GHGs to a specified level measured in percentages of the base year 1990. Between 2008 and 2012 countries are supposed to reduce the average emission of GHG by about 5.2 percent of the 1990 reference-level. Europe agreed to reduce the emissions of GHG by 8 percent in comparison to the emissions of 1990. The KP was coupled with the condition that at least 55 member states, which altogether produce more than 55 percent of the global emissions of ???? 2 , have to ratify the protocol before it can enter into force (Kyoto Protocol; 1998, p. 19). 2 # III. # Climate Policy in Germany The 55 percent rule was fulfilled when Russia ratified the KP in November 2004. Therefore, the countries and other governmental entities have ratified the KP. The United States, the largest single emitter of GHG signed but did not ratify the KP at the national level. Once international treaties are negotiated, countries have to implement policies to fulfill what has been agreed. The alternative is to free-ride on the international agreement. Germany has chosen a mixed strategy to reduce the emission of GHG. On the one hand, there is the market solution (implemented in Europe) of trade with certificates related to GHG emissions. 3 Germany has the target to reduce emissions by about 21 percent in 2012 compared to 1990 baseline emissions. On the other hand, the government is using incentives to encourage the application of particular (allegedly) climate friendly technologies. For instance, the former "red-green" government coalition 4 From a theoretical point of view most GTs available, even today, are costly alternatives compared to conventional energy technologies (wind turns out to be an exception). The political argument for investment into GTs is to foster the development of GTs and to reduce global warming (EEG; 2009, section 1, purpose). There is an obvious connection between the problem of climate change and industrial policy, as feed-in tariffs are set on different levels what allows for the diffusion of more cost-intensive GTs. The range of feed-in tariffs in 2003 was from 6.5 Cent/KWh for electricity produced by using water and biogas up to 51.62 Cent/KWh for electricity produced with solar. passed the so-called "Renewable Energy Sources Act" (EEG) to support renewable energies by the use of technology specific feed-in tariffs. In what follows, we will focus on the promotion of GTs and its connection to climate change as this is an interesting case from a political economy perspective. # figure 3 This has led to a remarkable diffusion of GTs (compare and figure 4, Appendix, page 16). From 2000 to 2011 electricity produced with renewable energies increased from 6.4% The so-called 55 percent rule has important implications: It gives countries the opportunity to free-ride without nullifying the whole agreement. The free-rider problem is, thus, mitigated and it is more likely that the agreement will be implemented. # 3 The importance of defined property rights as an efficient solution for the externality problem has been highlighted by Coase's (1960) seminal paper. For theoretical considerations compare Baumol and Oates (1988). to 17% (BMU; 2011, p. 12). This is puzzling and needs an additional explanation. Another observation, that can be made, is that the German government takes an active role in KP came into force in February 2005. In 2011 188 international environmental negotiations. First of all, it can be seen that the German government established one of the highest GHG emission reduction targets within Europe. Second, at the G8 summit at Heiligendamm (Germany) in June 2007, the German government tried to use its role as an agenda setter to actively promote climate policies (e. g. Freytag and Wangler; 2011). There is further evidence that Germany as a member of the European Union is one of the leading industrial countries with respect to climate change and renewable energy policies (e. g. Weidner and Mez; 2008). With the recent event of the nuclear catastrophe in Fukushima (Japan) the current energy policy in Germany changed even more in favor of renewable energies. According to a new energy concept by the German government it is the aim to reduce GHG emissions until 2020 by about 40%, until 2030 by about 55%, until 2040 by about 70% and until 2050 by about 80-95% compared to 1990 baseline emissions (BMWI; 2011, p. 5). These GHG reduction targets are ambitious and are also surprising due to the fact that international policy coordination is confronted with difficulties. Interestingly, the German government tries to foster actively the export of green technologies. For this purpose, in 2002 the German Bundestag nominated the German Energy Agency to be responsible to promote actively the export of GTs. Under the label "Renewable Made in Germany" there is a whole concept of marketing for the related products and there is active support to create international networks, to create knowledge about potential export markets of GTs and to provide active services facilitating foreign market entrance (e.g. by active lobbying). The support by the German Energy Agency is not limited to German companies alone, criteria for support is in close connection to the job creation in the GT sector within Germany. 6 The findings of the previous sections can be summarized as follows: With respect to the climate change problem, there is the need for international policy coordination. This coordination, however, turns out to be difficult and perceived as a failure. If we follow this line of arguments, it is surprising that an industrialized country like Germany takes a leading position in climate policies despite the fact that coordination failures increase country specific marginal abatement costs. It seems that politicians in Germany have a long term time horizon by actively promoting the diffusion of GTs as this policy (if at all) will only have in the long run a positive impact on the world climate. This 6 To get more insights see DENA (2011), p. 14. behavior is somehow puzzling as the general wisdom suggests that politicians are rather short term oriented. # IV. Political Economy Considerations a) Behavioral Assumption From a political economy point of view politicians are considered as rational actors that are mainly concerned about re-election (Schumpeter; 1987b). Incentives to foster structural change in the energy sector are rather low as this is costly and reduces the political influence of conventional energy producing companies. This helps to understand the difficulties in particular countries to invest into climate abatement policies. Due to the free-riding of other industrialized countries, we also should not expect that politicians in Germany seriously support diffusion of GTs. Obviously, this is not the case. As stated in the previous section there was an observable diffusion of GTs and in the future they will be of increasing importance. The aim is to achieve a share of 35% by 2020, in 2030 the share shall be 50% and in 2050 the share of renewable energies of cross electricity consumption shall achieve 80% (BMWI; 2011, p. 5). Theory suggests huge difficulties for policies aiming to foster structural change in the energy system. Today the support for most GTs is still not profitable under current relative prices. The described empirical observation is therefore counterintuitive and needs an additional explanation. A standard political economy explanation refers to the median voter model (Black; 1948; Downs; 1957). The government follows the median voters' preferences which are increasingly directed to protect the climate. Therefore, the government invests relatively more than other countries into climate protection as this is in line with median voter preferences within the country. The likelihood of such a political preference for early investment into abatement policies is doubtful, due to international free-riding behavior and the relatively high investment costs that are related to GTs. If we take into account that international preferences are characterized through a game with national elections on a first stage and the delegation of representatives to international levels on a second stage, there is still some explanatory power related to the median voter theorem. The described model is known as strategic delegation model of IEA formation. In the underlying game voters delegate their decision power to agents representing the country at international negotiation tables. The agents, usually the government, then have the power to negotiate the terms and conditions of an international agreement. This setting is generally applied as a two-stage game within a two country setting. At the first stage voters (using majority rule) elect their preferred politician who, at the second stage, is responsible to negotiate the international treaty. Foreign election outcomes are taken as given for the election on the national level. This allows voters to select the candidate that represents most favorable their position in the international policy game. One basic feature of the underlying game is that it is rational for voters to elect a politician with different preferences than their own; with the result that international outcomes deviate from the median voter's 'true' preferences. It is rational for voters to strategically misrepresent individual preferences if the election outcome gives an advantage at international policy negotiations (see Persson and Tabellini; 2000, Chapter 12). There are different economic phenomena such as international tax policies and the provision of transboundary public goods to which the strategic delegation approach has been applied (e. g. et al.; 1994; Kempf and Rossignol; 2010; Persson and Tabellini; 1992; Roelfsema; 2007; Segendorff; 1998). Segendorff (1998) finds that voters will choose politicians that have stronger preferences for the private good compared to themselves. The idea behind is that this lowers the reservation utility and thus, weakens the bargaining position of the other agents participating in negotiations. They find a gap between cost and actual willingness to pay in particular for the USA what might serve as an explanation for the withdrawal of the USA from the Kyoto agreement. Buchholz et al. (2005) study the effect of strategic delegation with a focus on IEAs. They find that in the equilibrium the median voter in each country chooses a government that is less concerned about environmental problems compared to himself, with the intuition that this improves a country's position at the international bargaining stage. The results described allow to explain why investment into climate protection might be too low. This is different from the described German position within the international climate policy-game. However, models of strategic delegation are also helpful to explain why countries might support rather strict environmental standards on international levels. Roelfsema (2007) studies the effects of strategic voting within a two country setting and non-cooperative behavior with a focus on the Kyoto protocol. Two equilibria are possible. One where politicians are less concerned about the environment than the median voter and one in which politicians have a higher preference for the environment compared to the median voter. There will either be a 'race to the bottom' or a 'race to the top', depending on the strength of the environmental preferences of the median voter. Models of strategic delegation can help to explain why politicians in some particular countries are highly engaged for environmental protection also at international levels. In Germany it seems that strategic delegation leads to high preferences for international climate standards. Median voters' preferences might be different from those of the delegates. Strategic delegation allows delegates to promote long term environmental targets as long as in the short run partial gains at the regional level exist, like short term employment in the GT industries (generating directly observable growth in the GT industry). Politicians are not directly sanctioned by the voters if they convincingly argue that diffusion of GTs is related to future export sales. The job creation in a particular GT industry (Blanco and Rodrigues; 2009; Hillebrand et al.; 2005; Lehr et al.; 2008; Lund; 2009) very likely creates stable (or increasing) transfer flows to the particular GT industries (lock-in effect). Politicians can maximize their political support function (in the short run) with this job increase and at the same time justify these transfers by expected future payoffs (e. g. future exports) related to the investment. This relationship between short term employment and long term export expectations might be the main reason for the observable diffusion of GTs within Germany and the strong preference for high international standards to protect the climate. The described policy will only pay off in the future if other countries also adapt to the high German standards. This explains why the German government has to support a rather strict environmental position on international meetings. The aim is to prepare future export markets in order to make the (over) investment into GTs profitable. Thus, for investment into GTs it mainly holds in a one-shot game that free-riding behavior of other countries is problematic for the domestic government and its climate abatement targets. From a dynamic perspective, this free-riding behavior in the short run may further encourage governments for ambitious unilateral political action, as long as it can be expected that other countries over time have to increase their environmental standards, as well. Such an increase seems to be likely in the context of climate change with its long term time horizon. What still has to be answered is the reason for the observed heterogeneity between countries with respect to be able to start investment into GTs. One explanation might be that governments act ideology driven or that under particular circumstances they have the opportunity to implement partisan policies. As climate change requires structural change within the economic system, some governments are not able to overcome the resistance of the interest groups within the system in the short run. These governments are obviously forced to free-ride on international environmental agreements. Over time the government composition might change and policy reforms might be established. Especially partisan politics seem to be a good explanation why the GT sector in Germany could initially become possible. There was a kind of window of opportunity when the green party for the first time Strategic Trade Policy as Response to Climate Change? became part of the German government under the so called red-green coalition as the green party could express its preferences for climate friendly policies (from 1998 to 2005). # b) Strategic Interaction There are existing theoretical papers that use game theory to evaluate strategic interaction between countries in the case of environmental policy (e. In this paper we focus on the German case and try to explain the political calculus behind the climate policy of the German government. Without any policy induced demand for a certain GT j, there is no intersection between supply and demand and marginal production costs are assumed to be constant. Diffusion of GTs is not observable. Diffusion is related to the regulations within the energy system allowing GTs to diffuse. We further assume learning curve effects, thus, the cost curve has a negative slope (compare Madsen et al.; 2005; Nemet; To start with, we assume that only one countryin our framework the home country (H) -implements measures that allow for diffusion of GTs. The measure taken is a policy induced demand for renewable energy at a level that allows the GT industry to establish. There is no international trade in GTs as the foreign country (F) free-rides on climate change mitigation policies. The resulting effect is a comparative advantage for the national GT industry (first mover advantage) as it moves rightwards on the learning curve. Concentrating on the domestic consequences of supporting renewable energy beyond the market demand for GTs (under the assumption that F does not support the GT sector), the balance is negative. Because conventional substitutes for producing energy exist, the creation of the GT sector generates costs in H that can be translated into a reduction in the level of national GDP. In addition to the environmental regulation, these costs reduce the initial comparative advantages of other industries (that use energy as input and compete in international markets). Additional pressure comes from the short run free-riding strategy in country F. In other words: i H n H Y Y < 1 ( 1 n H Y stands for "new GDP" with policy induced demand for GTs and without exports, the latter for the GDP without policy induced demand for GTs). We get further insights when comparing both countries. Without any support being given to the GT sector the initial GDP of both countries is the same. This means that i F i H Y Y = ( i F Y stands for the GDP without any support for the GT industries in F ). H is the first who implements GTs. 8 i F n H Y Y < 1 If we compare the GDP levels of both countries after H has decided to implement a GT sector, in the short run we have the case that . This line of arguments is well known and can directly be applied as an explanation for the free-riding problem, resulting in an international prisoners' dilemma. We now turn to the open economy. Because we assume that H enters the market of GTs before F, it moves rightward on the cost curve. Hence, considering exports does lead to a change in the results. If F decides later to enter the GT market and starts its own production, it has to start at a higher point on the cost curve. Figure 2 shows that pr F c are expected to be higher than pr H c . The support for a certain GT industry in F could have different reasons. One striking argument is that knowledge creation about the problem of climate change makes free-riding over time more and more difficult to be maintained. Changes in F' s policy can be supported by international attempts of H' s government 8 We argue that this is due to the political process. Apart from this, both countries can be assumed to be symmetric. for the demand for a certain GT j with policy induced demand. We refer to j pid as diffusion of GTs that results from domestic political intervention. What we have in mind can be interpreted as command and control policies with characteristics similar to those of the EEG. Theoretically, however, j pid could also represent diffusion of GTs as a result of market-based instruments such as tradable certificates or subsidies. In any case, the parameter is exogenous and can be directly influenced by national legislation. It is highly sensible to use a framework of strategic trade policy to explain why H's government has strong incentives to support high environmental standards on an international platform. The first mover advantage stems from the chance to increase market power within markets with incomplete competition (e. g. Brander and Spencer; 1985). Thus, political support (or more generally a policy induced demand) can help the industry to exploit the rents that might be related to early market entrance. 9 Different scenarios are plausible. For instance, one could expect a scenario in which F decides in a later phase than H to implement a transfer scheme per unit of energy produced (e. g. a FIT) by a particular GT (what is captured by j pid ). We assume that producers located in F are also able to produce GTs, but they operate on a higher marginal cost curve. This allows the GT sector in H to enter the market in F as a Stackelberg leader (scenario 1). Alternatively, high environmental standards might be the result of supranational negotiations (scenario 2). The high environmental standards increase the demand for GTs indirectly. Results for plausible other scenarios are summarized in table 4 (Appendix, page 23). Based on the previous reasoning, it becomes obvious that politicians in H have strong incentives to (1) make use of industrial policy to support the national GT industry even though other countries free-ride, (2) to support high environmental standards at an international level and (3) to cooperate with the GT industry on international interests. We now look at the expectations related to exports of GTs (scenario 1 and scenario 2). The expected price-demand function is given by Note that we do not assume a monopolistic market in the GT sector in H. What we assume is that all GT industries in H are supposed to be symmetric and able to supply GTs at the same marginal costs and therefore, e H j ? represents aggregated profits. Politicians and representatives of the different GT industries in H are aware of their advantage in international competitiveness. Therefore, both groups expect to benefit from an increase in environmental standards in F. Obviously, gains are related to the export of GTs. We then get as an expected outcome that can be interpreted as potential extra gains for the GT industry in H (if F was free-riding in the short run and decides later to support diffusion of GTs without discriminating against H' s industry). This is one reason why there might be a strong interest in H to invest heavily in the diffusion of GTs and "to lobby" internationally for high environmental standards internationally. How does this result translate into H' s changes in GDP ( Y ) 10 ? We can substitute the calculated values . The model implies that exports of GTs can generate welfare gains which enter positively into the GDP of H compared to the first situation which 9 Only if countries subsidize their industries in order to be the first to enter into the market, a prisoners' dilemma is present and both countries would be better off without the subsidy (Brander and Spencer; 1985, p. 95) 10 Note that the welfare analysis is limited to the GDP and, therefore, ignores welfare gains due to the reduction of GHGs. In our study benefits of climate change protection are not taken into account. A cost-benefit analysis therefore would come to very different results. # Global Journal of Human Social Science Volume XIII Issue IV Version I Finally, just how realistic the expectation is that there is a long run net benefit for country H from subsidizing its GTs, has to be discussed. As table 4 (Appendix 4, page 23) shows, "only" in scenario 3, case (a), does the first mover advantage not lead to higher exports because of direct support in F for the GTs there. However, as * e F q is also bigger than zero, one can expect that the industry in F also gains. This implies less resistance in F. 12 1. GT industry j expects higher profits, All other scenarios are characterized by increasing exports. Thus, there are, at least, three political economy arguments that politicians in H use in support of the GTs, strategically: 2. national governments can reduce the political costs caused by the policy induced demand for GTs, 3. The GT industry in F can also generate profits which is important to reduce resistance against international standards. The intuition behind the framework presented is to analyze political incentives which we now try to incorporate into an econometric model. # V. # Econometric Model To test our theoretical argument, we propose an econometric model. With this model, we try to assess empirically whether the alleged strategy of the government and the GT interest groups is indeed observable in reality. The question is whether or not the Above a certain threshold, it might be the case that the gains are bigger than the losses, such that 1 2 n H i H e H Y Y Y n > > . 12 In addition, legal contracts for F might render scenario 3, if F is a WTO member and cannot just increase restrictions on GTs. That reduces incentives for opposition in F. This might also stiffen opposition in F as it cannot easily protect its own industry. link between climate policy and industrial policy has an influence on export expectations related to GTs (eventually leading to an increase of GDP beyond the free-riding status quo). This is, of course, difficult to estimate, as expectations cannot be modeled easily. We argue that expectations about future export sales and thus profits ( e H j ? ) are best expressed in patent applications and grants in foreign target countries ( HF PATENT ). The econometric model is, therefore, constructed in a way that it tries to proxy equation 4.1 ( j j j j j j j l e F pr H e F e H e e H e H c pid c q q A q ? + ? ? ? = ) ( ? ) econometrically. We build the model on the assumption that diffusion of GTs (as a result of pid) reduces marginal production costs. This relationship ) ( : Due to a lack of information, we have to ignore the costs of lobbying ) ( j l c . As our model makes use of future expectations, we do not have information on e H j q , e F j q , Strategic Trade Policy as Response to Climate Change? is described by 1 n H Y . 11 Thus, once the GT industry has been successful in establishing itself at the national level, the GT industry (in both, H and F) and the government (in H) have common interests at the international level. and e F j pid which is expected to be significantly higher than the observed variable j F pid . j In the following paragraphs, we describe in more detail our data-sources. The time frame of the dataset is from 1992 to 2002. 13 The institutional settings analyzed are the SEG (1990-1999) and the EEG (2000-2002). The four sources of the data are the German Patent Office, the International Energy Agency (IEA), Eurostat and the Federal Ministry for the Environment (BMU). The industries of interest are wind, solar, water & ocean, geothermal and biomass. The empirical approach we use to test the theoretical framework looks at the patents, with a priority on the German Patent Office (GPO) applied by German inventors and which are also protected at the European We are limited to this time span even though the data range is from 1990-2005. We drop the observations before 1992 as we assume that patenting abroad before 1992 was not related to diffusion of GTs under the SEG. Another problem is related to the huge time lag between patent application in Germany and the date when the patent is granted in a foreign country. As the dataset we use contains patent counts of patents that have already been granted in Germany and the foreign countries, after 2002 the dataset is biased. The reason for this is that there might be patents that have been applied for in foreign countries but have not been granted, so far. We therefore restrict the dataset to the observations until 2002. A summary of the data included in our dataset is provided in Appendix, page 24. )) ( : ( H HF HF H INCAP PATENT PATENT INCAP ? Thus, if there is a positive correlation between For the regression, we propose to use patent applications, Patent Office (EPO), Japanese Patent Office (JPO) and/or the American Patent Office (APO), respectively. Therefore, we are able to consider the protection of knowledge in different markets. The patent counts we use also contain information about the dynamics of patent application over time. The number of patents issued can, therefore, also be interpreted as diffusion of innovation and expectation for future export receipts. HF PATENT , as a dependent variable. HF PATENT measures patents filed to German inventors at the EPO, the JPO and the APO. As for the timing, we use the priority date which is the date of the patent application at the GPO. 14 If the patent is granted in the foreign country, protection begins with the priority date. The huge time lag that may occur by regressing patents applied in foreign countries on their priority dates is not as problematic as it seems to be at first glance. This is related to the patent cooperation treaty (PCT). Inventors, who desire patent protection in other countries, usually make use of the PCT. According to the PCT, there is only a time span of one year to name the foreign countries in which protection is desired. Note that this information is very important with respect to our assumptions about the time lags implemented in the regression analysis. For patents granted in a foreign country, the protection will go back to the application date in the home country. The rationality behind patenting abroad should be positively correlated with export expectations or the aim to sell licenses of a certain technology to the foreign country. 15 For the study, we use a predefined list of patent classes from table 5 (Appendix, page 25) to extract the patents of the overall sample. Even though key words have been used to find out whether these groups are exactly the international patent classification (IPC) classes where the technologies of interest will be patented, it might be that patents are applied in other groups which are not captured by our list. 16 mill industry, solar industry and biomass industry have generally increased after 1998. For the other two industries, there is no observable trend. The presented figures display the development since 1990-2005. It can be seen that, especially in the case of WIND, patent counts have decreased considerably since 2002. One possible explanation lays within the huge time lag we are confronted with when looking at patent applications that have been granted in foreign countries. We, therefore, drop observations after 2002 and assume that within a three year time span most foreign patent applications are granted. The previous arguments are now summarized to formulate our hypotheses. We use # HF # PATENT as a proxy for export expectations as described in our strategic trade policy framework. Strategic knowledge protection in foreign countries represents the first "mover advantage" from the theoretical part. We argue that feed-in tariffs in Germany are used strategically under the EEG to generate comparative advantages. # H INCAP is, therefore, used as a proxy to test whether it is true that the strategic use of feed-in tariffs did generate positive export expectations captured by HF PATENT . Hypothesis 1 (H1) is formulated as follows: H1: There is a positive relationship between installed capacity of GTs in Germany positively correlated with patents filed in this region in order to protect knowledge. This leads to hypothesis 2 (H2): H2: An increase in installed capacity abroad F INCAP has a positive impact on international patent applications. In addition to these two hypotheses there is the general assumption that there are significant differences with respect to region (r) and time ( t ). H3b: Most dynamics take place in Europe. 17 H3a: There are differences between EPO, JPO and APO because the markets are different from each other. # H INCAP H3c: International patent applications caused by are significantly higher under the EEG compared to the SEG. H3a and H3b capture the spacial dimension. H3c is related to the time dimension. To test H3c, we implement time dummies for the SEG and the EEG. We suppose a significant change in coefficients as Germany started to connect industrial policy with the climate change issue under the EEG. We now turn to the estimation of our econometric model. The core model that shall be estimated is This is somehow clear, because if H is the leader in a certain technology, the follower F cannot export to H as long as inventors in H have applied for a patent. Because patent applications are costly, it is plausible to assume that patent applications abroad go in hand with the commercial value of the invention related to the foreign marketplace. 16 Note that the extraction of the data has been done by an algorithm able to get rid of the problem of double counting of a certain patent. Therefore, double counting cannot be considered to be a problem in our study. 14 Because nearly all patent applications are first filed in the home country of the inventor (Popp; 2006, p. 52), we can look at patents with priority at the GPO applied for protection in other countries. 17 Europe has the highest share of renewable energies (6.9 percent) compared to the other countries of the analysis (Johnstone et al.; 2010, p. 134). # ) , ( The evidence presented at figures 5-9 (Appendix, page 22) shows that patents in the wind mill ELC are added to the core model as controls. 18 The dataset is constructed on three dimensions: (1) Time t, (2) Technology i and (3) Region r. A simple approach would be to estimate the regression for the EPO, JPO and APO separately. In this case there would be the estimation of three different panels. For each panel the estimation would be . , 1 6 1 5 F t ELC is a vector with electricity consumption per capita in region r and F t CPIE is a vector with the price index for energy. + H RuD , F APATENT , F CPIE and1 4 1 3 , 2 2 / 1 , 1 0 t i i F t F t F t F t H t i H t i CPIE ELC APATENT INCAP INCAP RuD ? ? ? ? ? ? ? ? ? + + + + + + + + ? ? ? ? ? (5. Because of collinearity of patent applications regarding r = EPO, JPO, APO, we integrate the third dimension with the same regression. In order to do so, we build region specific interaction terms. Fixed effects are integrated into the model by i ? in order to capture unobservable technology specific heterogeneity. All the residual variation is captured with the error term ?? ??,?? . Important for our model are the assumptions made about time lags and the implemented period dummies. Because our dataset allows for dynamic model specifications, time lags have to be implemented to be in line with economic theory. 19 # H RuD As the priority date indicates the application date in Germany, we expect a one year or a two year time lag for . For H INCAP no time lag is assumed. This assumption makes sense, as the diffusion of the technology in Germany can only take place when the technology is already developed. For We justify our assumptions on the time lags with reference to the PCT. According to the PCT, most of the patents applied at the national level extent to patent applications in foreign countries within a time frame of one year. We overcome this problem by just looking at those patents that already have been granted in Germany. This is a very pragmatic way of dealing with the problem of a time lag of four or five years between the patent application at a national patent office and the patent granting of a foreign patent office. As proposed by Johnstone et al. (2010), we use a negative binomial regression for estimation of the model from equation 5.1 but extend the panel by the third dimension (r). The events we "count" are the patent applications in different international levels indicated by r. The estimation is done for five technologies and 19 For a more detailed discussion on time lags related to patent data compare Hall et al. (1986). Brunnermeier and Cohen (2003) also make an econometric study and make the assumption that there is no lag at all. The result from Griliches (1998) also suggests that with respect to R & D the time lag can be assumed to be rather small. In what follows, we take a closer look on the estimation outcomes. The results of our reference model are presented in table 1, page 16 (estimation results under assumption of a one year time lag for H RuD ). Under the SEG and EEG, we find support for hypothesis 1. As seen, the evidence for hypothesis 2 is mixed but rather weak. Only for JPO such evidence is found. There is no evidence that can be found for hypothesis 3a and hypothesis 3b. To test hypothesis 3c we use a Chow-test and compare . We find significant differences for EPO (p = 0.0580) and JPO (p = 0.0713). For APO the difference is not significant under conventional statistical terms (p = 0.1220). However, if we look at the coefficients, we can see that the relationship under the EEG is smaller compared to the SEG what contradicts our hypothesis. We, therefore, have to reject H3c. INCAP remains significant, confirming hypothesis 1. It can be seen that the right specification of the lag structure for public R & D is crucial for the econometric model. The comparison between the different lag structures shows that for the EEG our findings remain significant. There is a robust finding for our strategic trade hypothesis for the time frame related to the EEG. In order to control for first order serial correlation, we show in table 7 (Appendix, page 26) a model estimated by a simple first differences ordinary least squares (OLS) model. We still get significant results for H INCAP 2002 2000? in JPO and APO. This demonstrates the relatively robust finding for hypothesis 1 (table 7, Appendix, page 26). If we run a Poisson model instead of a negbin model (Table 6, Appendix page 26) some of the results change and become significant but the overall picture remains the same. Even though the model is sensitive to model specification, different estimations have shown that H INCAP is a quite robust predictor for F PATENT under the EEG. As the theoretical model from section 4 mainly refers to this time period, the econometric model offers important insights related to our theoretical reasoning. # VI. # Conclusion We analyze the climate change debate from a perspective of political opportunity and economic © 2013 Global Journals Inc. (US) 2 20 # Global Journal of Human Social Science Volume XIII Issue IV Version I Different to the common view, we argue that free-riding of other countries encourages the German government to foster diffusion of GTs. The empirical evidence shows that for the time span analyzed, positive export expectations could be observed. The main driver we identify for this behavior is the installed capacity of GTs in Germany. This seems fairly plausible and can be interpreted as positive experience that helps also to stabilize international environmental agreements. The theoretical reasoning in combination with the empirical evidence suggests that one can expect Germany producing positive GT spillovers as long as this goes in hand with job creation on the national level in combination with future exports. International experience, however, also suggests that other countries will not open their markets easily. Instead, the German policies may be replicated and other countries may subsidize their own GT industry which renders the German policy unsuccessful. For this case we should expect that Germany reduces its ideal role in international climate policies. ? e H j = q e H j (A e ? q e H j ? q e F j ? c H pr j ) ? c l j ? 0. No additional exports. # Case (3a): The first mover advantage does not lead to exports. # Case (3b): ? e H j = q e H j (A e ? q e H j ? q e F j ? c H pr j ) ? c l j > 0. If the GT industry is so competitive that it was already exporting GTs to F without any subsidies ? In this case it can continue to export, if it is still able to compete with the GT industries j located in F. Case (3b): Decreasing exports GTs comto the case without local content clause. # Scenario 4 H competes with the GT industry located in another country (country I) in a "third" market in F. In this case F is not able to produce GTs but is forced to buy them (e. g. because of high international environmental standards). Case (4) There is competition between H and I. The underlying game depends on which cost curve H and I are operating. They can play Stackelberg, or if they have the same marginal costs, the market has the characteristic of a duopoly with simultaneous market entrance. Case (4) Increase in market size ? export of GTs. # Scenario 5 There is also the possibility that a firm located in H is making a direct contract with politicians in L Case (5a): ? e H j = qe H j pj ? c H pr j qe H j ? c l j > 0. qH j stands for "agreed quantity of GTs" which the GT industries j located in H can sell at the agreed price pj . Case (5a) F buys the technology from the GT industries j located in H. In this case the GT industry would sell a package of GTs to F ? Increase in market size ? export of GTs. Case (5b): ? e H j = qe H j pj ? c H pr j qe H j ? c l j ? ttr > 0. ttr stands for "technology transfer". Case (5b): The contract is combined with a local content clause ? Increase in market size ? export of GTs, but less compared to scenario (a). 21 The dataset contains patents which are granted in at the EPO, JPO and APO with priority in Germany (including the "Neue Bundes länder"). 22 The dataset includes patents and utility patents. The data we use comes from a freely available dataset of the European Patent (DOC-DB). 23 Patent counts about patents applied in region ð??"ð??" (?????????????? ?? ), source OECD: The variable ?????????????? ?? contains information about the overall number of patents applied in the specific territory (EPO, JPO, APO). This variable captures all patents applied for at the EPO, JPO and APO with the inventor's country of residence and fractional counts. The patent counts are based on the earliest priority date. The data mainly derives from EPO Worldwide Statistical Patent Database (April 2007). Information captured with ???????????? ???? , therefore, is industry specific (WIND, SOLAR, WATER, GEO, BIO) and country/territory specific (EP, JPO and APO). Information captured with ?????????????? ?? is country/territory specific (EP, JPO and APO). 25 The data for Germany is in million Euro on exchange rates from 2006. 26 Note that the list is extended in the case of patent classes for WATER, because the law for renewable energy which is analyzed for Germany also changed the institutional framework for energy produced with water. On the other hand, we excluded WASTE, because we focus on GTs and therefore, WASTE is not really considered as a renewable energy source. 22 Note that the date for the patents that are granted goes back to the date when inventors applied for the patent. Even though information about patents until 2006 is available, the analysis is restricted from 1992 to 2002. The information about the last three years is dropped to get rid of the problem that granted patents always go back to the priority date. Therefore, it is plausible to assume that the data from 2004 and 2006 contains a lack of information (Popp;2005, p. 5). 23 For further information see http://www.epo.org/patents/patentinformation/free.html. 24 For more detailed information see Organization for Economic Co-Operation and Development (OECD), Patent Database, June 2007. 25 For further information see http://www.iea.org/. 26 The data for Germany at the national level does not contain information about the expenditures of regional governments. 27 Compare BMU (2007). The data contains information about the installed capacity measured in megawatt-hours (MWh). It measures the overall installed capacity of the industry specific technology per year. The data comes from the Ministry of Environment. 27 Information captured with ?????????? ?? is at the German level and industry specific(WIND, SOLAR, WATER, GEO, BIO). German installed capacity of industry specific technology ?????????? ?? , source BUND: ?????????? ?? is used as a proxy for the induced demand implemented by institutional changes because of laws such as the EEG. Energy price index (???????? ?? ), electricity consumption (?????? ?? ) and installed capacity of renewable energies in the foreign country (?????????? ?? ), source IEA: ???????? ?? is a consumer price index for energy. ???????? ?? is country specific. Year 2000 is set to 100, taxes are consumption in KWh per capita.?????? ?? is country specific. ?????????? ?? measures the overall installed capacity of renewable energies in the foreign country. Information captured with ???????? ?? , ?????? ?? and ?????????? ?? is country/ territory specific (EP, JPO and APO). ![The coalition between the Social Democrats and the Green party from 1998 to 2005.](image-2.png "4") ![Strategic Trade Policy as Response to Climate Change?](image-3.png "B") ![g. Barrett; 1994; Rege; 2000; Ulph; 1996; Ulph and Ulph; ; 2004; Buchholz et al.; 2005; Dolado 2007).](image-4.png "") 1Strategic Trade Policy as Response to Climate Change?PATENTHFEPOJPOAPOH 1992 RuD 1 lag1999 ?-0.0049777-0.0033792-0.000487(0.0084893)(0.0104979)(0.0082184)H 2000 RuD 1 lag?2002-0.0181687-0.0207956-0.0241105(0.0131117)(0.0178787)(0.0147366)H INCAP 1999 1992?0.0002195 ***0.0003652 ***0.0003087 ***(0.0000659)(0.0000929)(0.0000816)H INCAP 2002 2000?0.000108 ***0.0002239 ***0.0002005 ***(0.0000263)(0.0000361)(0.0000313)lagINCAPF0.00001610.0008603 **-0.0000788(0.0000279)(0.0005283)(0.000058)lagAPATENTF-194e-06-0.00005940.0011413 **(0.0003891)(0.0002586)(0.0005508)F lagCPIE 1999 1992?0.0022767-0.00238750.0011234(0.0185545)(0.0178013)(0.020191)F 2000? lagCPIE20020.00924910.00096910.0070262(0.0158542)(0.0177275)(0.0170407)lagELCF-0.0084317-0.0087497 **0.0025994 ***(0.0054865)(0.0040787)(0.0008591)?032.48477(28.06769)Wald chi2214.33Nr. of observations:165Significance:*** ?1%, ** ?5%, * ?10%As a robustness check we present an additionalmodel ( 2of interest,H INCAP , under the SEG hypothesis 1 isonly confirmed for JPO. For EPO and APO it has to berejected. Under the EEG,H 2Strategic Trade Policy as Response to Climate Change?PATENTHFEPOJPOAPOlag2H 1992 RuD1999 ?0.0124475 *0.01121770.0169526 **(0.0072587)(0.0100838)(0.007521)lag2H 2000 RuD?20020.0071520.00371320.0048226(0.0107887)(0.0160282)(0.0126935)H INCAP 1999 1992?0.00009670.0002333 **0.000125(0.0000729)(0.000104)(0.0001025)H INCAP 2002 2000?0.0000872 ***0.0001909 ***0.0001545 ***(0.0000283)(0.0000389)(0.0000395)lagINCAPF0.00006750.00354970.0002025(0.0001194)(0.003328)(0.0003636)lagAPATENTF0.0028577-0.0010550.0038894(0.0040687)(0.0010724)(0.0037222)F lagCPIE 1999 1992?-0.0023775-0.009236-0.0082064(0.0250123)(0.0233636)(0.0258147)F 2000? lagCPIE20020.06497170.05472210.0583138(0.0790413)(0.0794015)(0.079379)lagELCF-0.0458226-0.03128070.0048331(0.0539446)(0.0298639)(0.0030676)?0147.7299(173.8297)Wald chi2163.21Nr. of observations:150Significance:*** ?1%, ** ?5%, * ?10%interest of countries benefiting from exporting GTs.International climate change policy is complementary toexport expectations for GTs. The theoretical welfareeffects of one country's industrial policy, therefore,strongly depend on the policy reaction of othercountries. 4D D D D )(Stackelberg Gamefrom ?????? ?? ?? and enters the foreign market as aStackelberg leader. The Stackelberg game can beThe profit maximization problem leads to?? ???? ?? ?? ???? ?? ?? ?? = ?? ?? ? ?? ?? ?? ?? ? 2?? ?? ?? ?? ? ?? ?? ?? ???? + ?????? ?? ?? ?? = 0?? ?? ?? = ?? ?? ??? ?? ?? ? =????? ?? ????? ?? ?? ???? ??????? ?? ?? 2.(B.1)?? ?? ??? ?? ?? ?? ?represents the response function for ??. ?? maximizes its expected profits with respect to ?? ?? ?? ?? bytaking equation B.1 into account. It follows???? ?? ?? ?? ???? ?? ?? ?? = ?? ?? ? 2?? ?? ?? ?? ?1 2?? ?? + ?? ?? ?? ?? +1 2?? ?? ?? ???? ?1 2?????? ?? ?? ?? ? ?? ?? ?? ???? + ?????? ?? ?? ???? ?? ?? ?? * =?? ?? +?? ?? ?? ???? ?2?? ?? ?? ???? +?????? ?? ?? ?? 2(B.2)© 2013 Global Journals Inc. (US)B 2013 23 ii.In our framework, the GT industry ?? in ?? benefits solved as follows: the GT industry in ?? and ?? are assumed to maximize profits. For ??the profit function is given by equation 4.1. © 2013 Global Journals Inc. (US)20 6??in JPO and APO. 6???????????? ????EPOJPOAPO???????????? 1992?1999 ???0.003891?0.0053209?0.0027717(0.0050035)(0.0070105)(0.0059082)???????????? 2000 ?2002 ???0.0218788***?0.0205298*?0.0242853**(0.0076388)(0.0113637)(0.0096594)?????????? 1992?1999 ??0.0001682***0.0003202***0.0002738***(0.0000476)(0.000074)(0.0000645)?????????? 2000 ?2002 ??0.0000832***0.0002117***0.0001901***(0.0000172)(0.0000279)(0.0000242)???????????????? ??4.78e?060.0005027*?0.000037(0.0000164)(0.0003115)(0.0000375)???????????????????? ???0.0001206?0.00014570.0007427**(0.000266)(0.000173)(0.000334)?????????????? 1992?1999 ??0.0190602?0.003186?0.0009674(0.025742)(0.0246747)(0.0269702)?????????????? 2000 ?2002 ??0.023799?0.00209670.0032361(0.02162)(0.0249417)(0.0220473)???????????? ???0.0035969?0.0058992**0.0023531***(0.003572)(0.0024022)(0.0005705)Wald chi2411.06Nr. of observations:165Significance: *** ? 1%, **? 5%, * ? 10% 7???????????? ????EPOJPOAPO??????1?????? 1992?1999 ???0.2735567?0.0252574?0.053742(0.1733541)(0.0849316)(0.1719576)??????1?????? 2000 ?2002 ???0.2446132?0.0247744?0.1965391(0.2085309)(0.1635901)(0.2077007)?????????? 1992?1999 ??0.00013680.00219670.0017532(0.0013486)(0.0013473)(0.0013839)?????????? 2000 ?2002 ???0.00071350.0012767 * *0.0013259 * *(0.0005851)(0.0005684)(0.0013259)???????????????? ??0.00094940.00242220.0013545(0.0106794)(0.0238106)(0.0095546)???????????????????? ??0.0060522?0.00364440.0060825(0.0559459)(0.0571749)(0.0470171)?????????????? 1992?1999 ???0.5386050.0867416?0.0706172(10.112697)(0.7568268)(0.5441484)?????????????? 2000 ?2002 ???0.36474330.12821160.0190248(10.356288)(10.207714)(0.8598966)???????????? ??0.0091146?0.04803790.0228105(0.0829318)(0.2854876)(0.1428906)?? 0?8.647436(88.44358)R-sq:0.3082F(27,108)1.89Nr. of observations:150Significance: *** ? 1%, **? 5%, * ? 10 © 2013 Global Journals Inc. (US) © 2013 Global Journals Inc. (US) * WBaumol WOates The theory of environmental policy Cambridge Cambridge Univiversity Press 1988 * On the rationale of group decisionmaking DBlack The Journal of Political Economy 56 1 1948 * Direct employment in the wind energy sector: An EU study MIBlanco GRodrigues Energy Policy 37 8 2009 * Erneuerbare Energien in Zahlen: Nationale undInternationale Entwicklung, Bundesministeriumfür Umwelt 2007 BMU * Erneuerbare Energien in Zahlen: Nationale undInternationale Entwicklung, Bundesministeriumfür Umwelt 2008 BMU * Erneuerbare Energien in Zahlen: Nationale undInternationale Entwicklung 2011 BMU * property=pdf, bere-ich=bmwi, sprache=en, rwb=true.pdf>. 9. of a breakthrough: the Kyoto Protocol as symbolic policy Bmwi European Journal of Political Economy 20 3 2011 Energy Concept: For an Environmentally Sound, Reliableand Affordable Energy Supplys * Export Subsidies and International Market Share Rivalry JABrander BJSpencer Journal of International Economics 18 1985 * Determinants of environmental innovation in US manufacturing industries SBrunnermeier MCohen Journal of Environmental Economics and Management 45 2 2003 * International Environmental Agreements and Strategic Voting WBuchholz AHaupt WPeters Scandinavian Journal of Economics 107 1 2005 * The Problem of Social Cost RHCoase Journal of Law and Economics 3 1960 * Efficiency -our focus: Introducing the German Energy Agency Dena 2011 German Energy Agency * Delegation in international monetary policy games JJDolado MGriffiths APadilla European Economic Review 38 5 1994 * An economic theory of democracy ADowns 1957 Harper and Row New York * Act Revising the Legislation on Renewable Energy Sources in the Electricity Sector and Amending Related Provisions, Federal Ministry for the Environment, Nature Conservation and Nuclear Safety Eeg 2009 * G8 and the Move to a Globalised International Economy AFreytag LWangler Diplomacy and Statecraft BMcKercher Routledge, New York 2011 * R & D and Productivity: The Econometric Evidence ZGriliches 1998 University of Chicago Press * Patents and R & D: Is There a Lag? BHall ZGriliches JHausman International Economic Review 27 2 1986 * The tragedy of the commons GHardin Science 162 859 1968 * The expansionof renewable energies and employment effects in Germany BHillebrand HGButtermann JMBegringer MBluel Energy Policy 34 2005 * The marginal impact of CO2 from PAGE 2002: An integrated assessment model incorporating the IPCC's five reasons for concern CHope Integrated Assessment 6 1 2006 * Renewable energy policies and technological innovation: Evidence based on patent counts NJohnstone IHascic DPopp Environmental and Resource Economics 45 1 2010 * Weltweiter Klimaschutz -Sofortiges Handeln spart hohe Kosten CKemfert 2005 DIW-Wochenbericht 72 * National Politics and International Agreements, Post-print and working papers HKempf SRossignol 2010 Université Paris1 Panthéon-Sorbonne * Kyoto Protocol to the United Nations Framework Convention in Climate Change, United Nations 1998 * Die Herausforderung globaler Klimaw-andel MLatif Perspektiven der Wirtschaftspolitik 11 s1 2010 * Renewable energy and employment in Germany ULehr JNitsch MKratzat CLutz DEdler Energy Policy 36 1 2008 * The skeptical environmentalist: measuring the real state of the world BLomborg 2006 Cambridge University Press Cambridge, New York * Effects of energy policies on industry expansion in renewable energy PLund Renewable Energy 34 1 2009 * Keeping warming within the 2°C limit after Copenhagen AMacintosh Energy Policy 38 6 2010 * Scale in technology and learning-by-doing in the windmill industry ESMadsen CJensen JDHansen Journal for International Business and Entrepreneurship Development 1 2 2005 JIBED) * Country-specific market impacts of climate change RMendelsohn WMorrison MSchlesinger NAndronova Climatic Change 45 3 2000 * Beyond the learning curve: factors influencing cost reductions in photovoltaics GNemet Energy Policy 34 2006 * Copenhagen Accord faces first test ANicoll JDelaney RStrategic Strategic Comments 16 1 2010 * Warming the world: economic models of global warming WNordhaus JBoyer 2003 MIT Press Cambridge * A regional dynamic general-equilibrium model of alternative climate-change strategies WNordhaus ZYang The American Economic Review 86 4 1996 * Governing the commons: The evolution of institutions for collective action EOstrom 1990 Cambridge University Press Cambridge * The Politics of 1992: Fiscal Policy and European Integration TPersson GTabellini The Review of Economic Studies 59 4 1992 * The Effect of New Technology on Energy Consumption TPersson GTabellini Resource and Energy Economics Popp, D. 23 4 2000. 2001 MIT press Political economics * Induced Innovation and Energy Prices DPopp American Economic Review 92 1 2002 * Using the Triadic Patent Family Database to Study EnvironmentalInnovation, Environment Directorate Working Paper DPopp ENV/EPOC/WPNEP/RD 2005. 2005 2 OECD, France, Paris * International innovation and diffusion of air pollution control technologies: The effects of NOX and SO2 regulation in the US DPopp Journal of Environmental Economics and Management 51 1 2006 * The Learning Curve in a CompetitiveIndustry EBRasmusen EPetrakis SRoy The RAND Journal of Economics 28 1997 * Strategic Policy and Environmental Quality: Helping the Domestic Industry to Provide Credible Information MRege Environmental and Resource Economics 15 3 2000 * Strategic delegation of environmental policy making HRoelfsema Journal of Environmental Economics and Management 53 2 2007 * Delegation and Threat in Bargaining JSchumpeter Kapitalismus Sozialismus Und Demokratie TübingenFrancke BSegendorff Games and Economic Behavior 23 2 1987b. 1998 * Public policies against global warming: a supply side approach HSinn International Tax and Public Finance 15 4 2008 * The Economics of Climate Change: The Stern Review NStern 2007 Cambridge University Press Cambridge, New York * Estimates of the damage costs of climate change, Part II. Dynamic estimates RTol Environmental and Resource Economics 21 2 2002 * Environmental Policy and International Trade when Governments and Producers Act Strategically AUlph Journal of Environmental Economics and Management 30 3 1996 * AUlph DUlph Global Warming, Irreversibility and Learning 1997 107 * Climate changeenvironmental and technology policies in a strategic context AUlph DUlph Environmental and Resource Economics 37 1 2007 * German climate change policy HWeidner LMez The Journal of Environment & Development 17 4 2008 * Strategic Trade Policy as Response to Climate Change?