“the Green Paradox Phenomenon: A European Union Empirical Application”

: The green paradox describes an undesirable and socially inefficient phenomenon 13 caused by the expansionary reactions of the Supply as a response to the various mechanisms that 14 combat climate change. This article seeks to understand and aggregate the different drivers of this 15 phenomenon portrayed in the literature, as well the empirical evidence associated and the proposed 16 solutions. For this purpose, compilation and systematization of the various scientific contributions 17 up to date in this context have been elaborated, which led to the identification of five major drivers 18 of the paradox: (i) Environmental Taxes, (ii) Green Supports, (iii) Uncertain Property Rights, (iv) 19 Temporal Lag and (v) Emission limits (Cap). Moreover, we evaluate its effective impact on the 20 European scenario in the last two decades, and we have obtained evidences that the hiatus in the 21 implementation of the European emission cap mechanism has generated a strong green paradox. 22 Moreover, a robust reflection regarding the economic and environmental adequacy of green supports 23 should be considered due to its questionable net benefits. Lastly, we offer some recommendations of 24 public environmental policies that escape the paradoxical phenomenon, through the enunciation of 25 the conditionalities of these provocateurs.


37
When we are confronted with (negative) externalities in the non-renewable natural resources market, we 38 assume that in the light of the seminal literature, the best alternatives would be to implement instruments such 39 as the taxes/subsidies of Pigou, or the market for CO2 permits (Carter, 2007). However, when we look more 40 closely at the state of the art of the benefits of these economic incentive instruments in this market, we 41 understand that some authors, such as Sinn (2008), have pioneered the existence of a Green Paradox, thus 42 deteriorating the increase in social welfare of the environmental measure. Moreover, it is also possible to 43 identify in seminal literature, the discussion of the "unintended effects", prior to this concept, through Strand 44 in 2007. Although he does not coin the term, he deductively demonstrates how the implementation of an 45 alternative technology can cause an acceleration of fossil fuel extraction at the present time, as fast as the rate 46 of innovation of this same technology is (Strand, 2007). In this way, Sinn conceptualizes the controversial 47 "green paradox" that gives rise to recent research around the concept. This aggregator phenomenon tends to 48 incorporate the expansionist reaction of the supply as a response to the various mechanisms for combating 49 negative externalities, also known as drivers or provokers of the paradox. It is in this context that the study, 50 mostly deductive, of the German economic advisor, alerts us to the strategic and rational behaviour of the 51 supply side which reacts offensively to these economic instruments in response to the announced appropriation 52 of its present and future rents (Sinn, 2008). This manoeuvre can be seen both in the increase of the extraction 53 (more usual) and/or in the geographical displacement of the extraction unit to another country with less 54 environmental regulation, also known as carbon leakage, as demonstrated (Eichner & Pethig, 2011), (Ritter & 55 Schopf, 2014), (van der Ploeg & Withageny, 2015), (Sen, 2016) and (Fujisaki, 2018). To this physical 56 remobilisation it can also be attributed the terminology "spatial green paradox" as opposed to the green paradox 57 whose consequence is only an increase in the production of the targeted resource (Ritter & Schopf, 2014).

58
Still with regard to the seminal study of Sinn, it is important to note that it considers the paradox to be 59 particularly problematic, not only because it culminates in the opposite effect to the desired, but mainly because 60 it leads to an increase in negative externalities generated at the present time, instead of smoothing over time 61 the amount of emissions released into the atmosphere. This situation becomes even more weary when we relate 62 this adverse phenomenon to the scientific expectation that in the future society will have better technologies 63 capable of mitigating the impact of these externalities, whereby socially it will not make much sense to 64 anticipate the extraction of these resources. From another point of view, the evolution of greener alternative aggregate the different drivers of this phenomenon portrayed in the literature, as well the empirical evidence 87 associated and the proposed solutions. The evaluation of the effective impact of these drivers is also made on 88 the European scenario in the last two decades, and we have obtained evidences that the hiatus in the 89 implementation of the European emission cap mechanism has generated a strong green paradox. Truly, the 90 major contributions of this article lie in the pioneering explicitness of the various drivers that promote the 91 paradox, as well as their conditionalities. This construction is especially innovative given the absence of a 92 compiler article on the present topic. Furthermore, there is a scarce literature that empirically study the green 93 paradox and its drivers for the European context. The structure of this article is as follows: following the present 94 introduction, we present a literature review with an inaugural description of the methodologies used in the 95 study of the paradox, the statement of the main drivers and the empirical evidence regarding those, and a brief 96 discussion of the eventual solutions to this phenomenon; in the third section we present the methodology and 97 data for the empirical approach, followed of the results presentation and discussion, and finally, the main

115
Nevertheless, since the conceptualisation of the paradox in 2008, the paramount importance of validating this 116 phenomenon through various empirical studies in different regions has been reiterated. This is a task that 117 however has been proved difficult for the scientific community (Österle, 2012). Nonetheless, it is crucial to of contemporary empirical studies whose approach to the green paradox incorporates some of the 125 conditionalities of the region studied. While Zhang et al. (2017) link fiscal decentralisation to the phenomenon, similar way, the expectation of an increase in this tax. This link is also accompanied in his research by green 137 financial support, or subsidies for cleaner technologies. Although not fully corroborated in his article, the 138 author suggests that the effects of these environmental aids would have a similar effect to the economic 139 disincentive measure, since both would create incentives for present extraction in favour of future extraction 140 (Sinn, 2008). Indeed, the presence of the expectations of economic agents, and in particular the owners of 141 natural resources, has been a relevant component in the work of Sinn and the subsequent authors. Besides, this 142 is unanimously regarded by researchers as valid and assertive, as this expectation derives directly from the 143 definition of efficient equilibrium in the instruments of Pigou (1920). The increase in marginal externality is 144 accompanied to the same extent by an increase in the tax/subsidy to be applied, and as the unitary 145 environmental impact of emissions is increasing, so will the amount to be applied by the economic 146 (dis)incentive measure (Klepper et al., 2006). Without prejudice to the main drivers mentioned here, the 147 seminal research also alerts to the problem of uncertain property rights. This issue, also related to the natural 148 risk aversion of economic agents, indicates that owners with greater insecurity regarding the maintenance of 149 their own resources choose rationally to extract more at present compared to those with less instability, a 150 concept first developed by Long (1975).

151
Currently, it is possible to find in the state of art of literature five great drivers of the green paradox. These 152 aggregate not only those mentioned above in the seminal articles of Sinn (2008), but also other dynamics whose

158
The most well-known and immediately related provocateur of the paradox is the tax, or the expectation of a 159 tax increase. This, as we have seen before, was first identified by Sinn, who offers the literature the general 160 traits for the future study of this driver. If on the one hand this instrument creates a depressive effect on the 161 present market, displacing supply in a northwest direction to faithfully represent its marginal costs, on the other 162 hand it conceives a clear incentive to present extraction in anticipation of an even greater decline in this market 163 in the future. In other words, it generates a reduction in the opportunity cost of retaining the resource in the 164 mineral deposit. The author concludes "unless the last effect (paradox) is dominated by the first (sector 165 depression) these policies cannot be proposed as a means to mitigate the greenhouse effect" (Sinn, 2008, p. 166 377).

167
Yet in the same sphere, some discussion has arisen about the characteristics of the tax that effectively encourage 168 early extraction. Distancing themselves from Pigou's (1920) definition of an efficient tax emerge the authors 169 Hoel (2010) and Edenhofer & Kalkuhl (2011) who study the relationship between the tax and the paradox 170 without the classic conception that it must necessarily equal the amount of negative externality in order to 171 achieve a social optimum. In this way, the analytical research listed above endorses the thesis that the actual 172 cause of the phenomenon described is not the tax, but in turn its characteristics and conditions.

173
Hoel shows how a low tax level at the present, together with a sufficiently quick growth of it, or its expectation, 174 surely generates conditions for the anticipation of extraction described in the paradox. However, the same does 175 not occur when the tax is sufficiently high, regardless of the agent's future expectations of its growth or the 176 format of the firm's cost function. This is due to the impossibility (and illegality) of the supply to carry average 177 costs higher than the price charged, and therefore virtually prevented from expanding its production from this Edenhofer and Kalkuhl also follow this idea, and beyond specifying the situations in which the tax causes the 186 paradox, they also analyse its relationship with the discount rate. Truly, the research mentioned here resembles 187 that of Hoel in what concerns the reduced risk of early extraction at a high tax level compared to a considerably 188 lower one (Edenhofer & Kalkuhl, 2011). However, this differs slightly from the above-mentioned literature by 189 specifying in even more detail the conditions inherent in provoking the green paradox through a tax. Whereas 190 upstream, Hoel points out that a low but fast-growing tax level is responsible for the paradoxical phenomenon,

191
Edenhofer & Kalkuhl concretely define the speed of this same growth. The specific conditions for the paradox 192 are thus described as: a relatively low level of tax and a permanent increase of it above the discount rate for 193 the resource owners (Edenhofer & Kalkuhl, 2011).

194
In addition to this discussion, Chen et al. (2020) support most of the conditions presented by Hoel and

195
Edenhofer on the relationship between the level of the tax and an early extraction. Thus, the recent research 196 cited illustrates this condition in quadratic form. Empirically linking the level of environmental regulation with 197 the CO2 emitted into the atmosphere the study concludes an inverted U relation. This trend is in line with the 198 conditions presented upstream, as reduced regulations encourage owners to extract increasingly in advance 199 until a peak extraction/emission is reached, which is only reversed by high regulations that effectively force 200 extraction to be reduced (which raises prices to sustainable values once more), or forces the closure of less 201 efficient enterprises, which in turn also generate a significant reduction in emissions (Chen et al., 2020).

203
Another relevant driver in the literature focuses on the assistance for "green" technologies. These incorporate 204 not only the financial side through subsidies, but also the allocation of resources for the innovation of these 205 technologies. In fact, this provocative of "unintended environmental effects" is highlighted even before the 206 conceptualization of the green paradox by Sinn (2008), when in the previous year Strand warned us about the 207 harmful results of this increasingly popular economic policy. This study, focusing its approach on the 208 relationship between the rate of innovation of an alternative technology and the extraction of fossil fuels, 209 deduces a positive relationship between the two, the higher the said rate of innovation is. This is a result of the 210 vision that the owners of energy resources have of an under development alternative technology, which turn 211 out to be an inherent threat to their business model (Strand, 2007).  context that Gerlagh (2011) offers us a useful classification of a green paradox as a way of responding to this 217 temporal problem of environmental costs and benefits in the short and long term. In this way, it is then possible 218 to distinguish between a weak green paradox and a strong green paradox. The first, of lesser environmental 219 severity, arises when "the anticipation of cheaper alternative energy technology increases emissions at the 220 present time". The second, of greater concern, arises when "anticipating cheaper alternative energy technology 221 increases cumulative emissions damage on a net present environmental level value (NPV)" (Gerlagh, 2011).

222
This distinction allows for a better understanding of the effective implication of green supports in the paradox.

223
Thus, the main concern should not arise from the emergence of a weak green paradox, since it only reflects the 224 immediate result of the measures, completely ignoring the future benefits of such innovation. On the contrary,

225
it is the occurrence of a strong paradox that needs to be highlighted. It is therefore feasible to decide rationally, 226 on the basis of this division, whether or not an alternative technology should receive public support. If its 227 emergence causes fossil fuel extraction to occur to such an extent that a strong paradox arises, then it 228 incorporates a negative externality that may even be taxed, as van der Ploeg & Withagen (2012) suggest. If, on 229 the contrary, only a weak paradox emerges, then it adds a positive externality that should be supported in order 230 to internalize the social benefits produced (Gerlagh, 2011).

231
Another way of exposing the effects of green support on the economy is illustrated by Grafton et al. (2012), green" effect, is enshrined in the strategic decision of the supply to reduce the equilibrium price of the above-236 mentioned fuels, and consequently increase the quantity traded. Thus, it is also fair to say that the green paradox 237 effectively only occurs from this driver when the weight of the indirect effect exceeds that of the direct effect 238 (Grafton et al., 2012).

239
In addition to what has been said so far about this provocateur, it is also important to note that it can be 240 expressed in many ways in the economy. The mechanism of encouraging early extraction through the 241 emergence of a substitute energy has been described by several authors and it is stressed in the literature that 242 this is not the only way in which the paradox is expressed in the green support aspect. In a short case study of 243 the Japanese region, Fujisaki reflects on the possibility that the introduction of renewable energy into national 244 electricity production may lead to the relocation of Japanese companies to other countries (Fujisaki, 2018). If 245 an eventual causal relationship between the two is confirmed, one of the major limitations of the above-246 mentioned research, the possibility of a spatial green paradox through carbon leakage to foreign states can be 247 observed from the stimulation of alternative energies.

248
Finally, its worthy to note the multifaceted character that the concept of green support can have. As we have 249 seen, it incorporates not only the technological innovation of energy alternatives, but also subsidies to them, or  property rights and the occurrence of a green paradox (Sinn, 2008). A determining factor in this link is the 266 difficulty of these fossil entrepreneurs to sell mineral deposits quickly, and to recover their sunk costs in the 267 geological evaluation and prospecting prior to the extraction. Thus, faced with a high risk, there is no other 268 option on the table than to hurriedly remove the resources from the subsoil and sell them at a lower price than 269 initially planned (and earlier), occurring moreover in a green paradox.   the fact that democracies are more likely to challenge the ownership rights of existing "clans" than dictatorial 281 regimes, which sometimes even live closely with them through close family and business relationships. In this green paradox" (Sinn, 2008, p. 371). Moreover, it is still possible to relate the inherent link between dictatorial 286 regimes and monopolies in the energy sector, in contrast to the substantially more competitive economic 287 environment that exists in most democracies. In fact, and taking into account Lewis (1976) seminal elations, 288 which indicate a propensity on the part of monopolists to extract non-renewable resources in a more lenient 289 way than economic agents in competitive markets, there is a strong possibility of the paradox to come naturally 290 with greater emphasis in modern democracies. Thus, the monopolist's action as a "resource saver" in distancing 291 itself from the competitive equilibrium imposes a benign limitation on the amount of resource marketed, and 292 hence of the externality generated (Lewis, 1976).

293
Finally, it is possible, in a similar way to green support, to find a multiplicity of expressions by which this 294 driver manifests itself in the economy. One of these is the exploitation of common goods. These are part of a 295 classification of goods, which considers the parameters of exclusivity and rivalry. In certain circumstances, 296 such as the discovery of common deposits, it is difficult to apply the condition of exclusivity effectively to this 297 property. In this way, and simultaneously to a clear association of this with an opportunity cost and consequent 298 rivalry, it becomes inevitable to prevent recurrent problems of the existence of a non-renewable common good.

299
Sinn (2008) warns that it is extremely important to define the property rights of these resources in a concrete 300 and definitive way, so that the owners of these resources can consummate the condition of exclusivity and 301 make it a private good. Indeed, the existence of a non-renewable common good will encourage its extraction 302 until the point where the next unit of resource will marginally cost more to extract/produce than the price of it.

303
The static balance will be in Price=Average Cost with industry profits being zero, as until this equilibrium 304 economy will bring new agents into the operation and/or additional extraction units. Thus, compared to the 305 scenario of defined property rights of a private good, we have an excessive and anticipated use of mineral 306 resources susceptible to a green paradox (Sinn, 2008).

325
The most obvious, classified by abundance effect, is characterised by the excessive extraction of the resource 326 in the period in which it is not constrained or limited. Thus, when a policy that creates obstacles to the extraction 327 of fossil fuels is announced, whether through a tax or an emission limit, "more of the resource will be extracted 328 in other (unrestricted) periods" (Di Maria et al., 2012, p. 105). In addition, the ordering effect arises, of 329 particular relevance when the targeted mineral resource has distinct properties depending on its age, such as 330 coal, whose carbon intensity increases as it forms underground. Thus, this effect states that it may be rational 331 for the owners of these deposits to protect the less polluting mineral deposits for the period targeted by the 332 measure, "dispatching" first and, prior to the measures, the extraction of the comparatively most polluting presented of a weak and strong green paradox, which underline the paramount importance of assessing as a 336 whole the net updated benefit, before sentencing the aforementioned measure (Gerlagh, 2011). Indeed, 337 although the existence of a time lag in this area does not contribute positively to social welfare, it should not 338 always be a reason to overturn the carbon tax policy, or emissions cap. At times, this gap, which is even legally 339 necessary, given the publication required in the Journal of the Republic before the law came into force, causes 340 "merely" a weak paradox, with that measure as a whole making a long-term contribution to the global 341 mitigation of environmental damage, and a social positive improvement.  leakage, which causes a spatial paradox, or even with the temporal gap, which is in itself a provocateur of the 360 same, and when together with the cap, acts as an amplifier of its effects.

361
Starting with the relationship that this driver has with the green (spatial) paradox, the articles by Eichner &

373
However, it is important to point out that, similarly to previous drivers, the circumstance of a green (spatial) 374 paradox does not necessarily imply the embargo of the present mechanism. In reality, as Sen (2016)

389
Despite the results obtained by the above-mentioned authors, it is important to highlight the appeal they leave 390 when analysing their own results. In fact, the geographical coexistence of groups of regions with different 391 restrictive environmental standards will permanently encourage carbon leakage, the main solution to which 392 would derive precisely from the abolition of these disparities. It is therefore "imperative to bring developing

454
which considers biofossil production to be a suitable proxy for subsidies, which may to some extent call into 455 question the link exposed by the author.

488
Another order of contributions to the resolution of the paradox arises from the allocation of subsidies. In this 489 field, the proposals to subsidize the mineral resource underground stand out, since its permanence in the deposit 490 reflects an absence of negative externality in society that is worth encouraging. However, the authors promptly 491 warn that the implementation of such measures would be politically unthinkable (Sinn, 2008), (Edenhofer &

492
Kalkuhl, 2011). Edenhofer goes even further, saying that given substantial environmental concerns at the 493 present it may be more efficient to tax green alternatives and subsidise underground mineral resources in order 494 to avoid early extraction at all costs (Edenhofer & Kalkuhl, 2011).

495
In the problem of the mechanism for implementing an emissions cap causing a spatial paradox, the literature 496 reaffirms the need to sign common international agreements covering all countries, and not just those within a 497 common economic integration, as has been the case until then. In fact, this indication has been common to 498 almost all the authors of this theme, since it would be the only effective resolution to the "carbon leakage"

511
In order to validate the green paradox discussed, the present investigation promotes the application of an 512 empirical methodology that aims to assess the effective impact of the five mentioned provocateurs In this way, 513 a multiple econometric analysis is instilled in this study, which aims to examine the propensity of the where the subscript it in the variables symbolizes the i-th country, in year t, represents the perturbation term 526 and 1, 2… 6 the estimated coefficients. The selected variables are described in the subsequent sub-chapter, 527 and in Table 1.  Table 1 for a resume of all variables considered.

567
These authors promote a study of the impact that subsidies to bio-fossils had in the first instance on North

568
American oil production and ultimately on the country's CO2 emissions. For this, the total production of bio-569 fossils is promoted as a proxy for these subsidies due to the "superior availability of data for this indicator and

689
To analyze the stability/stationarity of the selected variables, this study proceeds to four unit roots panel tests: where -1 symbolizes the standard normal probability distribution, and the common term Xm is the p-value of 707 the respective test.

710
To assess the existence of a cointegrated (and stable) relationship in the long term, the present study also carries

743
Among the dynamic models, the one that is regularly adopted in the literature and whose effectiveness stands

756
In order to complete the empirical validation previously proposed, the results referring to the parameters of 757 equations (1) and (2) are presented in this chapter, concentrated on the impacts of provocateurs on oil 758 extraction, and on carbon intensity, respectively.

761
There is a homogeneous tendency with regard to the stationarity of the analysed variables. The null hypothesis 762 specified in the previous chapter is rejected for the first differences. In this way, the effective evasion of 763 possible spurious regressions resulting from unit root variables is inferred, confirming the legitimacy of the 764 subsequent analysis to cointegration.

765
In this same context, the particular character of the variables related to the Temporal Lag and the 766 implementation of the Cap & Trade mechanism is also emphasized, whose binary nature prevents a statistical 767 interpretation similar to the other variables. The full results can also be seen in Table 5.

774
The analysis referring to cointegration uniformly reveals a genuine and stochastic trend among the variables 775 of the analysed regressions. It is possible to observe the rejection of the condition of absence of cointegration, 776 referring to the null hypothesis, in the ADF and t-statistics tests, simultaneously with the non-rejection of an 777 effective relationship between the endogenous variables and the variable explained in the v-statistics tests, and 778 p-statistics. In this way, and in addition to the foregoing investigation, it is feasible to confer an estimated 779 validity on the regressions proposed below. The full results can be seen in Tables 6 and 7. 780 781

790
The estimated regressions, evidenced in detail in the previous chapter, when arise from the application of the 791 generalized method of moments (GMM), are accompanied by the Sargen-Hensen test (Hansen, 1982;Sargan,    Note: ***, **, * represent significance at a level of 1%, 5% and 10%, respectively.

861
Standard deviation statistics are indicated in parentheses.

957
In the case of the environmental taxes, there is an effective fulfilment of the legislator's intentions, with a 958 significant decrease in the carbon intensity resulting from this tax burden. Thus, and taking into account the 959 circumstances previously indicated by which the tax may incite the paradox, it is recommended to maintain 960 this robust tax level in the European framework, and when changes are needed, they should occur smoothly, 961 as has been happening so far. It is also essential that the main political actors resist, in a preventive way, to the 962 radical discussion on the increase of this type of taxes, given that the economic agents expectation can be 963 similarly crucial for the occurrence of the paradox.

964
The results regarding the green support, it is not so clear that all the legislators' intentions have been fully