On the Access Principle in Science: A Law and Economics Analysis

Faculty of Law, Ruhr University of Bochum, Bochum, Germany


4.1 The Organization of Science: Open Access vs. Copyright

This section provides a comprehensive analysis comparing both systems from a social welfare point of view and hence asking whether academic publishing should be organized by means of a universal closed or open access mode. The first subsection is an extensive view on both regimes and their impact in the light of the publishing game and hence the prevailing “publish or perish”—environment in scientific research. The impact on researcher’s private incentives to write high quality papers will be investigated as well as the social welfare effects when shifting towards an universal OA regime. Several robustness checks and a model extension to think outside the box of the model’s inherent contest character provide a broad picture on the superiority of either regime. The second subsection picks up on the distributive effects from shifting towards an “author pays” principle when introducing OA as the dominant publishing mode, briefly highlighting some possible distortions that may result in an obviously heterogeneous world. In this regard, especially the consequences for developing countries will be addressed, providing the analytical framework for investigating the international political economy of access to scientific knowledge in Sect. 4.2 of this chapter.

4.1.1 Academic Copyright in the Publishing Game: A Contest Perspective1


As we have seen previously, the literature addressing OA principles in academic publishing can broadly be structured in three lines of research: (i) studies on the economic impacts of alternative publishing models on the scholarly system as a whole, (ii) studies assessing the effects of open access on readership and citations, and (iii) studies investigating researchers’ attitude and behavior towards open access (see the literature review below). In this subsection, we will primarily address the first literature stream. Most of this literature shares the two fundamental assumptions that scientists are hardly motivated by (small) royalties, and that social welfare increases in the incentives set for scientists due to positive externalities. We do not doubt the importance of positive research externalities, but we argue that the rent-seeking character of research may nevertheless lead to incentives that are excessively high from a social perspective.2 This is the motivating force of our paper.

Starting point of this subsection is that, similar to competing for promotion or for prizes in professional sports, we also find elements of a zero-sum game in the academic publishing game. This bears the risk of rent-seeking activities, and hence ceteris paribus of incentives that are excessively high.3 In reality, this risk is likely to be reinforced by the multi-task character of academics’ obligations: whereas the research output is relatively easily measurable and consequently strongly incentivized, this is more difficult for teaching and administrative work, and hardly possible for the contribution to a productive work atmosphere by supporting other people.4 Thus, taking additionally into account that effort incentives might be relatively distorted towards research,5 it seems not far-fetched that incentives for research may be either too low or too high. In this paper, we do not take the multi-task character of the academic profession explicitly into account, but restrict attention to the contest perspective.

Specifically, we adopt the canonical Tullock-contest model6 for analyzing the publishing game between scientists, and we argue that such a model complements the common view on research in three aspects: first and already mentioned, we emphasize that incentives for research are only too low if the rent-seeking activities of academic contest do not exceed the positive externalities from academic publishing in general. Second, efforts in contest models do not only depend on stakes (that is in our context on readership, reputation and career effects), capabilities and effort costs, but also on the equilibrium behavior of competitors. Thus, the contest perspective adds new factors to the determination of effort in academic publishing which seem to have been widely neglected so far. Third, equilibrium effort levels in Tullock-contests are decreasing in the competition’s asymmetry. We will argue that this is important for the problem at hand, and our view is based on the argument that switching from a closed to an open access-mode is likely to increase the gap between researchers from top and mediocre universities.

Primarily two institutional assumptions drive our analysis. First and foremost, we argue that the most talented researchers tend to be at the best universities. Obviously, as a university’s ranking reflects the sum of its researcher performances (see Crane (1965), and more recently Goodall (2006, 2009)) likely justifies that these two aspects are mutually conditional.7 Second and in contrast to Shavell (2010), we assume that mediocre universities will often not fully pay for the submission fees under open access. This may be debatable. Nevertheless, considering Germany as an example, even the moderate submission fees under closed access are currently not fully covered by all universities. Many professors at public universities have rather small budgets they can allocate among different objectives, and submitting papers reduces the funds left for attending conferences or for hiring student assistants. Furthermore, some (top) universities cover also the considerably higher submission fees for fast tracks in journals while others do not. Given that submission fees with open access would be much higher, assuming that (some) authors would have to care about them seems reasonable.8 This aspect may particularly gain recognition when considering developing countries.9

In Feess and Scheufen (2013) we integrate these two institutional factors in a stylized contest-model with two types, a more talented academic working at a top institution covering the submission fees, and a less talented academic employed by a mediocre institution not covering the fees. We derive the following results: First, when moving from closed to open access, there are countervailing effects with respect to the effort incentives—open access leads to higher efforts if and only if higher readership outweighs the incentive-reducing impact of the submission fees borne by the low type. Notably, the high type’s equilibrium effort is also decreasing in the low type’s submission fees due to the contest’s larger heterogeneity. Second, if private effort incentives are too low, because the externality effect outweighs the rent-seeking effect, then it depends on the model’s parameters which of the two publishing modes is superior. Third, open access is always superior when private effort incentives are excessively high, and we will provide a clear-cut intuition for this result after it has been developed.10

Let us briefly discuss why we think that the contest perspective is important. As any model of strategic competition, the contest perspective implies that participants take their behavior mutually into account. To see why we believe that this is often the case, consider Management Departments in Germany as an example. In twentieth-century-Germany, publications in international journals played a minor role for career opportunities in the academic management profession, which depended mainly on academic ancestry and book publications. Very recently, this has radically changed and teaching loads as well as the allocation of internal funds are now largely contingent on the position in a ranking based on a standardized scoring system for journal publications. Roughly, a few hundred academics in Germany may currently be qualified for publishing at a regular basis in respected journals, and many of them seem in fact concerned about the research output of their colleagues. Moreover, anecdotal evidence supports the view that it is far from obvious whether efforts are strategic substitutes or complements: Some (strong) academics tell that they respond to fiercer competition by enhancing their effort, while others seem to focus on different things. This is nicely reflected in the Tullock-contest where it depends on types whether efforts are, in equilibrium, strategic substitutes or complements.11

The remainder of this section is organized as follows. First, we will present the general model. We then turn to a discussion of the impact of a pure closed versus open access regime from an individual and social welfare perspective. We will conclude by discussing the main insights gained and the robustness of the model in terms of different specifications of the contest model.

The Model

There are two differently talented scientists i = H, L competing for publishing an article by exerting quality effort e i . The asymmetry in capabilities is modelled by assuming that the papers’ quality is q i  = θ i e i where θ i expresses author i’s talent. Without loss of generality, we normalize θ L  = 1 and set θ H  ≡ θ > 1 to capture the high type’s predominance. The probability of getting a paper published depends on quality, but only in a stochastic way to account for the unpredictability of the publishing game. The respective winning probabilities are therefore expressed by an asymmetric Tullock-contest, that is, 
$$p_{H} = \frac{\theta e_{H}} {\theta e_{H}+e_{L}}$$
$$p_{L} = \frac{e_{L}} {\theta e_{H}+e_{L}}$$

As most academics including ourselves enjoy doing research even if the outcome’s quality may be questionable, effort costs can best be seen as opportunity costs. We assume that the two types’ effort cost functions are the same, which expresses the view that research capabilities are not systematically related to the talent required for administrative work, consulting or teaching.

We consider two regimes k = C, O denoting closed and open access, respectively. By g i k , we define the submission fee covered by author type i under regime k. With academic copyright (closed access), we assume that there are no submission fees, that is g i C  = 0, i = H, L. This is counterfactual, but all we need is that submission fees are higher with open access. The part of the submission fee borne by authors with open access is denoted g i O  ≡ g i , and for the reasons discussed in the introduction we assume that g H O  = 0 while g L O  ≥ 0.

The authors’ benefit from publishing depends on the reputation of the journal and on readership which we denote as r k . Realistically, we assume that readership is (weakly) higher with open access, that is r O  = 1 ≥ r C .12

Summing up, the Tullock-contest can be described by the two types’ objective functions

$$\displaystyle\begin{array}{rcl} V _{L}& =& \left (r^{k} - g_{ L}^{k}\right )\left ( \frac{e_{L}} {\theta e_{H} + e_{L}}\right ) - e_{L}{}\end{array}$$


$$\displaystyle\begin{array}{rcl} V _{H}& =& r^{k}\left ( \frac{\theta e_{H}} {\theta e_{H} + e_{L}}\right ) - e_{H},{}\end{array}$$

which will be maximized with respect to e i .

Under both systems, the authors maximize their objective functions as given in Eqs. (4.1) and (4.2), respectively. Taking the first order conditions and solving for the equilibrium yields

$$\displaystyle\begin{array}{rcl} e_{H}^{{\ast}}& =& \frac{\left (r^{k}\right )^{2}\theta \left (r^{k} - g_{ L}^{k}\right )} {\left (r^{k}\left (1+\theta \right ) - g_{L}^{k}\right )^{2}},{}\end{array}$$


$$\displaystyle\begin{array}{rcl} e_{L}^{{\ast}}& =& \frac{r^{k}\theta \left (r^{k} - g_{ L}^{k}\right )^{2}} {\left (r^{k}\left (1+\theta \right ) - g_{L}^{k}\right )^{2}}.{}\end{array}$$


Recalling that r O  = 1 ≥ r C and g L O  ≥ 0 while g L C  = 0, we get the results summarized in Proposition 1:

Proposition 1.


All privately optimal effort levels are increasing in readership and decreasing in the two types’ heterogeneity, that is, 
$$\frac{\partial e_{i}^{k}} {\partial r^{k}} > 0$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq3.gif”></SPAN>, <SPAN id=IEq4 class=InlineEquation><IMG alt= .



With closed access, the two types’ effort levels are identical, e H C = e L C .



With open access, the high type’s effort is higher than the low type’s effort if the low type’s submission costs are positive, that is if g O > 0. Both effort levels are decreasing in the low type’s submission cost, 
$$\frac{\partial e_{i}^{O}} {\partial g^{O}} < 0\ \forall i$$


All proofs are provided in the Appendix.

The first part of part (i) is obvious as the marginal benefit from effort is increasing in readership. The second part of part (i) is a standard feature of Tullock-contests which says that effort incentives are decreasing in the contestants’ heterogeneity. Intuitively, the bad type decreases her effort as the (marginal) probability of winning is lower, and the good type responds accordingly as she is likely to win even with relatively low effort.13

Part (ii) says that, even so effort levels are decreasing in heterogeneity, both effort levels are identical in equilibrium if the heterogeneity refers solely to abilities. The reason is that the effort-decreasing impacts of heterogeneity are the same for both types. Thus, effort levels are the same for closed access.

Things are different with open access if and only if the low type bears higher submission costs (part (iii) of the Proposition). As the heterogeneity then refers to costs and not to abilities, the impact on the two types is no longer identical—both types reduce their efforts, but the bad type to a larger degree so. Assuming heterogeneity with respect to cost structures hence has different consequences from assuming heterogeneity with respect to abilities. Note that the low type’s submission costs do not only reduce his effort, but also the high type’s effort due to the strategic effect described for part (i) of the Proposition.

We now proceed by comparing the effort levels under the two systems, and we state the results in Proposition 2.14

Proposition 2.

(i) The low type’s effort is higher with open access than with closed access if and only if 
$$\frac{r^{O}} {r^{C}} > \frac{\left (1+\theta -g\right )^{2}} {\left (1+\theta \right )^{2}\left (1-g\right )^{2}}$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq8.gif”></SPAN> <SPAN class=EmphasisTypeItalic>. (ii) The high type’s effort is higher with open access than with closed access if and only if</SPAN> <SPAN id=IEq9 class=InlineEquation><IMG alt= \frac{\left (1+\theta -g\right )^{2}} {\left (1+\theta \right )^{2}\left (1-g\right )}$$ ” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq9.gif”> . (iii) Both types’ efforts are more likely to be higher under open access if the heterogeneity θ and the low type’s submission costs g are low. (iv) The ratio between the high and the low type’s effort is higher for open access.

Part (i) of the Proposition expresses the trade-off of the two regimes for the low type’s effort incentives: On the one hand, open access leads to a higher readership and thereby to a higher incentive for effort. The strength of this effect depends on the readership-ratio r O r C . On the other hand, the effort incentive is lower since part of the publishing costs must be borne privately by the low type. The strength of this effect depends on g. More interestingly, part (ii) says that the high type’s effort may also be lower with open access even though the readership is larger, and even though the high type pays no submission fees by assumption. As outlined after Proposition 1, this follows from the fact that efforts are strategic complements, that is, the high type reduces her effort in response to the low type’s lower effort provision.

Part (iii) of the Proposition shows that the two types’ heterogeneity expressed by θ has less impact on effort provision under closed access. The reason is that the contest’s additional asymmetry caused by the low type’s submission costs aggravates the negative impact of the asymmetry in abilities. The impact of the submission costs itself is straightforward.

Part (iv) is likely to be important from a social welfare-perspective as one might presume that the high type’s effort is socially more valuable (see the next section). Formally, the result is a straightforward implication of the fact that the two types’ efforts are identical with closed access, while the high type exerts relatively higher effort under open access.

Social Welfare

So far, we have only considered the private incentives under the two regimes, but we have not yet extended to social welfare. We use a utilitarian welfare function which is additively separable in the utilities of readers and authors. The authors’ (net) utilities follow directly from substituting their equilibrium effort levels into their objective functions. We neglect the low type’s submission costs under open access as these are purely re-distributive. Thus, it remains to specify the utility of readers which we define in a reduced form as

$$\displaystyle{ U =\beta r^{k}\left (q_{ H}^{0.5} + q_{ L}^{0.5}\right ). }$$


Such a utility function for readers seems quite natural and displays the following features: first, the two terms in brackets express that the readers’ utility is increasing at a decreasing rate in the quality of the articles. Second, the readers’ utility is higher for open access as r O  = 1 ≥ r C . This is straightforward as the audience for articles of a given quality will be higher it they can be downloaded for free.15 Note that, similar to the low type’s submission costs, we do not incorporate the prices of articles as they cancel out (publishing houses get what readers and authors pay). Thus, all that counts for articles of a given quality is readership itself. Third, β > 0 is just a factor expressing the average utility of readers from articles of a given quality.

Adding up over the utilities of readers and authors yields the social welfare function

$$\displaystyle{ \mathit{SW } =\beta r^{k}\left (q_{ H}^{0.5} + q_{ L}^{0.5}\right ) + r^{k}\left ( \frac{\theta e_{H}} {\theta e_{H} + e_{L}}\right ) + r^{k} \frac{e_{L}} {\theta e_{H} + e_{L}} - e_{H} - e_{L}. }$$


Recalling that q H  = θ e H and q L  = e L , and that the winning probabilities add up to one, the social welfare function is

$$\displaystyle{ \mathit{SW } =\beta r^{k}\left (\left (\theta e_{ H}\right )^{0.5} + e_{ L}^{0.5}\right ) + r^{k} - e_{ H} - e_{L}. }$$


Note that, as usual in contest models, the winner’s identity does not matter from a social welfare perspective. The socially optimal effort levels are given by maximizing SW with respect to e L and e H . We get

$$\displaystyle{ e_{H}^{f} = \frac{\left (r^{k}\right )^{2}\theta \beta ^{2}} {4},e_{L}^{f} = \frac{\left (r^{k}\right )^{2}\beta ^{2}} {4} \text{, and }\frac{e_{H}^{f}} {e_{L}^{f}} =\theta }$$


It follows that the high type should exert higher effort due to her higher (marginal) productivity in research. This could also be interpreted in the sense that universities should assign lower teaching and administration loads to highly qualified researchers which is the case in some universities and countries, but not in all. Second, due to higher readership (r O  = 1 ≥ r C ), socially optimal effort levels are higher under open access.

The relation between the privately and the socially optimal effort levels under the two regimes is expressed by Lemma 1.

Lemma 1.

(i) For closed access, the high type’s (the low type’s) effort is too low iff 
$$\beta > \frac{2} {\left (r^{C}\right )^{0.5}\left (1+\theta \right )}\ \left (\beta > \frac{2\theta ^{0.5}} {\left (r^{C}\right )^{0.5}\left (1+\theta \right )}\right )$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq10.gif”></SPAN> <SPAN class=EmphasisTypeItalic>. (ii) For open access, the high type’s (the low type’s) effort is too low iff</SPAN> <SPAN id=IEq11 class=InlineEquation><IMG alt= \frac{2\left (1-g\right )^{0.5}} {1+\theta -g} \ \left (\beta > \frac{2\theta ^{0.5}\left (1-g\right )} {1+\theta -g} \right )$$ ” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq11.gif”> . (iii) If 
$$\beta > \frac{2\theta ^{0.5}} {\left (r^{C}\right )^{0.5}\left (1+\theta \right )}$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq12.gif”></SPAN> <SPAN class=EmphasisTypeItalic>, then both efforts are too low under both regimes. (iv) For</SPAN> <SPAN id=IEq13 class=InlineEquation><IMG alt= , both efforts are too high under both regimes if 
$$\beta < \frac{2\theta ^{0.5}\left (1-g\right )} {1+\theta -g}$$
. For 
$$g < 1 -\frac{1} {\theta }$$
, both efforts are too high if 
$$\beta < \frac{2\left (1-g\right )^{0.5}} {1+\theta -g}$$

Most generally expressed, Lemma 1 shows that the relationship between the privately and socially optimal incentives for research depends on whether the utility of readers outweighs the rent-seeking impact of career concerns. Under both open and closed access, privately optimal efforts are more likely to be too low if the good type’s productivity increases, that is, if θ is high. The reason is that a higher productivity of the good type is fully reflected in social welfare, but reduces both effort levels due the contest’s increasing asymmetry. Moreover, under open access, private efforts are more likely to be too low when the low type’s submission costs are high as those reduce efforts for two reasons, the cost effect and the strategic effect from the contest’s asymmetry.17 Parts (iii) and (iv) of the Lemma state conditions ensuring that both efforts are too high or too low under either system. The case distinction in part (iv) is required as either the high or the low type’s effort may define the threshold for β.

We can now turn to a comparison of social welfare under closed and open access. Substituting the effort levels into the respective social welfare functions yields

$$\displaystyle{ \mathit{SW }^{C} =\beta r^{C}\left (\left ( \frac{r^{C}\theta } {\left (1+\theta \right )^{2}}\right )^{0.5} + \left ( \frac{r^{C}\theta ^{2}} {\left (1+\theta \right )^{2}}\right )^{0.5}\right ) - \frac{2r^{C}\theta } {\left (1+\theta \right )^{2}} + r^{C} }$$


$$\displaystyle\begin{array}{rcl} \mathit{SW }^{O}& =& \beta \left (\left (\theta \frac{\theta \left (1 - g\right )} {\left (1 +\theta -g\right )^{2}}\right )^{0.5} + \left ( \frac{\theta \left (1 - g\right )^{2}} {\left (1 +\theta -g\right )^{2}}\right )^{0.5}\right ) \\ & & -\frac{\theta \left (2 + g^{2} - 3g\right )} {\left (1 +\theta -g\right )^{2}} + 1. {}\end{array}$$


Regardless of whether the qualities provided are above or below the socially optimal ones, there are two advantages of open access: Readership is larger and efforts of the two types are different. Still, for analyzing the welfare ranking of the two systems, we need to distinguish between the situations where efforts are below or above the socially optimal ones. Proposition 3 refers to the first case:

Proposition 3.

Suppose all efforts are too low, i.e. 
$$\beta > \frac{2\theta ^{0.5}} {\left (r^{C}\right )^{0.5}\left (1+\theta \right )}$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq21.gif”></SPAN> <SPAN class=EmphasisTypeItalic>. Then, open access is more likely to be welfare superior if r</SPAN> <SUP>C</SUP> <SPAN class=EmphasisTypeItalic>, θ and g are low.</SPAN> </DIV></DIV><br />
<DIV class=Para>Recall first that we know from Lemma <SPAN class=InternalRef><A href=1 that all efforts are too low if 
$$\beta > \frac{2\theta ^{0.5}} {\left (r^{C}\right )^{0.5}\left (1+\theta \right )}$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq22.gif”></SPAN>. Then, social welfare is increasing in the two quality efforts. This given, the impact of the three variables is straightforward: first, efforts under closed access are increasing in readership <SPAN class=EmphasisTypeItalic>r</SPAN> <SUP><SPAN class=EmphasisTypeItalic>C</SPAN> </SUP>which makes it less likely that open access is superior.<A id=Fn18_source href=18 Next, a higher θ is more beneficial under closed access as, with open access, even the high type’s effort is decreasing in the low type’s publication costs. And the higher θ, the higher is the optimal effort that should be provided by the high type. For similar reasons, open access becomes less favorable when submission costs g are high. Note that, given that both efforts are inefficiently low, 
$$g \rightarrow 0$$
is a sufficient condition for the superiority of open access.

We now turn to the case where quality efforts are above the socially optimal ones. Proposition 4 shows that the welfare ranking of the two regimes is then clear-cut:

Proposition 4.

Suppose all efforts are efficient or too high, i.e. 
$$\beta < \frac{2\theta ^{0.5}\left (1-g\right )} {1+\theta -g}$$
$$g \geq 1 -\frac{1} {\theta }$$
$$\beta < \frac{2\left (1-g\right )^{0.5}} {1+\theta -g}$$
$$g < 1 -\frac{1} {\theta }$$
. Then, open access is superior to closed access.

For an intuition, let us neglect the difference in readership for a moment. Then, open access always yields lower equilibrium efforts due to the asymmetry in the coverage of publication costs. Hence, as efforts are above the socially optimal ones by definition of the case considered, open access is superior even when readership is the same. And as higher readership is always beneficial and is higher with open access, open access is superior. Thus, while the ranking between the two systems is ambiguous if effort levels are too low, it is clear-cut if they are too high.

Discussion and Conclusion

In conclusion, our model shows that there may be countervailing effects from shifting towards a pure OA regime. Assuming that scientists are also motivated by career concerns, we find that private research incentives can, from a social perspective, be too high when the rent-seeking motive outweighs the positive externalities of research. If incentives are too high, open access is always superior. The reason is that readership is larger while effort incentives are lower due to the asymmetry in privately borne publication costs. If quality effort incentives are too low, the welfare ranking of the two regimes depends on whether the higher readership under open access outweighs the detrimental impacts of asymmetric costs on effort provision. Our paper adds to the literature by analyzing the contest character of academic publishing. Accordingly, the conclusions we derive differ from the non-strategic model in Shavell (2010) where incentives for writing a paper can never above the socially optimal ones. We do not argue that our contest model is the only perspective in the debate. However, we do think that we add an important puzzle to the debate in academic publishing by emphasizing that there may be rent-seeking incentives for researchers in the predominant “publish or perish” environment in academic career advancement.

Furthermore, we do understand concerns that our results may be driven by the special characteristics of our contest model in place. That is the reason why we test the robustness of our model with respect to different contest model specification in Feess and Scheufen (2013).19 In particular, we present two model specifications. First, we model heterogeneity between the two types of researchers by means of different cost functions, i.e. our two researchers do no longer differ in their productivity but their cost of effort.20 The argument is then as follows: our two researchers face different costs when exerting quality efforts in writing an academic work. For modelling this heterogeneity in the cost functions we assume that the high type benefits from a cost advantage.21 Integrating different cost functions in Feess and Scheufen (2013), we show that the results of our main model do not qualitatively change. Second, we integrate type-specific readership. The argument behind this model specification is easy to grasp: as reading an article is costly, the utility a reader gains from reading an article of a give quality may be type-specific and higher for the high type researcher. This seems reasonable due to differences in the reputation of authors.22 We integrate the argument of type-specific readership by introducing a parameter of reputation in our social welfare function, allowing that reading an article of an author with high reputation comes along with a higher benefit.23 Again, our results reveal that our original contest model is robust with respect to the different variants of the contest model.24

Turning to further research, natural questions seem whether a coexistence of the two systems outperforms each standing alone system, and whether uncoordinated market behavior would induce such a coexistence. Accordingly, if a unique open access is superior, one might ask if markets will enforce such a system anyway or if the lock-in effect of the established closed access system will prevent such an evolutionary process. A the natural follow-up from our contest perspective would be how different researchers self-select to different contracts, and how those who “architecture” the contest would try to attract the best publications.

4.1.2 Academic Copyright in the States Game: An International Perspective


Our previous analysis shows that there may be countervailing effects of moving towards an open access regime in publishing academic works. In this respect, especially the consequences of a shift towards the “author pays” model needs to be investigated in more detail. Shavell (2010) circumvents the consequences for authors of having to bear the publication costs by simply assuming that universities and grantors would have a motive to subsidize publication costs in the absence of copyright. This may be true for the best universities and especially for universities located in the US. However, especially for middle- and low-class universities and eventually for universities outside the US, an OA regime would likely increase asymmetries between universities and countries, respectively. This aspect may particularly gain recognition in the light of developing countries.

This chapter seeks to briefly address the consequences of an “author pays” model by focussing on the international perspective—especially accounting for the differences in funding research between least developed countries and the developed world. Extending on our previous model, we will first describe the effects in a very simplified model, stressing to possible distortion effects between authors when shifting towards an OA mode. However, we refrain from doing a complex welfare analysis. Instead, we will investigate the consequences of a distortion between authors by means of a simple simulation, comparing different systems in a global science community. In this regard, especially the (funding) situation of researchers situated in a low or least developed country will be addressed.

The remainder is organized as follows. First, we will shows the basic model. We then derive the privately optimal quality efforts for the two regimes. We continue by comparing the effects of a shift towards the “author pays” model in the two country case. We conclude with some important policy implications that follow from our analysis.

The Model

Extending on the model in Feess and Scheufen (2013), we assume that the quality of an academic paper depends on effort and type and is simply q i  = θ i e i , where 
$$\frac{\partial q_{i}} {\partial \theta _{i}} > 0$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq29.gif”></SPAN> and <SPAN id=IEq30 class=InlineEquation><IMG alt= 0$$ ” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq30.gif”>.25 We consider two regimes k = C, O denoting closed and open access, respectively. By g i k , we include an index that captures the degree to which submission fees are borne by the author i, where g i O  ≧ 0 and g i C  ≡ 0. With r k we implicitly account for readership, which is r O  > 1 for open access and normalized to 1 for closed access. Let 
measure the readership advantage of OA, such that 
$$\varepsilon = r^{O} - 1$$
.26 The utility of author i in regime k is

$$\displaystyle{ V _{i}^{k}\left (e_{ i},\theta _{i}\right ) = \left (r^{k} - g_{ i}\right )\theta _{i}e_{i} - e_{i} }$$

which will be maximized with respect to e and θ.

Privately Optimal Effort Levels

Taking the first order conditions and solving for the equilibrium yields

$$\displaystyle\begin{array}{rcl} e_{i}^{{\ast}}& =& \frac{\left (r^{k} - g_{ i}\right )\theta _{i} - 1} {\left (r^{k} - g_{i}\right )}{}\end{array}$$


$$\displaystyle\begin{array}{rcl} \theta _{i}^{{\ast}}& =& \frac{\left (r^{k} - g_{ i}\right )e_{i} + 1} {\left (r^{k} - g_{i}\right )}{}\end{array}$$


Recalling that 
$$g_{i}^{O} \geqq 0 \equiv \ g_{i}^{C}$$
$$r^{O} > 1 \equiv r^{C}$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq36.gif”></SPAN>, we get the results as summarized in Proposition <SPAN class=InternalRef><A href=5.

Proposition 5.

(i) The private effort incentives are increasing in readership and decreasing in the submission fee, that is, 
$$\frac{\partial e_{i}^{k}} {\partial r^{k}} > 0$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq37.gif”></SPAN>, <SPAN id=IEq38 class=InlineEquation><IMG alt= . (ii) The capability variable is increasing in the submission fee and decreasing in readership, that is, 
$$\frac{\partial \theta _{i}^{k}} {\partial r^{k}} < 0$$
$$\frac{\partial \theta _{i}^{k}} {\partial g_{i}} > 0\ \forall i$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq40.gif”></SPAN>, <SPAN id=IEq41 class=InlineEquation><IMG alt= . (iii) Both variables do not only depend on g i and r k , but also on one another.

All proofs are provided in the Appendix.

Part (i) highlights the general trade-off when shifting towards an OA regime. Whether an OA or a CA model will more likely increase researchers’ incentives will depend on whether the increase in readership is able to outweigh the burden of having to bear the publication costs. Accordingly, an OA regime will only be superior as compared to a CA regime if and only if 
$$\Delta r^{O} - g_{i} = r^{O} - 1 - g_{i} > 0$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq42.gif”></SPAN>. <SPAN class=EmphasisTypeItalic>Part (ii)</SPAN> stresses that the extend to which high (low) qualified authors may react to a shift towards an “author pays” model.<A id=Fn27_source href=27 Part (iii) is obviously true, as the marginal effort level is always higher for high qualified authors and vice versa. Most interestingly, there may be different impacts on the incentives to publish if we let g vary between authors. Accordingly, the Lemmas 2 and 3 further investigate under which circumstances we will observe a distortion between authors and how this may affect the social welfare assessment of the OA regime.

Lemma 2.

For closed access, there is no distortion between authors.

Obviously, as the readers pay for the publication costs under CA, authors do not face any other costs but their own effort in writing an article of high quality. Thus, only the productivity variable θ provides for a natural differentiation. Accordingly, in a closed access regime authors with higher productivity (better type) will ceteris paribus always derive higher utility levels. Other than that, there is no distortion between authors under the CA mode (Lemma 2).28

Lemma 3.

For open access, it depends on the distribution of g whether we will observe a distortion effect. (i) For g i = 0∀i, there is no distortion for OA which is hence always superior due to r O > 1. (ii) For g i identical 
$$\forall i$$
, there is still no distortion. However, the effort incentives may be smaller if 
$$r^{O}\longrightarrow 1$$
and if g i is high. (iii) For different g i we have a distortion within authors, since even highly qualified authors (i.e. those with high θ) might not submit if g i is high, since the effort incentive depends on both θ and g.

Most generally expressed Lemma 3 reflects on the consequences for the basic trade-off in an OA regime when publication costs are or are not fully covered by all universities. In this regard, OA may induce a distortion between authors depending on the direct cost effect that researcher will have to bear. Thus, it will depend on the distribution of g over the group of researchers 
$$I =\{ i = 1;\ldots;i = N)$$
whether an OA regime will induce a distortion between authors. Part (i) reflects the simple case as assumed in Shavell (2010), i.e. all universities will cover the publication costs that result from shifting towards an OA regime for academic publications. In this case, there is no distortion between authors as 
$$g_{i} = 0\forall i$$
. As a consequence, the OA advantage amounts to the extend to which OA promises higher readership and hence to 
$$r^{O} - 1 =\varepsilon > 0$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq47.gif”></SPAN>. If we believe readership under an OA regime to be at least weakly higher as compared to the traditional publishing mode, OA will always be superior. In <SPAN class=EmphasisTypeItalic>Part (ii)</SPAN> publication costs are not fully covered, but the degree to which researchers will have to bear the costs themselves is equally distributed. In this scenario, there is still no distortion between authors, since the benefit of shifting towards an OA mode is <SPAN id=IEq48 class=InlineEquation><IMG alt=. Nevertheless, since the author will now partly bear the publication costs, the impact on researchers efforts decreases. Only if the OA readership advantage outweighs the publication cost disadvantage, efforts will be higher under OA, i.e. if and only if 
$$\varepsilon > g_{i}$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq49.gif”></SPAN> OA will be superior. Finally, we do observe a distortion between authors if the publication costs are not equally distributed between authors and hence if some universities (at least partly) cover the costs for their researchers, while others do not (<SPAN class=EmphasisTypeItalic>Part(iii)</SPAN>). Here, some authors will refuse to exceed higher efforts under OA if <SPAN id=IEq50 class=InlineEquation><IMG alt=\varepsilon$$ ” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq50.gif”>. Since authors incentives depend on both θ and g, also highly qualified authors may refuse to submit their paper. Clearly a worse case scenario from a social welfare point of view. To see the consequences especially in the international research arena, we will now turn to a simple two country comparison.

Simulation: The Effect of the “Author Pays” Model in a Heterogeneous World

In a truly heterogeneous world, a shift towards an OA publishing mode may likely result in a distortion between authors. Most importantly, this may be an issue between researchers of countries at different stages of economic development. Especially since some countries may not be able or willing to provide funding for each journal publication. A simple simulation model may shed some light on the effect on researcher’s incentives located in different countries and choosing between the OA and CA regime.

Suppose that the utility of an author i of country j in regime k is 
$$V _{\mathit{ij}}^{k}\left (e,\theta \right ) = \left (r^{k} - g_{j}\right )\theta _{\mathit{ij}}e_{i} - e_{i}$$
, where the home country status of author i decisively regulates the portion of submission fees born by herself. Further assume a two country and two author case, say one researcher from a developed or industrialized (IC) country and one from a developing country (DC). While the researcher from the industrialized country receives a full waiver for the publication fees of her paper submission, the researcher in the developing country has to bear part of the submission costs herself. As a result, assume that 
$$g_{\mathit{IC}}^{O} = g^{C} = 0$$
and g DC O  = g O  > 0. Also recall that 
$$r^{O} = 1+\varepsilon > r^{C} = 1$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq53.gif”></SPAN>, where <SPAN id=IEq54 class=InlineEquation><IMG alt= denotes the readership advantage of the OA mode. For simplicity and without loss of generality we normalize θ and e to 1. The incentives scheme is summarized in Fig. 4.1.


Fig. 4.1
Distortion effects between countries: Scenario 1

Obviously, with only some states paying for the publications of their university researchers, a situation results where OA is only increasing researchers’ incentives to the extend of the difference between the positive impact of a higher readership under the OA regime, and the negative effects on those authors who will not receive sufficient financial resources. In the two country case this induces an environment that can be best described as follows: The researcher living in an industrialized country (IC) will benefit from OA by the full readership effect. Under the veil of ignorance she would choose an OA mode since 
$$\varepsilon > 0$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq55.gif”></SPAN>. In contrast, the researcher from the developing country (DC) will only gain net benefits if the readership advantage outweighs her costs for publishing and hence if and only if <SPAN id=IEq56 class=InlineEquation><IMG alt=\varepsilon$$ ” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq56.gif”>. Taking into account that most developing countries will unlikely have the financial means to cover any of these costs, researcher from these countries would prefer to publish their papers in CA journals, since 
$$0 >\varepsilon -g$$
” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq57.gif”></SPAN> for <SPAN id=IEq58 class=InlineEquation><IMG alt=\varepsilon$$ ” src=”/wp-content/uploads/2015/10/A330421_1_En_4_Chapter_IEq58.gif”>.

Now suppose that the two countries do not only differ with respect to g. Recalling the assumption in Feess and Scheufen (2013) that the best researchers will be employed by the best universities and that mediocre universities will most likely not fully pay the publication costs under OA, may even enhance our observation of a distortion effect between authors of different origin. We can easily find evidence for the fact that the best universities are located in the industrialized countries or especially in the US. The top 100 university ranking of 2012 lists 56 universities from the US.29 Not a single university from a developing country is listed in the top 500. Only a few universities from emerging economies like Brazil (with 6 universities) and India (with only the Institute of Science listed in the top 400) are represented. In fact, the ranking is highly dominated by US universities, with 8 (80 %) universities from the US listed in the top 10 or 36 (72 %) in the top 50. As a consequence, assuming that θ now differs between countries and that the best researchers will tend to be located in an industrialized country—that is we assume θ IC  > 1 = θ DC —is reasonable. This leaves us with the outcomes as summarized in Fig. 4.2.


Fig. 4.2
Distortion effects between countries: Scenario 2

Accordingly, the benefit of a researcher in an industrialized country is 
$$\varepsilon \theta _{\mathit{IC}}$$
when choosing OA, whereas OA for the DC researcher still comes along with an outcome of 
$$\varepsilon -g <\varepsilon \theta _{\mathit{IC}}$$
. The fact that eventually only the US may be endowed with the financial means to (1) fully subsidize OA publishing and (2) attract the best researchers from all over the world may even exacerbate the dilemma in the states game. Thus, the danger of reinforcing the digital divide between industrialized and developing countries or even inducing a brain drain—even from countries like Germany or other states endowed with a more rigid budget than the US30—to the US shows that a shift towards a global OA regime is in need for comprehensive investigation also on distributional issues.

Policy Implications

Our analysis suggests that the biggest obstacle in shifting towards an universal OA publishing mode may be the funding scheme. Most importantly, having the author to bear the publishing costs by means of an “author fee” would likely reduce the interest for publishing research results in the lack of sufficient funding. Unless discounts are available to authors from low or least developed countries or external funding is provided to cover the publishing cost, article processing charges could exclude authors from some nations or less well-funded research fields from publishing in OA journals.

As a consequence, the two worlds (OA and CA) create a dilemma that is somewhat twofold: While OA lowers the access barriers for researchers of countries who have been hardly able to subscribe to a single journal in the past, it necessarily creates a participation constraint as it sets a price for participation in the publishing game. We have seen that OA journals do not necessarily charge author fees, but are financed by means of other income sources such as grants, print subscriptions or advertisements.31 Moreover, many OA publishers have realized the dilemma of authors from developing countries and offer discounts or waivers to authors suffering from financial hardship. In this regard, PLoS offers a fee waiver policy that allows to waive or reduce the payment for authors from low or middle income countries. Eligible countries are distinguished in two groups32: Group one countries are not charged a publication fee.33 Countries that are listed as group two members are asked a reduced fee at a flat tax of $500.34

Furthermore, OA still restricts access to scientific knowledge based on sufficient means of IT infrastructure. That is, online access is subject to internet access. We will see later that the poor information and technology infrastructure of countries in the developing world, especially in Sub-Saharan-Africa and rural areas, may be the most influential factor that prevents from bridging the digital divide between the developed and developing world. Accordingly, policy makers will have to account for both the participation constraint (funding of publication costs) and the access constraint (sufficient IT infrastructure) when formulating and codifying OA policies on a global scale.

4.2 The Political Economy of Access to Scientific Knowledge

The implementation of OA publishing in a truly global science community and the organizational implications for academic publishing will be in need for a deep understanding and comprehensive analysis of the legal as well as politico economic transition ahead. This section will analyze the obstacles from a legal, economical and political perspective. First, we will further investigate the options and further needs for implementing OA policies in the international copyright framework. In this context, especially the role of developing countries as “leaders” for the implementation of a more flexible copyright framework and codification of soft law declaration will be highlighted. Second, the relevance and opportunities of OA publishing for developing countries and its implications from the perspective of development economics will be further investigated.

4.2.1 On the International Political Economy of Access in Science


All of our previous analysis shows that OA may eventually provide with a better suited publishing model as compared to the traditional model, especially as it may more directly account for the needs of researchers. Nevertheless, consciousness regarding possible distortion effects within authors and also between countries will be needed in formulating an appropriate policy agenda. The legal framework for a possible reform of the academic publishing market is primarily determined by copyright law. We have seen that researchers typically transfer their copyright—or more specifically their commercial right—to the publisher when being accepted for publication. As a result, it is the copyright that enables publishers to sell journals as a bundle of academic works.

This section analyzes the options for a reform of copyright in the international arena and alternative legal origins for implementing the principles of OA in academic publishing. As we have seen, OA comes along with different costs and benefits for researchers of different countries. As a result, a broader picture on how to design the future of academic publishing is required to account for the distinguished needs and specific circumstances of developed and low or least developed countries. The remainder is organized as follows. Section “The International Copyright Law Framework” will provide a general understanding on the international legal framework. In section “Implementing OA in the International Arena” we further investigate the options for a reform of copyright and contract law. We will conclude in section “Harmonizing Copyright Law on a Global Scale” by presenting an agenda on how to proceed, clearly highlighting the need for further harmonization of international copyright law standards.

The International Copyright Law Framework

Analyzing the options for legal reform in the international copyright framework requires a general understanding on the principles of international copyright law. We have seen that copyright is territorially based. Thus, there is nothing like an international copyright law. Indeed, the relevant treaties leave it to the national legislation to install a copyright system according to the specific national needs.35 The TRIPS Agreement and the WIPO Copyright Treaty may be rather seen as guidelines and minimum standards that national legislation shall fulfil. This sovereignty of national legislation has important implications for our later investigation on the option for copyright reform. As a result, this subsection shall recall the two basic traditions in dealing with a limiting of copyright protection. Furthermore, we will elaborate on the general applicability of the (Berne) three-step test for implementing limitations or exceptions of copyright at the international level.

Two Approaches: Continental Europe vs. US Copyright Law

As already outlined in Sect. 2.​1 of Chap. 2, there are two distinct approaches for limiting copyright in scope. On the one hand, the approach in continental Europe (henceforth the European model) which has evolved on the notion of authorship, highlighting that the rights of personhood and moral rights shall dominate policy making for copyrightable works. On the other hand, the US approach (henceforth the US model) which has been driven by the utilitarian notion of copyright. Thus, copyright design follows a balancing of the costs and benefits associated with any use. As a consequence, two distinguished systems or traditions of copyright build the framework for limiting copyright in scope and hence for introducing an OA exception in the predominant national copyright laws.36

The European model manages copyright limitations by means of a list of enumerated exceptions or a closed catalogue. With other words: the copyright laws in Europe explicitly list a number of exceptions where the user may circumvent the principle of authorization. But at the same time this approach formulates the understanding that all other activities not covered in this list are usually proscribed (Reichman and Okediji 2012, p. 1376). These codified exceptions are updated on a regular basis as new technologies change the environmental conditions and possible uses of copyright material. Moreover, courts have tended to interpret these closed catalogues narrowly (Geiger 2010b, pp. 519 et seq.).

The US model, in contrast, combines the rigid catalogue model with a broad and flexible “fair use” doctrine (sections 106–122 U.S.C.). Similarly, the English copyright law allows for the principle of “fair dealing”. Following the utilitarian notion, the “fair use” principle allows a flexible balancing of the benefits and costs connected with a particular use of copyright material. Recalling the reasoning of Fig. 2.​1 in Chap. 2, any use is considered to be fair and hence non-infringing as long as its benefits are able to outweigh its costs. The four statutory factors under section 107 U.S.C.—(1) the purpose and character of the use, (2) the nature of the copyrighted work, (3) the amount and substantiality of the portion used and (4) the effect of the use upon the potential market for or the value of the copyrighted work—assist in measuring the respective costs and benefits.

As a consequence, when investigating the opportunities for implementing the OA principle in domestic legislation, an understanding of the differences between the systems (civil law versus common law countries) as well as between national copyright laws is important. Despite these still prevalent differences, the (partial) harmonization of international laws by means of treaties (TRIPS, WIPO Copyright Treaty) has induced a melding of the two approaches. Most important in this harmonization process was the introduction of the (Berne) three-step test.

International Law: The Three-Step Test and Beyond

The first and most important platform for copyright harmonization throughout the twentieth century has been the Berne Convention. Established in 1886, several revisions and reforms (Paris 1896, Berlin 1908, Berne 1914, Rome 1928, Brussels 1948, Stockholm 1967 and Paris 1971) have since supplied with the primary harmonization platform, also for the implementation of limitations and exceptions in the signatory countries. For instance, the revision of 1948 in Brussels introduced an exception “for excerpts from literary and artistic works in educational or scientific publications” in Article 10(2) of the convention (Reichman and Okediji 2012, p. 1378). However, a more general term for the regulation of the limitations or exceptions in the national copyright laws was not implemented until the revision of Stockholm in 1967.

The Stockholm Revision Conference in 1967 formally incorporated an exclusive reproduction right into Article 9(1) and simultaneously subjected it to a three-step test under Article 9(2) (Reichman and Okediji 2012, p. 1379). The text of Article 9(2) says that “it shall be a matter for legislation in the countries of the Union to permit the reproduction of such works in certain special cases, provided that such reproduction does not conflict with a normal exploitation of the work and does not unreasonably prejudice the legitimate interests of the author” (Berne Convention, Article 9(2)).37 Thus, the test puts forward three steps that need to be fulfilled for limitations on exclusive rights. First, the exclusive right of the copyright owner may only be limited in certain special cases. Second, a use may be legitimate if it does not conflict with the normal exploitation of the work. Third, the use shall not unreasonably prejudice the legitimate interests of the author. In the years after the Stockholm conference the three-step test has been transplanted to other international treaties. Nowadays, the test appears also in the Agreement on Trade Related Aspects of Intellectual Property Rights (Article 13), the WIPO Copyright Treaty (Article 10) and the WIPO Performances and Phonograms Treaty (Article) (Schonwetter 2007; Reichman and Okediji 2012, pp. 1389 et seq.; Senftleben 2006, pp. 411 et seq.; Kur 2009, pp. 302 et seq.). The wording of the provisions experienced slight modifications compared to the original Berne three-step test. Accordingly, the text of Article 13 of the TRIPS Agreement states that “members shall confine limitations or exceptions to exclusive rights to certain special cases which do not conflict with the normal exploitation of the work and do not unreasonably prejudice the legitimate interests of the right holder” (TRIPS Agreement, Article 13).38 The wording almost reads the same compared to the original, however, referring in its last part to the interests of the right holder, not the author. Despite the incorporation of the test in all relevant international IP treaties, there is still considerable uncertainty as to the actual meaning of the test. In practice, only one court case exists that may guide our interpretation and application of the three steps as laid forward under Article 13 of TRIPS.39

A promising road in interpreting the three-step test more openly and also more in the light of a “more economic approach” is a recent proposal from the Max Planck Institute for Intellectual Property and Competition Law in Munich (published in Geiger et al. 2008). In a nutshell, the proposal emphasizes the need for a more nuanced balancing between interests of authors and the broader public. In doing so, Geiger et al. (2008) follow the lines of the preamble of the WIPO Copyright Treaty, particularly mentioning education, research and access to information. In particular, the proposal highlights six issues for the future applicability of the three-step test: First, the three steps should be considered as a whole in a comprehensive and overall assessment. There should be no prioritization of any one step or the need for an affirmative answer to all steps. Instead the test should forward a “judicial balancing” of the different factors in the tradition of the US fair use approach (Reichman and Okediji 2012, p. 1454).40 Second, limitations and exceptions should not be interpreted narrowly, but their objectives and purposes should explicitly be taken into account. Third, the criterion to restrict limitations and restrictions to certain special cases should allow (a) policy makers to also implement open ended limitations and exceptions as long as its scope is reasonable foreseeable or (b) courts to apply existing statutory limitations and exceptions or add further some where possible in the legal system (Geiger et al. 2008, p. 121). Fourth, the test should seek to foster competition, especially in secondary markets (Reichman and Okediji 2012, p. 1455). Fifth, in applying the test both the interests of the original and subsequent rightholders should enter the assessment. Sixth and most interestingly from our viewpoint, also the interests of third parties should be considered in the application of the test, including (a) interests stemming from human rights and fundamental freedoms, (b) interests in competition and (c) other public interests, with a particular emphasis on scientific progress as well as cultural, social and economic development (Geiger et al. 2008, p. 121; Reichman and Okediji 2012, p. 1454). Especially the last aspect will need particular emphasis when implementing OA in the international arena.

Implementing OA in the International Arena

In the following we will discuss two basic approaches for implementing the OA principles in the international arena. In particular, we find options for promoting the OA principles by a reform of (1) copyright and (2) contract law. In the end, both strategies may complement one another and hence forward an argument in favor of a mixed approach for the future of OA in the scholarly publishing market.

By Legislation

One option to promote open access publishing and to foster an enduring evolutionary process towards a better balancing between the interests of authors, publishers and the society as a whole is by means of a change in legislation. In this regard, we will concentrate on reforms of two promising legislative branches: First, the ability for implementing the OA principles by means of the introduction of limitations and exceptions in the international copyright system. In doing so, we will elaborate on the previously discussed three-step test, its interpretation and the openness of the balancing test as discussed by several legal scholars (Reichman and Okediji 2012; Hugenholtz and Okediji 2008; Ginsburg 2001; Geiger 2007, 2010b; Senftleben 2006; Geiger et al. 2008). Second, we will also discuss possible reforms for the prevailing licensing models in the international arena. We will see that the current licensing models show significant weaknesses in satisfying the multiple requirements in a purely global science community. In the end, we will discuss the arguments for and against the proposal by Metzger (2012) to apply a “lex mercatoria” approach.

Reform of Copyright

By discussing adjustments of the international copyright framework to accommodate the needs of science,41 Reichman and Okediji (2012) highlight two important arguments which may provide hope for overcoming the previously discussed rigidity of the international legal framework: First, both the TRIPS Agreement as well as the WIPO Copyright Treaty contain specific deference provisions that deliberately leave room to maneuver and to adjust the law according to the national needs and policy (Reichman and Okediji 2012, p. 1452). In particular, Art. 1.1 of TRIPS explicitly highlights that the “members shall be free to determine the appropriate method of implementing the provisions of this Agreement within their own legal systems and practice” (TRIPS 1994, Art. 1.1). Also Art. 14(1) of the WIPO Copyright Treaty stresses that its members “undertake to adopt, in accordance with their legal systems, the measures necessary to ensure the application of this treaty” (WIPO Copyright Treaty, Art. 14(1)).42 Second, the flexibility built into TRIPS and the WIPO Copyright Treaty provide with means to not only strengthen copyright, but also to limit its scope by introducing new limitations and exceptions along with other balancing features (Reichman and Okediji 2012, p. 1452). Okediji (2007) emphasizes that the TRIPS Agreement eventually sets—for the first time in the history of copyright—important limits on the scope of copyright protection. Both aspects make us confident to believe that the implementation of OA in the international arena should be feasible. Nevertheless, primarily two steps shall be necessary for a more flexible framework.

First and foremost, a reinterpretation of the three-step test beyond its prevailing narrow character is probably the most important obstacle towards limiting the scope of copyright in favor of an OA culture in academic publishing (Reichman and Okediji 2012, p. 1453). Nevertheless, there may already be evidence that the existing legal framework may exhibit flexibility features. Senftleben (2010) believes that the three-step test already sets forth open-ended factors and hence allows for the flexibility needed, without undermining the ability of national legislation to implement general exceptions for scientific and educational purposes. In particular, he stresses that “the cultural innovation cycle supported by copyright law requires both rights [freedom and protection] to be broad enough to spur investment and creativity, and limitations broad enough to provide sufficient breathing space” (Senftleben 2010, p. 526). Several recent decisions by German courts on the three-step test as codified under Art. 10 of the WIPO Copyright Treaty eventually show the space for a more liberal construction of such limitations, explicitly taking into account constitutional arguments (Geller 2010; Reichman and Okediji 2012, p. 1376). A WTO panel report from 2000 states the view of the USA that “Article 10 of the WIPO Copyright Treaty […] reflects the standard set forth in Article 13 of the TRIPS Agreement” (WTO 2000, 6.67, p. 25). This makes us confident to believe that a more flexible application of the three-step test—more in the light of Art. 10 WIPO Copyright Treaty—should be feasible. Nevertheless, the fact that EU copyright legislation has been ignoring the opportunities for a more flexible assessment still shows the path dependence in adapting copyright law to the challenges of the digital environment.43

A promising road in this regard seems the declaration by the Max Planck Institute as outlined before, providing guidance for judges applying the three-step test.44 The implementation of the six features on the agenda seems not that far from being realistic. Especially the sixth proposal—broadening the view of the three-step test to also include the interests of the general public, and not just the interests of the right holder as codified in the third step of the test under Art. 13 of the TRIPS Agreement—seems to not only explicitly consider the economic rationale of copyright as outlined in Sect. 2.​1, it also goes along with prevailing legal interpretations for limiting the scope of intellectual property rights. As a matter of fact, a similar extension has already been implemented with regard to patents.45 Here, Art. 30 of the TRIPS Agreement extends the three-step test to patent law while adding the phrase “taking account of the legitimate interests of third parties” (Reichman and Okediji 2012, p. 1454). The declaration of the Max Planck Institute takes up on this fact, stressing that the absence of such a phrase in the “copyright test” does not detract from the necessity of explicitly taking account of such interests. In contrast, the signatories of the declaration emphasize that it “indicates an omission that must be addressed by the judiciary” (Geiger et al. 2008, p. 119). Also in this vein can be seen the proposal by Senftleben (2010) to establish an EC fair-use doctrine on the basis of the three-step as codified under Art. 5(5) of the EC Copyright Directive. In particular, Senftleben (2010) proposes a reinterpretation of the three-step test by means of a “refined proportionality test” that allows for enough space for unauthorized use within reasonable limitations.46 In conclusion, only a more flexible approach—especially in the EU copyright jurisprudence—for adjusting the scope and hence the limitations and exceptions to copyright law will allow policy makers to foster OA publishing in the international arena.47

Second, Reichman and Okediji (2012) emphasize the role of developing countries for leveraging the objectives as layed down in the WIPO Development Agenda. At its heart the 45 recommendations of the agenda stress the goal of bridging the knowledge and technology gap between industrialized and developing countries (Hugenholtz and Okediji 2008, p. 8). In this regard, the drafted treaty on “Access to Knowledge” (A2K) may be a sufficient orientation on how to reach the goal to “protect and enhance [expand] access to knowledge, and to facilitate the transfer of technology to developing countries” (A2K Treaty, Art. 1–1).48 As outlined before, the draft treaty emerged throughout the efforts of Brazil and Argentina and was primarily intended to ease the transfer of knowledge to the developing world (Opderbeck 2007, pp. 113 et seq.). During a series of meetings in 2005 a draft treaty was prepared by representatives from developing countries as well as representatives from the UK and the US, including a catalogue of exceptions to copyright that essentially mirror the “fair use” or “fair dealing” concepts. In particular, the “A2K” treaty would generally limit copyright law akin to existing compulsory licensing provisions (Opderbeck 2007, p. 115; Helberger 2005). Furthermore, the “A2K” draft treaty includes sections on the limitation of digital right management (DRM) systems (article 3-6), copyright term extension (article 3-9) and compulsory licensing of copyrighted works in developing countries (article 3-12). Part 5 of the draft treaty further specifies ways for “expanding and enhancing the knowledge commons”. Accordingly, any work “resulting from government-funded research shall be publicly available at no charge within a reasonable time frame, subject to reasonable exceptions, for example, for classified military research, for patentable discoveries, and for works that generate revenue for the author, such as books” (article 5.2(a)). In addition, a knowledge commons committee (KCC) shall “promote cooperation and investment in databases, open access journals and other open knowledge projects that expand the knowledge commons” (article 5-1). Within two meetings in February and June 2006, finally, the committee chair proposed to move forward on proposals that had received consensus support. This proposal was rejected by the developing countries that claimed significant IPR reforms. In fact, the proposal followed primarily the interests of the US and the European states which were considered as a back-room maneuver by developing countries (Opderbeck 2007, pp. 116 et seq.).

For the future reform process and success in reaching the objectives of the WIPO Development Agenda, several scholars have been pointing to the special role of developing countries for claiming more flexible measurements. Reichman (2009) argues that developing countries should rather lead than follow in the process of reform, triggering a codification of users’ rights at the international level or at least a soft-law instrument to be adopted at WIPO (Reichman and Okediji 2012, p. 1457).49 Hugenholtz and Okediji (2008) argue that “a joint initiative between WIPO and the WTO could be an ideal and appropriate expression of a soft-law modality with real impact for collective action on an international instrument on L&E’s” (Hugenholtz and Okediji 2008, p. 49). For instance, Hugenholtz and Okediji (2008) point to the “Joint Recommendation Concerning Provisions on the Protection of Marks and Other Industrial Property Rights in Signs, on the Internet” and the “Joint Recommendation Concerning Provisions on the Protection of Well-Known Marks”, both adopted by the Assembly of the Paris Union for the Protection of Industrial Property and the General Assembly of the WIPO.50 Despite the fact that neither of these provisions are binding, especially the latter has been evolving towards an international standard as it has been incorporated in several bilateral agreements by the US (Hugenholtz and Okediji 2008). In this regard, the draft of the A2K treaty may provide a useful pattern in developing a similar provision concerning copyrights. Last but not least, Hugenholtz and Okediji (2008) point to the special role of the “Standing Committee on Copyright and Related Rights” (SSCR) to contribute to developing coherence in the international copyright framework. Working towards a coordination between SSCR and the TRIPS council would likely foster a harmonization process in the international copyright arena and could eventually forward an instrumental framework for reaching collective action despite the sovereignty of national states.

In conclusion, both a recodification of the international three-step test (in accordance to the US fair-use principles) in combination with a reform process that incorporates the needs of developing countries (by incorporating users’ rights provisions at the international level) is decisive for adjusting the international copyright framework to accommodate particular OA principles for academic publishing. Nevertheless, the promotion of an OA mode in academic publishing will not be effective by copyright reform alone. In fact, a reform of the available licensing models but also contractual deliberations may be required.

Reform of Licensing Models

Even within the copyright system as a “all rights reserved” regime, alternative modes have been developed which offer legal tools or instruments for a self-enforcement of a “some rights reserved” regime. These instruments use the general licensing options as codified under the law.51 In this regard, most OA journals but also self-archiving platforms revert to the so-called Creative Commence licenses. We have seen that the “Public Library of Science” (PLoS) and BioMed Central as the leading OA publishers, for example, apply the so-called Creative Commons Attribution License (CCAL). The CCAL allows authors to retain their copyright, but allows anyone to download, reuse, reprint, modify, distribute and/or copy articles from the respective journal. Springer Open Choice,52 as a commercial publisher, uses the Creative Commons Attribution Non-Commercial License, which allows readers to read, copy and distribute a work and to create derivative works for non-commercial purposes. In general, the CC licenses offer a means to specify certain usage rights. In doing so the CC scheme operates on the basis of four general licensing elements, whose combination form a scheme of six different licenses. Figure 4.3 summarizes the six licensing models (red frame) as well as the basic license elements (black frame) as offered by CC.


Fig. 4.3
The creative commons licensing scheme

Accordingly, there are four general licensing elements that form six different licenses on three levels. The most basic license on the first level is the CC Attribution License (CC BY). With this license the right holder lets others distribute, remix, tweak and build upon her work, even commercially, as long as attribution to the original creator is guaranteed. This “attribution” element also forms the basic element for all combinations on the second and third level in the licensing scheme. As a result, on the second level three licensing schemes can be distinguished that constitute combinations of the attribution element with each of the three other terms. First, the Attribution-No Derivative Works (CC BY ND) licence which allows others to distribute the work, even commercially, as long as the work is unchanged and attribution to the original creator is guaranteed. Second, the Attribution-Noncommercial (CC BY NC) licence allows others to distribute, remix and build upon the work only if the use is not intended to follow any commercial purpose and if the original creator is credited. Third, the Attribution-Share Alike (CC BY SA) licence lets others use—i.e. distribute, remix and build upon—the work even for commercial purposes as long as the credit of attribution to the original author is assured and the resulting new works are shared under the same licensing terms. With other words, this license allows to share the material as long as new works are also shared in return. The third level, finally, depicts combinations of three of the basic terms, always accounting for attribution as the basic element. As a consequence, two licenses can be distinguished. The Attribution-Noncommercial-Share Alike (CC BY NC SA) licence allows any non-commercial use of the work subject to attribution and the sharing of new works under the very same terms. The Attribution-Noncommercial-No-Derivatives (CC BY NC ND) licence instead allows any non-commercial use of the work as long as the original creator is credited and the original work is not adapted or changed in any way. As a matter of fact, a licence cannot feature both the Share-Alike and No-Derivatives terms as the Share Alike requirement applies only to derivative works.53

A special feature of the CC model is that it eventually seeks compliance with the international copyright framework. In particular, CC International has been working on an international license porting project since 2003, seeking to port CC licenses to different copyright legislations around the world (Maracke 2010, p. 6). In the porting process the original license is basically modified to reflect local nuances in two respects, the legal terms and the language. That is, the licensing terms are translated into the respective local language(s) and adapted to the legal culture to reach compliance with the national legal requirements. CC International offers guidelines for this process by means of a detailed ten-step program, explaining in detail how to proceed from the forming of a porting team to the final launch of the national version of the licenses.54 By now, more than 70 jurisdictions offer ported national versions of the CC licenses. Among the first countries to adapt the CC licensing terms to national requirements were Brazil, Germany and the Netherlands in 2004. All ported licensing schemes are upgraded on a regular basis, currently available in version 3.0. Version 4.0 is currently in preparation.

Despite the wide spread of ported licensing versions, several problems remain that are mainly concerned with issues of internationalization and the constraints that the principle of territoriality imposes on the applicable choice of law. In this regard, some ported licensing version even comprise a “choice of law clause” that refers to the law of the given jurisdiction. This is e.g. the case for the latest licensing versions in Austria, England, Germany, Romania, Switzerland and Wales. However, the licensing versions of France, Netherlands, Poland and Spain do not comprise such a choice of law clause (Metzger 2012, supranote 35).55 Elkin-Koren (2006) highlights the problems associated with the territoriality of CC licensing for managing licensed content on a global scale, emphasizing that “the lack of standardization in the licenses supported by this licensing scheme, further increase the cost of determining the duties and privileges related to any specific work. This could further increase the chilling effect of copyrights” (Elkin-Koren 2006, p. 17).56 Metzger (2012) frames the legal difficulties associated with ported licenses by means of a simple illustration: “A is a historian at the university of Bucharest. He has created a database of Jewish cemeteries in Central and East Europe consisting of some hundred entries with maps, photographs and descriptions in different languages. A wants to share the database with other interested researchers in Romania and abroad. After visiting www.​creativecommons.​org, he chooses the Creative Commons Attribution-ShareAlike Version 3 Romania. The license text is in the Romanian language. According to Section 8 lit. f) Romanian law is applicable. B from Berlin finds the database on the Internet. He makes a number of important entries on cemeteries in Germany and wants to make this modified version available on his private website. Unfortunately, B does not read Romanian. In this case, B would be worse off as compared to the use of the ‘unported’ license version because he would have to translate the license terms before reading them. It could even be that under German contract law, standard terms in languages which may not be expected to be understandable for contracting parties may be unenforceable, especially in the case of consumers” (Metzger 2012, p. 364). As any use via the internet tends to cross borders, ported licenses may impose extremely high costs and legal uncertainty (Maracke 2010, p. 12). CC takes the prevailing problems with ported licenses explicitly into account in the process of launching version 4.0. In this regard, the list of objectives and goals includes internationalization, seeking to “further adapt the core suite of international licenses to operate globally, ensuring they are robust, enforceable and easily adopted worldwide.”57

A strategy to avoid the various problems associated with national licensing versions and comprising a choice of law clause is to provide unported licenses. CC does provide such unported licenses which may be used for jurisdictions that do not offer a ported licensing version (Metzger 2012, p. 364). The idea behind unported licenses is to ease the global management of creative content by means of just one generic license text for worldwide use. In doing so, unported licenses refer to internationally accepted terminology as codified under the respective IP conventions. Accordingly, the legal code of the CC-BY 3.0 unported license, for instance, specifies under section 8 f. that the “rights granted under, and the subject matter referenced, in this License were drafted utilizing the terminology of the Berne Convention for the Protection of Literary and Artistic Works (as amended on September 28, 1979), the Rome Convention of 1961, the WIPO Copyright Treaty of 1996, the WIPO Performances and Phonograms Treaty of 1996 and the Universal Copyright Convention (as revised on July 24, 1971). These rights and subject matter take effect in the relevant jurisdiction in which the License terms are sought to be enforced according to the corresponding provisions of the implementation of those treaty provisions in the applicable national law” (CC, Art, 8 f.).58 Nevertheless, also unported licenses may come along with drawbacks, especially when having to decide which law to apply in front of the court. When licensor and licensee belong to different jurisdictions, we may be left with very similar problems as before. As a consequence, this shows an important reference to the previously raised arguments for a reform of copyright at the international level.

A workable solution for these problems has recently been provided by Metzger (2012) and can be somehow seen in the tradition of the previously discussed soft law declaration. Metzger (2012) argues that unported licenses should not be governed by national laws, but should follow a lex mercatoria approach. The basic idea behind the lex mercatoria theory is that throughout the history of international trade a body of internationally customary rules has evolved that effectively regulates trade independent from any specific national law. The old lex mercatoria basically consisted of unwritten law that was applied by the medieval courts of admiralty (Metzger 2012, p. 365).59 The modern lex mercatoria instead refers to a set of model rules such as the “UNIDROIT Principles for International Commercial Contracts”60 to govern international commercial transactions (Marrella and Yoo 2007, pp. 817 et seq.).61 Especially in this las characteristic of lex mercatoria—the international contractual practice—Metzger (2012) sees the required theoretical link between lex mercatoria and open source/access communities.62 Most importantly, such a lex mercatoria licensing system would provide a neutral terminology in the licenses beyond any choice of law clause or need for an application of national law. By contrast, the licensing terms would be based on the community structure. Universal rules and norms that follow the mechanisms and customs of academic publishing. Lex mercatoria licensing models would offer general principles comparable with standard contracts in the lex mercatoria debate. In this regard, the UNIDROIT principles could serve as fall back provisions (Metzger 2012; Marrella and Yoo 2007).

The applicability of the lex mercatoria approach for OA may hence depend on the question whether there are such customs or norms within the community of researchers. We may find guidance in the history of academic publishing. We have seen that copyright did not play a vital role in the relationship between publishers and authors before commercial publishers entered the academic publishing market in the 1960s and 1970s. In fact, the relationship between academic journals and copyright was merely occasional, as academic journals were primarily published by learned societies and academic institutions (Ramello 2010, p. 13).63 Consequently, there must have been a body of immanent rules that governed publisher/author transactions. Looking more closely at the community of researchers does in fact reveal a high level of social homogeneity. We have learned that researchers primarily publish their works for “reputation building claims” and for the purpose of reaching a socialization within their peer-community. Beyond this immanent reward structure, a mechanism of scientific communication by means of academic journals has evolved that has become standard. Furthermore, the overall process from writing the work, to the peer-review for quality selection and typesetting tasks is enforced by the community itself. These rules are not just social norms, they are enforced in practice and create the standard setting for measuring performance of academics worldwide. Here, it will be the task of academic association and scientific institutions to self-enforce these principles. Recalling the lessons from the history of academic journal publishing—with the tradition that academic associations published the works of their researchers—may be a good guidance for the future policy agenda.64

Nevertheless, Metzger (2012) clearly highlights the limits of the lex mercatoria approach as a framework for managing OA licensing on a global scale. Most importantly, the application of a lex mercatoria approach will be limited to cases where the parties are actually free in choosing the applicable law. Copyright issues of the licensing contracts—especially features such as scope and hence limitations and exceptions to copyright protection—will not be subject to lex mercatoria. Especially in this context, reforms for a deliberalisation of the international copyright framework—as previously discussed—will be necessary to complement a workable licensing scheme beyond the perspectives outlined here.

By Contract

A complementary approach to a reform of copyright is to contractually regulate the access, use and reuse of academic journal content. In particular, governments and nonprofit agencies who largely fund scientific research, especially in the OECD countries, could impose an OA mandate on researchers who received funding for their research results. With other words, researchers would be mandated to arrange for free and unrestricted access to their funded research outcomes as a condition to their funding contract(s). Here, both governments and funding agencies can force such contractual conditions. Governments can dedicate government funded works to the public domain (Reichman and Okediji 2012, p. 1469; Reichman and Uhlir 2003, p. 318). Such a mandate is for example enforced by US copyright, where 17 U.S.C. §105 denies copyright protection for works that are produced by government employees within the scope of their employment. But also funding agencies have the ability to condition receipt of funding sources on OA requirements.65 Such a mandate could require funded researchers to either deposit pre- or post print versions of their works in OA repositories or to submit their works to purely66 OA journals.67

In general, three types of “pure”68 OA mandates can be distinguished (Suber 2012, pp. 77 et seq.). First, loophole mandates. This form of OA mandate requires self-archiving of papers that do not conflict with the journal publishers’ copyright agreement. With other words, if a journal publisher explicitly bans authors from depositing their work in an OA repository, authors can use these loopholes to escape the mandate. Second, deposit mandates. Here, self-archiving is required as soon as the paper is accepted for publication. However, deposit mandates separate the timing of deposit from the timing of OA (Suber 2012, p. 79). In particular, the timing of OA depends on publisher’s permission and hence whether the publisher allows for a self-archiving after a certain period of time. As a result, also with this form of mandate a deposited work may remain non-OA. Third, rights-retention mandates. Just like deposit mandates the rights-retention mandates require self-archiving as soon as the paper is accepted for publication. However, the funding agency may require to retain the nonexclusive right to authorize OA throughout self-archiving platforms (Suber 2012, p. 80).69 Figure 4.4 summarizes the different categories of mandates, leaving us with a reference for further discussion in Chap. 5


Fig. 4.4
OA policies at funding agencies (extending on Suber 2012, pp. 77 et seq.)

Only gold members can continue reading. Log In or Register to continue