Abstract
I study the impact of ad-valorem and unit taxes on the intellectual property policies of two-sided digital platforms. First, I address the monopoly case, in which I show that the effects of taxes depend on which side they are levied on. If developers are taxed, I find that ad-valorem taxes reduce the platform openness and the exclusivity period granted to developers to exploit their innovations. The opposite is true when taxes are levied on users. On the other hand, the effect of unit taxes is ambiguous in general. Then, I extend the model to address the duopoly case, and I find that competition may increase welfare, but it is not guaranteed. The effects of taxes on welfare are similar in both regimes. In general, they are ambiguous, but I characterize those cases which are welfare-enhancing unambiguously. I conclude highlighting the potential impact of the Digital Service Tax (DST) proposed by the European Commission on platform openness and digital innovation in Europe.
Similar content being viewed by others
Notes
See Belleflamme and Toulemonde (2018) for a short review on this topic.
In recent work, they also find that taxes may affect the political views of newspapers, see Kind et al. (2013).
This framework is intuitively similar to Chowdhury and Martin (2017), in which an upstream monopolist provides platforms (newspapers) with a complementary good (comics) that only some users (readers) value. In contrast, in this model, the platform itself provides the complementary good (the APIs or SDKs) and sets how long it could be enjoyed.
The openness could be higher than 1, which implies that the platform is subsidizing the developer.
Note that \(\delta =e^{-rt}\).
I keep the same assumption.
Nowadays, Google or Apple themselves generate fewer innovations than Android or iOS ecosystems. It has become more important the investment carried out by independent developers on the platform than the investment carried out by the owner of the platform. See, for example, Uber, Tinder, Airbnb, or Whatsapp.
The presence of heterogeneous users does not change our conclusions because developers are the ones who set the prices for their apps, not the platforms. This is a realistic approach because Google Play and the App store do not set the price for third-party apps.
In the annex, I prove that this assumption does not modify the previous conclusions, but it makes the model analytically more tractable. This approach is also adopted in other works, such as Belleflamme and Peitz (2019).
I assume the market is complete, and all developers want to stay in the market.
Note that \(\sigma \) cannot be lower than 0, which implies the platform subsidizes the proprietary side. In this framework, a negative \(\sigma \) would imply that developers produce negative quantities, which is not possible. On the other hand, I assume developers always set a non-negative price for their apps.
Symmetrically, \(M_{-j}^{d}\).
This same assumption is also made on Armstrong (2006).
I solve the model without assuming any specific expectation formation process to allow clear comparisons with the Parker and Van Alstyne’s framework.
Current evidence highlights that digital products within a given category are highly differentiated in the eyes of the consumer; therefore, the demand for any product is hardly affected by other products, see Belleflamme and Peitz (2015), p. 557.
A similar approach can be found in Choi et al. (2010).
However, because of a larger \(\sigma \) and \(\frac{\partial \delta }{\partial \sigma }<0\), the \(\delta ^*\) would be lower than in the monopoly.
References
Armstrong M (2006) Competition in two-sided markets. RAND J Econ 37(3):668–691
Bacache M, Bloch F, Bourreau M, Caillaud B, Cremer H, Crémer J, Demange G, de Nijs R, Gauthier S, Lozachmeur J-M (2015) Taxation and the digital economy: a survey of theoretical models. France Stratégie
Belleflamme P, Peitz M (2015) Industrial organization: markets and strategies. Cambridge University Press, Cambridge
Belleflamme P, Peitz M (2019) Price disclosure by two-sided platforms. Int J Ind Organ 67:102529
Belleflamme P, Toulemonde E (2018) Tax incidence on competing two-sided platforms. J Public Econ Theory 20:9–21. https://doi.org/10.1016/j.ijindorg.2019.102529
Bourreau M, Caillaud B, Nijs R (2018) Taxation of a digital monopoly platform. J Public Econ Theory 20(1):40–51
Choi H, Kim S-H, Lee J (2010) Role of network structure and network effects in diffusion of innovations. Ind Mark Manag 39(1):170–177
Chowdhury SM, Martin S (2017) Exclusivity and exclusion on platform markets. J Econ 120(2):95–118
Foros Ø, Kind HJ, Sørgard L (2015) Merger policy and regulation in media industries. In: Anderson SP, Waldfogel J, Strömberg D (eds) Handbook of media economics, vol 1. Elsevier, pp 225–264
Jullien B, Sand-Zantman W (2019) The economics of platforms: a theory guide for competition policy. TSE digital center policy papers series, (1)
Kind HJ, Koethenbuerger M (2018) Taxation in digital media markets. J Public Econ Theory 20(1):22–39
Kind HJ, Koethenbuerger M, Schjelderup G (2008) Efficiency enhancing taxation in two-sided markets. J Public Econ 92(5–6):1531–1539
Kind HJ, Koethenbuerger M, Schjelderup G (2009) On revenue and welfare dominance of ad valorem taxes in two-sided markets. Econ Lett 104(2):86–88
Kind HJ, Koethenbuerger M, Schjelderup G (2010) Tax responses in platform industries. Oxf Econ Pap 62(4):764–783
Kind HJ, Schjelderup G, Stähler F (2013) Newspaper differentiation and investments in journalism: The role of tax policy. Economica 80(317):131–148
KPMG ETC (2018) Eu commission releases package on fair and effective taxation of the digital economy. Technical report, KPMG
Lam WMW (2017) Switching costs in two-sided markets. J Ind Econ 65(1):136–182
Parker G, Van Alstyne M (2018) Innovation, openness, and platform control. Manag Sci 64(7):3015–3032
Tremblay M (2018) Taxing a platform: transaction vs. access taxes. Mimeo. https://doi.org/10.2139/ssrn.2640248
Acknowledgements
I am grateful to two anonymous reviewers for their careful reading and numerous suggestions that have improved the manuscript. This work has also benefited from thoughtful comments by Lapo Filistrucchi, Luis Corchon, Carmelo Rodriguez, Lourdes Moreno, Covadonga de la Iglesia, and Giorgio Ricchiuti as well as seminar participants from the University of Florence, the Complutense University, and the European Institute of Technology. Part of this work has been done during a research visit at the University of Florence.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The author declares that he has no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendices
A: Parker and Van Alstyne model with horizontally differentiated developers
In the duopoly setting, developers are horizontally differentiated. To prove that this change does not qualitatively change the conclusions, let’s extend Parker and Van Alstyne’s model to accommodate such an assumption. Eq. 3.1 becomes,
where \(M^{d}= \frac{p^{e}k(\sigma V)^{\alpha }+\delta p^{e}M^{\alpha }_{e}k^{\alpha +1}(\sigma V)^{\alpha ^2}}{2 \lambda }\), the monopoly demand that the platform faces on the developer’s side. Likewise the original model, platforms profits are well behaved, and there exists a tuple \(\left\langle \sigma ^*,\delta ^* \right\rangle \) that maximizes \(\varPi _{j}\). Therefore, the optimal lengths of the exclusivity period and the degree of openness are respectively,
Note that the optimal \(\delta \) does not change, but the optimal \(\sigma \) is expressed differently. The transportation cost (\( \lambda \)) appears in the denominator, reducing the optimal \(\sigma \). On the other hand, in the numerator, \(v(1-\delta )\) is increasing the optimal \(\sigma \). This value was included in \(\pi \) in the original model because it considers that the platform and developers can perfectly predict the price. In this extended version, I distinguish between the expected price that developers take into account to form their profit expectations (\(p^e\)) and the price (\(v(1-\delta )\)) that platforms know that developers will set under their intellectual property policies. Lastly, by the implicit function theorem, it is possible to prove that \(\frac{\partial \sigma }{\partial \delta }<0\) like in Parker and Van Alstyne’s model,
With respect to taxes, the optimal \(\delta \) and \(\sigma \) become,
In this general model, the conclusions remain the same as in Sect. 3.1.
B: Optimal solutions for \(\delta \) and \(\sigma \): Proof Proposition 2
Taking Eq. 4.3, the first-order conditions on platform profits with respect to \(\delta _j\) are
If I rearrange terms in Eq. B.1, I arrive at the first case of Eq. 4.5. Note that the second-order conditions are
which implies that \(\varPi _j\) is concave in \(\delta _j\), and by solving the first-order conditions, I obtain the maximum. Also, note that this interior solution only exists if \(2\pi ^{d}_{j,2} + \pi ^{d}_{-j,2} > \varDelta _{1}+6 \lambda \). This expression can be easily derived from Eq. 4.5. Otherwise, \(\delta _j=0\), the corner solution. Taking Eq. 4.3, the first-order conditions on platform profits with respect to \(\sigma _j\) are
Again, if I rearrange terms in Eq. B.2, I arrive at the first case of Eq. 4.6. In this case, the second-order conditions are
it implies that \(\varPi _j\) is concave in \(\sigma _j\), and by solving the first-order conditions, I obtain the maximum. Also, note that this interior solution only exists if \(\delta _j>0\). Otherwise, I only have to substitute \(\delta _j=0\) in Eq. B.2, and \(\sigma \) would be equal to the second case of Eq. 4.6, the corner solution.
Eq. 4.6 shows that \(\sigma \) depends on \(\pi _{i,j}\), \(i=1,2\) that also depends on \(\sigma \). This situation may raise concerns about the uniqueness of the solution, but the proof is similar to the one in Parker and Van Alstyne (2018). Nonetheless, it is a little bit more tedious. For simplicity’s sake, I omit it here, but it is available upon request. Lastly, Eq. 4.5 depends on \(\sigma \) that depends on \(\delta \), by the implicit function theorem, \(\frac{\partial \sigma }{\partial \delta }<0\). Note that \(\sigma ^*\) is similar in both the duopoly and the monopoly frameworks.
C: Optimal solutions for \(\delta \) and \(\sigma \): Proof Proposition 3
Taking Eq. 4.8, the first-order conditions on platform profits with respect to \(\delta _j\) are the same as before. Therefore, the previous proof applies here. On the other hand, the first-order conditions on platform profits with respect to \(\sigma _j\) are
If I rearrange terms, I arrive at the first expression of Eq. 4.11. In this case, the second-order conditions also verify that the interior equilibrium is the maximum. Nonetheless, it also exists a corner solution that is reached when \(\delta _j=0\), as in the previous case. To prove whether or not \(\sigma _j\) is decreasing with respect to \(\tau ^{vat}\), I proceed in two parts. First, there is a direct and negative effect of \(\tau ^{vat}\) on \(\sigma _j\), but a change in \(\sigma _j\) triggers a change in \(\delta _j\) that influence \(\sigma _j\) again and so on. By the implicit function theorem, \(\frac{\partial \sigma }{\partial \delta }<0\), and by differentiating Eq. 4.5, \(\frac{\partial \delta }{\partial \sigma }<0\). Therefore, the impact of ad-valorem taxes on \(\sigma _j\) is negative when \(\delta ^*>0\). If \(\delta ^*=0\), there an unambiguously negative effect of \(\tau ^{vat}\) on \(\sigma _j\).
Lastly, if I take Eq. 4.9, the first-order conditions on platform profits with respect to \(\delta _j\) are
In this case, the second-order conditions also verify that the interior equilibrium is a maximum. Nonetheless, it also exists a corner solution that is reached when \((v-\tau ^{sp})2\pi ^{d}_{j,2} + \pi ^{d}_{-j,2} > \varDelta _{1}+6 \lambda v\). Lastly, if I take Eq. 4.9, the first-order conditions on platform profits with respect to \(\sigma _j\) are
If I rearrange terms, I arrive at the first expression of Eq. 4.14. The second-order conditions also verify that the interior equilibrium is a maximum, and it also exists the corner solution that is reached when \(\delta _j=0\), like in the previous cases. To prove whether or not \(\sigma _j\) is decreasing with respect to \(\tau ^{sp}\), the procedure is the same as before. The effect is ambiguous when \(\delta ^*>0\) and negative if \(\delta ^*=0\).
Rights and permissions
About this article
Cite this article
Sánchez-Cartas, J.M. Intellectual property and taxation of digital platforms. J Econ 132, 197–221 (2021). https://doi.org/10.1007/s00712-020-00717-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00712-020-00717-5