Abstract
Purpose
This paper aims to verify how the intellectual property (IP) box affects firms’ effective tax rate, growth and innovation activity outcomes related to intellectual property rights.
Design/methodology/approach
Implementing the innovation box regimes into the tax system intends to encourage firms to engage in more innovative activities. In UK, Italy and Poland, the IP box tax relief was introduced in 2013, 2015 and 2019, respectively. In return, companies may reduce their tax rate to increase their investment and innovativeness. With a panel model approach – system GMM and DiD with multiple time periods – it analyses data from the Orbis database for 2011–2019 of 673 firms from the gaming industry in 11 countries and hand-collected data on intellectual property rights protection. The authors study public and private companies from the gaming sector in leading European markets and all three countries that protect intellectual property rights of software (Japan, South Korea, the USA).
Findings
Recent reforms enable gaming companies to use preferential tax treatment for IP-related income and significantly impact a firm’s revenue growth.
Practical implications
Nevertheless, European gaming firms require time to leap the gap to the growth and innovativeness of countries that protect software.
Originality/value
The authors show that the IP box stimulates gaming firms to protect IP via wordmarks, figurative marks, trademarks and software patents that bring effects in five years. Despite the critics against IP box, the authors prove its lagged efficiency, especially in profitable and larger firms.
Keywords
Citation
Białek-Jaworska, A., Teterycz, A., Sichel, R. and Woźniak, M. (2023), "IP box effects in the gaming industry", Central European Management Journal, Vol. 31 No. 3, pp. 291-325. https://doi.org/10.1108/CEMJ-12-2021-0143
Publisher
:Emerald Publishing Limited
Copyright © 2023, Anna Białek-Jaworska, Agnieszka Teterycz, Ricardo Sichel and Michał Woźniak
License
Published in Central European Management Journal. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) license. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this license may be seen at http://creativecommons.org/licences/by/4.0/legalcode
1. Introduction
This paper aims to study how the introduction of innovation box regimes impacts a gaming firm’s effective tax rate (ETR), revenue growth and innovation activity outcomes related to intellectual property rights. Intangible assets and intellectual property are crucial in developing all multinational companies, especially gaming ones.
Over the last 20 years, many countries have decided to implement innovation box regimes into their tax system to reduce shrinking income tax revenues. Intellectual property box (IP box) is a preferential tax on income derived from selling products or services based on intellectual property rights. Thus far, 14 out of 27 European countries have applied these regulations. These tax regimes enable companies to reduce the tax rate on income emerging from patents, licenses, and other forms of intellectual property (Evers, Miller, & Spengel, 2015; Merrill, 2016). Intellectual property box regimes are considered back-end or output-based incentives that offer entrepreneurs reduced taxation on income earned from a successful innovation (Lester, 2021). The scope of the tax base and the preferential tax rate resulting from the IP box varies between countries ranging from 4.4% in Belgium to 13.9% in Italy, and an exemption of qualified intellectual property income from 50% to 80% (Chen, De Simone, Hanlon, & Lester, 2019). The lowest tax rate resulting from the IP box regime is in Malta (outside our research sample) and equals 0%. In various countries, the IP box relief may apply to the income from patents, intellectual property created by programmers such as software or domains, or industrial property like trademarks or trade names. The primary purpose of this tax incentive is to increase the attractiveness of conducting research and development (R&D) activity, encourage entrepreneurs to search for business potential in intellectual property rights and retain income from intangible assets that could be transferred out of the country to low-tax jurisdictions.
The gaming sector is constantly growing due to the increasing interest in and access to online games. In PwC’s report “Global Entertainment & Media Outlook 2020–2024” (2020), experts expect the gaming industry in the analyzed countries to grow at a compound annual growth rate (CAGR) of 3.9%–11.6% between 2019 and 2024. The global value of the gaming sector exceeds USD 159 billion, and in Poland, its revenue tops USD 596 million, with export constituting 96% of that sum (Rutkowski, Marszałkowski, & Biedermann, 2020). We used the panel data model approach (the Blundell-Bond system GMM estimator, DiD with multiple time periods, the conditional logit, and complementary log-log model) for the following proxies (dependent variables) to capture the effects of introducing the IP box in various countries: ETR, revenue growth, and IP protection use in the gaming industry. Our study encompassed the period from 2011 to 2019. We based it on data from Orbis, OECD, and KPMG databases. Our research sample covered 5384 observations of 673 listed and unlisted gaming companies from the main European markets (Germany, the UK, France, Italy, Spain, Poland, Belgium, and the Czech Republic) and all three countries where software intellectual property rights are protected (Japan, South Korea, and the USA). It is essential to remember that software is not patentable in Europe. A guideline from the European Patent Office states that computer programs are excluded from patentability under Art. 52(2) (c) and (3) if claimed as such. However, following the generally applicable criteria for Art. 52(2) and (3) (GII, 2), the exclusion does not apply to computer programs having a technical character. To have technical nature and thus not be excluded from patentability, a computer program must produce a “further technical effect” when run on a computer. A “further technical effect” is a technical effect that goes beyond the “normal” physical interactions between the program (software) and the computer (hardware) on which it runs. For example, the typical physical effects of the execution of a program, e.g. the circulation of electrical currents in the computer, are not sufficient to confer technical character to a computer program (T 1173/97 and G 3/08). In reference to Newzoo (2020), our surveyed countries were pacemakers in revenue generated from gaming activities (see Figure A1 in Appendix).
Moreover, not all of the analyzed countries introduced IP box regulations. The exceptions are Japan, the USA, South Korea, the Czech Republic, and Germany. In the UK, the scope of qualifying income is narrowly defined and limited to patents only. In other countries, income includes patents, software, know-how, models, trademarks, licensed IP, plant brands, etc. The foreign-derived intangible income (FDII) deduction in the USA applies to American taxpayers with income from export sales or services. It decreases ETR to 13.1%, compared with 21% CIT till 2026 and 16.4% after 2026 (Cunningham, 2018). In South Korea, qualifying assets are only software protected by copyright, industrial patents, trademarks, designs, models, processes, and formulas, subject to legal protection. Depending on the tax base, the reduced tax rate varies from 5% to 12.5% for transfer and from 7.5% to 18.7% for a license (OECD Dataset Intellectual Property Regimes, 2019).
Literature shows that IP box reduces ETR more in MNEs than domestic firms (Evers et al., 2015; Bornemann, Laplante, & Osswald, 2020) and boosts R&D and patent applications (Bradley, Dauchy, & Robinson, 2015; Evers et al., 2015; Mohnen, Vankan, & Verspagen, 2017; Chen et al., 2019; Haufler & Schindler, 2020). However, it does not touch the gaming industry not patenting in Europe, which experienced explosive but rapidly and frequently changing growth – the highest growth of sales occurs right after a successful game launch and gradually expires until the preorder boom of another game 6–8 years later. Our study contributes to the existing academic literature by confirming that IP box regulations reduce ETR by two to three percentage points and enable gaming companies to boost their revenue growth and protect their intellectual property. However, the effects of the innovation box take a few years to materialize, thus we observed lagged impacts. Nevertheless, owing to the IP box, lower tax rates stimulate them to benefit from IP protection via software patents and trademarks.
The remainder of this paper is structured as follows. Section 2 will cover the literature review and hypotheses development. We will describe the data used and the research design in Section 3. We will then present our results in Section 4. Finally, Section 5 will conclude the paper, discuss research limitations, and suggest directions for future research.
2. Literature review and hypotheses development
Intellectual property box regimes are a controversial tax policy tool because of their unclear impact on increasing business innovation and their potential to increase tax avoidance. Supporters of this regime argue that reducing the statutory tax rate for intellectual property is a policy tool crucial for boosting domestic innovation, which is underinvested (Zhong, 2018; Bornemann et al., 2020). Facing increasing tax rate competition, many countries used IP box incentives to (1) attract corporate investments and employment and (2) retain taxable income that would otherwise be shifted to lower-taxed jurisdictions (Lester, 2021). On the other hand, opponents believe that IP box regimes foster adverse effects of tax competition and favor income shifting due to the preferential tax rate in different countries without increasing innovative activity in domestic ones. In response to this contentious issue of the innovation box relief, the Organization for Economic Cooperation and Development (OECD) implemented the Base Erosion and Profit Shifting (BEPS) Action 5, which reduces the IP box tax benefit to innovation-related income exclusively developed within a country (OECD, 2015). Gaessler, Hall, and Harhoff (2021) confirm that the particular design of the patent box determines to what extent IP rights are reallocated. The requirement that further invention development occur within the country to enjoy the lower tax rate seems to mitigate transfers for purely tax reasons.
Analyzing previous empirical studies (Evers et al., 2015; Ohrn, 2016; Koethenbuerger, Liberini, & Stimmelmayr, 2018; Chen et al., 2019; Bornemann et al., 2020), we can conclude that the impact of implementing the IP box relief is multi-dimensional. Previous literature suggests that multinational enterprises (MNEs) have more opportunities to reap tax benefits than domestic companies. This is because MNEs can allocate their R&D and intangible assets in low-tax jurisdictions, distorting their locations (Karkinsky & Riedel, 2012). Intellectual property box regulations are designed to prevent the artificial relocation of R&D locations and intangible assets to combat this problem (Bradley, Robinson, & Ruf, 2021). Therefore, some studies examined how the innovation box affects locations of intangible assets or R&D, cross-border payments, and reported income (Karkinsky & Riedel, 2012; Koethenbuerger et al., 2018; Chen et al., 2019; Bornemann et al., 2020). Other research indicates that this tax policy tool boosts firms’ innovative activities and encourages R&D investments (Bradley et al., 2015; Evers et al., 2015; Chen et al., 2019; Haufler & Schindler, 2020). There is also empirical evidence that the innovation box notably supports a decrease in ETR (Evers et al., 2015; Bornemann et al., 2020).
Literature on the effect of IP boxes on innovation activity provides ambiguous conclusions. On the one hand, the IP box is positively correlated to innovation (Mohnen et al., 2017), as a 1% decrease in the corporate tax rate boosts new patent applications by 3% on average (Bradley et al., 2015). An average Dutch firm tends to use only a part of the tax advantage for extra R&D investment. Therefore, the innovation box additionality effect is lesser than the costs involved in foregone taxes (Mohnen et al., 2017). In Belgium, on the other hand, implementing the innovation box has not increased innovation activity extraordinarily, because patent-owning firms have not been further encouraged to increase their innovation after introducing the IP box regimes (Bornemann et al., 2020). The IP box regulations offer an additional tax incentive to develop intangible assets. In Belgium, implementing the innovation box decreased the marginal ETR of 1.9% on marginal R&D investments (Evers et al., 2015), and ETR by 7.2% to 7.9% when developing intellectual property (Bornemann et al., 2020). Still, MNEs receive more IP box tax benefits than domestic companies.
American payments to foreign affiliates increase when the IP box regulations qualify income as existing intellectual property, but this regime significantly impacts only newly developed R&D (Ohrn, 2016). A 1% change in the statutory tax rate results in an approximately 14.8% reduction in income shifting and an increase in profits by about 1.31 million euros because of higher tax benefits due to IP box regulations (Bornemann et al., 2020). Moreover, MNEs without income-shifting opportunities may expect a greater ETR reduction than MNEs that transferred profits (Bornemann et al., 2020). The innovation box system is associated with a lower sensitivity to reported profits to local statutory income tax rates. It means less income shifting across borders than observed in other jurisdictions. Moreover, income shifting is concentrated in countries implementing the innovation box with a relatively large tax advantage (Chen et al., 2019). The number of inventors in the destination countries increases when patent box benefits are conditional on the further development of patents that were transferred there (Alstadsæter, Barrios, Nicodeme, Skonieczna, & Vezzani, 2018).
Similarly, the size of the patent box differential is harmful both for patent filings and business R&D when there is no requirement for further development but insignificant otherwise. Thus, introducing a patent box does not influence aggregate innovative activity if further development of the inventions contained in transferred patents is required (Gaessler et al., 2021). Contrary to expectations (Gaessler et al., 2021), before OECD introduced the BEPS restrictions, we could observe a somewhat negative impact of the patent box introduction on local invention and R&D. However, nexus rules reduce firms’ incentives for acquiring developed intangibles and, by extension, reduce the amount of M&A activity. The inclusion of a strict nexus requirement effectively renders the IP box incentives ineffective from the perspective of stimulating M&A investment. On the one hand, demanding nexus limits the investment response, while on the other hand, it requires actual activity in exchange for tax benefits. Consequently, the nexus requirements have been a primary way for governments to regulate the amount and type of IP box benefits they provide (Lester, 2021), among others, in Poland. Thereby it affects cross-border tax competition.
However, the increasing costs of R&D capital and knowledge spillovers throw entrepreneurs into a gap between investments in tangible and intangible assets leading to underinvestment in innovative activities (Bornemann et al., 2020). Therefore, many countries introduce various tax incentives, such as tax credits for R&D expenditures, to help close the wedge between investment in tangible and intangible assets (OECD, 2020) and encourage entrepreneurs to increase their R&D activity. In our study, we propose three hypotheses.
First, introducing the innovation box in countries allows the reception of tax benefits. Gaming companies that conduct R&D activities and create qualified intellectual property rights may reduce their tax liabilities. Laplante, Skaife, Swenson, and Wangerin (2019) found that a one standard deviation increase in strategic R&D classification leads to a 1.7% reduction in GAAP ETR. However, a bigger firm with a higher sales volume has a higher ETR (Jones, Baker, & Lay, 2017; Thomsen & Watrin, 2018). Although the possible result is not clear, we expect that:
Gaming companies decrease the ETR after introducing IP box regimes in their countries.
Our second hypothesis examines whether innovation box regimes are efficient in supporting the revenue growth of gaming companies. Enforcing IP box rules allows gaming firms to increase their free cash flows through lower tax payments. Thus gained surplus cash may improve their investment capacity and boost R&D expenditures. Corporate productivity growth is positively correlated with R&D expenditures as R&D activity represents efforts towards achieving efficiency (Minasian, 1962, 1969). The value-added increases due to success, whereas R&D expenditures include both the cost of failed projects and the fruitful ones. Typically, returns to R&D amount to several hundreds to thousands of percent. However, increased productivity caused by R&D partially reaches consumers via lower goods’ prices. Finally, lower prices will tend to decrease the firm’s profits and the return rate (Minasian, 1969). Next, at another stage of the gaming firm’s life cycle, the extra money can be utilized to expand production capacity and quicken revenue growth when their video games enter the market. Therefore, we formulated the following hypothesis H2:
Gaming companies increase revenue growth after introducing IP box regimes.
In the third hypothesis, we examined what determines whether a gaming company protects intellectual property. Previous literature suggests that the number of patent applications is related to the corporate tax rate and indicates that an increase in the number of patent applications depends on the tax credit available through the IP box or favorable treatment of R&D expenditures (Bradley et al., 2015). Moreover, Bornemann et al. (2020) show that the number of patent applications and patents granted increased after implementing IP box regulations in Belgium. Thus, we formulated hypothesis H3 consistent with the policy goals of innovation boxes:
Introducing the innovation box relief positively impacts the R&D activities’ outcomes in the gaming sector.
3. Research design
Our study examined how introducing the IP box regimes impacts a gaming firm’s tax benefits, revenue growth, and innovation activity outcomes related to intellectual property rights. Therefore, we built a research sample to entail leading European (Belgian, Czech, French, German, Italian, Polish, Spanish, and British) and Japanese, South Korean, and American private and public gaming companies that substantially impact the international market. We chose a nine-year sample period from 2011 to 2019, which covers the implementation of the IP box at different times in the analyzed countries (see Figure A1 and Table A1 in the Appendix). We collect financial, employment, and ownership data retrieved from Orbis and KPMG databases. Moreover, we recovered aggregated information about the number of patents pending or granted from the Eurostat database and hand-collected individual data on registered word marks, trademarks, and figurative marks from the European Patent Office and the United States Patent and Trademark Office.
We begin with 12,872 firm-country-year observations. Because the gaming industry includes mainly private companies unlisted on stock exchanges, obtaining complete financial data was difficult and led to the estimation of unbalanced panel models. Thus, we replaced the missing data with an average value within the company. To test hypothesis H1, we use ETR measured by a ratio of tax expenses to profit before tax (PBT) (earnings before interest, taxes, depreciation, and amortization-EBITDA). Effective tax rate estimates tax benefits that companies could receive after introducing the IP box relief. To avoid situations in which a negative value of ETR would be challenging to interpret, we removed the observations that took a negative value for the profit before tax and income tax, similarly to Thomsen and Watrin (2018). The above exclusions reduced the sample to 2403 firm-year observations of 359 distinct companies concerning ETR based on PBT and 2543 firm-year observations of 362 firms using ETR based on EBITDA (see Table A2 in the Appendix). The impact of some variables on ETR can be ambiguous due to their relationship with both the denominator and the numerator. For example, depreciation or interest (forms of non-debt and debt tax shield) is related to the tax burden, while their impact may weaken after certain accounting transformations in the denominator, which is the reference point for the taxes reported. Therefore, we expected a strong negative relationship between capital intensity or leverage and ETR based on EBITDA. On the other hand, EBITDA lacks a simple interpretation and direct comparability with statutory tax rates.
To test hypothesis H2, we had to exclude observations from 2011, because otherwise, we could not calculate our dependent variable. The lagged revenues are required to calculate the Growth variable that measures a firm’s revenue growth. Therefore, our sample consisted of 2261 firm-country-year observations of 360 firms when drawing upon ETR based on EBITDA and 5384 firm-country-year observations of 673 individual firms without ETR limitations. Moreover, to test hypothesis H3, we used panel data for the entire research period for the conditional logit model and complementary log-log model for the IP protection dummy variable that captures successful innovation activities’ outcomes. Table A2 in the Appendix presents the distribution of firms by analyzed countries. The definitions of all variables are described in Table 1 (Table A3 in the Appendix).
Table A9 in the Appendix provides descriptive statistics. In our sample, the average value for the ETR based on profit before tax (ETR_PBT) was 22.9%, while the ETR based on EBITDA (ETR_EBITDA) was 16.9%.
Average ETR measures were lower than the average STR, which equaled 27.6%–27.8% (for details, see also Figure A4 in the Appendix). The average tangible fixed assets varied from 8% to 9.7% of total assets, while intangible assets were between 9.6% and 13%, respectively.
Figure A3 in the Appendix verifies the Kendall correlation.
We estimated the following models (1) and (1a) to investigate whether gaming companies benefit from IP box regimes (hypothesis H1) using system GMM and FE estimators, respectively. To measure the ETR, we used three proxies to verify the robustness of the results: first, a tax expense to profit before tax ratio (ETR_PBT); second, ETR_PBT scaled to non-negative values lower than 1 (H1); and third, tax expense in relation to EBITDA ratio (robustness check). IP BOX is a binary variable, which equals one for all years after introducing the IP box relief and zero otherwise.
We controlled the size, leverage, ROA, intangibility, capital intensity, and inventory variables. It is crucial to do so because larger companies have more tax planning opportunities to avoid taxes (Rego, 2003). However, larger firms also have higher political costs that act contradictory (Zimmerman, 1983; Gupta & Newberry, 1997). Higher leverage allows reducing the tax burden by deducting interest paid on debt from taxable income (McGuire, Omer, & Wang, 2012; Dyreng, Hanlon, Maydew, & Thornock, 2017; Thomsen & Watrin, 2018). We controlled intangibility to capture possibilities of income shifting (Dyreng, Hanlon, & Maydew, 2008). Given the tax preferences and total assets, an increase in ROA leads to an increased tax burden (Chen et al., 2019; Bornemann et al., 2020). Therefore, it is crucial to control profitability. Capital intensity captures tax optimization opportunities by selecting higher or accelerated depreciation for tax purposes (Bornemann et al., 2020). Finally, inventory reduces taxable income only at the time of sale (Bornemann et al., 2020). On the firm level, we controlled foreign shareholders’ participation in the equity (MNE), ownership structure, and the number of employees. In contrast, on the country level, we controlled the statutory tax rate (STR), good governance by a mean of Kaufmann indices (Kaufmann), and the number of patents pending or granted (patents).
We estimated the abovementioned models (eq. 1) at the firm-country-year level using the Blundell-Bond system GMM estimator dedicated to dynamic panel data (Blundell & Bond, 1998) and models described in equation (1a) with firm-country fixed effects on the diversity of patent activity across countries (Bornemann et al., 2020; Karkinsky & Riedel, 2012). The subscripts i, c, and t denote firm, country, and year.
Next, we adopted the state-of-the-art staggered difference-in-differences (DID) framework to calculate the IP box group-time average treatment effects on ETR (Callaway & Sant’Anna, 2021). The main idea behind this approach was first to estimate the individual cohort-time-specific treatment effects, allowing for treatment effect heterogeneity, and then aggregate the individual treatment effects to generate estimates of overall treatment effects. Therefore, we applied this specific DiD procedure with multiple time periods to check the robustness of results for testing H1 using alternative research designs. Our control group consisted of firms from countries without IP boxes for software, while the treated groups were from countries that implemented IP boxes after 2011 (we need information before policy implementation). For example, the UK, which introduced the IP box in 2013, Italy in 2015, and Poland in 2019 are treated groups defined by treatment timing in our DiD framework. Finally, to better interpret the results, we aggregated group-time average treatment effects by the length of exposure, group, and time period.
To address hypothesis H2, i.e. whether IP box regimes are associated with companies’ revenue growth in the gaming sector, we estimated models in eq. (2) in which revenue growth captures incremental changes in value. Moreover, we used several control variables such as size, leverage, employees, intangibility, capital intensity, inventory, mne, and ownership structure on a firm level and STR, patents, and Kaufmann on a country level. The company’s size proves that larger firms that produce on a large scale are likely to have more innovative activity and benefit from economies of scale. However, we used leverage to address the financial constraints of companies (Hall, Thoma, & Torrisi, 2007; Balsmeier, Fleming, & Manso, 2017), but also the opportunity to decrease the cost of capital (i.e. the weighted average cost of capital WACC) in the case of public (listed) companies.
Models described by eq. (2) were estimated at the firm-country-year level using the Blundell-Bond system GMM estimator dedicated to dynamic panel data (Blundell & Bond, 1998). The subscripts i, c, and t denote firm, country, and year accordingly.
In hypothesis H3, we examined what significantly impacts protecting intellectual property in the gaming industry. We tested whether introducing IP box regimes increases the gaming sector’s innovativeness by estimating the following conditional logit model (eq. 3) and complementary log-log regression (eq. 4).
A subscript i identifies a firm, and subscript c recognizes a country.
4. Results
Table 2 (and Table A4 in the Appendix) provides outcomes of testing hypothesis H1 of the innovation box regimes’ impact on gaming companies’ ETR. Columns (1)–(8) present the models’ estimation results using the two-step Blundell-Bond system GMM estimator. In (1)–(4), the dependent variable is the ETR_PBT, whereas in (5)–(8) – ETR_PBT_clipp. Because GMM two-step standard errors are biased, robust standard errors are recommended. Therefore, we adopted an estimation using a robust variance-covariance matrix in even-numbered models.
Results show that implementing the IP box negatively impacted the ETR. It enabled entrepreneurs in the gaming sector to reduce their tax liability by ca. 14 pp. It seems to have maximized benefits from the IP box because the distance from the IP box tax rate to the nominal varied from over 24 pp. in Belgium to 10 pp. in Italy. It reached 18 pp. in France, 15 pp. in Spain, and 14 pp. in the UK and Poland (see Table A1 in Appendix). However, in different countries, ETR caught differences between tax base and financial income and simultaneously between tax law and balance law (accounting act). Therefore, the ETR_PBT varied from zero to 233% of accounting profit, whereas ETR_PBT_clipp changed from zero to 98%. Our results are consistent with hypothesis H1.
Moreover, all estimated models (1)–(8) show that MNEs pay fewer taxes. The results were less substantial when we approximated the ETR by ETR_PBT_clipp scaled to values lower than one or based on EBITDA. However, the models (7–8) estimated on a subsample excluded Belgium, France, and Spain that implemented IP boxes before 2011 and findings of models (9–10) for ETR_EBITDA estimated using a fixed-effects estimator (in line with the Hausman test) support hypothesis H1. The latter results for ETR_EBITDA prove that implementing the IP box enables firms to reduce their tax burden by two pp. Firms set in countries with one pp. higher statutory tax rate pay higher taxes since their ETR based on EBITDA is higher by 0.20–0.224 pp.
This effect catches differences among the scope of tax-deductible costs and taxable revenues in various countries.
The ownership structure also significantly differentiates the ETR. Public companies pay fewer taxes as the ETR is lower by four pp., but gaming companies funded by mutual and pension funds pay more taxes (the ETR is higher by 2.3 to 2.5 pp.) (see Table A5 in the Appendix). Besides, we applied DiD with multiple time periods (Callaway & Sant’Anna, 2021) to provide robust results for testing H1. The UK (since 2013), Italy (since 2015), and Poland (since 2019) are treated groups, whereas countries without IP boxes for software (Japan, South Korea, the USA, Germany, and the Czech Republic) are a control group. Figure 1 shows group-time average treatment effects that support hypothesis H1 (see also Table A6 and Figure A2 in the Appendix). Post-treatment time period effects confirm that gaming companies decreased their ETR after introducing IP box regimes in the UK (2013 group) overall by four pp. and particularly from 2013 to 2015, whereas in Italy (2015 group) overall by 13.19 pp., and particularly in 2015 and from 2017 to 2019. Moreover, event study dynamic effects showed a decrease in ETR in the first year by five pp., in the second year by 10 pp., and in the fourth year by 9 pp. after implementing the IP box regime, overall by seven pp.
Moreover, calendar time aggregation estimates show an average negative effect of participating in the treatment in a particular period for all groups that participated in the treatment in that time period overall and for five years from 2013 to 2017. Finally, Figure 1 (and Table A6) shows adverse group-time average treatment effects for the UK and Italy that are contrary to Poland. This is because Poland requires more time to materialize the effects of the IP box introduced in the last year (2019) covered by our research sample.
Table 3 delivers the outcomes of testing hypothesis H2 using a system GMM dynamic panel data model (columns (1)–(2) for the total sample, columns (3)–(4) for the subsample excluded countries that introduced the IP box before 2011, columns (5)–(8) for narrower sample due to ETR restrictions, with the last two columns excluding Belgium, France, and Spain). It answers how implementing the IP box regime affects a gaming firm’s revenue growth (for more details, see Table A7 in the Appendix).
Gaming companies reduce their tax rate and tax liabilities thanks to IP box, but the tax savings and free cash flows require time to materialize and increase their revenue growth compared to their competitors from the countries with the highest revenue in the gaming industry. Although the parameter at the IP box variable is significant and negative in most models, the coefficients at the lagged or lead IP box dummy variable are positive and significant. Models estimated using system GMM on subsamples excluding Belgium, France, and Spain provide robust results. It means that gaming companies in countries implementing the IP box regulations after 2011 have a higher revenue growth by 0.16 to 0.37 pp., but the effects require time to materialize. This is partly because of higher inventory increases by unsold games. Thus, we have no basis for rejecting hypothesis H2. Larger companies with higher capital intensity and intangibility in countries with higher statutory tax rates and lower governance quality note higher revenue growth. Although the total number of patent applications (patents) at a country level positively influences revenue growth, their impact diminishes (or becomes adverse when lagged) after logarithmic transformation in GMM models (see Table A7 in the Appendix).
In hypothesis H3, we tested whether the introduction of the IP box is connected to boosting innovative activity outcomes. We measured the innovation by the scale of the intellectual property protection by gaming firms via trademarks, wordmarks, and figurative marks registered (models 9–11) or patent-pending (models 12–13 – only software patents registered in USPTO, models 9–11 – all IP protection). Table 3 (and Table A8 in the Appendix) presents the results estimated using conditional logit and complementary log-log models.
Our findings prove that implementing these regulations decreases the probability of intellectual property protection by registering in the Patent Office (models 12–13). However, IP box regulations encourage gaming companies to register their intangible assets as trademarks or protect their word or figurative marks. Still, the effects require time to materialize (models 9–11). Consequently, we have no basis for rejecting hypothesis H3 based on the findings of conditional logit. Moreover, our results show that larger firms with more employees are more likely to protect the intellectual property of their R&D activity’s outcomes, contrary to gaming companies funded by hedge funds. In contrast, those owned by financial companies, individuals, or families are more likely to use patent protection or register their word marks, figurative marks, or trademarks. The fact that software is subject to patent only in the USA, Japan, and South Korea where IP box regimes have not been implemented influenced our results. Thus, based on the parameter at the STR variable, we conclude that taxes stimulate firms to benefit from IP protection in the gaming industry, primarily via software patents and trademarks.
5. Conclusions
We examined how introducing innovation box regulations affects the gaming sector’s ETR, revenue growth, and innovation activity outcomes. We found evidence that IP box regulations are associated with lower ETR as gaming firms reduced their ETR by two pp. for earnings before interest, depreciation, and amortization. However, considering the debt tax shield, their ETR decreased on average by eight pp. (for ETR_PBT_clipp) and 9.5 pp. when we excluded Belgium, France, and Spain. Notably, immediately after entering the IP box in Italy, ETR dropped by 13.2 pp., but only by four pp. in the UK Furthermore, the results vary across the ownership structure. Public gaming companies had a lower ETR, but private firms set up by mutual and pension funds, nominees, trusts, and trustees had a significantly higher ETR based on EBITDA.
Moreover, introducing IP box regulation allowed gaming companies to increase their revenue growth by 0.1 to 0.7% points after two years, including firms with financial losses, compared to 0.37 pp. in the year before the introduction of the IP box. Profitable firms increased their revenue growth in one year by 0.377 pp. and 0.45 pp. when we excluded Belgium, France, and Spain. The benefits of this relief in the form of additional free cash flows require time to provide revenue growth sufficient to rise to the level of the top gaming companies from the countries that have not implemented the IP box. Therefore, revenue growth decreased immediately in the year when the IP box was introduced. Despite the IP box, the gaming industry still has lower sales growth than outside Europe. Our results show that larger firms grow faster than small firms in the gaming industry. Despite implementing the IP box, the gaming industry’s growth rate distance between developed and developing countries was still significant. In countries that have not introduced IP box, like the USA, South Korea, Japan, Germany, and the Czech Republic, firms grow faster than in countries where IP box regimes are in force (see Figure A5 in the Appendix). The largest and most recognizable digital game distribution platforms, such as Origin, Steam, or Epic Games, which allow users to purchase games of worldwide production, are based in the USA. There is only one such platform in Poland, GOG.com, set up by CD Project, the largest gaming company listed on the Warsaw Stock Exchange in 2019–2020.
6. Discussion
We confirm the IP box’s lagged effect on protecting intellectual property. It positively relates to firms’ intellectual property registration, but effects require at least five years to materialize. However, the IP box does not encourage firms to patent R&D outcomes properly. Our results are weaker than those provided by Bradley et al. (2015) and Bornemann et al. (2020), because software patenting is available only outside Europe, in the USA, Japan and South Korea. These countries require patenting software without having introduced IP box regimes. Meanwhile, European countries are still trying to leap the gap using the IP box, but they will need more time and R&D incentives to succeed. Finally, we found that lower taxes resulting from IP box stimulate firms to benefit from IP protection in the gaming industry, primarily via the game’s wordmarks, figurative marks, trademarks and software patents. This aligns with Bradley et al. (2015). However, effects require ca. five years to materialize. Moreover, although innovation scholars and the OECD are critical of the IP box, we provided evidence of its lagged efficiency, which is more significant and occurs earlier in profitable and large firms.
Our study has some limitations because IP box regulations vary between countries, the sample period was short, and there was no control for R&D subsidization although the number of subsidies is growing yearly (Haufler & Schindler, 2020).
Figures
Definitions of variables
Variable | Definition |
---|---|
Dependent variables | |
ETR_PBT | |
ETR_PBT_clipp | ETR_PBT scaled to non-negative values lower than 1 |
ETR_EBITDA | |
Growth | |
Patent_pp | Binary variable indicating if a patent is registered or pending, 1 – if yes, 0 – if no |
Patent_tn_pp | Binary variable indicating if a patent, a game’s wordmark, a firm’s trademark, or figurative mark is registered or pending, 1 – if yes, 0 – otherwise |
Test variable | |
IP BOX | The binary variable indicating when the IP box has been implemented equals 1 – for the year of introducing the IP box and after, and 0 – otherwise |
Control variables | |
Capital intensity | |
Intangibility | |
Inventory | |
Leverage | |
MNE | The binary variable indicating a firm with foreign shareholders equals 1 – if the firm’s foreign shareholders have more than 50% share in equity, and 0 otherwise |
Employees | |
Size | |
ROA | |
Patents | Number of patent applications or patents granted by the European Patent Office (EPO), patents granted by the United States Patent and Trademark Office (USPTO), and triadic patent families |
STR | Statutory Tax Rate |
Kaufmann | Average of six sub-index of the Worldwide Governance Indicators that rank countries concerning six aspects of good governance: voice and accountability, political stability and violence, government effectiveness, rule of law, and control of corruption |
Ownership | Categorical variable indicating the type of ownership structure that consists of 15 levels: bank; corporate; hedge fund; employees, managers, directors; financial company; private equity firms; public (listed companies); venture capital; insurance company; foundation; research institute; self-ownership; one or more named individuals or families; other unnamed shareholders aggregated, mutual and pension fund, nominee trust trustee; public authority state government |
Source(s): Own elaboration
IP box’s impact on the effective tax rate
Excl. BE FR ES | Excl. BE FR ES | Excl. BE FR ES | Excl. BE FR ES | |||||||
---|---|---|---|---|---|---|---|---|---|---|
WC-robust | WC-robust | WC-robust | WC-robust | FE | FE | |||||
ETR_PBT | ETR_PBT | ETR_PBT | ETR_PBT | ETR_PBT_clipp | ETR_PBT_clipp | ETR_PBT_clipp | ETR_PBT_clipp | ETR_EBITDA | ETR_EBITDA | |
L1.dependent | 0.0667*** | 0.0667 | 0.0551** | 0.0551 | 0.1696*** | 0.1696 | 0.1177** | 0.1177 | ||
variable | (0.024) | (0.058) | (0.027) | (0.097) | (0.049) | (0.105) | (0.050) | (0.117) | ||
IP BOX | −0.1433*** | −0.1433** | −0.1538*** | −0.1538** | −0.0803*** | −0.0803*** | −0.0954*** | −0.0954*** | −0.0201*** | −0.0202*** |
(0.041) | (0.055) | (0.039) | (0.072) | (0.021) | (0.029) | (0.021) | (0.032) | (0.008) | (0.009) | |
L1.IP BOX | −0.0482** | −0.0482* | −0.0910*** | −0.0910*** | ||||||
(0.022) | (0.028) | (0.022) | (0.031) | |||||||
MNE | −0.0738** | −0.0738** | −0.0665** | −0.0665* | −0.0785*** | −0.0785*** | −0.0721*** | −0.0721*** | ||
(0.032) | (0.036) | (0.031) | (0.036) | (0.020) | (0.023) | (0.020) | (0.025) | |||
L1.ln_Patents | 0.0256 | 0.0256 | 0.0688*** | 0.0688* | 0.0162 | 0.0161 | 0.0110 | 0.0110 | ||
(0.024) | (0.032) | (0.025) | (0.037) | 0.0110 | (0.015) | (0.011) | (0.016) | |||
Controls | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES |
Ownership | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES |
Time effects | YES | YES | YES | YES | YES | YES | YES | YES | NO | NO |
Observations | 1478 | 1478 | 1194 | 1194 | 1478 | 1478 | 1194 | 1194 | 2543 | 2543 |
Groups | 285 | 285 | 219 | 219 | 285 | 285 | 219 | 219 | 362 | 362 |
Instruments | 61 | 61 | 61 | 61 | 59 | 59 | 59 | 59 | – | – |
Wald test | 183.03 | 61.25 | 238.71 | 65.41 | 125.87 | 68.21 | 157.00 | 92.05 | ||
p-value | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | ||
Sargan test | 33.3707 | 33.3707 | 38.6599 | 38.6599 | 37.5799 | 37.5799 | 39.5516 | 39.5516 | ||
p-value | 0.3527 | 0.3527 | 0.1621 | 0.1621 | 0.1932 | 0.1932 | 0.1395 | 0.1395 | ||
AR(1) test | −3.6375 | −3.5081 | −3.4534 | −3.1186 | −4.8848 | −4.1753 | −4.5638 | −3.7709 | ||
p-value | 0.0003 | 0.0005 | 0.0006 | 0.0018 | 0.0000 | 0.0000 | 0.0000 | 0.0002 | ||
AR(2) test | 0.6358 | 0.6024 | 0.3291 | 0.2951 | 0.6331 | 0.5645 | 0.5623 | 0.4707 | ||
p-value | 0.5249 | 0.5469 | 0.7420 | 0.7679 | 0.5267 | 0.5724 | 0.5739 | 0.6378 | ||
Hansen J test | 7.3700 | 7.3700 | 6.8314 | 6.8314 | 138.6457 | 138.6457 | 32.3694 | 32.3694 | ||
p-value | 0.5987 | 0.5987 | 0.6547 | 0.6547 | 0.7175 | 0.7175 | 0.0003 | 0.0003 | ||
F-statistic | 2.07** | 1.89*** | ||||||||
Hausman test | 89.755*** | 66.377*** | ||||||||
Breusch-Pagan test | 126.51*** | 131.43*** |
Note(s): Robust standard errors in parentheses; *p < 0.1; **p < 0.05; ***p < 0.01
Source(s): Own elaboration
Effect of IP box on firm’s growth and innovative activity: system GMM, conditional logit, and complementary log-log models’ results
Two-step | One-step robust | Excl BE FR ES Two-step | Excl BE FR ES Two-step robust | One-step | One-step robust | Excl BE FR ES Two-step | Excl BE FR ES Two-step robust | Conditional (FE) logistic regression | Complementary log-log | ||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Growth | Growth | Growth | Growth | Growth | Growth | Growth | Growth | Patent_tn_pp | Patent_tn_pp | Patent_tn_pp | Patent_pp | Patent_pp | |
L1.y | −0.2551*** | −0.1978*** | −0.2702*** | −0.2702*** | −0.1935*** | −0.1935*** | −0.1178** | −0.1178 | |||||
(0.032) | (0.044) | (0.042) | (0.056) | (0.041) | (0.055) | (0.046) | (0.078) | ||||||
L2.y | −0.1295*** | −0.1019*** | −0.1137*** | −0.1137** | −0.1024** | −0.10248 | −0.0548 | −0.0548 | |||||
(0.027) | (0.034) | (0.037) | (0.047) | (0.041) | (0.060) | (0.038) | (0.079) | ||||||
IP BOX | −0.2271** | −0.2245* | 0.1062 | 0.1062 | −0.4597*** | −0.4597** | −0.2915*** | −0.2915** | 0.4381 | 0.4055 | 0.8955 | −0.7967 | −1.0370** |
(0.112) | (0.128) | (0.125) | (0.127) | (0.149) | (0.207) | (0.083) | (0.132) | (1.120) | (1.121) | (1.096) | (0.539) | (0.508) | |
F1.IP BOX | 0.3733** | 0.3733** | |||||||||||
(0.168) | (0.183) | ||||||||||||
L1.IP BOX | −0.0489 | −0.0671 | −0.0737 | −0.0737 | 0.3766** | 0.3766* | 0.4521*** | 0.4521* | |||||
(0.178) | (0.192) | (0.179) | (0.185) | (0.149) | (0.206) | (0.159) | (0.249) | ||||||
L2.IP BOX | 0.0740** | 0.0975** | 0.0374 | 0.0374 | −0.2084 | −0.2084 | −0.1050* | −0.1050 | |||||
(0.036) | (0.044) | (0.047) | (0.051) | (0.147) | (0.136) | (0.061) | (0.077) | ||||||
L3.IP BOX | 0.1338 | 0.1338 | |||||||||||
(0.090) | (0.092) | ||||||||||||
L4.IP BOX | −1.3437** | ||||||||||||
(0.637) | |||||||||||||
L5.IP BOX | 1.9494** | 1.9091** | 1.8510** | ||||||||||
(0.778) | (0.770) | (0.754) | |||||||||||
STR | −0.9434 | −0.6583 | −1.7416** | −1.7416** | 12.0923*** | 13.8349*** | |||||||
(0.708) | (0.896) | (0.829) | (0.874) | (3.281) | (3.392) | ||||||||
L2.STR | 2.2249** | 2.7544** | −0.5470 | −0.5470 | 3.8669*** | 3.8669* | 2.0844** | 2.0844 | 38.635** | 40.2068** | |||
(0.907) | (1.076) | (1.055) | (1.242) | (1.383) | (2.158) | (0.974) | (1.459) | (18.453) | (18.522) | ||||
L1.Patents | 0.3557*** | 0.3557*** | 0.2646*** | 0.2646*** | |||||||||
(0.118) | (0.103) | (0.073) | (0.097) | ||||||||||
L2.ln_Patents | −0.0943** | −0.1010 | −0.1343** | −0.1343** | |||||||||
(0.041) | (0.066) | (0.056) | (0.068) | ||||||||||
Controls | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES |
Ownership | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES |
Time effects | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES | YES |
Observations | 4038 | 4038 | 2865 | 2865 | 1061 | 1061 | 1064 | 1064 | 6057 | 6057 | 6057 | 6057 | 4200 |
Groups | 673 | 673 | 573 | 573 | 270 | 270 | 224 | 224 | |||||
Instruments | 58 | 58 | 50 | 50 | 52 | 52 | 56 | 56 | |||||
Wald test | 199.64 | 127.81 | 131.95 | 100.99 | 178.72 | 172.48 | 5626.30 | 747.87 | 448.95 | 513.95 | |||
p-value | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | |||
Sargan test | 27.8450 | 27.8450 | 9.7696 | 9.7696 | 30.9723 | 30.9723 | 35.5607 | 34.9131 | |||||
p-value | 0.1809 | 0.1809 | 0.8340 | 0.8340 | 0.0556 | 0.0556 | 0.0457 | 0.0531 | |||||
AR(1) test | −9.8449 | −9.3283 | −7.8938 | −7.0402 | −5.2057 | −5.2057 | −6.714 | −5.0413 | |||||
AR(2) test | −1.9307 | −1.432 | −1.3076 | −1.2333 | −1.2897 | −1.2897 | −1.6266 | −1.0117 | |||||
p-value | 0.0535 | 0.1521 | 0.1910 | 0.2175 | 0.1972 | 0.1972 | 0.1038 | 0.3117 | |||||
Hansen J test | 19.8998 | 19.8998 | 75.0158 | 75.0158 | 3.4997 | 3.4997 | 51.9128 | 51.9128 | |||||
p-value | 0.0107 | 0.0107 | 0.0000 | 0.0000 | 0.8353 | 0.8353 | 0.0000 | 0.0000 | |||||
LR | 31.92*** | 31.40*** | 26.75*** | ||||||||||
Log-likelihood | −71.68*** | −71.94*** | −74.26*** | −271.35*** | −335.58*** | ||||||||
Pseudo R2 | 0.1821 | 0.1792 | 0.1526 |
Note(s): Robust standard errors in parentheses; *p < 0.1; **p < 0.05; ***p < 0.01
Source(s): Own elaboration
Summary of European IP box
Country | Year of introduction | IP qualifying assets | IP box tax rate (%) | Statutory CIT rate (%) |
---|---|---|---|---|
Belgium | 2007 | Patents, Software | 4.44 | 29 |
France | 2000 | Patents, Software | 10 | 28 |
Italy | 2015 | Patents, Software | 13.95 | 24 |
Poland | 2019 | Patents, Software | 5 | 19 |
Spain | 2008 | Patents, Software | 10 | 25 |
UK | 2013 | Patents | 10 | 19 |
Source(s): OECD Dataset Intellectual Property Regimes (2019) Tax Foundation Report: Patent Box Regimes in Europe (2020); KPMG Corporate Tax Rate (2021)
Distribution companies by country
Country | IP box | Number of companies for testing hypothesis | |||
---|---|---|---|---|---|
H1 based on PBT | H1 based on EBITDA | H2 restrictions put on ETR | H2-H3 without ETR limitations | ||
Belgium | Yes | 28 | 28 | 28 | 28 |
Czech Republic | No | 20 | 16 | 16 | 26 |
France | Yes | 38 | 41 | 41 | 45 |
Germany | No | 15 | 15 | 15 | 55 |
Italy | Yes | 21 | 21 | 21 | 23 |
Japan | No | 74 | 75 | 74 | 76 |
Poland | Yes | 45 | 43 | 43 | 51 |
South Korea | No | 20 | 21 | 20 | 22 |
Spain | Yes | 25 | 26 | 26 | 27 |
Great Britain | Yes | 65 | 67 | 67 | 308 |
USA | No | 8 | 9 | 9 | 12 |
TOTAL | 359 | 362 | 360 | 673 |
Note(s): Five countries (the Czech Republic, Germany, Japan, South Korea, and the USA) did not introduce the IP Box regulations in our sample. In South Korea and the USA, tax regulations concerning R&D and intellectual property activities are implemented but differ significantly from the European Intellectual Property Box regimes
Source(s): Own elaboration
Definitions of variables
Variable | Definition |
---|---|
Dependent variables | |
ETR_PBT | |
ETR_PBT_clipp | ETR_PBT scaled to non-negative values lower than 1 |
ETR_EBITDA | |
Growth | |
Patent_pp | Binary variable indicating if a patent is registered or pending, 1 - if yes, 0 – if no |
Patent_tn_pp | Binary variable indicating if a patent, a game’s wordmark, a firm’s trademark or figurative mark is registered or pending, 1 - if yes, 0 – otherwise |
Test variable | |
IP BOX | Binary variable indicating when IP Box has been implemented equals 1 – for the year of introducing IP Box and after, and 0 – otherwise |
Control variables | |
Capital intensity | |
Intangibility | |
Inventory | |
Leverage | |
MNE | The binary variable indicating a firm with foreign shareholders equals 1 – if the firm’s foreign shareholders have more than 50% share in equity, and 0 otherwise |
Employees | |
Size | |
ROA | |
Patents | Number of patent applications or patents granted by the European Patent Office (EPO), patents granted by the United States Patent and Trademark Office (USPTO) and triadic patent families |
STR | Statutory Tax Rate |
Kaufmann | Average of six sub-index of the Worldwide Governance Indicators that rank countries concerning six aspects of good governance: Voice and Accountability, Political Stability and Violence, Government Effectiveness, Rule of Law, and Control of Corruption |
Ownership | Categorical variable indicating the type of ownership structure that consists of 15 levels: Bank; Corporate; Hedge fund; Employees, managers, directors; Financial company; Private equity firms; Public (listed companies); Venture capital; Insurance company; Foundation; Research Institute; Self-ownership; One or more named individuals or families; Other unnamed shareholders aggregated, Mutual and pension fund, nominee trust trustee; Public authority state government |
IP box’s impact on Effective Tax Rate (ETR_PBT) – system GMM results
Excl. BE, FR & ES | Excl. BE, FR & ES | Excl. BE, FR& ES | Excl. BE, FR& ES | |||||
---|---|---|---|---|---|---|---|---|
WC-robust | WC-robust | WC-robust | WC-robust | |||||
ETR_PBT | ETR_PBT | ETR_PBT | ETR_PBT | ETR_PBT_clipp | ETR_PBT_clipp | ETR_PBT_clipp | ETR_PBT_clipp | |
Coef | Coef | Coef | Coef | Coef | Coef | Coef | Coef | |
(Std. Err.) | (Std. Err.) | (Std. Err.) | (Std. Err.) | (Std. Err.) | (Std. Err.) | (Std. Err.) | (Std. Err.) | |
L1.ETR_PBT/L1.ETR_PBT_clipp | 0.0667*** | 0.0667 | 0.0551** | 0.0551 | 0.1696*** | 0.1696 | 0.1177** | 0.1177 |
(0.024) | (0.058) | (0.027) | (0.097) | (0.049) | (0.105) | (0.050) | (0.117) | |
L2.ETR_PBT | 0.0024 | 0.0024 | −0.0115 | −0.0115 | 0.0418 | 0.0418 | 0.0368 | 0.0368 |
(0.018) | (0.044) | (0.018) | (0.064) | (0.038) | (0.076) | (0.039) | (0.090) | |
IP BOX | −0.1433*** | −0.1433** | −0.1538*** | −0.1538** | −0.0803*** | −0.0803*** | −0.0954*** | −0.0954*** |
(0.041) | (0.055) | (0.039) | (0.072) | (0.021) | (0.029) | (0.021) | (0.032) | |
L1.IP BOX | −0.0482** | −0.0482* | −0.0910*** | −0.0910*** | ||||
(0.022) | (0.028) | (0.022) | (0.031) | |||||
L2.IP BOX | 0.0292 | 0.0292 | 0.0258 | 0.0258 | ||||
(0.027) | (0.030) | (0.026) | (0.038) | |||||
ROA | −0.0005 | −0.0005 | −0.0001 | −0.0001 | −0.0001 | −0.0005 | −0.0004 | −0.0004 |
(0.0004) | (0.0005) | (0.0004) | (0.001) | (0.000) | (0.0004) | (0.0001) | (0.0005) | |
MNE | −0.0738** | −0.0738** | −0.0665** | −0.0665* | −0.0785*** | −0.0785*** | −0.0721*** | −0.0721*** |
(0.032) | (0.036) | (0.031) | (0.036) | (0.020) | (0.023) | (0.020) | (0.025) | |
L1.Capital intensity | −0.0061 | −0.0061 | 0.0167 | 0.0167 | 0.0138 | 0.0138 | 0.0839 | 0.0839 |
(0.052) | (0.070) | (0.061) | (0.088) | (0.051) | (0.057) | (0.056) | (0.064) | |
Intangibility | 0.1989* | 0.1989 | 0.3485** | 0.3485 | 0.1031 | 0.1031 | 0.1219 | 0.1219 |
(0.105) | (0.194) | (0.148) | (0.319) | (0.064) | (0.081) | (0.089) | (0.138) | |
L1.Inventory | −0.0012 | −0.0012 | −0.0035 | −0.0035 | −0.0027 | −0.0027 | −0.0053 | −0.0053 |
(0.004) | (0.004) | (0.004) | (0.004) | (0.002) | (0.003) | (0.004) | (0.007) | |
Employees | −0.0553** | −0.0553 | −0.0927*** | −0.0927 | −0.0250* | −0.0250 | −0.0473** | −0.0473 |
(0.023) | (0.038) | (0.028) | (0.060) | (0.015) | (0.023) | (0.023) | (0.035) | |
Leverage | −0.0680* | −0.0680 | −0.0589 | −0.0589 | −0.0679** | −0.0679* | −0.0365 | −0.0365 |
(0.034) | (0.043) | (0.054) | (0.074) | (0.030) | (0.037) | (0.042) | (0.058) | |
L1.Size | −0.0070 | −0.0070 | −0.0052 | −0.0052 | −0.0051 | −0.0051 | −0.0005 | −0.0005 |
(0.005) | (0.006) | (0.005) | (0.008) | 0.0041 | (0.005) | (0.004) | (0.006) | |
Kaufmann | 0.0752 | 0.0752 | −0.0859 | −0.0859 | 0.0683 | 0.0683 | 0.0666 | 0.0666 |
(0.104) | (0.158) | (0.114) | (0.215) | 0.0664 | (0.090) | (0.074) | (0.115) | |
L1.ln_Patents | 0.0256 | 0.0256 | 0.0688*** | 0.0688* | 0.0162 | 0.0161 | 0.0110 | 0.0110 |
(0.024) | (0.032) | (0.025) | (0.037) | 0.0110 | (0.015) | (0.011) | (0.016) | |
Ownership effects | YES | YES | YES | YES | YES | YES | YES | YES |
Time effects | YES | YES | YES | YES | YES | YES | YES | YES |
Number of observations | 1,478 | 1,478 | 1,194 | 1,194 | 1,478 | 1,478 | 1,194 | 1,194 |
Number of groups | 285 | 285 | 219 | 219 | 285 | 285 | 219 | 219 |
Number of instruments | 61 | 61 | 61 | 61 | 59 | 59 | 59 | 59 |
Wald test | 183.03*** | 61.25*** | 238.71*** | 65.41*** | 125.87*** | 68.21*** | 157.00*** | 92.05*** |
Sargan test | 33.3707 | 33.3707 | 38.6599 | 38.6599 | 37.5799 | 37.5799 | 39.5516 | 39.5516 |
p-value | 0.3527 | 0.3527 | 0.1621 | 0.1621 | 0.1932 | 0.1932 | 0.1395 | 0.1395 |
Arellano-Bond test for zero autocorrelation in first-differenced errors | ||||||||
AR(1) | −3.6375 | −3.5081 | −3.4534 | −3.1186 | −4.8848 | −4.1753 | −4.5638 | −3.7709 |
p-value | 0.0003 | 0.0005 | 0.0006 | 0.0018 | 0.0000 | 0.0000 | 0.0000 | 0.0002 |
AR(2) | 0.6358 | 0.6024 | 0.3291 | 0.2951 | 0.6331 | 0.5645 | 0.5623 | 0.4707 |
p-value | 0.5249 | 0.5469 | 0.7420 | 0.7679 | 0.5267 | 0.5724 | 0.5739 | 0.6378 |
Note(s): *p < 0.1; **p < 0.05; ***p < 0.01
Source(s): Own elaboration
IP box’s impact on Effective Tax Rate (ETR_EBITDA) – Fixed-effects (FE) results
(1) | (2) | |
---|---|---|
IP_BOX | −0.0201*** | −0.0202** |
(0.0084) | (0.0092) | |
ROA | 0.0003** | 0.0003** |
(0.0002) | (0.0002) | |
MNE | −0.0069 | −0.0012 |
(0.0162) | (0.0167) | |
Capital intensity | 0.0031 | 0.0007 |
(0.0382) | (0.0386) | |
Intangibility | −0.0195 | −0.0213 |
(0.0333) | (0.0337) | |
Inventory | 0.0141 | 0.0102 |
(0.0245) | (0.0249) | |
Employees | −0.0048 | −0.0044 |
(0.0068) | (0.0070) | |
Leverage | −0.0339* | −0.0307 |
(0.0191) | (0.0199) | |
Size | 0.0013 | 0.0016 |
(0.0031) | (0.0031) | |
STR | 0.2002*** | 0.2240** |
(0.0938) | (0.1189) | |
Kaufmann | 0.0536* | 0.1628* |
(0.0335) | (0.1084) | |
Patents | −0.0088 | 0.0111 |
(0.0294) | (0.0378) | |
Ownership effects | YES | YES |
Private equity firm | 0.0088 | 0.0097 |
(0.0127) | (0.0130) | |
Public company | −0.0429* | −0.0407* |
(0.0215) | (0.0212) | |
Mutual & pension fund | 0.0229** | 0.0255** |
nominee trust trustee | (0.0143) | (0.0145) |
State-owned enterprise | 0.0210 | 0.0215 |
(0.0187) | (0.0181) | |
Country x Kaufmann | NO | YES |
Number of observations | 2,543 | 2,543 |
Number of groups | 362 | 362 |
R2 within | 2.53% | 4.55% |
F-statistic | 2.07** | 1.89*** |
Hausman test | 89.755*** | 66.377*** |
Breusch-Pagan test | 126.51*** | 131.43*** |
Note(s): *p < 0.1; **p < 0.05; ***p < 0.01
Source(s): Own elaboration
IP box’s impact on ETR – Group-time average treatment and event-study effects
Group-time average treatment effects | Dynamic effects | |||
---|---|---|---|---|
Group | Time | ATT(g,t) | Event time | ATT |
(Std. Err.) | (Std. Err.) | |||
2013 UK | 2012 | 0.0109 | −7 | 0.0024 |
(0.0178) | (0.0205) | |||
2013 UK | 2013 | −0.028*** | −6 | −0.0573*** |
(0.011) | (0.011) | |||
2013 UK | 2014 | −0.0601*** | −5 | −0.0066 |
(0.0234) | (0.0252) | |||
2013 UK | 2015 | −0.0365*** | −4 | −0.0223 |
(0.0151) | (0.0151) | |||
2013 UK | 2016 | −0.0402 | −3 | 0.0158 |
(0.0431) | (0.0166) | |||
2013 UK | 2017 | −0.0471 | −2 | 0.0106 |
(0.0306) | (0.0258) | |||
2013 UK | 2018 | −0.0508 | −1 | −0.0157 |
(0.0333) | (0.0208) | |||
2013 UK | 2019 | −0.028 | 0 | −0.0182 |
(0.0299) | (0.0568) | |||
2015 Italy | 2012 | 0.0045 | 1 | −0.0492** |
(0.0178) | (0.0181) | |||
2015 Italy | 2013 | −0.0242*** | 2 | −0.1022** |
(0.011) | (0.0472) | |||
2015 Italy | 2014 | 0.001 | 3 | −0.1339 |
(0.0228) | (0.0695) | |||
2015 Italy | 2015 | −0.0988*** | 4 | −0.0869** |
(0.0151) | (0.0338) | |||
2015 Italy | 2016 | −0.0383 | 5 | −0.0508 |
(0.0279) | (0.0302) | |||
2015 Italy | 2017 | −0.1679*** | 6 | −0.028 |
(0.0145) | (0.0321) | |||
2015 Italy | 2018 | −0.2276*** | Overall | |
(0.0196) | −0.067** | |||
2015 Italy | 2019 | −0.1266*** | (0.0304) | |
(0.0251) | ||||
Overall summary of ATT's based on calendar time aggregation | ||||
2013 | −0.028** | |||
(0.011) | ||||
2014 | −0.0601*** | |||
(0.023) | ||||
2015 | −0.0676*** | |||
(0.025) | ||||
2016 | −0.0393 | |||
(0.034) | ||||
2017 | −0.1075** | |||
(0.050) | ||||
2018 | −0.1392* | |||
(0.083) | ||||
2019 | −0.0275 | |||
(0.045) | ||||
Overall | −0.067** | |||
(0.027) | ||||
Overall summary of ATT's based on group/cohort aggregation | ||||
2013 UK | −0.0415*** | |||
(0.0182) | ||||
2015 Italy | −0.1319*** | |||
(0.016) | ||||
2019 Poland | 0.0721*** | |||
(0.0165) | ||||
Overall | −0.0337*** | |||
(0.0107) |
Note(s):*p < 0.1; ** p < 0.05; ***p < 0.01
*p < 0.1; **p < 0.05; *** p < 0.01 based on simultaneous confidence band
Source(s): Own elaboration
Effect of IP box on firm’s revenue growth– system GMM results
Excl. BE, FR, ES | Excl. BE, FR, ES | Excl. BE, FR, ES | Excl. BE, FR, ES | |||||
---|---|---|---|---|---|---|---|---|
Two-step | One-step robust | Two-step | Two-step robust | One-step | One-step robust | Two-step | Two-step robust | |
L1.Revenue_Growth | −0.2551*** | −0.1978*** | −0.2702*** | −0.2702*** | −0.1935*** | −0.1935*** | −0.1178** | −0.1178 |
(0.032) | (0.044) | (0.042) | (0.056) | (0.041) | (0.055) | (0.046) | (0.078) | |
L2.Revenue_Growth | −0.1295*** | −0.1019*** | −0.1137*** | −0.1137** | −0.1024** | −0.1024* | −0.0548 | −0.0548 |
(0.027) | (0.034) | (0.037) | (0.047) | (0.041) | (0.060) | (0.038) | (0.079) | |
IP BOX | −0.2271** | −0.2245* | 0.1062 | 0.1062 | −0.4597*** | −0.4597** | −0.2915*** | −0.2915** |
(0.112) | (0.128) | (0.125) | (0.127) | (0.149) | (0.207) | (0.083) | (0.132) | |
F1.IP BOX | 0.3733** | 0.3733** | ||||||
(0.168) | (0.183) | |||||||
L1.IP BOX | −0.0489 | −0.0671 | −0.0737 | −0.0737 | 0.3766** | 0.3766* | 0.4521*** | 0.4521* |
(0.178) | (0.192) | (0.179) | (0.185) | (0.149) | (0.206) | (0.159) | (0.249) | |
L2.IP BOX | 0.0740** | 0.0975** | 0.0374 | 0.0374 | −0.2084 | −0.2084 | −0.1050* | −0.1050 |
(0.036) | (0.044) | (0.047) | (0.051) | (0.147) | (0.136) | (0.061) | (0.077) | |
L3.IP BOX | 0.1338 | 0.1338 | ||||||
(0.090) | (0.092) | |||||||
Capital intensity | 0.1217* | 0.1543** | 0.1441** | 0.1441** | 0.7530** | 0.2346 | 0.2652 | |
(0.0686) | (0.0720) | (0.0678) | (0.0700) | (0.310) | (0.297) | (0.410) | ||
L1.Capital intensity | 0.7530** | 0.5245* | 0.5245 | |||||
(0.324) | (0.311) | (0.431) | ||||||
Intangibility | −0.0931 | −0.1027 | −0.0770 | −0.0770 | 0.0081*** | 0.0081* | 0.0206*** | 0.0206*** |
(0.150) | (0.160) | (0.209) | (0.217) | (0.002) | (0.005) | (0.002) | (0.004) | |
L1.Intangibility | 0.2789 | 0.4047** | 0.3029 | 0.3029 | ||||
(0.171) | (0.184) | (0.239) | (0.262) | |||||
Inventory | 0.0328 | 0.0361 | −0.0835 | −0.0835 | ||||
(0.080) | (0.092) | (0.084) | (0.089) | |||||
L1.Inventory | 0.0939 | 0.1144 | 0.1314* | 0.1344* | 0.0001 | 0.0001 | 0.0002*** | 0.0002** |
(0.069) | (0.078) | (0.072) | (0.075) | (0.0001) | (0.0001) | (0.000) | (0.0001) | |
Leverage | −0.2072** | −0.2033** | −0.1441* | −0.1441* | ||||
(0.097) | (0.010) | (0.084) | (0.084) | |||||
L1.Leverage | 0.1843** | 0.1752* | 0.0035 | 0.0035 | 0.2782** | 0.2782*** | 0.1588 | 0.1588 |
(0.079) | (0.095) | (0.077) | (0.079) | (0.131) | (0.075) | (0.223) | (0.296) | |
Size | 0.0965*** | 0.1212*** | 0.0978*** | 0.0978*** | 0.2936*** | 0.2936*** | 0.2742*** | 0.2742*** |
(0.022) | (0.024) | (0.025) | (0.026) | (0.032) | (0.078) | (0.041) | (0.073) | |
STR | −0.9434 | −0.6583 | −1.7416** | −1.7416** | ||||
(0.708) | (0.896) | (0.829) | (0.874) | |||||
L1.STR | 2.0789* | 2.0789 | 1.6387 | 1.6387 | ||||
(1.213) | (1.337) | (1.051) | (1.391) | |||||
L2.STR | 2.2249** | 2.7544** | −0.5470 | −0.5470 | 3.8669*** | 3.8669* | 2.0844** | 2.0844 |
(0.907) | (1.076) | (1.055) | (1.242) | (1.383) | (2.158) | (0.974) | (1.459) | |
ln_Patents | −0.0838 | −0.1319 | −0.1675 | −0.1675 | ||||
(0.086) | (0.143) | (0.106) | (0.131) | |||||
L1.Patents | 0.3557*** | 0.3557*** | 0.2646*** | 0.2646*** | ||||
(0.118) | (0.103) | (0.073) | (0.097) | |||||
L2.ln_Patents | −0.0943** | −0.1010 | −0.1343** | −0.1343** | ||||
(0.041) | (0.066) | (0.056) | (0.068) | |||||
Country | 0.0136 | −0.0093 | 0.2549* | 0.2549 | ||||
(0.064) | (0.109) | (0.151) | (0.348) | |||||
Kaufmann | −1.2608*** | −1.2608** | −0.5324* | −0.5324 | ||||
(0.373) | (0.582) | (0.308) | (0.379) | |||||
Ownership_Corporate | 0.0262 | 0.0516 | 0.0390 | 0.0390 | −0.1286* | −0.1286** | −0.0892* | −0.0892 |
(0.044) | (0.048) | (0.058) | (0.061) | (0.068) | (0.054) | (0.046) | (0.063) | |
Ownership_Hedge_fund | 0.0654 | 0.0730 | 0.0743 | 0.0743 | 0.1186 | 0.1186 | −0.0516 | −0.0516 |
(0.077) | (0.079) | (0.092) | (0.093) | (0.216) | (0.086) | (0.094) | (0.103) | |
Ownership_Private_equity_firm | −0.0085 | 0.0246 | 0.0559 | 0.0559 | 0.0454 | 0.0454 | 0.0200 | 0.0200 |
(0.050) | (0.057) | (0.069) | (0.071) | (0.098) | (0.053) | (0.052) | (0.068) | |
Ownership_Public company | 0.0506 | 0.0078 | 0.1648 | 0.1648 | 0.2180* | 0.2180 | ||
(0.188) | (0.201) | (0.195) | (0.129) | (0.113) | (0.149) | |||
Ownership_Venture_Capital | −0.0339 | −0.0900 | 0.0906 | 0.0906 | −0.0251 | −0.0251 | −0.0813 | −0.0813 |
(0.097) | (0.119) | (0.061) | (0.063) | (0.136) | (0.082) | (0.084) | (0.109) | |
Ownership_Insurance_company | −0.0624 | −0.0822 | −0.1301 | −0.1301 | −0.0465 | −0.0465 | −0.0121 | −0.0121 |
(0.124) | (0.140) | (0.137) | (0.152) | (0.125) | (0.088) | (0.081) | (0.101) | |
Ownership_Foundation_research_instit | 0.0567 | 0.0475 | 0.0751 | 0.0751 | 0.0843 | 0.0843 | 0.0527 | 0.0527 |
(0.073) | (0.078) | (0.084) | (0.102) | (0.247) | (0.094) | (0.064) | (0.068) | |
Ownership_Self_ownership | −0.0658 | −0.0806 | −0.1261 | −0.1261 | −0.1074 | −0.1074 | −0.0541 | −0.0541 |
(0.065) | (0.075) | (0.078) | (0.083) | (0.123) | (0.065) | (0.055) | (0.065) | |
ownership_Family firms_individuals | 0.0164 | 0.0194 | −0.0356 | −0.0356 | 0.0938 | 0.0938 | 0.0695 | 0.0695 |
(0.041) | (0.045) | (0.064) | (0.065) | (0.071) | (0.083) | (0.054) | (0.075) | |
ownership_Unnamed_shareholders | −0.1191 | −0.0917 | −0.3404 | −0.3404 | −0.1681 | −0.1681 | −0.2445 | −0.2445 |
(0.188) | (0.219) | (0.224) | (0.281) | (0.211) | (0.141) | (0.169) | (0.171) | |
ownership_Mutual & pension fund | 0.0226 | 0.0450 | 0.0375 | 0.0375 | 0.0264 | 0.0264 | −0.0063 | −0.0063 |
(0.082) | (0.092) | (0.087) | (0.094) | (0.101) | (0.085) | (0.065) | (0.079) | |
ownership_State-owned enterprises | 0.1647 | 0.1639 | 0.0269 | 0.0269 | 0.0446 | 0.0446 | 0.1130 | 0.1130 |
(0.150) | (0.158) | (0.201) | (0.201) | (0.156) | (0.123) | (0.122) | (0.157) | |
year2014 | 0.0328 | 0.0363 | 0.0627 | 0.0627 | −0.0588 | −0.0588 | ||
(0.039) | (0.048) | (0.039) | (0.041) | (0.077) | (0.104) | |||
year2015 | −0.0262 | −0.0408 | 0.0052 | 0.0052 | 0.0361 | 0.0361 | 0.0020 | 0.0020 |
(0.037) | (0.043) | (0.037) | (0.044) | (0.073) | (0.096) | (0.068) | (0.091) | |
year2016 | −0.0125 | −0.0165 | −0.0379 | −0.0379 | −0.0600 | −0.0600 | −0.0592 | −0.0592 |
(0.027) | (0.033) | (0.024) | (0.025) | (0.065) | (0.078) | (0.056) | (0.073) | |
year2017 | 0.0032 | 0.0336 | −0.0160 | −0.0160 | −0.0304 | −0.0304 | −0.0342 | −0.0342 |
(0.025) | (0.029) | (0.020) | (0.021) | (0.056) | (0.075) | (0.047) | (0.064) | |
year2018 | 0.0675*** | 0.0652*** | 0.0097 | 0.0097 | 0.0249 | 0.0249 | ||
(0.022) | (0.025) | (0.050) | (0.054) | (0.043) | (0.054) | |||
Number of observations | 4,038 | 4,038 | 2,865 | 2,865 | 1,061 | 1,061 | 1,064 | 1,064 |
Number of groups | 673 | 673 | 573 | 573 | 270 | 270 | 224 | 224 |
Number of instruments | 58 | 58 | 50 | 50 | 52 | 52 | 56 | 56 |
Wald test | 199.64*** | 127.81*** | 131.95 | 100.99*** | 178.72*** | 172.48*** | 5626.30*** | 747.87*** |
Sargan test | 27.8450 | 27.8450 | 9.7696 | 9.7696 | 30.9723 | 30.9723 | 35.5607 | 34.9131 |
p-value | 0.1809 | 0.1809 | 0.8340 | 0.8340 | 0.0556 | 0.0556 | 0.0457 | 0.0531 |
Arellano-Bond test AR(1) | −9.8449 | −9.3283 | −7.8938 | −7.0402 | −5.2057 | −5.2057 | −6.714 | −5.0413 |
p-value | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 0.0000 |
AR(2) | −1.9307 | −1.432 | −1.3076 | −1.2333 | −1.2897 | −1.2897 | −1.6266 | −1.0117 |
p-value | 0.0535 | 0.1521 | 0.1910 | 0.2175 | 0.1972 | 0.1972 | 0.1038 | 0.3117 |
Note(s): *p < 0.1; **p < 0.05; ***p < 0.01
Source(s): Own elaboration
IP box effects on innovative activity – conditional logit and complementary log-log model
Conditional (fixed-effects) logistic regression | Complementary log-log regression | ||||
---|---|---|---|---|---|
Excl. BE, FR, ES | |||||
Patent_tn_pp | Patent_tn_pp | Patent_tn_pp | Patent_pp | Patent_pp | |
Robust | Robust | ||||
IP BOX | 0.4381 | 0.4055 | 0.8955 | −0.7967 | −1.0370** |
(1.120) | (1.121) | (1.096) | (0.539) | (0.508) | |
L4.IP BOX | −1.3437** | ||||
(0.637) | |||||
L5.IP BOX | 1.9494** | 1.9091** | 1.8510** | ||
(0.778) | (0.770) | (0.754) | |||
MNE | 0.6486 | 0.6240* | |||
(0.417) | (0.358) | ||||
Employees | 1.0440 | 1.0384 | 0.9971 | 0.3519*** | 0.2671** |
(0.681) | (0.676) | (0.654) | (0.097) | (0.122) | |
Size | 0.3998 | 0.3741 | 0.3769 | ||
(0.458) | (0.456) | (0.454) | |||
STR | 12.0923*** | 13.8349*** | |||
(3.281) | (3.392) | ||||
L2.STR | 38.6350** | 40.2068** | |||
(18.453) | (18.522) | ||||
L2.Leverage | −0.8419 | −1.1042 | −0.9845 | −0.8497 | |
(1.155) | (1.123) | (0.747) | (0.806) | ||
Ownership_Family firms_ individuals | 1.0952* | 1.0798* | 0.6720 | 0.3641 | 0.1923 |
(0.649) | (0.646) | (0.605) | (0.290) | (0.252) | |
Ownership_Bank | 1.0711** | 0.8280* | |||
(0.492) | (0.442) | ||||
Ownership_Financial_company | 1.6664*** | 1.7358*** | |||
(0.474) | (0.383) | ||||
Ownership_Hedge_fund | −0.8673** | −1.1651*** | |||
(0.396) | (0.322) | ||||
Ownership_Private_equity_firm | −1.1688 | −1.1582 | −0.8898 | 0.1261 | 0.2084 |
(0.794) | (0.792) | (0.790) | (0.335) | (0.349) | |
Ownership_State-owned enterprise | 1.0867 | 1.1031 | 0.8362 | 0.6390 | 0.3495 |
(1.403) | (1.404) | (1.340) | (0.400) | (0.377) | |
Ownership_Mutual pension fund | −0.5471 | −0.5060 | −0.3299 | −0.0858 | 0.2777 |
(0.858) | (0.850) | (0.839) | (0.352) | (0.305) | |
Ownership_Corporate | −0.2135 | −0.0346 | |||
(0.373) | (0.367) | ||||
Ownership_Employees_managers_directors | 0.0623 | 0.3788 | |||
(0.479) | (0.504) | ||||
Ownership_Foundation_research_Institute | −0.5761 | −0.6452 | |||
(0.459) | (0.473) | ||||
Ownership_Insurance_company | −0.3702 | −0.2396 | |||
(0.348) | (0.345) | ||||
Ownership_Unnamed_ shareholders | −0.1798 | −0.7624 | |||
(0.686) | (1.116) | ||||
Ownership_Public company | 0.6354 | 0.6466 | |||
(0.527) | (0.601) | ||||
Ownership_Self_ownership | −0.1292 | −0.2778 | |||
(0.335) | (0.354) | ||||
Ownership_Venture_Capital | 0.4334 | 0.4837 | |||
(0.369) | (0.338) | ||||
year2013 | −9.3018*** | ||||
(1.267) | |||||
year2014 | −8.7229*** | ||||
(1.222) | |||||
year2015 | −8.1185*** | −8.5180*** | |||
(1.047) | (1.153) | ||||
year2016 | −0.5780 | −0.6209 | −0.0699 | −7.7972*** | −8.0648*** |
(0.534) | (0.533) | (0.471) | (0.983) | (1.075) | |
year2017 | −0.7836 | −0.7665 | −0.4397 | −7.6886*** | −8.0202*** |
(0.518) | (0.513) | (0.491) | (0.969) | (1.089) | |
year2018 | −0.9021** | −0.8774** | −0.7456* | −7.8413*** | −8.0833*** |
(0.448) | (0.443) | (0.443) | (0.936) | (1.027) | |
year2019 | −7.7613*** | −8.0730*** | |||
(0.942) | (1.038) | ||||
Number of observations | 6,057 | 6,057 | 6,057 | 6,057 | 4,200 |
LR | 31.92 | 31.40 | 26.75 | ||
p-value | 0.0025 | 0.0017 | 0.0084 | ||
Wald test | 448.95*** | 513.95*** | |||
Log-likelihood | −71.68*** | −71.94*** | −74.26*** | −271.35*** | −335.58*** |
Pseudo R2 | 0.1821 | 0.1792 | 0.1526 |
Note(s): *p < 0.1; **p < 0.05; ***p < 0.01
Source(s): Own elaboration
Descriptive statistics
Min | p25 | Median | p75 | Max | Mean | St. Dev | |
---|---|---|---|---|---|---|---|
H1 (panel data, N = 2,403) | |||||||
ETR_PBT | 0.000 | 0.078 | 0.198 | 0.315 | 0.988 | 0.229 | 0.199 |
ROA | −95.622 | 3.437 | 10.699 | 25.933 | 100.000 | 15.605 | 22.423 |
STR | 0.190 | 0.200 | 0.295 | 0.333 | 0.407 | 0.276 | 0.066 |
Kaufmann | 0.470 | 0.884 | 1.266 | 1.366 | 1.582 | 1.142 | 0.295 |
Employees | 0.000 | 0.693 | 1.771 | 3.135 | 9.679 | 2.000 | 1.778 |
Size | −6.787 | 4.272 | 5.491 | 7.208 | 15.006 | 5.961 | 2.655 |
Capital_Intensity | 0.000 | 0.010 | 0.031 | 0.084 | 1.000 | 0.081 | 0.133 |
Intangibility | −0.001 | 0.000 | 0.008 | 0.123 | 0.946 | 0.105 | 0.186 |
Inventory | 0.000 | 0.389 | 0.855 | 0.975 | 1.000 | 0.673 | 0.356 |
Leverage | 0.000 | 0.000 | 0.000 | 0.034 | 1.000 | 0.069 | 0.164 |
Patents | 0.000 | 0.039 | 0.142 | 0.372 | 1.000 | 0.212 | 0.200 |
Robustness Check of H1 (panel data, N = 2,543) | |||||||
ETR_EBITDA | 0.000 | 0.014 | 0.143 | 0.269 | 0.999 | 0.169 | 0.169 |
ROA | −95.622 | 0.994 | 8.892 | 22.285 | 95.938 | 12.270 | 21.802 |
STR | 0.190 | 0.210 | 0.297 | 0.333 | 0.407 | 0.278 | 0.066 |
Kaufmann | 0.470 | 0.864 | 1.249 | 1.366 | 1.582 | 1.132 | 0.296 |
Employees | 0.000 | 0.693 | 1.825 | 3.296 | 9.806 | 2.100 | 1.798 |
Size | −2.139 | 4.401 | 5.712 | 7.572 | 15.098 | 6.161 | 2.442 |
Capital_Intensity | 0.000 | 0.012 | 0.033 | 0.091 | 0.980 | 0.086 | 0.136 |
Intangibility | −0.001 | 0.000 | 0.017 | 0.179 | 0.095 | 0.130 | 0.203 |
Inventory | 0.000 | 0.360 | 0.783 | 0.966 | 1.000 | 0.647 | 0.353 |
Leverage | 0.000 | 0.000 | 0.000 | 0.048 | 0.992 | 0.074 | 0.168 |
Patents | 0.000 | 0.039 | 0.142 | 0.372 | 1.000 | 0.218 | 0.206 |
H2 (panel data without ETR limitations, N = 5,384) | |||||||
Revenue_Growth | −5.712 | 0.000 | 0.000 | 0.000 | 5.789 | 0.017 | 0.531 |
Assets_Growth | −5.979 | −0.150 | 0.000 | 0.267 | 5.858 | 0.051 | 0.821 |
ROA | −99.585 | 0.000 | 0.000 | 8.901 | 100.000 | 3.334 | 21.750 |
STR | 0.190 | 0.190 | 0.230 | 0.300 | 0.400 | 0.249 | 0.061 |
Kaufmann | 0.470 | 1.106 | 1.370 | 1.435 | 1.582 | 1.242 | 0.276 |
Employees | 0.000 | 0.000 | 1.200 | 2.700 | 10.000 | 1.619 | 1.655 |
Size | −6.787 | 3.543 | 5.049 | 6.749 | 15.098 | 5.032 | 3.086 |
Capital_Intensity | 0.000 | 0.008 | 0.032 | 0.097 | 1.000 | 0.097 | 0.167 |
Intangibility | −0.001 | 0.000 | 0.000 | 0.074 | 0.985 | 0.096 | 0.195 |
Inventory | 0.000 | 0.547 | 0.907 | 0.985 | 1.000 | 0.737 | 0.324 |
Leverage | 0.000 | 0.000 | 0.000 | 0.000 | 1.000 | 0.060 | 0.164 |
Patents | 0.000 | 0.140 | 0.142 | 0.253 | 1.000 | 0.203 | 0.184 |
H2 (panel data with restrictions put on ETR, N = 2,261) | |||||||
Revenue_Growth | −6.686 | −0.032 | 0.000 | 0.171 | 6.284 | 0.074 | 0.554 |
Assets_Growth | −5.000 | −0.030 | 0.030 | 0.300 | 5.000 | 0.109 | 0.622 |
ROA | −95.622 | 0.998 | 8.817 | 22.128 | 95.938 | 12.255 | 21.920 |
STR | 0.190 | 0.200 | 0.296 | 0.333 | 0.400 | 0.275 | 0.064 |
Kaufmann | 0.470 | 0.864 | 1.266 | 1.370 | 1.582 | 1.133 | 0.300 |
Employees | 0.000 | 0.693 | 1.825 | 3.296 | 9.806 | 2.104 | 1.800 |
Size | −2.139 | 4.407 | 5.775 | 7.636 | 15.098 | 6.184 | 2.453 |
Capital_Intensity | 0.000 | 0.011 | 0.032 | 0.090 | 0.980 | 0.085 | 0.135 |
Intangibility | −0.001 | 0.000 | 0.016 | 0.177 | 0.946 | 0.129 | 0.203 |
Inventory | 0.000 | 0.362 | 0.785 | 0.967 | 1.000 | 0.648 | 0.354 |
Leverage | 0.000 | 0.000 | 0.000 | 0.048 | 0.992 | 0.074 | 0.168 |
Patents | 0.000 | 0.039 | 0.142 | 0.372 | 1.000 | 0.213 | 0.200 |
H3 (conditional logit data, N = 6,057) | |||||||
IP Protection | 0.000 | 0.000 | 0.000 | 0.000 | 1.000 | 0.038 | 0.191 |
Employees | 0.000 | 0.000 | 1.200 | 2.657 | 9.806 | 1.617 | 1.651 |
Size | −6.788 | 3.526 | 5.030 | 6.699 | 15.098 | 5.013 | 3.071 |
STR | 0.190 | 0.190 | 0.240 | 0.3000 | 0.407 | 0.253 | 0.062 |
References
Alstadsæter, A., Barrios, S., Nicodeme, G., Skonieczna, A. M., & Vezzani, A. (2018). Patent boxes design, patents location, and local R&D. Economic Policy, 33(93), 131–177. doi: 10.1093/epolic/eix021.
Balsmeier, B., Fleming, L., & Manso, G. (2017). Independent boards and innovation. Journal of Financial Economics, 123(3), 536–557. doi: 10.1016/j.jfineco.2016.12.005.
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143. doi: 10.1016/S0304-4076(98)00009-8.
Bornemann, T., Laplante, S. K., & Osswald, B. (2020). The effect of intellectual property boxes on innovative activity & effective tax rates. Wu International Taxation Research Paper Series, 2018–03. doi:10.2139/ssrn.3115977.
Bradley, S., Dauchy, E., & Robinson, L. (2015). Cross-country evidence on the preliminary effects of patent box regimes on patent activity and ownership. National Tax Journal, 68(4), 1047–1072. doi: 10.17310/ntj.2015.4.07.
Bradley, S., Robinson, L., & Ruf, M. (2021). The impact of IP box regimes on the M&A market. Journal of Accounting and Economics, 72(2-3), 101448. doi: 10.1016/j.jacceco.2021.101448.
Callaway, B., & Sant’Anna, P. H. (2021). Difference-in-differences with multiple time periods. Journal of Econometrics, 225(2), 200–230. doi: 10.1016/j.jeconom.2020.12.001.
Chen, S., De Simone, L., Hanlon, M., & Lester, R. (2019). The effect of innovation box regimes on income shifting and real activity. Working Paper No. 3453. Stanford University, Graduate School of Business, Stanford, CA. doi: 10.2139/ssrn.3486428.
Cunningham, C. C. (2018). Tax Reform Deduction: What Foreign-Derived Intangible Income means to C-Corporations. BDO. Available from: https://www.bdo.com/insights/tax/tax-reform-deduction-what-foreign-derived-intangible-income-means-to-c-corporations# (accessed 5 May 2021).
Dyreng, S. D., Hanlon, M., & Maydew, E. L. (2008). Long-run corporate tax avoidance. The Accounting Review, 83(1), 61–82. doi: 10.2308/accr.2008.83.1.61.
Dyreng, S. D., Hanlon, M., Maydew, E. L., & Thornock, J. R. (2017). Changes in corporate effective tax rates over the past 25 years. Journal of Financial Economics, 124(3), 441–463. doi: 10.1016/j.jfineco.2017.04.001.
Evers, L., Miller, H., & Spengel, C. (2015). Intellectual property box regimes: Effective tax rates and tax policy considerations. International Tax and Public Finance, 22, 502–530. doi: 10.1007/s10797-014-9328-x.
Gaessler, F., Hall, B. H., & Harhoff, D. (2021). Should there be lower taxes on patent income?. Research Policy, 50(1), 104129. doi: 10.1016/j.respol.2020.104129.
Gupta, S., & Newberry, K. (1997). Determinants of the variability in corporate effective tax rates: Evidence from longitudinal data. Journal of Accounting and Public Policy, 16(1), 1–34. doi: 10.1016/S0278-4254(96)00055-5.
Hall, B. H., Thoma, G., & Torrisi, S. (2007). The market value of patents and R&D: Evidence from European firms. Academy of Management Annual Meeting Proceedings, 8, 1–6. doi: 10.5465/ambpp.2007.26530853.
Haufler, A., & Schindler, D. (2020). Attracting profit shifting or fostering innovation? On patent boxes and R&D subsidies. CESifo Working Paper No. 8640. Available from: https://ssrn.com/abstract=3720397 (accessed 22 April 2021).
Jones, S., Baker, M., & Lay, B. F. (2017). The relationship between CSR and tax avoidance: An international perspective. Australian Tax Forum, 32(1). Available from: https://ssrn.com/abstract=2954291 (accessed 22 April 2021).
Karkinsky, T., & Riedel, N. (2012). Corporate taxation and the choice of patent location within multinational firms. Journal of International Economics, 88(1), 176–185. doi: 10.1016/j.jinteco.2012.04.002.
Koethenbuerger, M., Liberini, F., & Stimmelmayr, M. (2018). Is it just luring reported profit? The case of European patent boxes. CESifo Working Paper Series No. 7061. Available from: https://ssrn.com/abstract=3211147 (accessed 22 April 2021).
Laplante, S. K., Skaife, H. A., Swenson, L. A., & Wangerin, D. D. (2019). Limits of tax regulation: Evidence from strategic R&D classification and the R&D tax credit. Journal of Accounting and Public Policy, 38(2), 89–105. doi: 10.1016/j.jaccpubpol.2019.02.003.
Lester, R. (2021). Tax accounting research on corporate investment: A discussion of the impact of IP box regimes on the M&A market by Bradley, Ruf, and Robinson (2021). Journal of Accounting and Economics, 72(2-3), 101451. doi: 10.1016/j.jacceco.2021.101451.
McGuire, S. T., Omer, T. C., & Wang, D. (2012). Tax avoidance: Does tax-specific industry expertise make a difference?. Accounting Review, 87(3), 975–1003. doi: 10.2308/accr-10215.
Merrill, P. (2016). Innovation boxes: BEPS and beyond. National Tax Journal, 69(4), 847–862. doi: 10.17310/ntj.2016.4.06.
Minasian, J. R. (1962). The economics of research and development. In The Rate and Direction of Inventive Activity: Social and Economic Factors. Princeton University Press. doi: 10.1515/9781400879762-004.
Minasian, J. R. (1969). Research and development, production functions, and rates of return. The American Economic Review, 59(2), 80–85.
Mohnen, P., Vankan, A., & Verspagen, B. (2017). Evaluating the innovation box tax policy instrument in The Netherlands, 2007–13. Oxford Review of Economic Policy, 33(1), 141–156.
Newzoo (2020). Top 10 countries/markets by game revenues. Available from: https://newzoo.com/insights/rankings/top-10-countries-by-game-revenues/ (accessed 22 April 2021).
OECD (2020). Compendium of R&D tax incentive schemes: OECD countries and selected economies. Available from: http://oe.cd/rdtax (accessed 22 April 2021).
OECD (2015). Countering harmful tax practices more effectively, taking into account transparency and substance. Paris: OECD Publishing. Action 5–2015 Final Report. doi: 10.1787/9789264241190-en.
OECD (2019). Intellectual property regimes, Available from: https://qdd.oecd.org/subject.aspx?Subject=IP_regimes (accessed 22 April 2021).
Ohrn, E. (2016). The effect of IP box regimes on international IP payments and foreign research and development. Working Paper. Grinnell College. Grinnell, Iowa. Available from: https://ericohrn.sites.grinnell.edu/files/IP_Box/IP_Box_8_2016.pdf (accessed 10 May 2021).
PwC (2020). Global entertainment & Media Outlook 2020–2024. Available from: https://www.pwc.com/gx/en/industries/tmt/media/outlook.html (accessed 10 May 2021).
Rego, S. O. (2003). Tax-Avoidance activities of U.S. Multinational corporations. Contemporary Accounting Research, 20(4), 805–833. doi: 10.1506/VANN-B7UB-GMFA-9E6W.
Rutkowski, E., Marszałkowski, J., & Biedermann, S. (2020). The game industry of Poland — report 2020. Polish Agency for Enterprise Development, Warsaw. Available from: https://www.parp.gov.pl/storage/publications/pdf/The_Game_Industry_of_Poland_report_2020v4.pdf (accessed 10 May 2021).
Thomsen, M., & Watrin, C. (2018). Tax avoidance over time: A comparison of European and U.S. Firms. Journal of International Accounting, Auditing and Taxation, 33, 40–63. doi: 10.1016/j.intaccaudtax.2018.11.002.
Zhong, R. I. (2018). Transparency and firm innovation. Journal of Accounting and Economics, 66(1), 67–93. doi: 10.1016/j.jacceco.2018.02.001.
Zimmerman, J. L. (1983). Taxes and firm size. Journal of Accounting and Economics, 5, 119–149. doi: 10.1016/0165-4101(83)90008-3.
Further reading
Dischinger, M., & Riedel, N. (2011). Corporate taxes and the location of intangible assets within multinational firms. Journal of Public Economics, 95(7-8), 691–707. doi: 10.1016/j.jpubeco.2010.12.002.
Grubert, H. (2003). Intangible income, intercompany transactions, income shifting, and the choice of location. National Tax Journal, 56(1, Part 2), 221–242. doi: 10.17310/ntj.2003.1S.05.
Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The worldwide governance indicators: Methodology and analytical issues. World Bank Policy Research Working Paper No. 5430. Available from: https://ssrn.com/abstract=1682130 (accessed 22 April 2021).
Acknowledgements
The research presented in the article was carried out as a part of the project financed by the National Centre for Research and Development in Poland no. GOSPOSTRATEG-VI/0029/2021-00 “The monitoring of innovation performance of firms and regulatory impact assessment: developing tools to support economic policy”, acronym FIRMINREG. The authors thank Gabriel Ralile de Figueiredo Magalhães, Rita de Cássia Claudiano Cabral Pinto and Maria Inês Lopa Ruivo for their data collection efforts.
Since acceptance of this article, the following author(s) have updated their affiliation(s): Ricardo Sichel is at the Universidade Federal do Estado do Rio de Janeiro – UNIRIO, Rio de Janeiro, Brazil.