Master’s Thesis
Distinguishing patent transactions between separate organizations
and within corporate groups, contributing to the strategic
environment of patent transactions.
Zeger Wever (s2955067)
MSc Business Administrations – Strategic Innovation Management
Supervisor: Dr. P. Arqué-Castells
Co-Assessor: N.E. Fabian
1 Table of contents Abstract ... 2 1. Introduction ... 3 2. Literature review ... 5 3. Hypotheses development ... 6 4. Data ... 9 5. Methodology ... 11 6. Results ... 12
7. Discussion and conclusion ... 12
Appendix 1. Tables and graphs... 15
2 Abstract
This paper aims to validate current literature on the increase of technology markets and patent transactions. The generalizability of the literature is questioned by addressing a potential distinction in patent transactions between separate organizations and within corporate groups. Reviewed literature justifies the assumption on differences in the amount of patents per transaction and the technological proximity between the transacting partners. Secondary data samples were constructed from the USPTO Patent Assignment Dataset to descriptively analyse both assumptions. Average testing confirms a lower amount of patents per transaction for corporate groups, while higher technological proximity for corporate groups contradicts the initial hypothesis.
3 1. Introduction
According to The United States Patent and Trademark Office (USPTO) the grants of utility patents in the U.S. has increased 6,7% on average per year in the period 2008-2018 (USPTO, 2019). Besides the number of granted patents the total amount of patent transactions also follows an increasing trend (De Marco, Scellato, Ughetto and Caviggioli, 2017). According to Serrano (2010) 13.5% of all granted patents are traded at least once over their lifecycle. The value of these patents transaction in the field of technology is 9% of the total non-defence R&D spending (Arora, Fusfori and Gambardella, 2001). Next to the increase in patent transactions literature on technology markets has become more common as well (Arora et al., 2001; Figueroa and Serrano, 2019; Hagiu and Yoffie, 2013). In addition to specific numbers regarding patent transactions and volume the USPTO gathered and assigned information about the organizations and technological fields wherein the patents were traded (Serrano, 2010). This enabled studies on the variables of the strategic environment such as organization size (Figueroa and Serrano, 2019), location (De Marco et al., 2017) and technological fields (Arora et al., 2001).
Literature on the basic principles of external knowledge acquisitions tends to simplify the incidentals involved in the transaction of patents (Teece, 1986; Pisano, 1990). Contradictory, Arora et al. (2001) depict technology markets as complex environments wherein transactions are characterized by detailed contracts and often take place in technological alliances. The complexity of patent transactions is enhanced by the versatility of objectives that patents can serve. Patents can be utilized internally, to (cross)license or to block others from engaging in the patented technology (Giuri et al., 2007). Differences in the utility of patents could justify differences in patent transaction between separate organizations or within corporate groups. Where transactions between separate organizations are expected to have the purpose of acquiring knowledge, transactions between corporate group are more likely to serve cost, tax or strategy related issues (Heald, 2003; Trajtenberg, 1990).
4
To elaborate on existing literature in technology markets it is essential to address the actual transfer of information within a patent transaction. The likelihood of patent transactions serving different purposes when the transacting partners belong to the same corporate group undermines the classic concept of knowledge sharing but might form a base for new research directions. It is assumable that patent transactions within corporate groups can occur with less restrictions and therefor deviate from the patent transactions between separate organizations. In order to investigate whether there are differences in the first place the following research question is formulated.
Do patents transactions between separate organizations differ in nature from patent transactions within corporate groups?
To test the research questions and the underlying motives this research focuses on two aspects of the patent transaction: technological proximity and the number of patents traded within one transaction. Putting the technological sectors of the trading organizations in perspective with the technological factors of the patents reveals potential prerequisites of patent trading actors. Understanding how technological proximity is related between patent transactions within corporate groups or between separate organizations indicates if a distinction between both sorts of transactions is justifiable in the first place.
Technological proximity covers facets of information sharing, spillover effects and absorptive capacity. The latent theoretical charge makes technological proximity a favourable variable to determine the distinction between both types of patent transactions. Patent transactions arise from the extension of the organizations complementary assets. The defined scope and intended spillover effects of separate organizations narrow the potential transaction partners, increasing the technological proximity. In contrast the structure of corporate groups enables different branches to cooperate fostering spillover effects regardless of the technological proximity.
The amount of traded patents within a transaction shifts the focus from the patent level to the organizational level. Confirming differences between corporate groups and separate organizations will solidify the distinction between the two types of transactions. Transacting partners within corporate groups will be less prone to information asymmetries easing the draw of contracts. Less boundaries enables multiple patent transactions to occur with lower amount of patents per transaction. Whether potential differences are incidentals to corporate groups, or organizations adopt certain structures to foster these differences will be additional research topics that to this paper.
5
randomized samples feature standardized organization names which enabled dividing the data in patent transactions between separate organizations and within corporate groups.
Average tests were executed to test both hypotheses confirming the lower amount of patents per transaction for corporate groups. In contrast with the initial hypothesis technological proximity turned out lower for patent transactions between separate organizations. The contribution of this hypothesis is however not without value since the distinction between both sort of patent transactions remains present.
2. Literature review
Organizations are not limited to their own practices in when it comes to innovation. Oppositely innovation is endorsed by knowledge from outside the organization. Albeit to benefit from the recombination of knowledge (Grimpe and Kaiser, 2010; Griffith, Lee and Straathof, 2017) or to economize on investments in research or complementary assets (Hagedoorn, 2002; Teece, 1986). Acquisition of external knowledge can take place in a variety of ways including the transaction of patents. Incidentals of knowledge acquisition through the transfer of patents are the direct and indirect data resulting from the reported transactions. Trends and connections in patent data have formed the fundamentals for subsequent research and literature development.
The importance of external knowledge acquisition is affirmed by Arora et al. (2001) who emphasize the contribution of external knowledge acquisition in the innovation strategies of organizations. In this paper the increase in patent transactions is assimilated to the growth of technology markets. Combining the importance of external knowledge acquisition with the reproducibility of patent data has led to follow-up research in the strategic environment of patent transactions. Where Arora et al. (2010) already indicated distinctions for knowledge acquisitions in different technological fields, De Marco et al. (2017) expanded the scope of the theoretical field by adding locational aspects of the transacting organizations to patent trading patterns. Figueroa and Serrano (2019) later included organizations size in justifying the engagement in patent transactions.
6 3. Hypotheses development
In order to create a sustainable competitive advantage an organization should feature resources that have value, are rare and are neither imitable nor substitutable (Barney, 1991). The appropriability of these resource characteristics can be protected by the issuance of patents (James, Leiblein and Lu, 2013).The ability to generate rents from these resources, however, depends on the organizations complementary assets (Teece, 1986). Patent application is costly and generating rents can be impeded when subsequent investments have to be made in complementary assets. Licensing or selling patents to other organizations can increase the rents generated from these patents. Contrary, organizations can follow the make-or-buy principle by choosing to acquire or license patents when this is economically more viable than investing in the development of the technology itself (Arora et al., 2001). A major part in the decision to trade or licence patents is affected by the spillover of information. Positive spillover effects arise from the recombination of knowledge and can be beneficial for both licensor and licensee. The licensor generates rents from patents and the licensee receives advantages of knowledge recombination. However, negative spillover effects can occur when sharing knowledge leads to a change in the market position of the rivalling licensor and licensee (Bloom, Schankerman and Van Reenen, 2013).
The severity of knowledge spillover consequences is intertwined with the collaboration governance between the trading organizations. The relation between the trading organizations as well as the knowledge that is being shared and the intentional spillover effects determine the modes of collaboration (Faems, Janssens, Madhok and Van Looy, 2008). Knowledge transfer through patent transactions can be delimited by implementing detailed contracts, which is not more than usual according to Arora et al. (2001). When two separate organizations are operating in close proximity regarding their technology and markets, non-competition clauses and non-disclosure agreements could be built into contracts to prevent negative spillover effects. This process of market-based transactions is costly and time consuming, however it is required to prevent opportunistic behaviour from the transaction partner (Faems et al., 2008; Powel, 1990).
7
The value of patent information is largely determined by the complementary assets of organizations (Teece, 1986; James et al., 2013). The licensee or acquiring organization has to be able to implement the acquired patents in the organization in order to benefit from recombination of knowledge (Grimpe and Kaiser, 2010). In order to do so absorptive capacity is required. Optimal utilization of externally acquired knowledge is more likely to take place when this knowledge can be processed and implemented through internal knowledge systems (Cassiman and Veugelers, 2006). A licensee or acquiring organization is expected to have internal knowledge capabilities that are compatible and similar to the fields of knowledge where the patent arose from and therefor having a high technological proximity with the transaction partner.
As patents contain highly specified information which requires similar complementary assets to be implemented in the organization (Teece, 1986) transferring patents can be considered illogical when the technological proximity of the transaction partners is lower. Transacting knowledge processes (Mudambi and Tallman, 2010) or even engaging in joint ventures (Deitz, Tokman, Richey and Morgan, 2010) are means of knowledge transactions that bridge the differences in complimentary assets and are therefore more likely to take place when the technological proximity is lower.
The rivalry in the market that restricts both intended and unintended spillovers is either less or not present when both parties belong to the same corporate group. On the contrary, partners involved in transactions within a corporate group could stimulate unintended spillovers without appropriability concerns. Patent transactions within the corporate group are expected to be less prone to information asymmetries than patent transactions between organizations since the partners in the former serve the same parent. Easing the transaction process in combination with the elimination of negative spillover effects enables additional patent transactions even when technological proximity between partners is high. At the same time it enables the addition of patents as a complement to other modes of knowledge sharing when the technological proximity is lower, whereby the spread of technological proximity is expected to be wider in patent transactions within the corporate group.
8
activities ensured by financial coverage of the CHQ in exchange of intellectual property. Organizing the corporate group in specialized entities can structure efficient R&D (Foss, 1997) and economize on tax expenditures at the same time (Heald, 2003). The artificial patent transactions that follow these organizational structures are unlikely to be found when partners do not belong to the same corporate group. Especially in corporate groups that specialize in multiple technological sectors the probability of patent transactions between two entities with different technological backgrounds are likely to take place.
Summarized low technological proximity is expected in patent transactions between separate organizations since patents serve as an extension for their complementary assets. Other modes of knowledge sharing appear to be more suitable when the technological proximity is low, therefor patent transactions between separate organizations are not expected when the technological proximity is lower. These effects are mitigated when the transacting partners belong to the same corporate group since information asymmetries and negative spillover effects will be less present. Additionally low technological proximity is expected due to artificial patent transactions in specifically designed corporate groups, which leads to the formulation of the first hypothesis:
H1: Patent transactions within corporate groups take place at lower technological proximity than patent transactions between different organizations.
In addition to the aforementioned market-based transactions, patent transactions between firms can also be based on hierarchy and network (Powell, 1990). Contracts and mutual expectations form key elements in transactions between separate organizations while dependency and routines will add more value in transactions between affiliates. Negotiation costs and information asymmetries will decrease when organizations involved in the transactions strive for the goals of the corporate group instead of their own benefits.
9
occurrence of having lower amounts of patent transacted in single transactions is expected to be higher in corporate groups. Subsequently the second hypothesis is formulated:
H2: The amount of patents traded in one transaction is lower within a corporate group than between different organizations.
Corporate groups can structure their organizations in specific ways that separate organizations cannot. Following the CHQ theory of Foss (1997) corporate groups can separate their organizations in specialized divisions that strive for the benefits of the CHQ. Patents can flow from R&D subsidiaries to CHQ in exchange for royalties or reduction of losses. These patents can subsequently flow to holding or blocking subsidiaries in exchange for royalties to be paid to the CHQ. Artificial patent transaction comprising the illogical flow of patents is likely to occur when these transacted patents are not serving the purpose of knowledge sharing but are exclusively used as transferrable assets serving strategical purposes (Heald, 2003; Trajtenberg, 1990). When they are reported as patent transactions to the USPTO a distortion of the ratio of number of patents per transactions could occur. The presence of lower amounts of patents per transactions as argued in previous paragraphs is reinforced by this duplicate reporting.
4. Data
This paper elaborates on the converted USPTO PAD of Arqué-Castells and Spulber (2017). Patents filed under the USPTO are not limited to US organizations only, they are however legally protected in the U.S. Arqué-Castells and Spulber (2017) assigned data from Compustat, SDC platinum, Osiris and the NBER Patent Data Project DYNASS file to Compustat GVKEYS. IDDR identifiers link patent transactions to standardized assignors (or_iddr) and assignees (ee_iddr) names. Subsequently these or_iddrs and ee_iddrs are linked to GVKEYS that comprise standardized organization names. Additional provided data by Arqué-Castells and Spulber (2017) consists of the amount of transacted patents within one transaction and the classification of transacted patents in the NBER categories (Hall, Jaffe and Trajtenberg, 2001).
10
categories (Hall et al., 2001). The main dataset uses the Real Frame ID number (rf_id) (USPTO, 2015) to describe each individual transaction. Furthermore the main dataset consists of the ee_iddr and or_iddr per transaction as well as their standardized corporate group (ee_gvkey and or_gvkey) and the year wherein the transaction is reported (eyear).
In order to test both hypotheses subsamples have to be constructed by merging the additionally provided data to the main dataset. There will be two subsamples since merging conditions differ and the main dataset is not equally covered by the additional provided data. To create subsample 1 additional data on transacted patents is used. The additional provided dataset on transacted patents consists of 1,254,791 patents accompanied by the ee_iddr or or_iddr through which the patents were transacted, the NBER technology class and the year wherein the patents were traded.
Patents are individually linked to NBER categories (Hall et al., 2001). This classification is more difficult for organizations because they can operate in multiple technological fields and therefor fit in multiple categories. Technological proximity however, can be proxied by calculating the amount of overlapping patent categories between assignors and assignees. To create a technological proximity variable a subset of variables has to be created first. The patents can be classified into 418 different NBER technology classes varying from 2 to 850 (Hall et al., 2001). These classes were rearranged from group 1 to 418 in order to make them sequential and were labelled as “technology_class”. The variable “total” counts the number of transacted patents for each ee_iddr or or_iddr. The variable “total_tech” counts the number of transacted patents per technology class for each each ee_iddr or or_iddr. To enable subsequent stages in the analysis the dataset is expanded by adding the remaining technology classes to each ee_iddr and or_iddr regardless of the fact if patents were transacted in that technology class.
Dividing “total_tech” by “total” and multiplying it by 100 results in a percentage on the weighted value of technology classes per ee_iddr and or_iddr. This percentage is labelled as the “subshare” variable. The sum of these “subshares” squared is labelled as “sumsubshare_squared”. Eventually the variables ee_iddr, or_iddr, technologyclass, subshare and sumsubshare_squared are preserved.
11
or_iddr. After adding the variable “technologicalproximity” to the main dataset 96,516 patent transactions remained.
The USPTO (2015) is not able to confirm whether transactions where all variables except rf_id’s are the same concern different patents or involve the same patents but are reported by different entities. Additionally the randomized sample was derived from a population where duplicates were present. Therefor duplicate transactions are disregarded leaving 21,903 patent transactions. To distinguish patent transactions between separate organizations and corporate groups the variable “relation” was created labelling a transaction “within” if the ee_gvkey and or_gvkey matched (14,978 observations) and “between” if the ee_gvkey differed from the or_gvkey (6,925 observations).
To create subsample 2 the additional provided data on the amount of patents per transaction was merged to the main dataset using rf_id as the key variable. The additional dataset consisted of the rf_id which was merged and the variable “npatents” describing the number of patents per transaction. Of all 110 466 patent transactions the amount of transacted patents were reported resulting in a fully covered merge. However, as the unedited main dataset was used duplicate transactions had to be removed again leaving 24,480 patent transactions. The variable “relation” as described in the previous paragraph was added to divide the sample in 16,928 observations were the ee_gvkey and or_gvkey matched and 7,552 observations were the ee_gvkey and or_gvkey differed.
5. Methodology
This paper aims to address the complete strategic environment of patent transactions including the dimension of organizational relations. The acknowledgement of the USPTO by previous research in the strategic environment as well as the broad scope covered by the patent transactions registered to the USPTO contribute to the validity of the USPTO PAD data in testing both hypotheses. The amount and diversity of patent data gathered by the USPTO ensures the generalizability of the results. The coherent downfall of USPTO data is that parameters would have to be applicable to all patent transactions limiting the amount of specified information. The data therefor limits the testing of relational patterns, however it suffices in descriptive tests.
12
mean of both subsamples was rounded up to the next whole year (2004) and the tests were re-executed for the samples upward of this mean.
6. Results
Variance tests for the “Npatents, 2004-2013” and “Technological proximity, 1978-2013” subsamples turned out unequal therefor t-test with variances assumed to be unequal were executed for these subsamples, for the other subsamples variances were assumed to be equal. The results are shown in table 1.
The t-test results provide no support for hypothesis 1. In fact, the results debunk the hypothesis by showing significant lower technological proximity for patent transactions between separate organizations. The slight increase in both means for the subsample “Technological proximity, 2004-2013” does not alter these results. Although the initial hypothesis was not supported, finding the opposite still contributes to the main research question since the distinction between both organizational relations remains present.
The amount of patents traded per transaction was found to be lower for corporate groups as stated in hypothesis 2. Additional to the significance (ρ <0.01) the provided results show a substantial difference in patents per transaction for the sample as a whole. As shown in table 1, this difference becomes even greater when the subsample for the years 2004-2013 is utilized, resulting in more than twice the amount of patents per transaction for separate organizations compared to transactions within corporate groups.
Notable is the increase in the different standard deviations for the subsample 2004-2013 whereas they were near to equal for the sample as a whole. In line with the hypothesis development the t-test were executed on the transaction level, disregarding the number of patents per transaction. Therefore the results are affected by organizations trading single patents in high amounts of transactions.
7. Discussion and conclusion
13
The results contribute to the dismemberment of patent transaction data across the strategical environment wherein the transactions take place, adding the dimension of inter- and extra organizational patent transactions. The importance of this dichotomization for the literature depends on the intention of future research and the measurements used to execute this research. A notice was found in this paper concerning the technological proximity. Technological proximity was calculated by transaction count and therefore significantly affected by the high amount of single patent transactions within corporate groups. Alternatively the technological proximity could be weighted by the number of traded patents per transaction. In defence of the former, the development of the hypothesis was based on the transaction level. Additionally multiplying the technological proximity by the number of patents would have to be justified since a single patent transaction with high technological proximity would not automatically equal a transaction with multiple patents and low technological proximity. As this research adds a factor to the strategic environment of patent transactions, theoretical implications are linked to the amount of overlap between the theory and the acknowledged dimension. Subsequently the level on which patent data is analysed contributes to the valuation of this implication. Research topics regarding patent flows or transaction volume can be highly affected by the dichotomy of patents transactions depending on the value attached to the relational background of the transacting partners.
In order to research a potential division in patent data between separate organizations and within corporate groups this research is focused on analysing a population wherein other dimensions are assumed to be homogenous. The downside of this holistic view is that it lacks interaction with other dimensions of the strategic environment considering the location and size of organizations as well as the technological fields wherein they operate. Now that a distinction has been addressed follow-up research could address subcategories in which this distinction is more or less clearly present.
14
15 Appendix 1. Tables and graphs
Table 1. T-test results
Obs. Mean SE. SD. t ρ
Technological proximity. Between.1978-2013 6,925 .417 .004 .307 -61.6 <0.01
Technological proximity. Within.1978-2013 14,978 .697 .003 .326
Technological proximity. Between. 2004-2013 3,077 .456 .006 .317 -38.6 <0.01
Technological proximity. Within. 2004-2013 8,658 .712 .003 .315
Npatents. Between.1978-2013 7,552 18.3 1.315 114.3 4.7 <0.01
Npatents. Within.1978-2013 16,928 11 .879 114.4
Npatents. Between. 2004-2013 3,401 20.2 2.616 152.6 3.8 <0.01
16 References
Arqué-Castells, P., & Spulber, D. (2017). Matching assignees and assignors in the USPTO Patent Assignment Dataset to Compustat firms.
Arora, A., Fosfuri, A., & Gambardella, A. (2001). Markets for Technology and their Implications for Corporate Strategy. Industrial and Corporate Change, 10 (2), 419-451.
Barney, J. (1991). Firm Resources and Sustained Competitive Advantage. Journal of Management, 17 (1), 203-227.
Bloom, N., Schankerman, M., & Van Reenen, J. (2013). Identifying technology spillovers and product market rivalry. Econometrica, 81 (4), 1346-1393.
Cassiman, B., & Veugelers, R. (2006). In Search of Complementarity in Innovation Strategy: Internal R&D and External Knowledge Acquisition. Management Science, 52 (1), 68-82.
Deitz, G., Tokman, M., Richey, G., & Morgan, R. (2010). Joint venture stability and cooperation: Direct, indirect and contingent effects of resource complementarity and trust. Industrial Marketing
Management, 39 (5), 862-873.
De Marco, A., Scellato, G., Ughetto, E., & Caviggioli, F. (2017). Global markets for technology: Evidence from patent transactions. Research Policy, 46 (9), 1644-1654.
Faems, D., Janssens, M., Madhok, A. & Van Looy. (2008). Toward an integrative perspective on alliance governance: connecting contract design, trust dynamics, and contract application. Academy
of Management Journal, 51 (6), 1053-1078.
Figueroa, N., & Serrano, C. (2019). Patent trading flows of small and large firms. Research Policy, 48 (7), 1601-1616.
Foss, N. (1997). On the rationales of Corporate Headquarters. Industrial and Corporate Change, 6 (2), 313-338.
Giuri, P., Mariani, M., Brusoni, S., Crespi, G., Francoz, D., Gambardella, A., Garcia-Fontes, W., Geuna, A., Gonzales, R., Harhoff, D., Hoisl, K., Lebas, C., Luzzi, A., Magazzini, L., Nesta, L., Nomaler, O., Palomeras, N., Patel, P., Romanelli, M., & Verspagen, B. (2007). Inventors and invention processes in Europe. Research Policy, 36 (8), 1107-1127.
Griffith, R., Lee, S., & Straathof, B. (2017). Recombinant innovation and the boundaries of the firm.
International Journal of Industrial Organization, 50, 34-56.
Grimpe, C., & Kaiser, U. (2010). Balancing Internal and External Knowledge Acquisition: The Gains and Pains from R&D Outsourcing. Journal of Management Studies, 47 (8), 1483-1509.
Hagedoorn, J. (2002). Inter-firm R&D partnerships: an overview of major trends and patterns since 1960. Research Policy, 31 (4), 477-492.
17
Hall, H., Jaffe, A., & Trajtenberg, M. (2001). The NBER patent citations data file: lessons, insights and methodological tools. National Bureau of Economic Research, working paper 8498.
Heald, P. (2003). A Transaction Cost Theory of Patent Law. Ohio State Law Journal, 66 (3), 473-509. Jaffe, B. (1986). Technological Opportunity and Spillovers of R & D: Evidence from Firms' Patents, Profits and Market Value. The American Economic Review, 76 (5), 984-1001.
James, S., Leiblein, M., & Lu, S. (2013). How Firms Capture Value From Their Innovations. Journal of
Management, 39, (5), 1123-1155.
Mudambi, S., & Tallman, S. (2010). Make, Buy or Ally? Theoretical Perspectives on Knowledge Process Outsourcing through Alliances. Journal of Management Studies, 47 (8), 1434-1456. Pisano, G. (1990). The R&D Boundaries of the Firm: An Empirical Analysis. Administrative science
quarterly, 35 (1), 153-176.
Powel, H. (1990). Neither Market Nor Hierarchy: Network Forms of Organization. Research in
Organizational Behavior, 12, 295-336.
Serrano, C. (2010). The dynamics of the transfer and renewal of patents. RAND Journal of Economics, 41 (4), 686-708.
Teece, D. (1986). Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy. Research policy, 15 (6), 285-305.
Trajtenberg, M. (1990). A penny for your quotes: Patent citations and the value of innovations. The
RAND Journal of Economics, 20 (1), 172-187.
USPTO. (2015). The USPTO Patent Assignment Dataset: Descriptions and Analysis. USPTO Economic Working Paper. Retrieved from:
https://www.uspto.gov/sites/default/files/documents/USPTO_Patents_Assignment_Dataset_WP.pdf USPTO. (2019). U.S. Patent Statistics Chart Calendar Years 1963 – 2018. Retrieved from:
