The effect of patent acquisition on subsequent patenting activity
The importance of the patent market has increased in the high technology market due to open innovation and social efficiency. As most high-tech firms cannot rely on their own repository of technologies, they need to employ external technologies. These can be acquired through mergers and acquisitions (M&A) or from the patent markets, which is often referred to as open innovation. Because of insufficient comprehensive data and the complexity and contingency provision of the contract, there have been only a few empirical studies on patent transactions. This study seeks to understand the patenting activities of patent buyers after acquisition. By merging patent bibliographies and patent assignment information, the original dataset was prepared to investigate the patent acquisition activities of the patentees who were granted patents from 2007 to 2011 and their patent application activities from 2012 to 2016. The factors affecting patent production and acquisition and the relationship between patent acquisition and subsequent production were examined. The results show that the patent acquisition activity depends on the applicant type and technology field, and affects the future patent production of the applicant. Thus, observing a competitor’s patent acquisition activities gives insights into the future direction of its technology development.
The patent market has rapidly grown in recent years [, , ]. A study estimates the increase of licensing revenues at approximately $100 billion worldwide between 1950 and 2003 . The global patent revenue is reported to have increased from $15 billion in 1990 to $100 billion in 2000 . Several reports suggested the growth of the patent market in other aspects, such as the secondary patent market (from less than $1 billion in 2013 to $12.5 billion in20171) and the licensing market (from $135 billion $ in 2005 to $300 billion in20142).
Part of the reason for the growth of the patent market is the recognition of patents as a valuable source of income rather than a simple legal right protection for inventions [6,7]. The income from the patents, often represented as private and social gains by reallocation of the patent rights to other firms, is generated by patent transactions, which takes place through different forms such as licensing agreement, sales, or transfer. Another reason is that most of the high-tech industries nowadays operate in extremely complex technological environments. Firms cannot rely upon the repository of technologies developed by themselves, and hence they have to go beyond their boundaries in pursuit of external sources of technologies , which is often referred to as open innovation . External technologies can be acquired by buying from the patent markets [10,11] as well as through co-operation and alliance [12,13], mergers and acquisition (M&A). M&A allows not only to obtain tangible resources such as expert  but also to absorb technological knowledge and capabilities of the target company. Therefore, M&A for technical purpose has a substantial effect on subsequent internal innovation activities of the acquirer [15,16].
According to the previous study the reasons and driving forces for patent transactions are reported . The former is categorized into six groups: Strategic, monetary, managerial and firm-specific, patent-specific, transaction cost, and exogenous, and the latter is found to be most relevant to “the demand side” and “strategic behavior” with 40 companies in several industries. It is also found that the patents included in a firm’s core business are not likely to be sold, whereas the ones that are unrelated to the firm’s core business, with a short residual life and minor awareness of suitable licensees are likely candidates for sale.
The creation of monetary benefits through the enforcement of patent rights is the core of the business model of “non-practicing entities (NPE)”. They do not produce patented products or services, but they have a significant impact on the patent market and affect industry innovation in terms of facilitating the integration and movement of knowledge .
Considering patents as valuable assets, which can be traded, the price of them is a significant factor to maximize commercial gains. Hence there have been studies, though very limited, on examining the relationship between the age of patent and price. According to a recent study  the relationship between the patent age and selling price was examined using some 500 US patent families which were sold in US market (patent auction) in different technology fields. The authors categorized patents into single invention and portfolio lots, each of which consists of several different cases, and performed change point analysis to demonstrate the change of patent price with age. They reported that the price of patent sold after the change point (3700 days from grant) was higher especially for single invention lots for computers and communications fields. The mean price of patent sold after the change point was 496 thousand dollars and was much higher than the mean price of younger patents (180 thousand dollars). Their findings could have implications on the bundling strategy for technology transfer manager or patent holder’s technology.
Patent transactions are important in high-technology markets for developing the market structure, which consists of large and small firms. The latter are, in general, specialized in new technologies but are short of mass production and marketing capacities as compared with the former. By transferring the technologies of small firms to their larger counterparts, social gains are realized efficiently . Patent holding companies, in other perspective, create patent pools for standard essential patents (SEP) and allow cross-licensing transactions to pursue innovation more efficiently . The representative companies that have recognized the commercial value of patents and have been active with patent transactions are IBM and Texas Instruments .
A limited number of empirical studies on patent transactions have been reported due to the scarcity of comprehensive data and the complexity and contingency provision of the contract. Studies based on patent auctions [7,, , , , , ] and patent reassignments [, , , , ] are the representative ones. The study on patent transactions is in the early stages and most of the studies are therefore empirical, with a limited range of data and a variety of research interests. Some of the selected research interests are described below.
Serrano  contributes the use of the data on the transfer and renewal of patent ownership to the empirical literature on patent transactions. Change of patent ownership, more specifically the dependence of the rates of patent transfer and renewal on the types of patentees and technology fields, is analyzed in this study using the legally required records in the assignment text provided by the US Patent and Trademark Office (USPTO). The relative importance of patent transfer and renewal with respect to the firm size for each technology field is another main interest of this study. The obvious finding is the difference in the relative importance of small and large patentees in some technology fields. For example, small patentees show considerably (3.0 times) larger sales rate than larger ones in the computer and telecommunication field, while this ratio is only 1.37 in the chemistry.
Figueroa and Serrano  find that small firms are more engaged in patent licensing than larger ones, but this trend cannot always be applied to patent trade. They examined the determinants of small and large firms’ decisions on patent sale and acquisition using the combined dataset of patent assignment, patent renewal fee, and patent application and grant data. Their findings address several important issues: Small firms sell and disproportionately acquire more patents than large firms, especially those originally registered to other small firms. The employment of patent citation as a potential proxy of patent transaction is notable, in an extension of the usage of patent citation as the proxy of patent value . Patents with high citation count (received) have a higher chance of being sold to large firms. Small firms disproportionately cite other small firms’ patents. The patents with a higher proportion of citations made by small firms have a higher chance of acquisition by a small firm.
A study by Galasso, Schankerman, and Serrano  focuses on the effect of patent trade on litigation. Other focus of their study is the importance of the commercialization and enforcement gains from patent trade. The data for the study is a combination of US patents by individual inventors registered during 1983–2000 period, trade, and litigation information. Some of their findings address that the reduced litigation risk of the individually owned patents indicates that the enforcement gains are more important than market gains, and that the patents with a high probability of reassignment are expected to return large gains from patent transactions. The other finding is that characteristics such as the portfolio size of the buyer and the fit of the patent of interest to the buyer’s patent portfolio in terms of technology are related to the impact of the patent trade.
Drivas and Economidou  examined the geographic effect on knowledge flows via two different channels, market and non-market, where knowledge flows via patent transactions and patent citation, respectively. They used the reassignment data of US-issued patents to examine the knowledge transfer based on whether the patent transaction is affected across the states and sectors in the US. According to their results, the geography in terms of distance and contingency restricts patent transaction and hence the knowledge flows much more than in patent citation. Patent citations can be employed to capture actual knowledge flow from one party to another, other than trade of goods or inventors’ mobility. In other words, firms that bought the patent could also previously cite it as a token of interest in the knowledge covered by it.
Caviggioli  examined the strategies behind the patent acquisitions via two different channels (patent market and M&A). The target of their study were ten firms sampled in three industries (i.e., automotive suppliers, semiconductors, and computer networks), which were the holding ones with respect to the number of patent transactions in each industry. According to their study, acquired patents are more complex than internally developed ones, and of higher technological merit, closer to basic research, and hence more technologically focused. Patents acquired via patent market cover less complex technologies than ones acquired via M&A, according to characteristics such as the lower count of backward citations, claims, and inventors.
Although the number is limited, previous studies have suggested ways to utilize patent transaction data from various perspectives, such as to investigate the characteristics of the technology market , to track the flow of knowledge between organizations , and to analyze the management of patent portfolio such as patent renewal, sales, and litigation preparation [27,31]. A few studies compared external technologies acquired through patent transactions with internal technologies for a limited number of companies [18,29]. The current study aims to contribute to the existing literature by associating the external knowledge acquired through the patent transaction directly with the internal innovation from the perspective of the patentee organization, using comprehensive data instead of a case study. In other words, this study explores the patent activities of patentees after the acquisition of external patents. The patent bibliographies and the patent assignment information were merged to prepare the original data set to investigate the patent acquisition activity of patentees from 2007 to 2011 and patent application activities from 2007 to 2016. The results show that patent acquisition depends on the type of patentee and the technology field of the patent. The effect of patent acquisition on the quantity and quality of subsequent patent production varies by the type of patentee and the technology field.
This paper is organized as follows. Section 2 introduces the data construction process. Section 3 describes the patent acquisition activities of patentees with the descriptive statistics on explanatory variables. Section 4 analyzes the relationship between patent acquisition and subsequent patent production. Section 5 discusses the findings and presents the conclusions.
2. Data construction
This study aims to understand the patenting activities of patent buyers<sup>3</sup>, and thus requires two types of data: patent bibliographies and patent assignment information. Patent bibliographies such as applicant names and International Patent Classifications (IPCs) are extracted from PATSTAT<sup>4</sup> provided by the European Patent Office (EPO). The applicant names are standardized by a simple unification of upper/lower-case letters and elimination of spaces and punctuations (e.g. commas and periods). Number of utility patents (granted) are measured for each applicant. Citation count is measured for each patent to be used as a measure of patent quality. The information on patent acquisition is retrieved from the Patent Assignment Text by USPTO, which includes the elements such as a unique identifier for each assignment (reel and frame number), the names of seller (assignor) and buyer (assignee), the date it was recorded at USPTO, the date the parties signed on the private agreement, associated patent numbers, conveyance text, and other information associated with the patent right transfer. The conveyance text, which is the indicator of the reason of patent transaction, is standardized following the process mentioned more in detail afterwards. The names of buyers are matched with the applicant names of the granted patents to examine the patent production activities of the applicant/buyer depending on the experience of patent acquisition from outside the firm.
This study focuses only on the utility patents granted in the US. The granted patent provides exclusive rights to the inventor and the applicant. Therefore, it is appropriate to analyze the granted patents in order to properly distinguish whether patents with exclusive rights are produced from inside or from outside. However, it takes considerable time<sup>5</sup> from the patent filing to being granted. Knowledge secured by internal production or by inflow from outside should be regarded as affecting the production of patents filed afterwards<sup>6</sup>. However, since it is difficult to measure the impact of each patent separately, it is appropriate to set the period to measure the impact. To examine the subsequent patenting activities, the overall timeframe (2007–2016) is divided into two, with a five-year<sup>7</sup> period for each. The first (2007–2011) and second (2012–2016) time frames are referred to as Phase 1 and 2, respectively.
Various reasons for patent right transfer are recorded in the “conveyance text” field of Patent Assignment database. As the conveyance text is in a free-text format rather than a set of pre-defined texts, it is difficult to identify the reason for each patent right transfer. The conveyance text needs to be mapped into a pre-defined set of reasons for ease of further analysis.
Marco et al. proposed an approach for this task in their USPTO working paper , which is used in this study after minor modifications and a few additional treatments. The conveyance texts are mapped by keyword search and pattern matching into the following ten reasons of patent right transfer: correction (to a prior record); release (of a security agreement); merger; change of name; government (interest agreement); employer assignment; security (agreement); other (license, joint research agreement, etc.); assignment; missing. A complete list of conveyance types and corresponding matching keywords are given in Appendix A. These reasons are listed in the order of the following classification process. A set of conveyance texts is classified based on a keyword search corresponding to a specific reason. For example, the keywords such as “correc*, conver*, amed*, etc.” are searched, and the matched documents are classified to the first conveyance type, “correction”. Remaining documents are treated with the second set of keywords for the second conveyance type, “release”. This process is repeated with the remaining pairs of conveyance text and keyword sets. The conveyance texts which do not match any of the keywords are classified as “missing”.
In the next step, administrative events among the various types of reasons, such as change of patentee’s name or address, security interest, and correction, need to be screened to find the “effective” patent right transfer. In this process, identifying “employer assignment (patent right transfer within an organization from inventing employee to employer organization)” is especially important. According to the approach of Marco et al. the employer assignment is identified when a record meets the following conditions:
“a) the earliest transaction recorded for the property (rf_id with earliest execution date); and b) the property was transferred alone (i.e., no other properties were listed in the PTO-1595 cover sheet); and c) the execution date is prior to the patent application disposal (grant or abandonment) date (or December 31, 2014 for pending applications); and d) keyword searching identifies the conveyance text as an assignment.”
USPTO provides the descriptions and analysis of the patent assignments that is used in this study as a reference material for mapping the conveyance text to the reasons of patent right transfer.<sup>8</sup>
Effective patent right transfer is generally recorded as “assignment” or “merger” besides “employer assignment”. Within these two types of patent right transfer, however, the size of the patent bundle needs to be considered to identify the patent trade for technological purposes. Though “merger” might not represent technology transfers in general , M&A of small firms very often occur for acquisition of technology. The patents traded in large bundles might not represent technological transfer as is in the case of a merger between two large firms. For these reasons, all transactions involving more than ten patents are excluded .
The result of data processing is summarized in Fig. 1. A total of 930,111 utility patents (P1) are granted in Phase 1, which are filed by 96,995 applicants (N1) after the name standardization. A total of 378,020 patents (T1) are acquired by 45,217 buyers (N4) after the name standardization in Phase 1. The name-matching process between the applicants of the granted patents (N1) and the buyers of the patent assignments (N4) in Phase 1 gives 20,193 unique names (N2) existing in both of the databases. These are those who have both the patent acquisition and production experiences: acquiring 312,272 patents (T2) from outside the firm either through merger or change in ownership and producing 692,864 patents (P2). Only 19,504 applicants (N3) out of N2 acquire 108,487 patents (T3) in small acquisition and produce 689,414 patents (P3).
Future Technology Analysis Center, Korea Institute of Science and Technology Information (KISTI), Seoul, South Korea Future Technology Analysis Center, Korea Institute of Science and Technology Information (KISTI), 66 Hoegi-ro, Dondaemun-gu, Seoul, 02456, Republic of Korea