Download Complementary Technologies and Returns to Disclosure ... and more Exercises Local Area Network (LAN) in PDF only on Docsity!
Complementary Technologies and Returns to Disclosure during Standard Setting
Abstract
In industries with many technological components that need to be interoperable, coordination is
increasingly achieved via firms’ disclosure to Standard Setting Organizations. We study whether
and how disclosure generates returns to the firm during standard setting. Departing from the
convention of focusing on disclosed standard essential patents (SEPs), we examine the role of the
firm’s non-disclosed complementary technologies in generating returns. Main results show that
firms with more non-disclosed complementary technologies experience greater returns to stock
prices over disclosure events, and these technologies gain more in value (indicated by patent
citations) over disclosure, at rates higher than even that of the disclosed SEPs. Our findings
suggest that, in these systems, a firm is using its larger technological portfolio to appropriate value
from coordinating a smaller part of it with others via standard setting, and that a systemic view of
the firm’s portfolio is important in understanding strategy within these systems.
Keywords: Complementary technologies, standard setting, disclosure
1. INTRODUCTION
In many industries, firms operate within systems consisting of complementary technological components that need to be interoperable (Cargill 1989, Kaplan 1996). These technologies are complementary in the sense that they rely on each other to function as a system and thus are super-additive in value (Milgrom and Roberts 1995, Baldwin and Woodward 2009). An example is the set of technologies in a mobile communication network related to chipsets, customer premise equipment, transmission, switchboard, etc. When they are supplied independently by different firms without clear presence of a ‘platform leader’ (Bresnahan and Greenstein 1999), coordination in development becomes crucial, without which the system risks ‘forking’ and suffering from incompatibility issues (Rosenkopf and Tushman 1998). This coordination is increasingly managed by standard setting organizations (SSOs). SSOs enable firms to voluntarily come together to develop and agree on standard specifications that ensure compatibility across components (Updegrove 2007, Leiponen 2008). Coordination is achieved when firms reach consensus on these specifications and formally disclose to SSOs their standard essential patents (SEPs) – patents on which any implementer of the standard will by definition have to infringe – or at least disclose that they possess relevant SEPs, along with agreements to license the SEPS to any implementer (Bekker et al. 2012 , Lerner and Tirole 2015). As a firm partakes in this standard setting exercise, one of its central concerns is ‘whether and how disclosure and becoming part of industry standard generates positive returns to the firm ’. This question is by no means straightforward and to date still insufficiently addressed. While recent research shows empirically that the SEPs themselves gain in intrinsic value after disclosure (Rysman and Simcoe 2008, Bekkers et al. 2017), as measured by increased citations received or litigation rates, it remains in question if this increased value accrues to the disclosing firm. The ‘openness’ of technologies resulting from standard setting (Weiss and Sirbu 1990) does not necessarily translate to improved financial wellbeing of the disclosing firm.
returns to the firm as the SEPs get disclosed during standard setting, and what happen to them after standard formation. There is a practical basis to justify this focus on non-disclosed complementary technologies. Even though, in a stylized depiction, the complementary technologies with which the one firm is trying to coordinate development in the system are all owned by other firms, in reality, a firm active in standard setting would usually possess some of these complementary technologies itself. Examples that come to mind are Qualcomm and Nokia. In our sample (described later), more than two-thirds of disclosing firms own multiple of these complementary technologies themselves, not all of which are disclosed to SSOs. In 2013, Nokia successfully sued HTC in UK High Court and won preliminary US ITC ruling against HTC for infringing on its non-disclosed non-SEP, and sought injunction and financial compensation. By this point, Nokia had asserted more than 50 non-disclosed non-SEPs against HTC. The notion that ‘ a firm discloses its SEP during standard setting to subsequently profit off its non-disclosed complementary technologies ’ appears prima facie valid but has not received empirical validation thus far. It furthers prior research suggesting that a firm uses its influence to determine what becomes SEP (Leiponen 2008, Bekkers et al. 2011), though the implicit firm objective here is less about increasing the value of what is disclosed, and more about profiting from their non-disclosed counterparts upon disclosure of SEPs. Our first proposition is that the more non-disclosed complementary technologies a firm possesses, the greater the returns in firm value from the firm’s disclosures during standard setting. To test this, we examine disclosures made by firms to 13 SSOs during standard setting in the information and communication technology (ICT) industry 1988-2010, and use both panel fixed effect regressions with Tobin’s Q as a measure of firm value as well as a stock market event study. Main findings are as follows. In general, a firm’s disclosure does not seem to increase its stock market returns; in fact, evidence suggests that such effect of disclosure is negative. However, its non-disclosed complementary technologies significantly increase its returns over the disclosure event. In the event study, one standard deviation increase in complementary technologies increases the cumulative abnormal returns over a three-day
window by 0.48 percent on average, translating into approximately $328 million increase in market value over three days. Imposing firm, industry, standards area, and year fixed effects, and varying event windows yield similar findings. A challenge with this empirical design is that possession of complementary technologies can be an endogenous firm choice, and a firm with such possession may be systematically different than another firm without, in ways that are unobservable and cannot be captured even with fixed effect models (time-varying unobserved attributes). To check if this problem is generating biased estimates, we further use inverse probability-weighted regression adjustment (IPWRA) models (Wooldridge 2007 , Elfenbein et al. 2010), which allow us to account for the firm’s propensity to have complementary technologies, and then reduce the abovementioned problem by essentially comparing treatment effects (of having complementary technologies) on a set of treated observations relative to the same set of observations themselves as if they have not been treated. Findings remain similar. We also stress-test the logic of the proposition as follows. If it is true that the firm’s non-disclosed complementary technologies increase overall firm value over the disclosure event (marginal effect) because they are more compatible with what became the industry standard, then this positive marginal effect on firm value should diminish when the complementary technologies are also compatible with rivals’ technologies similar to the firm’s disclosed SEPs in the first place. Likewise, if the positive marginal effect arises because the non-disclosed complementary technologies become more competitive against rival technologies over the disclosure event, then this marginal effect should be lessened if the firm is operating in a less competitive industry in the first place, especially if the firm is a larger firm within this industry structure. Evidence supports both tests. Our second proposition is that the firm’s non-disclosed technologies complementary to its disclosed SEPs become more valuable after standard setting. To test this, we conduct the analyses at the patent level, using subsequent citation received as an indication of value (Rysman and Simcoe 2008). Analyses reveal three main findings: non-disclosed complementary non-SEPs received more citations after the disclosure event than before (13.3 percent yearly increase), their rate of increase in citations received is higher after
enforceability, changes in value, terms of licensing, etc., our findings suggest that managers should also focus on appropriation conditions surrounding their firm’s complementary technologies as they disclose the SEPs. Our findings may have policy implications as well. In supervising standard setting activities, the objective of institutions such as the U.S. Department of Justice and Federal Trade Commission (FTC) is usually to ensure that firms are not locking in users at monopolistic prices. Cases in point are FTC’s lawsuits against Dell in 1996 and more recently against Rambus in 2005. There have also been efforts to insert protective mechanisms in the standard setting process to prevent disclosing firms from subsequently reneging on licensing at FRAND rates once lock-ins occur (Updegrove 2007, Sidak 2015). Our findings raise the possibility (though by no means indicative) of a different potential problem: even if a firm’s assertion of market power over its disclosed SEP is curtailed, the question remains as to whether it may still create socially undesirable lock-ins via its non-disclosed complementary non-SEPs that are not subjected to oversight. By retaining control over licensing and use of these non-SEPs, the firm may yet create distortions from socially optimal equilibrium prices and usage as it maximizes profits. We clearly do not offer evidence of such distortions in this paper, but simply raise a possibility that may be worth examining.
2. RELATED LITURATURE ON STANDARDS We begin by reviewing related literature on formation of compatibility standards^2 by SSOs and how it could generate positive returns for the disclosing firm. A standard functions as a mechanism for selection and diffusion of the firm’s technologies included in it (Antonelli 1994). It increases demand of these technologies by mitigating potential hold-up problems as perceived by users and reducing transaction costs and purchase risk (Weiss and Sirbu 1990). By enabling coordination across producers, it triggers a network effect that further raises users’ valuation of the firm’s technologies (Demsetz 1993). It also (^2) Compatibility standards – technical specifications that enable interoperability between various components in a technological system – are different from performance standards, which specify minimum required levels in quality, safety, etc., or reference standards in pharmaceutical drugs (Farrell and Saloner 1988).
increases supply of complementary components in the system by allowing adopters and other complementors to access technological specifications and underlying IP, decreasing their design and production costs, and reducing barriers to entry into markets for products based on the firm’s standard technologies (Farrell and Saloner 1985). Firms are likely aware that participating in standard setting can potentially generate positive returns, judging from the strategic behavior they exhibit in the standard setting process (Lerner and Tirole 2006, Simcoe et al. 2009 ). They ‘forum-shop’ for SSOs that provide most favorable terms to them (Chiao et al. 2007) and exert their influences in determining which IPs become SEPs (Leiponen 2008, Bekkers et al. 2011). There is increasing concern in recent years that firms are purposefully disclosing low quality patents in late stages of the process to be included in standards, so as to extend their monopoly and ex post bargaining power (Lerner and Tirole 2015, Kang and Bekkers 2015). Participants in SSOs would attempt to safeguard and advance their own high-stake vested interests, sometimes ending up creating coordination difficulties as a result (Farrell and Saloner 1988, Farrell and Simcoe 2012). For instance, Simcoe (2012) finds that strategic behavior by participants delayed the adoption of several important internet standards. Even though the merits of standard setting are well understood, there has been relatively few empirical studies that directly examine whether participating in standard setting and disclosing SEPs to SSOs generate positive returns, especially to the disclosing firm itself. Rysman and Simcoe (2008) was an instrumental step in this direction: using received patent citations as indicators of value, they find that consensus standards not only select SEPs that are valuable to begin with, but that post-adoption, these SEPs become even more valuable. But they stop short of showing whether the disclosing firm gets to reap this increase return. Recent research provides some evidence that the disclosing firm gains from participation, though only under some conditions. Aggarwal, Dai, and Walden (2011) find that announcement of a standard increases a participating firm’s stock returns, but as the number of other participating firms increases, the firm’s stock returns decreases while idiosyncratic risk rises. Pohlmann et al. (2015) find a positive correlation between a firm’s return on assets and ownership of SEPs, but this correlation rapidly
also disclose.^6 At many SSOs, firms disclose their IPs via declaration letters or emails to the SSO. Letters can take either form: specific disclosures that list patents and IP claims over essential technologies, or blanket disclosures that indicate ownership of IPs but do not list specific patents or patent applications.^7 Disclosures to SSOs can be costly to the disclosing firm. It is not uncommon for a firm to spend millions of dollars on standard related activities (Chiao et al. 2007). SSOs’ IP policies typically dictate that the disclosing firm must make SEPs available to be licensed on a FRAND basis (Bekkers et al. 2012). This applies even to IPs that were not previously mentioned by the firm during the standard setting process, as long as the firm had participated in the committee for a certain minimum amount of time (e.g. more than 60 days). These licensing obligations preclude the firm from exclusive use of the technologies covered by its SEPs, and stifle its ability to block or slow rivals’ development along similar technological trajectory (Clarkson and Toh 2010). Through disclosure, valuable information about its patent portfolio and technology directions may be revealed to rivals (Chiao et al. 2007). Even during the workgroup discussions before disclosure, knowledge beyond that specified in essential patent documents can spillover to rivals (Rosenkopf et al. 2001).^8 This facilitates rivals’ imitation and allows them to potentially beat the firm in applying the underlying technologies in ways beyond that specified in the standard. Returns from licensing SEPs may not fully compensate for the cost of disclosures. Pricing SEPs and setting licensing rates ‘fairly’ are challenging tasks (Lerner and Tirole 2015).^9 Most SSOs’ IP policies commit firms to FRAND licensing terms. Such ex post negotiations would bring rates to ‘middle-ground’ (^6) It is also possible for a participant to disclose another participant’s technology, though based on our analysis of disclosures, this is very rare. (^7) The common understanding is that firms use blanket disclosures when they own large patent portfolios with high search cost of determining the exact relevant patents or claims to be listed in the disclosure (Updegrove 2007). However, scholars debate on whether high search cost is the key reason for using blanket disclosures. Bekkers and Martinelli (2013) suggest that firms use blanket disclosures when they have lower quality patents. (^8) From a conversation with an industry participant, we learn that firms are aware of such spillovers and try to mitigate them by only sending managers with substantial standards setting experience to participate in workgroup discussions. (^9) While theoretically, the disclosing firm may ex ante expect licensing rate for its SEP to surpass perfect competition level since there would be few or no competing technologies once the standard is set, in practice, to avoid antitrust concerns, the licensing rate is not ex ante considered or discussed during the selection process (Lerner and Tirole 2015). Rather, pricing negotiations only begin ex post , when the standard specifications are determined and the selection of SEPs have occurred. Many SSOs adopt strict IP disclosure and licensing rules to deter malignant pricing strategies (Bekkers et al. 2015), as encouraged by the Antitrust Division of the U.S. Department of Justice, which often end up unfairly favoring licensees over SEPs owners (Sidak 2015).
levels that may not favor the disclosing firm (Sidak 2015). Some standards utilize patent pool, which is an arrangement among multiple patent holders to aggregate their SEPs to be licensed out as a packet (Merges 2001). It avoids the need for multiple party-to-party negotiations and allows a licensee to access all patents in the pool for one fee. SEP owners typically split the licensing revenue based on their proportion of patents in the pool, and firms with more important IP may receive less revenue than they otherwise would have under party-to-party negotiations. Moreover, royalty-free licensing is being used in many important standards (e.g. mobile browsing). This, coupled with the costliness of disclosure, raises the question of whether there are other mechanisms that the firm is using to generate returns over disclosure.
4. COMPLEMENTARY TECHNOLOIGES AND RETURNS FROM DISCLOSURE We go on to discuss how the disclosing firm’s non-disclosed complementary technologies acts as a mechanism to generate returns for the firm over the disclosure of SEPs. Two technologies are complementary when their joint value created in functioning together is larger than the sum of values they create in functioning individually (Milgrom and Roberts 1990; 1995 ). Complementarity usually requires that the technologies are compatible and rely on each other to properly operate (Baldwin and Woodward 2009). The technologies represent different parts of a broader combined solution to a problem, and are often used or built on jointly. For example, Qualcomm developed a ‘high rate packet data transmission’ technology (patent: US6173007) that sends data over a wireless network channel. This technology enables efficient ‘packeting’ of data to better utilize the communication channel and became an essential technology in the UMTS standard. However, when used by itself, it does not allow for fair allocation of data speed to all users in the network. To solve the problem, Qualcomm created a complementary technology for generating optimal data packet lengths (patent: US6064678) to be used jointly with the former. This complementary technology is programmed into base station software and optimizes the data transmission system performance by ensuring that all users get their fair share of data throughput.
components in the GSM network as well. Importantly, this signal clarity enhancement technology is not disclosed as part of the GSM standard, and hence not subjected to the cost of disclosure. Second is the increased competitiveness of the firm’s complementary technologies against rivals with competing technologies of similar functionalities. As the firm often co-designs both the disclosed and nondisclosed technologies together as part of a larger blueprint, its complementary technology is frequently the best way to harness the functionality of its disclosed technology, relative to rivals with similar substitutes. The firm has a head start against rivals who have yet to develop substitute technologies compatible with the new standard. With a portfolio of technologies that are already compatible, it is more attractive to licensees and potentially has greater bargaining power over them. It can embed its complementary technologies in other components in the system, and extend its advantage of being part of the standard to other parts of the system. Using the Nokia example: upon Nokia’s radio signal technology becoming part of the GSM standard, Nokia already has its signal clarity enhancement technology that is designed specifically as a complement, that rivals may not be able to replicate quickly, that is attractive to licensees, and that Nokia embeds into its phones to make the phones more attractive to users.^10 The complementary technologies also allow the firm to have a head start against rivals as the industry expands on the use of the standard technologies in other areas of application. For instance, in the late 2000s, many industries began to utilize wireless communications networks, in which Qualcomm has a strong foothold. Qualcomm was able to build on its existing complementary technologies to expand its wireless applications into industries such as automotive, home appliance, and health care.^11 (^10) To further illustrate the concept: Nokia was in an opposite situation with the CDMA based wireless communication standard (compared to the TDMA based standard described in the earlier example). Within the CDMA standard, it lacked complementary technologies needed to differentiate its CDMA mobile phones. In order to compete in the CDMA based phone market, Nokia would have needed to invest in developing the whole set of complementary technologies, or license them from rivals like Qualcomm and Motorola. Nokia deemed this to be unprofitable, and consequently exited the CDMA-based phone market to focus on the TDMA-based market instead. See “Nokia to exit CDMA after scrapping Sanyo JV plans” via http://www.itwire.com/it-industry-news/strategy/4724-nokia-to-exit- cdma-after-scrapping-sanyo-jv-plans. [Last accessed, March 2017.] (^11) See Qualcomm Incorporated press release, May 14, 2015: https://www.qualcomm.com/news/releases/2015/05/14-
- [Last accessed, March 2017.]
Based on the above, we arrive at the following two main predictions. (1) The more non-disclosed complementary technologies a firm possesses, the greater the returns in firm value from the firm’s disclosures during standard setting. Next, if it is indeed the non-disclosed complementary technologies that are driving up returns to the firm, then these technologies themselves should be increasing in value over the disclosure. Thus, (2) the firm’s non-disclosed technologies complementary to its disclosed SEPs become more valuable after standard setting.
5. EMPRICAL ANALYSIS 5.1 Data & Complementary Technologies Measure We conduct the empirical analysis in the information and communication technology (ICT) industry 1988-2010. Standards are essential for many products within the ICT industry to function properly, and many firms within the industry engage in standard-setting activities across multiple standards and SSOs. Examples of standards include: wireless telecommunication standards (e.g. GSM, GPRS, CDMA, WCDMA, LTE), local area networks (e.g. Wi-Fi), mobile browsing (e.g. xtml), and audio-video compression (e.g. MPEG-4). We use several data sources in the analysis. The Disclosed Standard Essential Patents (dSEP) database (Bekkers et al. 2012) provides data on IP disclosures to SSOs. The dSEP Database includes date of disclosure, entity disclosing, name of the standard or technical committee, name of the SSO, and any disclosed patents. We collect patent data from the U.S. Patent and Trademark Office (USPTO), citation data from the National Bureau of Economic Research (NBER) (Hall et al. 2001) and the Patent Network Dataverse (Lai et al. 2011 ), firm financials from Compustat and 10Ks, stock prices from the Center for Research in Security Prices (CRSP), and information on information on mergers, acquisitions, and stock buybacks from Thomson SDC. To create the disclosure dataset, we use the dSEP database to identify 4,609 disclosures events to 13 different SSOs in 1988-2010. Each disclosure event consists of letters or emails to a given SSO on a
Technology takes the value of one if the patent meets the above criteria. In firm level analyses, Complementary Technologies is a count of the total number of patents meeting the above criteria. To illustrate our measure, consider the following example.^12 Nokia’s SEP, “Support of multiplicity of radio interfaces over an interface between a base station system and mobile switch ” (US 5,664,004), specifies a method for allowing a mobile radio system to support multiple radio interfaces over a single signaling interface. The patent was disclosed to ETSI’s GSM standard. Nokia’s complementary technology, “Method and an arrangement for setting up a data call, and an adapter equipment” (US 6,192,055), combines with the above described SEP to allow for data services simultaneously with voice communication, thus increasing the functionality and value of a digital communication system. To identify complementarity between the two patents, we observe their joint citation by Intel’s “Wirelesses local loop with intelligent base station”^13 (US 7,050,799) which combines the multiple radio interface technology of 5,664,004 and adapter equipment specification of 6,192,055 in the specification of a wireless local loop system. There may be underestimation in our measure of Complementary Technologies , as it does not capture complementarities of SEPs with non-patented technologies. However, focusing on a single industry (^12) This footnote details two additional examples. We begin with the complementary patents described in the first paragraph of Section 4. Qualcomm’s SEP, “High rate packet data transmission” (US 6,173,007), specifies how to send data over a wireless network channel, but by itself, does not allocate throughput fairly. Qualcomm’s complementary technology, “Method for assigning optimal packet lengths in a variable rate communication system” (US 6,064,678), optimizes the data transmission system performance to ensure all users get their fair share of data throughput. The complementarity between the two patents can be identified by Lsi Logic Corporation’s “Accelerated message decoding” (US 6,567,390), which builds on how both of Qualcomm’s packet data transmission technologies co-function. Another example comes from Nokia. Nokia’s SEP, “Method for decreasing the frame error rate in data transmission in the form of data frames” (US 6,178,535), helps reduce errors that occur in transmission that can render speech signals unintelligible. To take advantage of the above method, Nokia has several complementary technologies that can be implemented in a mobile handset so that it can transmit a signals in a way in which, when combined with the SEP, errors can be reduced and voice clarity can be improved. Patents “Data transmission method and transmitter” (US 5,949,790), and “Spreading factor determination” (US 6,678,314) detail these complementary technologies. To identify complementarity, we rely on a patent such as Nokia’s “Coupled error code protection for multi-mode vocoders” (US 6,681,203), which draws on all three above mentioned patents to reduce faulty data frames sent from and by detecting errors received by handsets. (^13) A local loop is a simple local mobile system that connects a set of fixed-positioned wireless terminals to a wired telephone system (also known as a plain old telephone system or POTS).
where patenting propensity is similar across firms and using within-firm variance both help minimize the problem. In Section 5. 7 , we report the results of several robustness tests with the measure as well as discuss the results using an alternative measure. 5.2. Tobin’s Q Analysis We begin our analysis by tracing how firm performance changes in response to its disclosure of SEPs. Our sample consists of firms for which ICT standards are relevant, which we define as those that patent in 89 USPTO technology classes related to communications equipment as per the NBER concordance system. After matching with Compustat and CRSP, we have an unbalanced panel of 394 firms with observations that span the years 1988 - 2010. We measure firm performance using Tobin’s Q, a common measure of firm value in the economics, finance, and management literatures (Venkatraman and Ramanujam, 1986; Morck, Shleifer, and Vinishy 1988; Villalonga 2004). We calculate the firm’s Tobin’s Q as the sum of the end-of-year market value of equity and reported book value of total liabilities divided by the total book value of assets. To measure IP disclosure, we count the number of disclosure events the firm has in a year ( Disclosure Events ). Table 2 displays the estimates from the following equation: 𝑄𝑖𝑡 = 𝛽𝐷𝑖𝑠𝑐𝑙𝑜𝑠𝑢𝑟𝑒 𝐸𝑣𝑒𝑛𝑡𝑠𝑖𝑡 + 𝑋𝑖𝑡 𝛤 + 𝛼𝑖 + 𝜆𝑡 + 𝜀𝑖𝑡 where 𝛼𝑖 denotes the firm fixed effect, 𝜆𝑡 denotes the year effect, and Xik is a matrix of variables that control for event and firm factors. We describe these variables in the note under of Table 2. We estimate the equation using a linear fixed effects model In Model 1, we find a negative and significant relationship (-0.069; p-value 0.000) between Disclosure Events and Tobin’s Q. Thus, while we would intuitively expect firms to benefit from being part of industry standards, our finding suggests that disclosures to SSOs per se may not generate positive returns to disclosing firms. To assess the role of complementary technologies, we split Disclosure Events into two variables: Disclosures With Complementary Technologies (i.e. Complementary Technologies >0) and Disclosures
In terms of market reaction, we have a priori reasons to believe that the market does react to IP disclosures in ways we propose, based on the following. First, the notion that market has a deep knowledge of firms’ R&D activities and technology portfolios, and that it reacts to events affecting them, is grounded in existing literature. For instance, Girota, Terwiesch, and Ulrich (2007) shows that the market reacts to pharmaceutical firms’ phase III drug development failures. They find that this reaction depends on how the failed drug relates to other projects in the firm’s pipeline and the underlying resources used in its development. Similarly, others have shown that stock prices react to changes in R&D and marketing expenditures (Erickson and Jacobson 1992, Chakravarty and Grewal 2011), product quality awards from suppliers or third parties (Hendricks and Singhal 1996), delays in new product introductions (Hendricks and Singhal 1997), new product announcements and trademark applications (Fosfuri and Giarrantana 2009), rivals’ R&D alliances (Oxley, Sampson, and Silverman 2009), and product recalls (Thirumalai and Sinha 2011). Closer to our context, prior research also finds that investors react to standards related news. Aggarwal et al. (2011) find that the stock market reacts to announcements of firm’s participation in information technology standards, while Aggarwal et al. (2006) find that investors react differently to the characteristics of the proposed standard (e.g. open vs. proprietary). Second, in our context, market analysts appear to actively monitor the progress of standards and firms’ standard-setting activity, and are aware of these firms’ standards-related non-SEPs. Equity analysts often mention firms’ SEPs and standards related activity in reports (Deshpande et al. 2013, Garcha et al. 2014 , Garcha et al. 2014, Sur et al. , 2015, Garcha et al. 2016, Davuluri 2017, Kelleher, 2017). Their stock recommendations are affected by developments in standards or at SSOs. For example, a change in a SSO’s IP policy triggered revised outlooks on several affected firms (Sur et al. 2015).^16 Analysts have clear understanding of how disclosing firm’s SEPs and non-SEPs relate. For instance, JP Morgan equity analysts explicitly report on how Nokia could leverage its non-SEPs to derive revenue from adopters of communication standards (see Deshpande et al. 2013), and a similar theme was echoed by Credit Suisse (^16) In this instance, a change to the IEEE patent policy triggered an immediate updated research report and revised outlooks for Broadcom and Marvell (see Sur et al. 2015).
analysts in the following months (see Garcha et al. 2014 and Garcha et al. 2014). Going further, we informally interviewed several top-ranked equity analysts specializing in the ICT industry. Our conversations revealed that they are keenly aware of what firms are behind which standards and what other compatible technologies and non-SEPs these firms have that could be used to leverage post-standardization revenue streams. We also informally consulted senior R&D executives at one of the most active firms in standard setting, to get a sense of how they are portraying their revenue model to the market. They clearly depicted a portfolio approach to their standard setting activities, strategizing around not only the SEPs, but also the complementary IP which they license out as a packet. Third, we check for corroborating signs of market reaction to disclosure events besides the changes in stock price. If the market does react to these events, we should see corresponding increases in trading volume. We indeed find evidence of significant increases in trading volume over disclosure events in our data. We detail this evidence at the end of this section. To estimate the impact of the disclosure event, we calculate daily abnormal returns using the Capital Asset Pricing Model. We calibrate the CAPM on 250 trading days prior to the beginning of the event window, using the S&P 500 as the market index.^17 The main dependent variable, CAR , is the cumulative abnormal return in the t-1, t0, t+1 event window, where t0 is the day of the disclosure. We focus on the typical three-day window because it tends to capture most of the market’s reaction to the event (MacKinlay
- and avoids picking up market reactions to other confounding news. Later in the analyses, we provide estimates for other window lengths to check for robustness of findings. We begin by analyzing CAR for the full sample of 2,732 disclosures from firms with available stock price data in the 1988-2010 period (Table 3). We test whether the mean CAR equals zero using a t - test (Test 1). Mean three-day CAR is negative and significantly different from zero (-0.30 percent; p-value 0.002). As we increase the window length, average CAR continues to be negative, but starts to lose its significance (^17) The CAPM regresses the firm’s risk adjusted returns on the benchmark return. The risk-free rate is specified using the daily interest factor from that day’s three-month Treasury Bill rate. As a robustness check, we calibrate the CAPM models using 500 daily trading days. All results remain robust.
