The impact of government policy on SEP profits: an event

The impact of government policy on SEP profits:
an event study
Fiona M. Scott Morton
Yale University and NBER
PRELIMINARY AND INCOMPLETE: SUBJECT TO REVISION
DRAFT
June 2015
[email protected]. Toomas Laarits, Albert Shin and Jialu Chen provided excellent
research assistance.
A. Introduction
The convergence of media, computing, and communications functions into a single small
mobile device means that the typical smart phone is very complex. These handsets typically embody
large amounts of IP, some of which is patented. Many of the communications and computing
products and services most valuable to consumers require a common standard for interoperability.
For example, UMTS, or 3G, is a telecommunications standard incorporated into many mobile
handsets that allows those handsets to send signals to, or receive them from, towers with UMTS
equipment installed. A standard will embody specific technology, which industry participants all
employ, that allows interoperability. The creation of such a standard and the choice of technology to
include in it is the responsibility of what is known as a standard setting organization, or SSO.
An SSO is comprised of member organizations. Members of an SSO are typically
organizations that generate revenue, or otherwise have a stake, in that industry, though membership
is voluntary. SSOs typically have established procedures by which they choose a technological path
for a new standard. These procedures can take years, as SSO members identify new areas that
require standards and assemble working groups to move a standard-setting process forward. During
this time the SSO committees choose functionality to include in the standard. They identify and
consider a number of feasible technological paths for the standard. The technology under
consideration might be in the public domain, or it might be claimed as proprietary by a firm with
one or more patents. The owner of those patents will typically advocate that its solution be chosen
by the SSO over competing alternatives. Ultimately the SSO chooses just one technology for each
functionality – the one its committee thinks is the best. If that chosen standard cannot be
implemented without using (infringing upon) a particular patent, the patent is said to be a “Standard
Essential Patent,” or SEP.
SSOs’ rules will include an Intellectual Property Rights (IPR) policy. Such polices establish
what member firms must disclose concerning the IPR they own, and at what point during the
standard-setting process; IPR policies also specify what licensing obligations members must commit
to should they wish their IPR to be included in the standard.
A typical restriction requires members to commit to license any IPRs they own that are
essential to the standard on Reasonable And Non-Discriminatory terms, or Fair, Reasonable, And
Non-Discriminatory terms (collectively, F/RAND terms). One problem with these commitments is
that SSOs typically specify very little as to the meaning of “fair” or “reasonable.” This leads to
different interpretations by licensors and licensees down the road when parties are attempting to
secure licensing agreements. Given the vague nature of most F/RAND agreements, there are
frequently disputes among different parties about what a “reasonable” royalty might be for a
particular portfolio of intellectual property.
It is important that the SSO’s IPR policy constrain SEP prices because of the substantial
market power necessarily enjoyed by the owner of an SEP in a successful standard. There is no
viable substitute for the SEP IPR after the standard is established, and therefore its owner can
demand very high royalties. Moreover, this market power is achieved through the joint action of
SSO members that might otherwise be competing for consumers to adopt their separate
technologies. This joint action is often acceptable for society because it trades off technology
competition among SSO members for quicker adoption, less risk, and economies of scale in a
successful standard.
The risk to society that the F/RAND commitment is designed to prevent is that once a
standard is in place, an SEP owner can use its resulting market power to engage in “hold-up.” Holdup occurs when licensees have invested in the standard by making the product, installing useful
infrastructure, building production capacity, and, generally, investing in complementary assets of
various kinds. At this point the SEP owner approaches such firms with an onerous licensing
demand. Assuming the patent is indeed essential and valid, the firm’s product must practice the
patent in order to be interoperable. The licensee is in a poor bargaining position because the next
best option is to re-build the factories and infrastructure to conform to a different standard (and get
others to join in that project), which is often prohibitively expensive and unrealistic.
In such a bargaining setting, unmodified by FRAND, the SEP owner can obtain payment far
in excess of the value of the technology alone, and appropriate the profits due to the later
complementary investments of others. This is hold-up. Such behavior raises licensing costs in the
industry, distorts markets for innovation and investment, and discourages adoption of standards.
Many competition agencies and other regulatory bodies or officials have noted that hold-up harms
competition and welfare.
In theory, F/RAND commitments prevent these inefficient outcomes by constraining IPR
owners to charge reasonable royalties commensurate with the value of the technology before it was
included in a standard. This “ex ante incremental value” approach to defining FRAND is long
standing and approved by both US antitrust agencies. 1 CITE. The flaw in the current process is that
the method for resolving a dispute over what constitutes a reasonable royalty is to go to court and
carry out a complete, and costly, trial. In the United States, parties in patent trials may request a jury,
amplifying the uncertainty and risk already inherent in litigation. Along with whole-product royalties,
exclusion orders, threats of royalty-stacking, and other techniques for requesting large sums in court,
the patent owner can threaten to place a very large dollar amount in front of the jury. This
encourages the licensee to settle for an amount in excess of the reasonable rate. In other words, the
current processes for resolving disputes on standard essential patents allow the hold-up problem to
persist.
B. The vagueness inherent in F/RAND
Courts, regulators, and competition authorities have called attention to this problem for
many years but have not been able to fix it. Arguably SSOs are best-positioned to eliminate holdup,
sitting as they do at the root of the problem. The fundamental cause of anticompetitive hold-up is
the ambiguity in the F/RAND commitment, which is under the control of the SSO. SSOs could
reform their IPR policies to create more clarity as to what F/RAND constitutes and design
alternative, less-expensive, methods to determine the level of a F/RAND royalty in the event of a
dispute. 2 Only one has done so. On February 2015, IEEE became the first major SSO to
successfully reform its IPR policy. The proposed changes to the IEEE IPR Policy were approved at
three different levels of committees and then finally by the IEEE Board of Governors. The IEEE
reforms have several important components. 3
The F/RAND Commitment Travels with the Patent
The new IEEE IPR policy requires that the FRAND commitment remains binding when the patent
is sold. In the absence of this requirement, a successful business strategy would be to agree to
Joint guidelines DOJ-FTC Joint Guidelines on Intellectual Property licensing. See the DOJ Business Review
Letter at: http://www.justice.gov/atr/public/busreview/311470.pdf
1
See Kuhn, Kai-Uwe, Fiona Scott Morton, and Howard Shelanski (2013) “Standard Setting Organizations
Can Help Solve the Standard Essential Patents Licensing Problem” CPI Antitrust Chronicle, March 2013
(special issue).
2
3
The IEEE policy is here: http://standards.ieee.org/develop/policies/bylaws/approved-changes.pdf
FRAND and then later sell the patents to a licensor who thought that he was not so bound. This
reform protects the integrity of the FRAND commitment, thereby limiting patent damage awards
and injunctions based on SEPS, by increasing the number of FRAND-encumbered patents and thus
protecting more implementers who relied upon the FRAND commitment.
No Bundling of Non-SEPs with SEPs
The new IEEE IPR policy prohibits the owner of SEPs from bundling them with other IPRs as a
condition of licensing those SEPs. Parties may choose broad cross licensing voluntarily, but it may
not be a requirement. The SEP owner is required to offer a simple license covering all of the SEPs
that read on a given standard. This change protects a licensee from bundling or tying that might
raise a royalty rate above the F/RAND level or otherwise enable evasion of the F/RAND
commitment.
Clarification of What Constitutes a “Reasonable” Royalty Rate
Importantly, IEEE takes a stand on defining what constitutes a “reasonable” royalty rate. In
particular, the first paragraph below states that the value associated with a patent claim excludes the
value resulting from the inclusion of that claim’s technology in the standard. This principle has been
advanced by the U.S. competition agencies for many years because it eliminates holdup and
associated inefficiencies. 4 There are three considerations the IEEE says should be taken into
account (but are voluntary) when determining a reasonable royalty. The first is that the value should
be measured based on its contribution to the smallest saleable unit in which the standard is
embodied. Secondly, that a reasonable royalty takes into account all the other patent holders who
claim a royalty. Lastly, IEEE says that licensors should not point to royalty rates negotiated under
the threat of an injunction and use them as a ‘competitive’ benchmark.
Limits on Seeking or Enforcing Injunctions
The new IEEE IPR policy takes a significant step forward on a second type of assertion cost by
limiting the use of injunction and exclusion orders. As long as the implementer is complying with
adjudication procedures that will eventually produce appropriate monetary damages for the SEP
holder, the SEP holder may not pursue an injunction.
Joint guidelines DOJ-FTC Joint Guidelines on Intellectual Property licensing. See the DOJ Business Review
Letter at: http://www.justice.gov/atr/public/busreview/311470.pdf
4
Taken together, these reforms resolve many problems with the F/RAND commitment. For
example, they greatly shrink the range of plausible F/RAND licensing demands. Unfortunately, the
process of creating, editing, and then voting to approve these reforms happened slowly over a long
period of time, so I am unable to use the IEEE change as an “event” in the empirical portion of the
paper.
Other SSOs show no sign of reform. The difficulty is that some SSO members benefit from
the status quo and would be financially harmed by reform. Since SSOs operate "by consenus," any
such member can block change. To date, no other major SSO has reformed its IPR policy
concerning SEPs.
Competition authorities recognize that the hold-up problem exists and harms the
functioning of competitive markets. However, the DOJ, FTC, and European Commission have
been on different timetables in terms of taking enforcement actions. This policy response, or lack
thereof, is part of the uncertainty I exploit in this paper. Over time, officials from these
organizations publicly explain why abuse of a F/RAND commitment is harmful to competition.
Over time some of them prosecute cases of F/RAND abuse. The European Commission has
prosecuted two cases in the last four years (Samsung and Motorola) under their competition laws,
while DOJ has engaged in no enforcement at all. The FTC has appended two RAND abuse cases to
other matters, Google and Bosch, under Section 5.
C. Events
Given continuing lack of F/RAND clarity and the welfare loss it causes, those parties being
harmed have appealed to other organizations for help: courts, competition authorities, and elected
authorities. Because of the number of years without SSO reform, and the increasing importance of
technological standards, these organizations have had time to take decisions that affect the business
model of those who monetize SEPs. These decisions have either determined what F/RAND is
directly, or altered the bargaining power the SEP holder has in a private negotiation by constraining
or enabling other options. Such changes, whether positive or negative, should affect the long term
revenue outlook for a patent holder whose business model includes montetizing SEPs through
holdup. If stock markets are efficient, then stock prices incorporate all available information
concerning the profitability of the firm going forward. The stock price of a firm holding SEPs and
monetizing them through holdup, should therefore move when regulators change expected payoffs
from SEPs. For example, if a judge were to rule that engaging in SEP holdup violated the law, and
that ruling was expected to apply to future similar cases, then holdup would become a less profitable
strategy, and the stock price of a firm with a business model of holdup would be expected to fall.
In this paper, I study 9 events that took place in the 2013-2014 period in three jurisdictions,
the US, Europe, and China. I chose events that were likely to have affected the cost and legality of
charging above-F/RAND royalty rates. Limiting F/RAND abuse would affect the future
profitability of firms whose business model is based on F/RAND abuse. Those firms stock prices
are therefore expected to fall in the event that a regulator limits F/RAND abuse. Firms that
regularly set royalties at what regulators view as true F/RAND rates would not see their profits or
stock prices affected.
Choosing the events that move stock prices is also tricky. The chosen events had to generate
news suddenly enough so that the impact on the stock price of licensing firms could be measurable.
The event also had to clearly either benefit or harm the F/RAND abuse business model. All the
events involve public actors of some sort, either courts, regulators, elected executives, or
competition authorities. Below I explain why and how each event would be expected to alter the
profits from licensing an SEP.
I employed a Chinese-speaking graduate student as a research assistant to evaluate the events
in China. In particular, she evaluated which announcements about the Qualcomm investigation were
“news” and what the correct windows for the events were. The transparency of the regulatory
process and outcomes in China is considerably different than in the US and the EU. It is not clear
what information stays confidential and what is leaked over time. In addition, the agencies are very
young and have little settled process, and the community of antitrust professionals is small. For this
reason I will analyze the full set of events, and then repeat the estimation on the set of only US and
EU events, where I am more confident that the event study methodology can be applied correctly. A
summary of the events is contained in Table 1; the Chinese events are shaded in blue.
1. ROBART DECISION: 25 April 2013, Pacific time, Thursday
BAD NEWS for SEP holders: MMI found to be demanding rates far above RAND
United States Judge Robart determined that the RAND rate in this dispute was $1.8M, very close to
Microsoft’s rate and less than 0.5% of MMI’s demand of $400M. For a court to determine a
FRAND rate was unprecedented at this time. The methodology and the finding served as precedent
for any party negotiating a F/RAND rate going forward. The decision was also news that going
forward SEP owners would have less leverage; the decision lowered litigation risk, and reduced
expectations for what monetary damages a court might award. The judge’s decision was released on
April 19, 2013; the parties were given until the release on April 25th to redact information before
decision was made public.
2. USTR VETO: 3 Aug 2013, Saturday
BAD NEWS for SEP holders: The US President through the US Trade Representative vetoes an
ITC exclusion order against Apple products based on SEP infringement
Public comments by the Antitrust Division of DOJ and the FTC urged the ITC not to allow an
exclusion order for an SEP for the reasons outlined above. 5 The relevant statute the ITC enforces
explicitly states that the interest of competition and consumers should be considered. Despite the
agency advice, the ITC ruled to exclude Apple products in an SEP dispute. The process is unusual in
that the USTR (part of the executive branch) is then permitted to over-rule the ITC’s decision. This
action is not common and is taken when the President feels policy considerations warrant it. The
USTR veto on August 3rd and the accompanying explanation indicate that the USTR is likely to
overrule the ITC on similar exclusion orders on SEPs in future. 6 It is possible the veto was not fully
anticipated by market participants. In addition, the accompanying explanation may have been news
that future exclusion orders would be hard to get.
3. HOLDERMAN DECISION: 3 OCT 2013
BAD NEWS for SEP holders: Innovatio found to be demanding rates far above RAND
United States Judge Holderman determined that the RAND rate in this dispute was $.09 per chip,
0.2% of Innovatio’s demand of $40 for a handheld scanner. This decision was further news that
going forward SEP owners would have less leverage; the decision lowered litigation risk, and
reduced expectations for what monetary damages a court might award. The judge’s decision was
released on October 1, 2013; the public release was on October 3rd.
4. HESSE SPEECH: 8 Nov 2013, Friday afternoon in California
GOOD NEWS for SEP holders: The US DOJ Deputy Assistant Attorney General for Antitrust
who specializes in IP makes a speech indicating the Antitrust Division of USDOJ will not be
preventing SEP hold-up under the antitrust laws. (Monday Nov 11 was Veterans Day when markets
are closed. Window open through Nov 12.)
The speech is entitled “The Art of Persuasion: Competition Advocacy at the Intersection of
Antitrust and Intellectual Property.” The Antitrust Division can enforce the competition laws or it
http://www.ftc.gov/sites/default/files/documents/advocacy_documents/ftc-comment-united-statesinternational-trade-commission-concerning-certain-wireless-communication/1206ftcwirelesscom.pdf
5
6
http://www.ustr.gov/sites/default/files/08032013%20Letter_1.PDF
can engage in what is known as ‘competition advocacy,’ which involves educating other agencies and
parties on the benefits of competition. This speech, and several others, describe the DOJ’s role in
working with SSOs and their members to persuade them to reform and engage in less SEP holdup;
this speech puts emphasis on the lack of plans for enforcement actions. Many affected parties
thought the DOJ would find SEP holdup to violate antitrust laws and file a lawsuit. This event is a
clear indication that the Division will not take that path, but will instead simply hope that SSOs
become more procompetitive. 7
The content of the speech could have been news to market participants given that the President had
recently issued a report clearly indicating that his administration would be working against patent
abuses. The leaders of the Antitrust Division are appointed by the President, so market participants
might have expected close coordination with the administration.
5. NDRC INVESTIGATES QUALCOMM: 25 Nov 2013
BAD NEWS for SEP holder: China's government seeking leverage in royalty negotiations
Chinese regulatory agency National Development and Reform Commission launched a sudden antitrust investigation into Qualcomm alleging royalties for wireless technology were too high. Royalty
revenues in China made up half of Qualcomm’s sales in 2013. The company said it was not aware of
any antitrust violations but would cooperate with the probe by NDRC.
6. EC ANNOUNCES QUALCOMM INVESTIGATION: 26 August 2014
BAD NEWS for SEP holder: EC investigation of Qualcomm’s licensing practices
The European competition authority opens an investigation of Qualcomm’s licensing of its SEPs.
7. FTC ANNOUNCES QUALCOMM INVESTIGATION: 5 Nov 2014
BAD NEWS for SEP holder: Qualcomm disclosed it was being investigated by the FTC
Qualcomm disclosed that the Federal Trade Commission is investigating its business practices. FTC
notified Qualcomm about the probe on "potential breach of FRAND commitments" in midSeptember.
8. NEWS OF NDRC PENALTY QUALCOMM: 6 December 2014
BAD NEWS for SEP holder: Rumors surface that the Chinese competition authority will fine
Qualcomm
CITE “In sum, I believe the Division’s competition advocacy, in combination with the efforts of other
agencies and the courts, is bringing some much-needed guidance on issues posed by standards-essential patents
that patent owners have committed to license on F/RAND terms. We will continue our work in this area
with the hope that the policies governing the use of such standards-essential patents will become even more
procompetitive.”
7
Under the anti-monopoly law, the NDRC can impose fines of between 1 and 10 percent of a
company's revenues for the previous year. It is rumored that the NDRC had completed the antitrust investigation into Qualcomm. Qualcomm is said to be facing $1 billion fine and a cut in
licensing fees in China.
9. NDRC FINES QUALCOMM: 10 Feb 2015
BAD NEWS for SEP holder: NDRC reached the decision to fine Qualcomm 6.088 billion yuan
($975 million).
After more than a year of investigation, regulators from NDRC decided that Qualcomm had
violated the Chinese anti-monopoly laws and fined Qualcomm 6.088 billion yuan ($975 million).
Qualcomm said it would not dispute the huge fine. After the fine, Qualcomm would keep operating
under its current business model with some concessions. Qualcomm makes the majority of its profit
from licensing patents for chips to other companies. Under its previous operating procedures,
Qualcomm would bundle a range of patents together, requiring companies interested in obtaining
licenses to use 3G and 4G smartphone chips to also purchase other patents from the company's
stables. Now Chinese the settlement forces the company to offer those 3G and 4G chip licenses
without tying. The company would also have to adopt a licensing rate in line with royalty rates for
the rest of the world.
Table 1: Summary of events
1
Robart decision
SEP owner
Motorola
Venue
US court
Decision or topic
RAND is low
Impact
Fall
2
USTR veto
Samsung
Holderman decision
Innovatio
Exclusion orders on
SEPs are limited
RAND is low
Fall
3
Executive
branch
US court
4
Hesse speech -
All
Executive
branch
Rise
5
NDRC opens
Qualcomm case
Qualcomm
6
EC opens Qualcomm
investigation
Qualcomm
7
FTC investigates
Qualcomm
NDRC may fine
Qualcomm
Qualcomm
NDRC determines
liability and fines
QCOM
Qualcomm
Chinese
competition
authority
EU
competition
authority
US antitrust
authority
Chinese
competition
authority
Chinese
competition
authority
USDOJ will not use
antitrust to enforce
FRAND
NDRC investigating
QCOM abuse of RAND
8
9
Qualcomm
Date
April 25,
2013
August 3,
2013
October
1, 2013
Nov 8,
2013
Window
April 1926, 2013
August 35, 2013
October
1-3, 2013
Nov 812, 2013
Fall
Nov 25,
2013
Nov 2526, 2013
EC investigating QCOM
abuse of FRAND
Fall
Aug 26,
2014
Aug 2627, 2014
US investigating QCOM
abuse of FRAND
NDRC may fine QCOM
for abuse of RAND
Fall
Nov 5,
2014
Dec 6,
2014
Nov 5-6,
2014
Dec 6-9,
2014
NDRC determines
QCOM abused RAND;
large fine
Fall
Feb 10,
2015
Feb 1011, 2015
Fall
Fall
D. Company Selection
Choosing the set of firms to study required developing limiting conditions. First, the firms had to
primarily operate in the area of communications, because we know that a large fraction of the
important patents in this industry are SEPs. Secondly, all firms had to earn significant revenue from
licensing, which was also a measurable portion of sales. Thirdly, the firms had to have a reputation
for engaging in aggressive licensing as shown by appearance in front of regulatory bodies such as
courts or competition agencies. Lastly, the firms had to be publicly traded. The firms that I use for
the event study are listed in Table 2 below.
The companies that were selected are some of the major players in the standards essential
patent industry – especially the 4G-LTE industry. There are many more firms that likely have
exposure to SEP licensing revenue. However, they also have other lines of business, for example,
selling devices or operating systems, and the changes to SEP revenue would be unlikely to move the
firm’s stock price in a measurable way. For example, Apple owns many patents, but not many are
SEPs. Microsoft owns many SEPs but it does not earn most of its revenue from licensing, rather, it
sells software. Many firms that obtain most of their revenue from patent licensing are not publicly
traded (e.g. Intellectual Ventures).
We determined a company’s dependency on patent licensing using public reports such as 20Fs (foreign) or 10-Ks (domestic). It would have been better to report a measure of SEP licensing
revenue, but we could not come up with a reliable metric. Because both Nokia and Ericsson have a
smaller share of licensing revenue than the other firms, we repeat the analysis omitting those two
companies.
Table 2: firm names, ticker symbols, and available information about licensing revenues
Company
Ticker
Qualcomm
QCOM
InterDigital
IDCC
Rambus
RMBS
Patent Licensing Revenue
Total Revenue 2014
29%
USD 26 billion
revenues in fiscal 2014 , 2013 and 2012 , respectively.” 2014
annual report
USD 415 million
“QTL revenues comprised 29% ,30% and 33% of total consolidated
97%
$403 million in patent licensing revenues (2014 annual
report)
88%
USD 297 million
Nokia
NOK
Ericsson
ERIC
“Royalties from patent licenses accounted for 88%, 92%
and 89% of our consolidated revenue for the years ended
December 31, 2014, 2013 and 2012, respectively” (2014
annual report)
4.5%
578 sales Nokia Technologies. “While the majority of our
current revenues are derived from licenses related to
Standard Essential Patents”
4%
10b/227b SEK (2014 annual report)
EUR 12,732 million
SEK 227 billion
(USD 35b)
E. Methodology
We conducted a standard event study using the Capital Asset Pricing Model. The CAPM models the
return of a security as a constant plus the return of the market portfolio scaled by that security’s
correlation with the market. For any security i, the model is
𝑅𝑅𝑖𝑖𝑖𝑖 = 𝛼𝛼𝑖𝑖 + 𝛽𝛽𝑖𝑖 𝑅𝑅𝑚𝑚𝑚𝑚 + 𝜀𝜀𝑖𝑖𝑖𝑖
𝑠𝑠𝑠𝑠𝑠𝑠ℎ 𝑡𝑡ℎ𝑎𝑎𝑎𝑎 𝐸𝐸(𝜀𝜀𝑖𝑖𝑖𝑖 ) = 0 𝑎𝑎𝑎𝑎𝑎𝑎 𝑉𝑉𝑉𝑉𝑉𝑉(𝜀𝜀𝑖𝑖𝑖𝑖 ) = 𝜎𝜎𝜀𝜀2𝑖𝑖
where 𝑅𝑅𝑖𝑖𝑖𝑖 and 𝑅𝑅𝑚𝑚𝑚𝑚 are the period-t returns on security i and market portfolio respectively. 8
We used stock market return data from the Bloomberg dataset. The S&P High Tech Index (ticker:
SPXEWTUP) was used as the market portfolio. First we estimated 𝛼𝛼 and 𝛽𝛽 using return data
ranging from 200 days before the event to the day before (including non-trading days). During this
200 day period, any of our other five events were excluded in estimating 𝛼𝛼 and 𝛽𝛽. Using the index
and estimated coefficients, we are able to calculate expected returns on each day for each stock. We
then obtained abnormal returns during the event windows by taking the difference of the predicted
and actual returns. Event windows are set to be as small as possible while including the time of the
news and the first or current day of operation of the New York stock exchanges. For example, if
information is released during the trading day on the east coast, I assume stock prices react on that
day. The longest event window is around the Robart release, when the decision might have leaked
before the opinion was public. Another slightly longer-than-normal window occurs with a Friday
8
MacKinlay, A. Craig, “Event Studies in Economics and Finance”, Journal of Economic Literature March 1997 p. 18
afternoon release of information followed by a Monday stock market holiday. The abnormal returns
can be expressed as 𝑒𝑒𝑖𝑖𝑖𝑖 :
𝑒𝑒𝑖𝑖𝑖𝑖 = 𝑅𝑅𝑖𝑖𝑖𝑖 − 𝛼𝛼𝑖𝑖 − 𝛽𝛽𝑖𝑖 𝑅𝑅𝑚𝑚𝑚𝑚
To obtain additional power, we calculated a total excess return by firm and also a total excess return
by event. Standard errors were calculated for event day t = 0 using
𝜀𝜀𝑖𝑖0 = 𝑒𝑒𝑖𝑖0 /𝑆𝑆𝑖𝑖
where 𝑆𝑆𝑖𝑖 is the standard deviation of the regression residuals prior to the event. 910
The tables that follow include tests of all events and firms pooled together, as well as firms
separately with events pooled, and events separately with firms pooled.
F. Results
Below are listed the abnormal returns and the p-value for each of 4 samples. I begin with the full
sample, then eliminate the Chinese events, then repeat that combination for the firms with a high
proportion of licensing revenue.
Tables 3a-3c: Abnormal returns for 5 firms and 9 events
Table 3a: Results by event
1
Cumulative
abnormal returns
-0.14929
-1.33
p_value
----------0.098806
2
-0.06807
-0.96
0.182194
3
0.032794
0.38
0.643641
4
-0.01974
-0.23
0.412165
5
0.023167
0.33
0.623739
7
0.019223
0.27
0.603289
8
-0.09625
-1.35
0.104875
9
0.081101
1.14
0.857778
10
0.101146
1.42
0.905271
event
t_stat
Table 3b: Results by firm
MacKinlay, A. Craig, “Event Studies in Economics and Finance”, Journal of Economic Literature March 1997 p. 18
When we estimated 𝛼𝛼 and 𝛽𝛽 and calculated the standard errors by compiling all events, the prediction period ranged
from 200 days before the first event to the last event.
9
10
t-stat
p-value
Eric
Cumulative
abnormal
returns
0.059215
0.857309
0.799741
Idcc
-0.01753
-0.14684
0.442298
Nok
0.048285
0.333566
0.629067
Qcom
-0.073
-1.39337
0.088714
Rmbs
-0.09289
-0.74221
0.232906
Firm
Table 3c: Pooled results:
Cumulative
abnormal
returns
-0.07592
Firm / event
all
Std dev abnormal
returns
t-stat
p-value
0.022533
0.3142
0.376971
Tables 4a-4c: Abnormal returns without the Chinese events: 5 firms and 6 events
Table 4a: Results by event
event Abnormal return
t-stat
p-value
1
-0.14929
-1.32511
0.098806
2
-0.06807
-0.95528
0.182194
3
0.032794
0.375773
0.643641
4
-0.01974
-0.22618
0.412165
6
0.019223
0.269769
0.603289
7
-0.09625
-1.35076
0.104875
Table 4b: Results by firm
Firm
Abnormal return
t-stat
p-value
eric
0.033809
0.569351
0.711488
idcc
-0.1259
-1.22641
0.118894
nok
0.052864
0.424781
0.661673
qcom
-0.12123
-2.69133
0.008027
rmbs
-0.12088
-1.12349
0.138902
Table 4c: Pooled results
Firm
/event
All
Abnormal return
t-stat
p-value
-0.28134
-1.35424
0.089646
Below I repeat the analysis for the three firms with proportionately significant licensing revenue.
Tables 5a-5c: Abnormal returns for high licensing firms: 3 firms and 9 events
Table 5a: Results by event
Event
Abnormal return
t-stat
p-value
1
-0.14676
-1.74156
0.051752
2
-0.1098
-2.06023
0.047203
3
-0.01334
-0.20434
0.421593
4
-0.02348
-0.35965
0.364208
5
0.043082
0.808347
0.772193
6
0.031879
0.598152
0.7121
7
-0.10651
-1.99849
0.051068
8
0.073578
1.380537
0.887026
9
0.067926
1.274498
0.870748
Table 5b: Results by firm
Firm
Abnormal return
t-stat
p-value
idcc
-0.01753
-0.14684
0.442298
qcom
-0.073
-1.39337
0.088714
rmbs
-0.09289
-0.74221
0.232906
Abnormal return
t-stat
p-value
-0.07592
-0.3142
0.376971
Table 5c: Pooled results
Firm /
event
All
Tables 6a-6c: The three high licensing firms and no Chinese events: 3 firms and 6 events
Table 6a: Results by event
Event
Abnormal return
t-stat
p-value
1
-0.14676
-1.83595
0.043846
2
-0.1098
-2.1719
0.040973
3
-0.01334
-0.21542
0.417415
4
-0.02348
-0.37915
0.357223
6
0.031879
0.630572
0.721993
7
-0.10651
-2.1068
0.044491
Table 6b: Results by firm
Firm
Abnormal returns
t-stat
p-value
idcc
-0.1259
-1.22641
0.118894
qcom
-0.12123
-2.69133
0.008027
rmbs
-0.12088
-1.12349
0.138902
Table 6c: Pooled results
Firm /event Abnormal returns
t-stat
All
-2.36838 0.010886
-0.36801
p-value
G. Conclusions
We see that events in the United States and Europe that are hypothesized to have an effect on
future profits from asserting SEPs do affect stock prices. At a 10% level of significance, the events
affect the returns of a set of 5 firms that license SEPs aggressively. On average, a negative event
lowers stock prices by 2.8%. If we restrict the sample to firms with more than 25% of revenue from
only patent licensing, the result is significant at the 1% level and the magnitude of the effect is
slightly larger, at 3.7%. These magnitudes seem plausible; these firms obtain a good proportion of
their revenue from SEPs, but any adjustment will affect the flow rather than the stock (which is
much bigger), and some patents may be less exposed to the news than others.
Chinese events that should in theory have an impact on SEP prices, do not. It may be that
we cannot measure news correctly in this region. Alternatively, the young enforcement agency in
China may not have a strong enough track record of establishing policies that a single event will
contain information about future profitability of another firm. For example, it may be that political
influence determines the outcomes of antitrust investigations, and so the topical content of a
decision is not informative about another firm’s profit potential. Or the tradition of precedent may
not apply with any certainty. Judicial precedent in the United States and policy-making in Europe
are both fairly consistent across their respective jurisdictions. In those geographies, a future case on
the same topic would be expected to result in an outcome similar to the earlier case.
There have been some courts and commentators who argue that the harm of holdup does
not exist. For example “While there may be a hypothetical risk of holdup, we have evidence that it
is not a threat in this case, or in this industry.” 11 It is, of course, very difficult to demonstrate that
certain innovations or products were not invented or launched because of anticompetitive royalty
demands. It is also difficult to document the amount and number of cases where royalties above the
FRAND rate were agreed upon by the licensee because of holdup by the SEP owner; such contracts
tend to be confidential and appropriate benchmarks may be confidential also. However, stock prices
are public. When a court in the United States rules that a F/RAND rate is less than 0.5% of the
original demand, and provides a road map for future litigation, this reduces the ability of SEP
holders to engage in future holdup. The fall in stock prices of firms that are known to aggressively
license SEPs provides evidence of holdup. Those stock prices represent the flow of future profits. If
the news that future license revenues will be constrained somewhat closer to the FRAND rate
reduces stock prices, then that means the original prices of those licenses were set above FRAND,
which is evidence of holdup.
11
ITC InterDigital Decision at p. 123 discussing the telecommunications industry.