An Economic Perspective on Patent Licensing Structure and Provisions
Varner, Thomas R., Business Economics
While economists have provided theoretical discussions of factors that shape the structure and provisions of technology licenses, there have been relatively few empirical studies that test these economic theories. This is likely due to the difficulty of obtaining large data sets of complete license agreements. By providing an overview of the characteristics of over 1,400 patent licenses collected from SEC filings, this paper provides empirical insights into the fundamental economic relationships that determine certain characteristics of patent licenses. This study confirms what economists have believed for some time-namely that there are differences in the structure and provisions of patent licenses across industries and across types of licensors. These findings support the view that differences in the structure and provisions of technology licenses can be attributable in part to fundamental economic differences between industries, as well as differences in the strategic incentives of parties negotiating agreements.
Business Economics (2011) 46, 229-238.
Keywords: patents, technology licensing, risk allocation strategies, royalties, sublicensing
The structure and provisions of patent licenses present a number of important issues for firms and individuals concerned with the economics of technology transfer. For example, if a technology license includes financial consideration, what is the motivation for some parties to prefer a running royalty and others to prefer a fixed fee, and why do many technology licenses include both forms of consideration? Why do some agreements include a grant of the licensee's stock or equity and others do not? Why do some agreements include separate sublicensing royalty terms or include provisions that reduce the base royalty rates in the event of selected contingencies (for example, royalty stacking provisions) and others do not?
Given these types of questions and the importance of technology licensing to businesses, one would think that economists working in the business community would be regularly engaged in reviewing, analyzing, and providing insights into the structure and provisions of technology licenses. After all, economists can draw on a long history of economic literature relevant to technology licensing, including such fields as property rights theory, contract theory, transaction-costs analysis, game theory under asymmetric information, and agency theory. However, based on my professional experience and conversations with other economists, there are few business economists whose responsibilities include reviewing their firm's technology licenses.
One possible explanation for the general lack of involvement by economists in technology licensing is that there are relatively few empirical studies of large data sets of licenses across multiple industries. This is in large part due to the exceptionally difficult task of obtaining enough licenses-many of which are protected under confidentiality provisions-in which to perform meaningful analyses.
Some rare examples of empirical research on technology licenses include Anand and Khanna , which considered a set of 1,365 licenses comprised of a variety of agreement types. The focus of their study was cross-industry differences among the licensing parties and ex-ante vs. ex-post relationships between the parties. Sakakibara  examined factors that affected royalty rates among 661 patent licenses in Japan. Gans, Hsu, and Stern  considered a sample of "more than 200 technology licensing deals" obtained from a proprietary data set, and Lyarskaya  considered 237 technology agreements collected from a survey of licensing professionals.
This paper draws on a unique data set of almost 1,500 patent licenses across multiple industries filed with the U.S. Securities and Exchange Commission (SEC) over a 17-year period (1994-2010). Roughly 25 percent of the agreements in this data set were filed with the SEC in redacted form and subsequently obtained in unredacted form under the Freedom of Information Act (FOIA). (1)
In a previous paper [Varner 2010], I presented findings of my analysis of a set of approximately 3,000 technology licenses. This broader data set included not only patent licenses but also other types of agreements, including joint venture agreements, product bundling or re-branding agreements, distribution agreements, settlement agreements, manufacturing and supply agreements, employment agreements, and acquisition agreements. One of the main findings of Varner  was that specified royalty rates are related to agreement type. For example, distribution agreements exhibit among the highest royalty rates observed in the data set, and bare patent licenses-those granting only rights to listed patents-exhibit among the lowest royalty rates in each industry category.
This paper considers a data set comprised of patent licenses and assignment agreements, that is, licenses that grant rights only to patents ("bare patent" licenses), licenses that grant rights to patents plus associated technology know-how ("patent-plus-know-how" licenses), and agreements that provide for full assignment or transfer of patent rights (patent assignment agreements). I focus this analysis on the financial structure of these agreements (such as fixed fees, running royalties, and equity) and provisions that modify their financial terms.
Analysis of a given license is shaped by circumstances of the specific licensed technologies and products, the parties, and the market conditions, among other factors. However, common economic factors also can influence the structure of these agreements across different industries and circumstances. For example, if product sales are costly to observe, then the parties may favor a fixed payment over a running royalty to reduce monitoring costs. A risk-averse licensor who wants to avoid the risks of uncertain future royalty payment may prefer a fixed payment structure, while a risk-averse licensee who wants to avoid the risk of a fixed payment if sales turn out lower than expected may prefer a running royalty structure. (A number of strategies that shift risks between the parties, such as adding provisions for royalty reductions, minimum annual royalties, and equity grants, are investigated later in this paper.)
The structure of agreements also may be affected by differences between the parties' expectations about the commercial success of the licensed products. If a licensor expects higher sales than does the licensee, the parties may agree to a running royalty structure in which the licensor forgoes a fixed fee in exchange for (what it believes will be) a higher eventual payout from a commercially successful product.
A licensee's preference for flexibility in choosing whether or not to incorporate a licensed technology into its products also may be a factor in determining the structure of an agreement. If the licensee's expected use of a technology is uncertain at the time an agreement is negotiated, then the licensee may prefer a fixed fee structure so that future product decisions are not affected by higher marginal costs associated with a running royalty. Economic analysis of a licensed product's price elasticity of demand, the degree of product substitutes, and the pricing options available to the licensee also may provide insights into the selection of a particular agreement form.
Analysis of provisions that modify the financial structure of agreements also can yield insights into the economics of licenses. For example, some licenses specify a range of royalty rates (or "tiered" royalties) subject to the level of sales, rather than a flat royalty rate. An agreement may specify a 3 percent royalty on net sales less than $10 million, then a 2 percent royalty on net sales in excess of $10 million. In this example, the royalty rate decreases with increasing sales. My analysis shows that the vast majority of high-tech and medical device patent licenses in the data set follow this trend. That is, royalty rates decrease with increasing sales and decrease over time (that is, royalties are "down-tiered"). However, for observations in the pharmaceutical industry, approximately 65 percent of those agreements in the data set that specify a tiered royalty exhibit a royalty rate that increases with increasing sales (that is, royalties are "up-tiered"). I discuss possible reasons for this observed difference in tiered royalty structures across industries later in this paper.
Economic analysis can yield insights into other provisions as well, such as royalty reductions contingent on future outcomes-for example, conversion from an exclusive grant to a nonexclusive grant, entry of competing products in the market, or the combination of licensed products with other products. This paper shows that the frequency of these provisions differs based on the nature of the licensing entity (commercial vs. noncommercial) and the nature of the industry involved (high-tech vs. life sciences).
The remainder of this paper is divided into the following sections: Section 1 describes the methodology used in the selection, coding, and analysis of licenses in the data set; Section 2 discusses findings of the analysis; and Section 3 presents conclusions and suggestions for further research.
Selection of license agreements and materiality
The data set used for this study consists of 1,458 patent licenses and patent assignment agreements included as exhibits in filings to the SEC. The data set consists of 474 bare patent licenses, 820 patent-plus-know-how licenses, and 164 patent assignment agreements. I excluded from the data set agreements that include a grant for patent rights but whose focus was not specifically on a patent grant. Among the agreements I excluded were university research agreements, consulting and employment agreements, merger and acquisition agreements, product re-branding or re-marketing agreements, and joint venture agreements. These other agreements, although certainly worthy of economic study in their own right, often include provisions that address issues and contingencies not directly related to patent technology licensing, which is the focus of this paper.
I also excluded from the data set patent settlement agreements resulting from litigation since such agreements often arise from different circumstances than nonlitigated patent licenses. Patent licenses arising out of patent infringement litigation were identified by searching for parties to the agreements in an electronic database of U.S. federal court records. (2)
All of the patent licenses in the data set were obtained by searching company filings submitted to the SEC including registration statements, annual and quarterly reports, proxy statements, and current reports. The SEC requires that registrants disclose in their filings "material definitive agreements not made in the ordinary course of business, (3) While all of the agreements in the data set were submitted by filers to the SEC, some of the agreements predate the filer's initial registration by months or years, and thus may involve parties in which neither was an SEC registrant at the time of the agreement.
I have reviewed numerous portfolios of firms' technology licenses as part of consulting engagements (performed under nondisclosure agreements) and testifying engagements (performed under protective orders of a court). My experience in reviewing these portfolios suggests that the data set used for my analysis in this paper may not be representative of this larger set of undisclosed agreements. Thus one should not necessarily infer that agreements in this data set are representative of all patent license agreements, and findings based on this analysis may not be relevant to the universe of undisclosed agreements that do not meet the SECs threshold for material definitive agreements.
Many of the agreements filed with the SEC are available only in redacted form, that is, they may be marked by the filer as "Confidential Treatment Requested." In some industries, such as the pharmaceutical industry, over 75 percent of the technology agreements filed with the SEC are granted confidential treatment. Notwithstanding the confidential treatment granted by the SEC, many of these agreements can be obtained in unredacted form through FOIA requests to the SEC after the grant of confidential treatment has expired. Included in this study's data set are approximately 350 patent licenses in unredacted form previously granted confidential treatment and obtained through FOIA requests to the SEC. (4)
Classificatin of parameters
Once the patent licenses and patent assignment agreements were collected, I summarized certain information from each agreement. This information included: (1) the type of agreement (bare patent, patent-plus-know-how, or patent assignment agreement); (2) caption information (names of parties and date of agreement); (3) the type of licensor and licensee (commercial or for-profit business, individual, educational, nonprofit, or government entity); (4) the type of technology being licensed and the nature of the licensed products; (5) exclusivity provisions; (6) the amount and nature of fixed fees, maintenance fees, and milestone payments; (7) the maximum and minimum running royalty rates; (8) sublicensing royalty rates; (9) minimum annual royalties; (10) the presence of equity consideration; and (II) provisions related to royalty reductions.
For purposes of this paper, I define a "fee" as any fixed financial payment and a "royalty" as a payment contingent on the level of sales. A royalty may be expressed as a percentage of sales (more commonly defined by net sales as opposed to gross sales), a percentage of profits, or dollars per unit of licensed product sold.
I searched for technology license agreements that included a reference to "patents" among all SEC filers over the period 1994 through 2010 and broadly grouped the agreements into high-tech (computer software, computer hardware, electronic components, instrumentation, and telecommunication industries), life sciences (medical device and pharmaceutical industries), and other industries.
Of the 1,458 agreements in the data set, approximately half are dated 1999 or earlier. The agreements are skewed toward earlier years for two reasons. First, while an agreement may be submitted to the SEC shortly after its effective date, an agreement also may be submitted as part of a registration statement (SEC Form S-l), in which case the agreement may predate the filing by several years. Second, the data set includes previously redacted agreements in which a grant of confidential treatment has expired-a period that is often five or more years after the agreement is first filed with the SEC. Thus, although approximately 25 percent of the data set is composed of previously redacted agreements, the data set includes a higher proportion of these agreements in earlier years.
2. Findings of Analysis
Basic data set statistics
Table 1 summarizes licenses in the data set by type of agreement and industry. Almost half (47 percent) of all patent licenses and patent assignment agreements in the data set are in the pharmaceutical industry and almost two-thirds (47 percent+17 percent = 64 percent) are in the life sciences. Less than a quarter (22 percent) of the agreements are in high-tech industries, with the remaining 14 percent in industries outside of the high-tech and life sciences fields.
Table 1. License Type by Industry (Percentages) Type of All Hightech Medical License Industries tech Device Pharma Other (1) (2) (3) (4) Patent-Plus- 56 12 9 27 7 know-how Bare patent 33 7 6 16 3 Patent 11 2 2 3 3 assignment All licenses 100 22 17 47 14 (5) (1.) High-tech industries include computer software, com-piiter Hardware, electronic components, telecommunication, and instrumentation firms. (2.) Medical device industries include surgical and medical instruments, orthopedic, prosthetic and surgical appliances and supplies, dental equipment and supplies, X-ray and related apparatus, electromedical and electrotherapeutie apparatus, and ophthalmic goods firms. (3.) Pharmaceutical industry includes pharmaceutical preparations, diagnostic substances, medicinal chemicals and botanical products, biological products, and biological research firms. (4.) "Other" industries include all other industries outside of high-tech and life sciences. These include, among others, chemicals, autpmotive, industrial, financial services, utilities, arid food and beverage firms. (5.) Note that totals may not sum due to rounding.
Table 2 summarizes licenses by type of licensor (virtually all licensees in the data set are commercial, or for-profit, entities). Almost two-thirds (63 percent) of all patent licenses and patent assignment agreements in the data set are between commercial entities, and 20 percent are between an educational institution and a commercial entity. Most of the nonprofit licensors in the data set are hospitals, and the majority of government licensors involve the U.S. Public Health Service.
Table 2. License Type by Licensor (percentages) Type of All Firms University Individual(s) Nonprofit Gov't License Licensor (1) (2) (3) (4) (5) Types Patent-plus- 56 38 10 5 3 <1 know-how Bare patent 33 17 9 3 2 2 Patent 11 8 <1 3 <1 <1 assignment All licenses 100 63 20 11 5 2 (6) (1.) "Firms" refers; to commercial or for-profit business entities. (2.) "University" refers to any educational institution. (3.) "Individual(s)" refers to one or more individuals listed as the licensor on license agreement. (4.) "Nonprofit" is typically a nonprofit research entity or a hospital. (5.) "Gov't" refers to a U.S. or foreign government entity.; (6.) Note that totals may not sum due to rounding.
Approximately 95 percent of all agreements observed in the data set specify some form of financial consideration, such as a fixed fee, a running royalty (typically expressed as a percentage of gross or net sales, or as dollars per unit of product sold), and/or company shares. Table 3 summarizes the frequency of different financial structures among the licenses observed. Among agreements without a fixed fee or royalty, approximately one-third (30 percent) include an equity consideration. Among the entire data set of patent license and patent assignment agreements, approximately 25 percent include an equity consideration.
Table 3. Financial Structure of Patent Licenses (Percentages) Type of All Fixed Fee (1) Royalty Fixed No Fee or License Forms + Royalty (2) Only Fee Only Royalty Patent- 56 30 21 3 3 plus- know-how Bare patent 33 19 9 2 2 Patent 11 2 3 4 2 assignment All licenses 100 50 33 10 6 (3) (1.) Fixed Fee refers to a fixed dollar ($) amount paid up-front or at a specified time. (2.) Royalty refers to running royalty payment based on sales (for example, 1 percent of sales, or SI per unit sold). (3.) Note that totals may not sum due to rounding.
Running royalty rates are a common form of consideration in many patent licenses. Table 3 shows that 83 percent (50 percent + 33 percent) of all patent licenses and patent assignment agreements in the data set include some type of running royalty rate. For those agreements that include a running royalty, over 90 percent specify a royalty expressed as a percentage of sales, with the remainder being dollars per unit of product sold, percentage of profits, or percentage of costs.
I reter to all three types of agreements patent-plus-know-how, bare patent, and patent assignment agreements-as "patent licenses" for the remainder of this paper. The remaining sections of this paper analyze economic considerations for the financial structure and provisions of these patent licenses.
A number of researchers have developed models to identify factors that affect the magnitude of running royalty rates. Examples of such factors include the number and quality of the licensed patents, the amount of sales or degree of profitability of the licensee, whether the licenses were granted on an exclusive basis, the nature of the licensed products, past licensing experience of the parties, dispute resolution provisions, and the nature, industry, and size of the licensing parties.
My previous research [Varner 2010] identified three factors as being statistically significant determinants (at greater than 95 percent confidence level) of royalty rates in patent licenses: (1) whether the patent license includes "know-how"; (2) whether the licensor is a for-profit, commercial entity; and (3) whether the patented technology relates to the life sciences industry. Inclusion of a grant of stock in the licenses firm to the patent holder had a slightly negative effect on royalty rates but was of relatively low statistical significance.
Other factors with relatively low statistical significance in relation to the royalty rate include the presence of separate sublicense royalty rates, fixed fees, milestone payments, royalty reduction provisions, and minimum annual royalty provisions. However, any one of these factors, or a combination of these factors, may be particularly relevant in a given license negotiation depending on the strategies of the parties, the technologies involved, the availability of substitute technologies, and markets for the licensed products, among other factors.
Regarding the structure of royalty rates in patent licenses, I observe that there are some, albeit not striking, differences across industries. Table 4 summarizes the frequency of different types of running royalties by industry. These include a percentage of sales, a defined dollar amount per unit sold, or other types of running royalties, such as percentage of profits or percentage of manufacturing costs.
Table 4. Frequency of Types of Running Royalties (Percentages) Type of Running All High-tech Life Sciences Other Royalty Industries Industry (1) Insustry (2) Industries Percentage of 91 81 95 74 sales $ per units 7 15 3 21 sold Other royalty 2 4 2 5 forms Total 100 100 100 100 (1.) High-tech industry includes computer software, computer hardware, electronic components, instrumentation, and telecommunication firms. (2.) Life sciences industry includes medical device and pharmaceutical firms.
The dominant form of running royalty for patent licenses in the life sciences industry is clearly a royalty based on a percentage of sales (95 percent are percentage of sales), whereas patent licenses in high-tech and other industries have a broader mix of running royalty forms.
Tiered royalty rates
Almost one-fifth (19 percent) of all patent licenses in the data set that specify a running royalty call for a sliding scale of royalties or ''tiered'1 (sometimes called "mixed") royalty rates that are based on the magnitude or time horizon of sales. For example, a patent license may specify an "up-tiered" royalty of 1 percent for the first $10 million in sales, and a royalty rate of 2 percent for sales in excess of $10 million.
There are a number of possible reasons for parties to structure a technology license with a tiered royalty instead of a flat royalty rate. Tiered royalty rates can be viewed as a means of shifting the risks of commercial success between the licensor and licensee. (Whereas milestone payments are usually tied to the technical success of a product or technology, tiei ed royalties usually relate to the commercial success of a product.)
Tiered royalties may arise if licensing parties have different expectations about a licensed product's future sales. If a licensor's expectation of product sales is higher than the licensee's expectation, then the licensor may be willing to grant a lower initial royalty rate in exchange for a higher royalty rate if actual sales turn out to be high. Also, if a licensor is relatively risk averse compared with the licensee, then the parties may agree to a higher royalty rate for lower sales and a lower royalty rate for greater sales.
Tiered royalties also may arise in response to expected changes in profit margins for licensed products. If a licensed product can command a high price initially, but eventually will face price competition, then the parties may agree to a relatively high royalty rate for initial sales and a lower royalty rate for later sales. One might expect to see this pattern of "down-tiered" royalties for high-tech products with rapid technology improvements and relatively short economic life spans. Conversely, one might expect to see "up-tiered" royalties for products that have longer economic life spans or higher up-front costs (such as initial marketing or advertising costs).
Table 5, a summary of agreements with tiered royalty rates, shows that the frequency of up-tiered vs. down-tiered royalty structures varies across industries. Tiered royalty agreements in the pharmaceutical industry exhibit a general trend toward increasing royalty rates with higher sales and increasing royalty rates over time (that is, royalty rates are up-tiered), whereas patent licenses in all other industries show a strong trend toward decreasing royalty rates with increasing sales and decreasing royalty rates over time (that is, royalty rates are down-tiered).
Table 5. Tiered Royalty Rates (Percentages) Based on Based on Sales Time Royalties Royalties Royalties Royalties increase with decrease with increase with decrease with increasing increasing increasing increasing sales sales sales sales Industry (up-tiered) (down tiered) (up-tiered) (down tiered) Pharmac 64 36 60 40 Utical (n=69) (n=38) (n=6) (n=4) industry All other 12 87 12 88 industries (n=9) (n=62) (n=3) (n=23)
One explanation for these observed differences may be that new pharmaceutical products often exhibit lower profit margins due to relatively high marketing and advertising costs early in a product's launch phase, while other industries (especially high-tech industries) often exhibit higher initial profit margins. If profit margins are low for early sales due to added marketing or advertising expenditures, then the parties may agree to a lower royalty rate for initial sales and a higher royalty rate for later sales; that is, the parties will agree to an up-tiered royalty rate structure.
Royalty rate reductions
Almost one-third (30 percent) of all bare patent and patent-plus-know-how licenses that specify a running royalty rate include provisions for a royalty rate reduction-a contingency to reduce the royalty rate under certain conditions. For example, an agreement may specify a royalty reduction if the licensee is subsequently required to obtain a patent license from a third party in order to sell a licensed product. These "royalty stacking" provisions are more commonly found in agreements in the pharmaceutical industry than in other industries (27 percent of all pharmaceutical patent licenses include a royalty stacking provision, whereas only 6 percent of patent licenses in non-pharmaceutical industries include such a provision).
A patent license also may specify that a royalty will be reduced if the licensed product is combined with another product. Combination product royalty reductions are a way to apportion a licensed patent's value in the event that a licensed product is later combined into a single saleable unit with another product that is not covered by the patented technology. Combination product royalty reductions can take a number of different forms, but the most common form observed in the data set is one in which the base royalty rate is multiplied by the ratio of the licensed product's stand-alone price (A) to the sum of the stand-alone prices of the licensed and unlicensed products (A + B). That is, the royalty rate would be modified by the factor A/(A + B). Other forms of combination product royalty reductions include ratios based on manufacturing or fully allocated costs, the number of active components, and the number of functions performed.
Combination product royalty reduction provisions again are more commonly found in agreements in the pharmaceutical industry than in other industries (13 percent of all pharmaceutical patent licenses in the data set include a combination product royalty reduction provision, whereas only 5 percent of patent licenses in nonpharmaceutical industries include such a provision). Agreements also may include provisions reducing royalties if the licensed patents are later found to be invalid or unenforceable, if a "generic" product enters the market, or if the licensor chooses to convert the license from an exclusive basis to a nonexclusive basis. Tables 6 and 7 summarize the frequency of royalty reduction provisions in bare patent and patent-plus-know-how licenses broken out by industry type and licensor type.
Table 6. Frequency of Royalty Reduction Provisions by Industry (Percentages) Industry Any Royalty Royalty Stacking Combination Product Reduction Provisions Provisions Provision (1) (2) Pharmaceutical 40 27 13 Medical device 20 8 5 High-tech 12 8 5 Other 17 6 5 industries All industries 30 18 9 (1.) Royalty stacking provisions refer to royalty reduction if the licensed is subsequently required to obtain a patent license from a third party in rder to sell a licensed product. (2.) Combination product provisions refer to a royalty reduction if the original licensed product is combined with an unlicensed product into one saleable product. Table 7. Frequency of Royalty Reductions by Type of Licensor (Percentages) Licensor Type Any Roylty Reduction provision Nonprofit 57 University 45 Government 31 Commercial 22 Individual(s) 16 All licensor types 30
Royalty reduction provisions often take the form of a percentage reduction in the base royalty rate and include a maximum specified royalty reduction. For example, among licenses in the data set that include a royalty stacking provision, almost two-thirds (64 percent) specify a maximum royalty reduction of 50 percent of the base royalty rate. Among patent licenses in the data set that include a royalty reduction contingent on a subsequent finding that the licensed patent is invalid or unenforceable, three-quarters specify a royalty reduction of 50 percent of the base royalty rate. Some patent licenses include a royalty reduction provision if the license is converted from an exclusive basis to a nonexclusive basis. Among patent licenses that specify a royalty reduction in the event that a licensor converts an exclusive license to a nonexclusive license (either by licensing to a third party or by entering the market itself), over two-thirds specify a royalty reduction of 50 percent.
Sublicensing royalty rates
Many patent licenses include provisions that permit a licensee to sublicense the subject technology to third parties. Under a sublicense grant, the licensee, acting as an intermediary between the patent holder and a third party, becomes a sublicensor and the third party becomes a sublicensee. Approximately one-quarter (23 percent) of all patent licenses that specify a base running royalty rate also specify separate sublicensing royalty rates that the licensee must pay to the licensor for revenues it receives from a sublicensee. In a sublicensing arrangement, a licensee may incur low marginal costs associated with its sublicensing revenue (that is, marginal profits on sublicensing revenues to the licensee are relatively high), and all of the licensee's revenues from the sublicensee may be attributable to the licensed technology.
The median sublicensing royalty rate observed among patent licenses that include a sublicensing provision is 25 percent, with almost half (48 percent) between 20 and 30 percent. Only 11 percent of patent licenses in which the licensor was a commercial entity included sublicensing royalty rates, whereas almost three times as many (32 percent) of those in which the licensor was a noncommercial entity included sublicensing royalty rates. One explanation for this observation may be that noncommercial entities may lack the market contacts to realize a technology's full potential, and thus are more inclined to rely on a licensee (which is almost always a commercial entity) to serve as an intermediary in the licensing process.
Analysis of selected license provisions: Exclusivity, equity, and minimum annual royalties
There are a number of differences across industries in licensing provisions observed in the patent agreements in the data set. Analysis suggests that patent licenses in the life sciences industries are almost twice as likely to be granted on an exclusive basis as licenses in the high-tech industry (84 percent vs. 46 percent), almost twice as likely to specify payment of minimum annual royalties (36 percent vs. 19 percent), and almost three times as likely to specify royalty reduction provisions (35 percent vs. 12 percent).
The data set also suggests that differences exist between patent licenses in which the licensor is a commercial entity and licenses in which the licensor is an educational institution. University licensors more frequently license on an exclusive basis than do commercial licensors (94 percent vs. 66 percent). University licensors also are more than twice as likely to grant royalty reduction provisions (45 percent vs. 22 percent), and almost three times as likely to require payment of minimum annual royalties (60 percent vs. 21 percent) than are commercial licensors.
Inclusion of some of these licensing provisions may be partially explained by strategies to shift risks from one party to another. For example, patent licenses granted on an exclusive basis are more than twice as likely as licenses granted on a nonexclusive basis to include a grant of equity (28 percent vs. 11 percent). An equity grant to the licensor gives the licensor the potential for a higher return on its license in the event that the licensed products are a commercial success, without burdening the licensee with higher marginal costs.
Patent licenses granted on an exclusive basis also are more likely (35 percent vs. 26 percent) to include minimum annual royalty provisions as are nonexclusive patent licenses. Failure by the licensee to pay a minimum annual royalty allows a licensor to license its technology to another firm in the event of commercial failure of the licensed product by the original licensee, thus reducing the licensor's risks. Other provisions that could be analyzed in terms of shifting risks between the parties include assignability provisions, bankruptcy provisions, and license or grant-back provisions related to technologies subsequently derived from the original licensed technology.
In this paper, I examine the financial structure and selected provisions of over 1,400 patent licenses and patent assignment agreements from a unique data set comprised of material agreements filed with the SEC. This data set is comprised specifically of patent licenses (that is, bare patent and patent-plus-know-how licenses) and patent assignment agreements as opposed to agreements in which patents are not the primary focus (for example, joint venture agreements, merger or acquisition agreements, distribution agreements, or supply agreements).
The main finding of this paper is that basic differences in the structure and provisions of patent licenses exist across industries and across types of licensors, and that many of these differences can be analyzed and explained using basic economic principles. For example, economic analysis of risk allocation strategies, product flexibility, differing perceptions of the likelihood of commercial success, or demand side concerns such as price elasticity of the licensed products are possible explanations for different provisions and financial structures of patent licenses. Such analyses may be useful either in considering the appropriate structure of the patent license (for example, fixed fee, running royalty, or some combination of the two) or the impact of an exclusive grant vs. a nonexclusive grant. At the same time, such analyses also may be useful in considering the impact of selected licensing provisions such as royalty reduction provisions, minimum annual royalties, or sublicensing provisions.
My purpose in studying patent licenses and analyzing their differences across industries and types of licensors is to encourage economists in the business community to take a more active role in analyzing and providing input into their firms' technology licenses. Although this paper examines only patent licenses, the type of economic analyses considered also can be useful in analyzing other types of licenses that involve technology development or transfer, such as joint venture agreements, cooperative research and development agreements, and merger or acquisition agreements. The economic concepts discussed in this paper-for example, strategies for managing risk, analysis of asymmetric information and differing product expectations, and consideration of future technical or commercial contingencies-also are well suited to this broader set of technology licenses.
(1) The SEC permits a registrant to redact for a specified period of time those portions of its filings that are "trade secrets."
(2) wwwpacergov.) (PACER: Public Access to Court Electronic Records).
(3) For example. Form 8-K General Instructions, SEC 873, Item 1.01 Entry into a Material Definitive Agreement.
(4) I wash to acknowledge the significant contribution of my wife, Laura Varncr, in assisting with the considerable effort in organizing and processing these FOIA requests made as part of this research.
Anand, Bharat N.., and Tarun Khanna. 2000. "The Structure of Licensing Contracts." The Journal of Industrial Economics, 48(1): 103-35.
Cans, Joshua S., David H. Hsu, and Scott Stern. 2008. "The Impact of Uncertain Intellectual Property Rights on the Market for Ideas: Evidence from Patent Grant Delays." Management Science, 54(5): 982-97.
Lyarskaya, Natalia. 2009. "Design of Technology Licensing Agreements: New Empirical Evidences,1' Working paper, http://www.epip.eu/conferences/epip04/files/ LYARSKAYA_Natalia.pdf.
Sakakibara, Mariko. 2009. "An Empirical Analysis of Pricing Patent Licensing Contracts/' Working paper, http://ssrn.com/abstract= 1515163.
Varner, Thomas R. 2010. 'Technology Royalty Rates in SEC Filings." les NouveUes, The Journal of the Licensing Executives Society, 45(3): 120-7.
* Thomas R. Varner is a Vice President with Economists Incorporated in San Francisco, CA. A focus of Varner's consulting practice is on patent, copyright, and trademark licensing and litigation. He has conducted extensive research into thousands of technology and trademark licenses, patent settlement terms, and patent litigation outcomes. Varner received his Ph.D. from Stanford University and MS and MBA degrees from the University of California at Berkeley.
THOMAS R. VARNER *…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: An Economic Perspective on Patent Licensing Structure and Provisions. Contributors: Varner, Thomas R. - Author. Journal title: Business Economics. Volume: 46. Issue: 4 Publication date: October 2011. Page number: 229+. © 1999 The National Association of Business Economists. COPYRIGHT 2011 Gale Group.