From a distinctive sign to an exchangeable asset: exploring the U.S. market for trademark licensing

ABSTRACT

A remarkable growth in the value of trademark licencing has been recently recorded. Our paper contributes to the understanding of this under-explored phenomenon using a dataset newly released by the USPTO. Our study analyses the evolution of licencing activities in the U.S. during the 2003–2017 period, the characteristics of these trademarks and agreements, and certain features of the licencing parties involved. We found that licencing activities varied considerably during these years. They were usually signed between two parties only, and, on average, they involved more than one trademark. Excluding under-reporting effect, the analyses reveal that a large portion of heterogeneity in licencing activity is due to the NICE international classes associated with each trademark. Indeed, trademark licencing agreements appear to be unevenly distributed across these classes, suggesting that this activity and the way it is carried out is correlated with the market to which the licenced trademark refers.

KEYWORDS:

• Market for trademark
• trademark licencing
• trademark classes
• licensors
• licensees

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Albeit we acknowledge that the concept of ‘markets for brands’ (Frey, Ansar, and Wunsch‐Vincent 2015; Castaldi 2019a) is in assonance with that of ‘markets for technologies’ (Arora, Fosfuri, and Gambardella 2001), we prefer to adhere to the more recent contributions in the patent-based literature that explicitly refer to ‘the market for patents’ (Gambardella, Giuri, and Luzzi 2007) as they specifically look at patent licencing as we do with trademark licencing. Hence in the present paper, we decided to narrow down our perspective and focus on trademarks transactions, which represent a subset of the so-called ‘market for brands’ (Frey, Ansar, and Wunsch‐Vincent 2015; Castaldi 2019a), as they are often an integral part of brands transactions (WIPO 2013: 62).

² Under a licencing contract, the licensor grants the right of use of the property for a certain period, in a specific field within a specified geographical area, under binding conditions (e.g. exclusive terms), to the licensee who agrees to use the trademark right in full compliance with these conditions in exchange for monetary and/or non-monetary compensation.

³ Global Licencing Industry Survey 2018, available at: https://www.licensingitalia.it/en/2018-lima-global-licencing-industry-survey/, accessed on June 2019.

⁴ ‘Top 150 Global Licensor’, LICENCE GLOBAL. The Licencing Industry’s Thought Leader, available at: https://www.licenseglobal.com/resource/top-150-global-licensors-2018, accessed on June 2019.

⁵ We also try to provide stronger evidence than mere descriptive statistics on the role of trademark classes for the licencing process by running two econometric exercises. The main purpose of these exercises is to show the persistence of heterogeneity across classes even when controlling for other structural aspects. All in all, we find robust evidence that trademark classes capture a significant degree of variation both in the probability of licencing and in the speed of trademark licencing. For full references, see the Empirical Analysis section.

⁶ See Frey, Ansar, and Wunsch‐Vincent (2015) and WIPO (2013); Graham et al. (2013) and Graham, Marco, and Myers (2018) for a full-fledged analysis of the reasons for the lack of complete data on the licencing phenomenon.

⁷ In other words, trademarks are the remedy introduced to mitigate a failure of the market – namely, information asymmetry (Akerlof 1970)– which characterises the relations between producers and consumers in terms of their knowledge of the characteristics and quality of goods and services (Landes and Posner 1987; Ramello 2006; Ramello and Silva 2006; WIPO 2013). In modern markets, especially those of consumer goods, the neoclassical assumption of perfect information – where buyers have full knowledge of the quality and characteristics of all the products and services offered – is not fulfiled in practice (WIPO 2013).

⁸ For a further analysis of trademark functions and impacts, refer to the article ‘Empirical studies of trade marks – the existing economic literature’ by Schautschick and Greenhalgh (2016), which offers a wider review of trademark-related literature.

⁹ The USPTO Trademark Assignment Dataset is publicly available for download at: https://www.uspto.gov/learning-and-resources/electronic-data-products/trademark-assignment-dataset. We used the 2017 version, which is the most recently updated. A complete description of this dataset is provided by Graham, Marco, and Myers (2018).

¹⁰ According to the 2013 WIPO report, the analysis of available deals shows that, across sectors, average royalty rates in trademark licencing agreements vary from less than 5% to more than 25% of (gross or net) sales. The highest average rates are found within the Celebrity and Character category, while the lowest average royalty rates relate to Corporate/Product and Fashion trademarks.

¹¹ Trademark licencing is one of the few strategies that companies can use to gain access to markets in which commercial, regulatory or cultural barriers preclude – or strongly discourage – the use of autonomous entry methods such as exports and direct investments (Clegg and Cross 2000; WIPO 2003).

¹² In almost any jurisdiction, trademark protection is linked to its effective use; therefore, businesses need to show stable commercial use of a trademark in order to preserve the rights they have over it. Since the licensee’s use of the brand is to the benefit of the licensor, the licence allows him/her to defend his/her rights in all countries and for all product categories for which the trademark has been regularly registered but not directly used by the original owner (Frey, Ansar, and Wunsch‐Vincent 2015).

¹³ Conveyance text captures non-standardised information from the coversheet describing the interest conveyed or transaction recorded (Graham, Marco, and Myers 2018).

¹⁴ We identified corrective registrations as corrections, re-recordations or amendments (and the like), and manually corrected them by linking these registrations, when possible, to the original recordations (through the unique licence code).

¹⁵ Since the data gathered so far for 2018 might underestimate the extent of the phenomenon – due to the lag between execution and recordation date of the most recent agreements (Graham, Marco, and Myers 2018) – we have opted to omit this year.

¹⁶ As suggested by Graham, Marco, and Myers (2018), we used a given contract’s execution date (or the acknowledgement date if the execution date field was not populated) – not the recordation date – to be the reference starting date of the agreement. Since transactions are recorded on a per-assignor basis and multiple execution dates are possible for a single transaction, we used the most recent execution date under the assumption that it denoted the moment the transaction was complete.

¹⁷ The Trademark Case Files Dataset contains detailed information on approximately 9.1 million trademark applications filed with, or registrations issued by, the USPTO between 1870 and 2018 (https://www.uspto.gov/learning-and-resources/electronic-data-products/trademark-case-files-dataset-0). A complete description of this dataset is provided by Graham et al. (2013).

¹⁸ Status was captured using a three-digit code indicating whether an application was abandoned or pending or whether a registration was live, cancelled or expired (Graham et al. 2013).

¹⁹ The NICE classification, named after the ‘Nice Agreement’ (1957), is an international standard to classify goods and services on the basis of their trademarks. It is made of 45 classes: the first 34 refer to goods while the remaining 11 refer to services. https://www.wipo.int/classifications/nice/en/.