“Who’s on First?”
Biopharma Licenses from the Seller’s Perspective
Over the past decade all biopharma industry participants have become licensors of biomedical IP from time to time. How do deal structures differ on the basis of who is the seller?
Covering the alliances of the biopharma industry used to be straightforward. The NIH funded biomedical research at academic institutions which, under the auspices of the Bayh-Dole Act, licensed intellectual property (IP) generated with such funding to biotech companies or, sometimes, to top pharma. Biotech companies added IP of their own and licensed it, along with the IP from universities, to top or mid-tier pharma or, sometimes, kept a program for themselves in an effort to become a top or mid-tier pharma.
Today’s biopharma industry is far more complex, and classifying biopharma licenses is akin to Abbott & Costello’s 1940s comedy routine ”Who’s on First?” — here’s a clip. Not only do research institutions continue to generate and license biomedical IP, so too do all other industry participants – biotechs license each other frequently, and mid-tier and top pharma are often licensors of IP as well.
One upside of this complexity is the possibility to observe and emulate best practices in the structure of biopharma licenses. If all industry participants are licensors of biomedical IP from time to time, how do deal structures differ on the basis of who is the seller?
To understand the extent to which industry customs differ on the basis of who is the seller of biomedical IP, we analyzed approximately 800 biopharma licenses with respect to four deal elements: field of use, sharing of pre-launch sublicense revenues, licensee diligence, and duration of royalty obligations.
For purposes of this analysis, we looked at biopharma licenses commenced since January 2007 which have been (i) SEC-filed as material contracts, (ii) obtained by BioSci in unredacted form, typically via Freedom of Information Act (FOIA) requests, and (iii) tagged for specific contract provisions by BioSci’s analysts. We excluded acquisitions and asset purchases, but included distribution, joint venture, co-promotion and co-market agreements, as these latter deal types contain a license component. We found 834 such alliances, all of which are listed in the All Sellers spreadsheet accompanying this article. This spreadsheet also contains data regarding therapeutic area, technology and deal economics for each alliance.
With respect to the four deal elements of interest, we applied these additional selection and classification criteria:
- Field of Use – There are 617 tagged instances wherein the licensed field of use is a defined term in the contract. These are classified as either (i) all fields of use, (ii) field limitations by disease, (iii) filed limitations by type of product (e.g. oral, antisense), or (iv) field limitations by disease and product type.
- Share of Pre-Launch Sublicense Revenues – There are 265 tagged instances wherein the licensee is obligated to share a percentage of revenues obtained in the event the licensed IP is sublicensed prior to commercialization. Most such instances include what constitutes sublicense revenues (e.g. upfront payments, milestones) as a defined term. We excluded any instances wherein the only sublicense revenues to be shared were based on product sales.
- Licensee Diligence – There are 627 tagged instances wherein the licensee is required to meet one or more standards of diligence to retain the license for the full contract term. These standards of diligence are classified as (i) commercially reasonable efforts (CRE) or diligent efforts, in each instance a defined term, (ii) a development timeline with specific performance requirements, (iii) development expenditure requirements, (iv) commercialization expenditure requirements (e.g. sales force size, promotional expense), and/or (v) minimum payment obligations during the post-launch period.
- Royalty Term – There are 724 tagged instances wherein the duration of royalty payment obligations is explicitly described in the contract. These are classified as (i) life of patent, (ii) later or earlier of (a) life of patents, (b) years from (x) first commercial sale (FCS) or (y) agreement signing or (c) loss of regulatory or market exclusivity, including generic entry, or (iii) some other royalty duration (e.g. perpetuity, life of product).
As with most BioSci analyses, the deal-specific contract terms included in this analysis are accessible to all readers via the embedded “Tags” in each of the four deal element spreadsheets described above. Subscribers to BiosciDB.com may also access the deal press release, financial notes disclosures and/or full contract(s) by following the “Deal” link.
Field of Use
As shown in Figure 1, roughly half of biopharma licenses have either a limitation on use by disease (24%) or by type of product (21% – e.g. diagnostic, monoclonal), while the other half have either no restrictions (42% – all uses) or are the most restrictive licenses (13% – limitations by disease and product type). It is surprising that a majority of biopharma licenses have some sort of field of use restriction, insofar as the freedom to operate and to exclude imitators have long been described as the basis of the biopharma industry’s heavy reliance on IP.
Figure 2 shows how field of use limitations vary by type of licensor. “All uses” licenses are most common for IP licensed either from Biotechs or Universities. Top Pharma licensors (i.e. the 20 pharma companies with most sales in 2019) issued more licenses with disease restrictions, and such limitations by disease were prominent in licenses from Mid-Tier Pharma as well. University licensors were least reliant on disease-based field restrictions, but were more reliant on limitations based on type of product.
Sharing of Pre-Launch Sublicense Revenue
Figure 3 shows the average and median share of sublicense revenue owed to the licensors of biomedical IP in the event that such IP is sublicensed prior to the commercialization of product. The rationale for such provisions is that a licensee who sublicenses (or “flips”) licensed biomedical IP to another entity owes a substantial share of any payment so obtained to the original (“upstream”) licensor, particularly in the instance that the licensee has added little value to the IP prior to its sublicense. Of the 265 biopharma licenses having such provisions, 111 (42%) contain a range of sharing percentages, wherein the share owed to the upstream licensor typically decreases on the basis of advanced stages of development, lapse of time or amount invested by the licensee prior to sublicense. To capture the variability implicit in such instances, we have included the lowest share of sublicense revenues in the “Share Low” calculations, and the highest share in the “Share High” calculations of Figures 3 & 4.
With respect to all biopharma licenses having sublicense revenue sharing provisions, Figure 3 shows that such sharing is 15-20% on a median basis, or roughly 18-25% on average, depending on whether one uses the “Share Low” or “Share High” calculations. However, it should be noted that the exclusions from what is defined as “sublicense revenue” vary substantially from license to license, so one should review the tagged definitions of sublicense revenue in specific instances prior to relying on any median or average figures.
Figure 4 shows that sharing of sublicense revenue tends to be higher when the IP licensor is a Biotech as compared to a University if the licensee does all it can to lower its obligation to share sublicense revenues. For the “Share Low” calculations, the median share owed to a Biotech licensor was 15%, versus 12.5% to a University licensor. On average, the Biotech licensor was entitled to 17.6%, versus 13.8% to a University licensor. Interestingly, the median and average shares owed to Biotech and University licensors are the same in the “Share High” calculations, suggesting that both groups of licensors apply equivalent sharing standards in the instance of “flipping.” There were too few instances of pre-launch sublicense revenue sharing provisions in licenses from Top Pharma or Mid-Tier Pharma to evaluate meaningfully.
Figure 5 shows the five standards of diligence typically found in biopharma licenses. Of the five, a defined standard of Commercially Reasonable Efforts (CRE) or Diligent Efforts was most common, and was utilized in 45% of the licenses. This was the only diligence standard in 56% of the CRE instances, and was combined with more or more diligence standards in the remainder of instances. One or more development timeline obligations was the next most frequent diligence standard in 27% of deals, followed by minimum annual sales payments (17%) and commercialization expenditures (7%). Obligations regarding development expenditures for the licensed IP were least common, occurring in only 40 licenses and representing 4% of all instances of diligence standards.
With respect to types of biopharma licensors, Figure 6 shows that CRE was typically used by all but University Licensors. The latter were most reliant on development timeline obligations, which were quite common for all other licensor types as well. Minimum annual payments were utilized least by Top Pharma licensors, who utilized commercialization expenditure obligations more often than did other types of licensors.
Figure 7 shows that of the 724 tagged instances having a specific duration of royalty obligation (as compared to profit split, transfer price or payment per unit), the most common standard was for the life of patents or “X” years from first commercial sale (FCS), which occurred in 44% of instances. Specifically, the “later of” life of patents or “X” years from FCS was most common, occurring in 282 licenses, or 40% of all tagged royalty term provisions. Next most common was life of patents only, at 33% of instances, followed by “X” years from date of license signing, at 13% of instances.
Figure 8 shows that “later of life of patents or ‘X’ years from FCS” was typically used by all but University Licensors. The latter were most reliant on life of patents only. Non-University licensors also utilized “’X’ years from FCS” quite extensively, although this standard was rarely used by University licensors.
Finally, Figure 9 shows the average and median durations, in years, of royalty obligations for each of the four standards not based on life of patents only. Interestingly, these average and median durations don’t vary significantly for each type of licensor, so they are simply shown here for all biopharma licensors.
The number and type of licensors of biomedical IP have increased significantly over the past decade. While many IP licensors have followed industry customs derived primarily from research institutions, corporate licensors have recently established different customs for the treatment of common license structures. BioSci believes it’s beneficial to all biopharma licensors to evaluate and emulate best practices in the structure of biopharma licenses.