Research Collaboration Agreement (RCA)

What a research collaboration agreement is

A research collaboration agreement (RCA) — sometimes called a collaborative research agreement (CRA) or, when it is specifically between two or more academic institutions, an inter-institutional agreement (IIA) — is the contract that governs a joint research project carried out by two or more independent legal entities working together toward shared research goals. It sits alongside the sponsored research agreement (SRA) and the material transfer agreement (MTA) as one of the three contract types a technology transfer office (TTO) or sponsored-programs office reaches for most often when researchers want to work across institutional boundaries, but it covers a distinct situation: genuine joint conduct of research, with more than one party actively performing work, rather than one party simply funding another’s project or transferring a physical material.

The defining feature is mutuality. In a classic sponsored research agreement, a sponsor funds a single performing institution to carry out a defined statement of work and receives reports, IP rights, or other deliverables in return — the relationship is directional. In an RCA, two or more parties each contribute effort, resources, expertise, data, or funding toward a shared research objective, and the agreement has to allocate rights, obligations, and outcomes among genuine co-participants rather than between a payer and a performer.

Inter-institutional vs. industry-academic RCAs

RCAs appear in two common configurations, and the label on the document itself varies by institution even when the underlying structure is the same:

• Inter-institutional (academic-to-academic): two or more universities, or a university and a government or nonprofit research organization, jointly conduct a study — for example, a multi-site project where each institution’s investigators contribute distinct expertise, each institution’s IRB or ethics board has independent oversight of its own site, and funding may flow from a shared external sponsor to each institution directly, or from one lead institution out to the others by subaward. Universities increasingly require a written agreement for this even when no money changes hands between the institutions, specifically to avoid later disputes over data ownership, authorship, and rights to any resulting IP.
• Industry-academic collaboration: a company and a university jointly design and carry out a research program, each contributing something beyond money — the company might contribute proprietary compounds, equipment, or scientist time, and the university contributes personnel, facilities, and expertise, with both sides sharing risk and both sides expecting a stake in what results. This is what distinguishes it from a standard industry-sponsored SRA, where the company is functionally a funder rather than a co-investigator.

In practice the line between an “industry-academic RCA” and a straightforward SRA is not always sharp, and many institutions use one master collaboration-agreement template that flexes to cover both patterns depending on how the specific project is structured — see CASRAI’s guide to industry-university research partnership structures for how sponsored research agreements, master/framework agreements, and consortium models are negotiated in the industry-funded case specifically.

Key terms typically negotiated

IP ownership and licensing: background vs. foreground IP

Every RCA has to separate background IP — intellectual property each party already owned before the collaboration began and brings into it — from foreground IP, IP newly created while performing the joint work. Background IP ownership is rarely contested; the agreement typically just confirms it stays with the party that already owned it and grants the other party a limited license to use it for purposes of the collaboration. Foreground IP is the clause negotiated hardest, because in a true collaboration more than one party may have contributed to a resulting invention, which raises joint-inventorship and joint-ownership questions that a single-performer SRA does not have to resolve in the same way. Common structures include joint ownership with an agreed cost-sharing arrangement for patent prosecution, allocation of ownership based on which party’s personnel actually invented a given result (with cross-licenses back to the other party), or a first-negotiation/first-refusal right to license the other party’s share. See CASRAI’s background IP vs. foreground IP comparison for how these terms are defined and typically allocated. Where the project involves federal funding, background patent rights and any Bayh-Dole-eligible subject inventions follow the statutory framework regardless of what the collaboration agreement says about foreground IP generally — the Act only applies to the specific inventions conceived or first reduced to practice with federal funds.

Publication rights

Because both parties may have confidential or commercially sensitive interests, RCAs almost always include a publication clause giving each party advance notice before the other publishes results, plus a short review window (commonly on the order of 30-60 days for a general confidentiality check, extendable by an additional 30-90 days if a party needs time to file a patent application) rather than an open-ended right to block publication. University policies generally cap the total permissible delay and prohibit indefinite suppression, since unrestricted publication of research results is a core academic value the institution won’t fully sign away.

Data sharing and ownership

The agreement needs to specify who owns data generated jointly, what each party may do with shared datasets during and after the project, and how any human-subjects or otherwise restricted data will be handled — often by cross-referencing a separate data use agreement (DUA) rather than trying to fully resolve data-sharing mechanics inside the collaboration agreement itself. For inter-institutional academic collaborations this section also typically addresses which institution’s data-management and retention policy governs the combined dataset.

Funding, cost-sharing, and administration

Unlike a single-direction SRA, an RCA has to describe how costs are shared or allocated across parties: whether each institution funds its own participation independently, whether one party subawards a portion of external sponsor funding to the other, or whether the parties are jointly applying for a single grant with a lead applicant. This section typically also names a lead or coordinating institution responsible for overall project administration and reporting to any external funder.

Liability, indemnification, and term/termination

As with most research contracts, RCAs include standard clauses on each party’s liability for its own personnel and activities, indemnification, insurance requirements, publication and confidentiality survival after termination, and the conditions under which either party can exit the collaboration early.

How an RCA differs from a sponsored research agreement

The practical test institutions use: if one party is funding a defined scope of work that another party alone will perform, with the funder receiving reports, data, or IP rights back, that’s an SRA. If two or more parties are each actively performing research work and jointly own or co-generate the outcomes, that’s an RCA. Some institutions use a single “collaboration and sponsored research agreement” template that includes both a statement of work and a joint-participation structure, which is why the two labels sometimes blur in practice — but the underlying legal question the agreement has to answer is the same: is this funding-for-performance, or joint conduct of research by co-investigators?

How an RCA differs from a material transfer agreement

A material transfer agreement governs the one-time or ongoing transfer of a physical research material (a reagent, cell line, plasmid, animal model, or dataset treated as a discrete item) between institutions, typically with no joint research, no shared statement of work, and no funding flow attached — the recipient agrees to use the material only for specified research purposes, not to redistribute it, and to handle any resulting IP under terms the MTA specifies. The moment a project moves beyond “send me the material and I’ll tell you what I find” into genuinely joint experimental design, shared personnel effort, or co-authored/co-owned results, institutions typically require an RCA (or an SRA, if the relationship is funder-to-performer) instead of, or in addition to, the MTA. See CASRAI’s guide to the MTA process for how that narrower agreement type works.

Worked examples

Example: inter-institutional collaboration. Two universities are named as co-investigator sites on a single federally funded multi-site study. Each has its own PI, its own IRB approval for its own site’s human-subjects activities, and its own share of the award (received either directly from the funder or by subaward from the lead institution). Before work starts, the two institutions sign a collaboration/inter-institutional agreement covering joint publication order and CRediT-style contribution rules, shared data ownership and access, and how any jointly generated IP will be allocated. This is a textbook RCA: genuine joint performance, not funding-for-performance.

Example: industry-academic collaboration. A company and a university lab jointly design a research program: the company contributes a proprietary compound library and a portion of the funding, the university contributes assay development and personnel time, and both sides expect to co-own resulting foreground IP with negotiated licensing rights. The agreement negotiated here is a collaboration agreement, not a plain SRA, because the company is contributing more than money and both sides are performing research.

When it isn’t actually a research collaboration agreement

Counter-example: a biotech company sends a university lab a proprietary antibody, asking only that the lab test it in an assay and report back the results, with no joint experimental design, no company personnel involved in performing the work, and no shared funding. Even though the two organizations are “collaborating” in a loose sense, the contract this needs is an MTA (governing the antibody transfer and any resulting data-use terms) rather than an RCA, because there is no genuine joint conduct of research by co-investigators — one party is simply supplying a material and receiving results.

Related terms

See also: sponsored research agreement, material transfer agreement, data use agreement, subaward, Bayh-Dole Act, background IP vs. foreground IP, and CASRAI’s guide to industry-university research partnership structures.