Research Highlights Article
June 16, 2026
Boundaries of the firm
Limiting the cannibalization of existing patented products
Source: l i g h t p o e t
When firms develop new products, conventional economic theory holds that the decision involving whether to perform in-house R&D or to outsource comes down to a standard set of considerations: relative competence, the cost of negotiating contracts, and the incentive problems that arise when two parties must split the gains from a joint effort.
In a paper in the American Economic Journal: Microeconomics, authors Thomas Jungbauer, Sean Nicholson, June Pan, Michael Waldman, and Lucy Xiaolu Wang argue that this list is incomplete. They find that firms may consider using internal R&D to avoid the development of new products that would compete with existing profitable portfolios.
"By outsourcing, I'm adding an uncertainty factor," Jungbauer explained in an interview with the AEA. "Maybe I would like to have the developed product come out in a particular way, but by outsourcing, there is increased uncertainty. The firm I'm outsourcing to will care a lot less about not cannibalizing my existing products than I do."
The authors add to Grossman, Hart and Moore's well-known property rights theory of the firm insights (1986 and 1990) from the durable goods literature, such as Waldman's 1993 analysis of planned obsolescence. They ask how a firm's product portfolio affects the boundaries of the firm when a new product can cannibalize sales of the firm's existing products.
“No one has really systematically investigated whether a firm’s current portfolio affects my interactions with other firms when innovating," Jungbauer said.
The researchers’ theoretical model uses a standard framework in which patented products occupy locations in product space, and a firm developing a new product chooses both a target location and an investment in location precision that determines how close to that target the realized product will land.
Because the fine-grained decisions that determine a product's location cannot be specified or enforced in a contract—a property known as “noncontractibility”—a firm that outsources cedes control over this precision to a firm that has low motivation to protect the originator's existing patents. The authors’ model yields four predictions.
A firm is less likely to outsource when (i) the firm owns a patented product nearby in product space, (ii) the existing patent has a longer remaining life, (iii) the firm’s potential market share for its existing product is larger, and (iv) in-house development is likely to produce fewer "pivots"—drifts into different product (sub-)categories than originally intended.
Even if we control for the expertise, we still find that the product portfolio has a substantial effect. Don't get me wrong, expertise one hundred percent matters. I'm just saying we are adding an additional novel explanation—avoiding self-cannibalization.
Thomas Jungbauer
To test these predictions, the researchers used the Pharmaprojects dataset, which tracks the development of pharmaceutical compounds worldwide between 1989 and 2004, supplemented with data from Intercontinental Medical Statistics sales, to construct market share measures. The final sample contained 11,493 compounds, of which roughly 78.5 percent were classified as in-house development at the compound-year level.
Using logit regressions that predicted whether each compound would be produced in-house or outsourced based on a range of relevant characteristics, the authors found support for all four predictions.
Pharmaceutical companies were more likely to do R&D in-house when they already held patents for compounds in the same therapeutic class, especially when the patents had a longer remaining life or a larger market share, and the compounds developed in-house saw fewer development pivots.
The analysis incorporated a number of robustness checks. In particular, the researchers controlled for each firm’s accumulated experience in a therapeutic class, addressing the most obvious alternative explanation.
"Even if we control for the expertise, we still find that the product portfolio has a substantial effect," Jungbauer said. "Don't get me wrong, expertise one hundred percent matters. I'm just saying we are adding an additional novel explanation—avoiding self-cannibalization."
The findings may carry implications beyond organizational economic theory.
For product development managers, the paper suggests evaluating outsourcing strategically rather than purely on cost and capabilities, since keeping development in-house can protect a valuable portfolio.
For regulators, it adds nuance to the "killer acquisitions" literature, which warns that dominant firms may acquire rivals to discontinue them: firms may steer development in-house simply to avoid competing with themselves, an impulse that reduces wasteful duplication. At the same time, in-house R&D shuts out the possibility of using more efficient outside developers and may suppress socially valuable pivots into new therapeutic territory.
♦
“The Organization of Innovation: Incomplete Contracts and the Outsourcing Decision” appears in the May 2026 issue of the American Economic Journal: Microeconomics.