Licensing and Innovation Regimes in Pharmaceutical R&D

Licensing and Innovation Regimes in Pharmaceutical R&D

Licensing in Pharma: Competitive Efficiency for Incremental Drugs vs. Lemons-Type Frictions for Novel Ones

When drug companies license new medicines, it works well for predictable, "me-too" drugs because they can accurately judge quality. However, for groundbreaking, highly novel drugs, the market struggles to price risk correctly. This leads to inefficiencies similar to a "market for lemons."

In the pharmaceutical industry, biotechnology firms often out-license drug candidates to large companies. This creates a natural environment for information asymmetry (a situation where one party has more information than the other). The originator usually understands the science better. Meanwhile, the licensee brings the commercial muscle. Does this imbalance cause the market to fail, or does competition fix it? A new study from Michele Liberatore and Massimo Riccaboni suggests the answer depends entirely on how "new" the invention actually is.

Does the market for technology actually work?

The authors investigate whether pharmaceutical licensing markets efficiently allocate innovative assets. They look for signs of adverse selection (a market failure where low-quality goods displace high-quality ones). This is a classic economic problem. It was famously illustrated by George Akerlof’s "market for lemons." In that scenario, buyers fear they are purchasing low-quality goods. They offer lower prices. This eventually drives high-quality sellers out of the market.

The central question is whether the pharmaceutical industry has solved this problem. Can firms use sophisticated contracts and due diligence to manage risk? Specifically, the researchers ask if the market's ability to screen for quality depends on the "innovation regime" (the specific type of innovation being traded). They propose that predictable, incremental projects might be screened effectively. However, frontier innovations might suffer from informational frictions (barriers caused by unequal access to information).

Cracks in the aggregate view

Until now, empirical evidence on pharmaceutical licensing has been contradictory. Some studies report that in-licensed projects have higher clinical success rates than internally developed ones. This seems to disprove the "lemons" theory. Other studies find that inexperienced originators receive heavily discounted payments. This suggests that information asymmetries are very much alive.

The authors argue that these conflicting findings exist because researchers have looked at the industry in the aggregate. By pooling predictable "me-too" drugs with highly uncertain "first-in-class" therapies, previous studies missed the underlying segmentation. As shown in, the industry is not a monolith.

It is a spectrum of risk. The researchers suggest that the market might function perfectly for one segment while failing for another.

Testing the innovation spectrum

To untangle these regimes, the researchers developed a theoretical model. In this model, information precision (how clearly a buyer can judge quality) is tied to the type of innovation. They tested this model using a massive dataset from Evaluate Pharma R&D. The data contains over 192,000 product-level observations.

The study employs a "Double Machine Learning" (DML) framework. DML is a statistical method that uses machine learning to control for hundreds of complex, non-linear confounders (hidden factors that influence both the decision and the outcome). This allows the researchers to isolate the true effect of licensing on success and profit. To ensure their findings were not just correlations, they added a second layer: "DML-IV." This method uses "exogenous pipeline shocks" (unexpected events outside a manager's control) as a tool. Specifically, they used sudden Phase III clinical trial failures at a licensee firm. These failures trigger "rushed" licenses. These rushed deals act as a natural experiment. The timing of a trial failure is driven by biology, not by a manager's desire to sign a specific contract.

By constructing a continuous proxy for project risk using machine learning, the authors mapped every project onto a spectrum of novelty. This allowed them to see exactly where the market begins to break down.

Winners, losers, and the risk-return trade-off

The results reveal a stark divide in how the market treats different types of innovation. First, the authors find evidence of "positive selection." In-licensed projects generally show higher success probabilities than internally developed ones. This trend is visible in .

This suggests that the market is successfully picking winners in the aggregate.

However, the real story emerges when looking at the money. For incremental, low-risk projects, the authors report a "competitive risk-return trade-off." While these licensed projects are more likely to succeed, they generate lower net returns than in-house projects. This is seen in .

This is the hallmark of an efficient market. The "insurance" of a higher success probability is paid for by accepting lower profitability. Neither strategy dominates; they simply serve different roles in a balanced ecosystem.

But for novel, high-risk projects, this balance evaporates. At the most innovative end of the spectrum, the return penalty disappears. The authors find that licensed frontier innovations retain higher success probabilities without a corresponding drop in net returns. More strikingly, the causal analysis of "rushed" licenses shows a different problem. For novel projects, the market fails to price risk correctly. This leads to significant return losses without any compensatory boost in success rates. This breakdown is visualized in . The protective trade-off seen in incremental drugs is missing for the most novel ones.

Implications for the frontier of R&D

The study demonstrates that market efficiency is not a universal constant. It is a function of information. If these findings hold, they suggest that the most critical segment of pharmaceutical R&D is problematic. The frontier of novel science is precisely where market imperfections are most concentrated.

There are two major implications. First, for practitioners, the data suggests that "rushing" a license due to pipeline pressure is dangerous for novel assets. A rushed license for a predictable drug might not hurt much. However, doing so for a first-in-class therapy can lead to a significant loss of value. This happens because the due diligence process is too compressed to overcome scientific uncertainty.

Second, for policymakers, the results highlight a potential issue. If the market cannot efficiently price and reward the most novel, high-risk projects, investment may shift toward safer, incremental improvements. The paper does not explore whether these patterns hold in other high-tech sectors. It remains to be seen if these frictions are a universal feature of all complex technology markets.

Figures from the paper