Predicting regulatory drug approvals for the U.S. pharmaceutical industry

Maciver, Gillian (2020) Predicting regulatory drug approvals for the U.S. pharmaceutical industry. PhD thesis, University of Glasgow.

Abstract

This thesis intends to predict outcomes from the drug discovery process. These predictions are then utilised to develop a potentially profitable investment strategy. It explores whether certain patent and firm-level publishing characteristics can prove informative to this predictive process. This study hypothesises and investigates (empirically tests), for the first time, the links between drug approval granted by the U.S. Food and Drug Administration (FDA) and a range of firm-level patent and publishing characteristics. Using hand-collected data for 290 drugs approved by the FDA and their corresponding 632 patents I analyse how certain selected characteristics influence the likelihood of FDA approval. Findings suggest that predicting the probability that drug-related patents result in FDA drug approvals or project failures are enhanced by employing five patent-based and two firm-publication indicators. At the time of patent publication, inventor team size, citations to scientific journal articles, independent patent claims and dependent patent claims all have a positive and significant role in the prediction process. Independent patent claims are especially informative at the time of patent publication, whilst patent citations received are found to be the strongest indicator of approval success, three years after patent publication. Firm publication and university co-authoring also possess a significant, although negative, influence on the likelihood of drug approval.
The model’s success in predicting approval-linked patents, correctly predicting 19.75% and 17.82% of these patents for in-sample and out-of-sample analyses, is matched by its success as the basis of a profitable investment strategy. The high levels of correctly predicted approval-linked patents enable the investment portfolios, constructed using the model’s predictions, to generate significant positive long run abnormal returns. I find that, contrary to the market efficiency hypothesis, my portfolios are able to generate in-sample approval date buy-and-hold abnormal returns of 60.40%.
With the high levels of information, asymmetry and uncertainty that exists within the drug discovery process these results will prove beneficial to practitioners, academic researchers and investors.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: patent indicators, pharmaceutical industry, Food and Drug Administration (FDA), drug development process, drug approval, prediction, shareholder wealth effects
• Subjects: H Social Sciences > HG Finance
• Colleges/Schools: College of Social Sciences > Adam Smith Business School > Accounting and Finance
• Supervisor's Name: Siganos, Dr. Antonios
• Date of Award: 2020
• Embargo Date: 14 May 2023
• Depositing User: Mrs Gillian Maciver
• Unique ID: glathesis:2020-81359
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 14 May 2020 15:02
• Last Modified: 14 May 2020 15:02
• URI: https://theses.gla.ac.uk/id/eprint/81359