Do Pharmaceutical Mergers Harm Drug Innovation?

Table of Contents

Tables and Figures

Introduction

1. Theoretical Framework and industry
1.1. Overview of M&A Theory and the pharmaceutical industry
1.2. Review of R&D and the drug approval process

2. Empirical Analyses
2.1. Empirical specification and data
2.1.1. Empirical specification
2.1.2. Data
2.2. Results - Mergers & Acquisitions and R&D intensity
2.3. Mergers & Acquisitions and R&D efficiency

3. Small Acquisitions and their Impact on Industry Efficiency

4. Discussion, Limitations and food for thought

Appendix

Tables and Figures

Figure 1: Average R&D cost per pharmaceutical drug in 2005 US$

Figure 2: Development of R&D and M&A expenditures

Figure 3: The pharmaceutical research and development process

Figure 4: Number of mergers over US$ 500 million in the dataset

Table 1: Anticipated effects of M&As on R&D

Table 2: Regional differences in the drug approval process and regulation

Table 3: Impact of M&A on R&D intensity

Table 4: Impact of M&A on R&D intensity incl. Country Dummies

Table 5: Impact of M&A on R&D efficiency, time periods

Table 6: Summary of key variables

Table 7: List of all mergers in the Dataset

Table 8: Thomson Dataset - Number of Companies per country

Table 9: Dataset 2 (Thomson) - Impact of M&A on R&D intensity 1980 to 2012, no control variables

Table 10: Dataset 2 (Thomson) - Impact of M&A on R&D 1980 to 2012 with control variables, Full table

Table 11: Dataset 2 (Thomson) - Impact of M&A on R&D 1980 to 2012 incl. Country Dummies, full table

Table 12: Impact of M&A on R&D efficiency, full table

Table 13: Impact of M&A on R&D efficiency, time periods full table

Introduction

The number and volume of pharmaceutical companies’ mergers and acquisitions have reached exorbitant magnitudes in the last decade, with estimations of over 1,300 deals worth US$ 690 billion between 2000 and 2009.¹ The biggest transaction, valued at US$ 66.7 billion², occurred in 2009 when Pfizer bought Wyeth. The sheer dimension of transactions in this sector, being subject to significant public interest, poses questions about the motivations behind these deals, which are mostly believed to be countermeasures for patent expirations, as well as the impact on drug innovation. Even though, the regular object of interest regarding mergers is the shareholder value, in this specific case drug innovation is of particular interest as it does not only influence the future prospects of the pharmaceutical companies but also long term public health. Hence, the objective of this thesis is to discover what the impact of pharmaceutical mergers on R&D intensity and efficiency of the acquiring organization and on the pharmaceutical industry as a whole is.

Accordingly, I present a short overview of general Mergers & Acquisitions (M&A) theory, the pharmaceutical industry and the historic development of M&A within the industry. Subsequently, I briefly explain and discuss R&D practices, drug approval processes and differences in regulation between selected countries. Afterwards, I present and analyze R&D spending and intensity of major pharmaceutical companies and its correlation with M&A’s. Additionally, I connect R&D spending to patent applications, controlling for real innovations and remakes as far as possible, in order to examine the effects on R&D efficiency more closely. The empirical analysis is constrained to major firms mainly due to data availability problems. The subsequent qualitative section however, will discuss smaller acquisitions, the research efficiency of small and mediums sized companies as well as their impact on the industry’s development efficiency. Major pharmaceutical players frequently buy start-ups or small, innovative research companies to refresh their patent portfolio or acquire promising intellectual property. I elaborate whether these deals increase overall drug innovation efficiency, by allowing research by potentially more efficient small companies with the prospect of a sell off. Finally, I expose limitations of the paper, recommendations for future research and food for thought about policy implications and behavior of pharmaceutical companies. The presented topic is not only relevant for pure academic purposes but also for policy makers and institutions involved in the pharmaceutical sector. If a significant part of mergers affects drug innovation, steps should be undertaken to prevent the negative effects or avert these mergers at all. It is of public interest that drug innovation is as efficient as possible and this should be prioritized over profits from a regulator’s perspective. Thus and because of little existing research on this topic, it is pertinent for a master thesis and adds to the existing academic literature.

1. Theoretical Framework and industry

In this section, I will give an overview of motivations for and performance of M&A in general and in the pharmaceutical industry especially. I present the main findings of recent literature and determine the avenue of my research in this paper. Additionally, I explain the characteristics of Research and Development (R&D) in the pharmaceutical industry, including regulations and processes regarding drug approval and patenting.

1.1. Overview of M&A Theory and the pharmaceutical industry

A tremendous amount of M&A’s occurred during the last decades. Between 1997 and 2007 alone, the global M&A deal value exceeded US$ 29 trillion (Financier Worldwide, 2008). The outcomes of many of these M&A transactions have been ambiguous, as short-term event studies show little shareholder value creation overall. Target companies, however, gain significantly whereas acquirers mostly lose shareholder value. This raises the question why companies undertake acquisitions in the first place. According to Sudarsanam, who summarizes existing research about the major motivations for horizontal mergers, economies of scale and scope, new growth opportunities, market power and network externalities incite companies to seek M&As (cf. Sudarsanam, 2010: 123–151). But these considerations do not explain the waves in which mergers occur, between 1900 and 2010 six waves arose. In order to understand these waves, additional decision factors for M&A activity, such as hubris and other behavioral biases as well as interdependencies between firms (e.g. bandwagon effect, mini-max regret theory)³, have been investigated. Furthermore, external shocks can cause acquisitions (cf. Powell & Yawson, 2005). Examples are the airline deregulation or technological advances in information technology, causing excess capacity which in turn leads to industry consolidation.

For this paper it is important to understand the genuine characteristics and motivations of pharmaceutical companies as well as the effect of the industry’s performance and efficiency on public health. A pharmaceutical company can be seen as being composed of two production activities; R&D with inputs such as labor, capital and technologies resulting in new drugs on the one hand and production, marketing and sales with self-developed, in-licensed or acquired compounds being the main input on the other hand (Danzon, Epstein, & Nicholson, 2004). Thus, any pharmaceutical company has the option to either invest in R&D or acquire compounds in order to sustain the input for revenue generating production. R&D was traditionally and still is the major investment for pharmaceutical companies, i.e. 15.3% of sales in the European pharmaceutical industry (European Federation of Pharmaceutical Industries and Associations, 2012) compared to well below 10% in most other industries. However, since the 1980s, M&A activity in the pharmaceutical industry has increased significantly. Several changes in the pharmaceutical market led to this turn towards acquisitions. First, while profits and prices were rising in the 1980s, R&D costs also increased significantly (2003; DiMasi, Hansen, & Lasagna, 1991). Second, many patents expired which put a tremendous pressure on the major companies to quickly refill their drug pipeline. Third, strong price competition from generic drug producers emerged, partly due to the expiration of major patents but also because of regulatory changes, such as the Waxman-Hatch act, which, since 1984, required generics to only prove bioequivalence without clinical tests for safety and efficacy (Grabowski, 2007) Fourth, the pressure on pharmaceutical companies increased even more, when stringent price controls and increased buyer power in form of managed care organizations in the US were introduced (Grabowski & Kyle, 2008).

From 1979 to 2005 the total average development cost per drug skyrocketed eightfold. Notable in particular is the over-proportionate increase of clinical cost vs. pre-clinical cost. This is partly due to more stringent clinical study requirements on the one hand and the rising difficulty of demonstrating improvements over drugs already on the market on the other hand. Other reasons such as the “Better than the Beatles problem” by Scannel et al. are explained in Section 2.2 (Scannell, Blackley, Boldon, & Warrington, 2012). The pre-clinical cost still increased by over four times during this period.

illustration not visible in this excerpt

Figure 1: Average R&D cost per pharmaceutical drug in 2005 US$

Source: Data from DiMasi, Hansen, & Grabowski (2003) and DiMasi & Grabowski (2007) compiled and translated to 2005 US$

Seeing this dramatic cost explosion, I am not surprised to find, that according to a KPMG survey (KPMG, 2011), the main driver for M&A activity is “Access to new technology and products”. Clearly, executives intend to use acquisitions in order to fill up their product pipeline, substituting or supplementing R&D investments. Figure 2 shows proof of increasing M&A expenditure and that, since the late 1990s, this went along with a decrease in R&D expenditure in relation to revenues. This is a clear indication for the substitution of R&D through M&A.

illustration not visible in this excerpt

Figure 2: Development of R&D and M&A expenditures

Source: Data from newly constructed dataset, Thomson Datastream, Thomson One, Zephyr

This leads to the question whether M&A really is more efficient than R&D for large pharmaceutical corporations and above that, what the effect on the pharmaceutical industry’s R&D efficiency as a whole is. This leaves me with two major research questions: First, whether pharmaceutical corporations should pursue large mergers under the rationale of filling up their product pipeline or invest these funds in their own R&D in order to induce growth and rising profit, especially considering significant premiums paid in many transactions. I analyze the empirical impact of large mergers on R&D intensity and efficiency at the firm level in Section 3. Afterwards, I discuss the direction of causality – Does an extraordinary decrease of R&D efficiency solely cause M&A activity or are mergers also one of the reasons for plummeting R&D efficiency? This requires a basic understanding of the drug development process and its regulation (Section 2.2).

Second, should regulators, in order to improve drug innovation and availability of treatments for diseases, encourage pharmaceutical M&As or limit them? This requires the inclusion of small companies in addition to major pharmaceutical companies in the analysis which specialize in the development of drug compounds in the pre-clinical stage. The prospect of being bought out allows small research firms to obtain funding for their drug development. If their efficiency is higher than the large corporations’, this type of M&A might increase overall industry innovation. Thus, a discussion of these small transactions supplements the analysis of major deals for regulators.

1.2. Review of R&D and the drug approval process

The research and development process that may finally lead to a cure for a disease is lengthy and expensive. (for the following: PhRMa, 2007, U.S. Food and Drug Administration, 2012) The first step is the pre-discovery phase, in which scientists analyze a disease, identify its underlying cause, and determine the proteins and/or genes which are altered and have to be targeted in order to treat the disease. This process alone can take years and the proteins found are not always suitable drug targets. They have to provide “docking stations” to which a drug compound can be attached and have to be affected by such a compound. The pre-discovery research is usually a joint effort of public research institutes such as universities as well as small and large pharmaceutical corporations. In the next step, the drug discovery stage, researches search for a “lead compound” which might ultimately become the new medicine. Common ways to find a lead compound are searching for existing molecules in nature, creating new molecules from scratch and the use of High-throughput screening (HTS). HTS is a relatively new approach, possible through advancements in processing power and robotics, which automatically tests hundreds of thousands existing and archived molecules as possible drugs for a target (molecule which is connected to a disease). This early phase research takes four to six years on average and only 250 of each five to ten-thousand compounds entering this stage make it to the next stage, the pre-clinical testing. During the pre-clinical testing, researchers examine whether potential drug compounds are safe and efficacious and especially how they will behave in the human body. In vitro and in vivo (animals) testing procedures are used. About five compounds on average show a promising effect and proceed to the clinical test stage, if their investigational drug application is accepted by the regulatory body. The clinical trials involve several phases with an increasing number of test subjects - from a small group of healthy subjects (phase I) in order to test the safety, over hundreds of diseased patients to prove efficacy (phase II) to a large group of several thousand patients with the disease (phase III), testing the required dose. This extent makes the clinical stage the, by far most expensive, stage of the drug development process. On average only one out of five tested compounds ever achieves the final approval as a drug by the regulatory body. The full process takes between 10 and 15 years and costs as much as US$ 1.3 billion.

illustration not visible in this excerpt

Figure 3: The pharmaceutical research and development process

Source: PhRMa, 2007

After understanding the drug development process, it seems instinctive to assume that the technological advancements in chemistry, robotics and computing power of the recent decades improved the efficiency of this process. Why then is it that we observe skyrocketing R&D cost per approved drug from 1979 to 2005? As discussed above and further investigated in Section 3, one of the factors could be the increased amount of M&As in the pharmaceutical industry. In order to understand the magnitude of this impact, several other causes of declining R&D efficiency are explored below.

Scannell et al. (2012) identify four primary causes for the decline. First the “better than the Beatles” problem which means that with an increasing number of existing drugs it gets harder and harder for new drugs, treating the same disease, to be successful. That is because a new drug, representing an iterative improvement, must not only prove higher efficacy and similar or higher safety but also has to compete with the generic existing solution. Thus, if the additional benefit is not high enough, patients or insurance firms will not pay enough to justify development and marketing. These challenges pressure pharmaceutical companies to focus their research on new areas, associated with high risks and difficulty but lower post-development competition, which ultimately leads to deteriorating R&D efficiency (also comp. Ruffolo, 2006). Additionally, many companies distribute R&D resources to the development of lifestyle drugs⁴ and thereby decrease the available funds for life saving drugs. Often lifestyle drugs promise higher profits and lower risks than life saving drugs. Another related factor is the “low hanging fruit” problem which means that most of the easy-to-discover drugs have already been discovered and it gets increasingly difficult to exploit new drug targets. However, the sheer amount of possible drug targets, which is huge, compared to the number of already addressed targets, makes this problem appear to currently have a minor impact (cf. Hu, Schultz, Sheu, & Tschopp, 2007). Second, a problem arises from the rising burden of proof for new drug approvals, Scannell calls this “the cautious regulator” (cf. Ruffolo, 2006 and Grabowski & Vernon, John M., Thomas, Lacy G., 1978). This happens not only due to higher moral standards and care but also due to the fact that a high availability of safe and efficacious drugs allows regulators to raise the bar for new compounds. Third, the “throw money at it” tendency led to rising R&D investments in recent decades (Scannell et al., 2012). Today, pharmaceutical companies try to increase R&D efficiency by cutting cost in certain parts of the R&D division. However, a poor understanding of factors determining R&D efficiency might lead to lower R&D cost without an increase in productivity. Last, the “basic research – brute force” bias addresses the changes in the research approach during the last decades (Scannell et al., 2012). Scannell suspects that brute force screening methods such as HTS are overestimated and that traditional discovery methods (cf. Bajorath, 2002); e.g., animal-based screens and iterative medicinal chemistry, might in fact be more productive. All findings of the empirical impact of M&As on R&D efficiency have to be interpreted and set in relation with the mentioned causes in mind. It will be examined whether acquisitions accelerate the R&D efficiency decrease or are simply a reaction to the effects of the mentioned causes. The empirical analysis will include time dummies in order to control for their impact. Furthermore, an indirect effect of mergers on R&D efficiency by impacting one of the other causes has to be considered. To conclude with this section, I explore the qualitative impact which pharmaceutical mergers can have on R&D and whose realized impact I will test in the empirical analysis as well as regulations in selected countries. It is not be possible to empirically measure the individual impact of each causal relation with the used dataset, only the joint influence is reflected in the coefficients of the merger variable.

I distinguish between temporary and long-term or permanent effects. The most important temporary impact is the disruption of R&D by the need to integrate the acquired R&D department (Ruffolo, 2006); i.e., streamlining the organizational structure, overcoming geographical barriers or combination of IT systems. During this post-merger integration process, neither R&D department can work as efficiently as before due to these disturbances. Which level of efficiency can be reached afterwards, highly depends on the sophistication of the integration - only if it is executed properly, benefits can be reaped. It is even possible to face a decrease in efficiency because of the higher organizational complexity. The according post-integration long-term effects have different magnitudes following the acquirer’s integration efforts.

illustration not visible in this excerpt

Table 1: Anticipated effects of M&As on R&D

Sources: Ornaghi (2009); Grabowski & Kyle (2008); Comanor & Scherer (2013); Koenig, Michael E. D. & Mezick (2004)

The first effect, fewer different approaches (see Table 1), means that two separate companies are likely to pursue two different avenues of research to solve a similar problem. Hence, the probability that either of the companies has chosen the right way is greater than if one merged entity only pursues one avenue. On the other hand, companies can obtain resources, such as compound databases, knowledge or high profile researches, in a merger or acquisition which they cannot attain otherwise. The impact of economies of scale is rather small, at a point where the best equipment is available only the enlargement of knowledge and compound databases slightly increases research capability. However, economies of scope play a larger role. The more areas of research a company engages in, the more knowledge and innovation spillovers increase the likelihood of success. Yet, it is important to keep in mind that there are limits to economies of scale and scope. Once a certain level is reached, no further benefits can be reaped. Thereby it is likely, that the major pharmaceutical companies have already reached this threshold and cannot improve further. Considering the entirety of effects (compiled in Table 1), there are both negative and positive impacts from a theoretical perspective. What the joint impact in individual cases is depends on the post-merger integration, the size and type of the acquisition and other factors. If economies of scale and scope are already exhausted, I expect an overall negative impact of mergers.

In the course of this paper, I use the constructed dataset to test the impact of regional differences on R&D intensity and revenue growth. In order to draw conclusions from the results, I present an overview of key characteristics of four regions: the United States, Japan, Germany, and Switzerland. I examine these four countries because only those are represented by three or more companies in the dataset. The findings, shown in Table 2, indicate that Japan lags behind in the efficiency of the drug approval process and that Japanese regulation fails to encourage development of drugs which are competitive in foreign markets. This is mainly due to the preference of low risk and low efficacy drugs over highly potent ones. Additionally, the sluggish approval process favors the relocation of R&D activities to more efficient regimes and the lack of early clearance of patented drugs for experiments delays the marketing of generic drugs and decreases price and innovation pressure. Accordingly, I expect Japanese pharmaceutical companies to lag behind the other regions in terms of revenue growth and R&D intensity. The situation in the United States, Germany and Switzerland is quite similar. The most notable difference is the formation of prices, which is regulated in Germany and Switzerland but not in the US. There are also slight differences in the orphan drug development incentives. Due to the higher drug prices in Germany and especially in the United States, which encourage R&D investment, I expect a slightly higher R&D intensity in these countries compared to Switzerland. Apart from that, I do not expect significant differences due to the country of origin. Another noteworthy impact of price formation is that the use of reference pricing schemes, such as the one in Germany, leads to a focus of R&D on new therapeutic areas rather than on incremental improvement of existing drugs. While this is desired by regulators, it does increase drug development costs and risks. (Pammoli, Magazzini, & Riccaboni, 2011)

illustration not visible in this excerpt

Table 2: Regional differences in the drug approval process and regulation

Finally, a vital task of regulatory institutions is the incentivization of orphan drug development. Orphan drugs can be defined as “those drugs intended to treat either a rare disease or a more common disease where manufacturer cannot expect to make profits.” (Thielke, D., Thyssen, J.P., Hansen, B.J., 2006), which includes not only rare diseases but also diseases where the majority of patients cannot afford a treatment, e.g. many tropical diseases. Only incentives such as federal grants, tax exempts or special consultation allow for the development of orphan drugs, as it would otherwise not be profitable. As seen in Table 2, in most developed countries many incentives have already been established in the recent decades. Still, the introduction of additional measures, such as a tax exemption in the European Union would further benefit the generally positive development. Furthermore many countries, e.g. India and New Zealand, are expected to implement measures soon. But not only incentives for the development of orphan drugs are important, also appropriate reimbursements by health insurances to allow for cost-covering drug prices are indispensable. As a result of the existing measures, the global orphan drugs market has grown tremendously from US$ 54.5 billion in 2005 to US$ 84.9 billion in 2009 and is expected to reach US$ 112.1 billion by 2014. Yet, it is important to closely observe the effect of measures such as the fixed, longer marketing exclusivity. It is possible that companies use this to earn obscene profits and sustain the exclusivity by continuously adding new patents around the main compound. In this way, Jazz Pharmaceuticals could raise prices by 40% annually to an extent that the treatment of one person costs US$ 65,000 today.⁵ Hence, some mechanism of limiting an orphan patent holding company’s profits, once the R&D costs are paid off, is desirable. Nonetheless, it can be hoped that the available treatment for orphan diseases improves significantly in the future. Today, only about 300 of over 5000 orphan diseases are treatable. (Sharma, Jacob, Tandon, & Kumar, 2010)

[...]

¹ About.com: http://pharma.about.com/od/BigPharma/tp/Big-Pharmas-Merger-Mania.htm

² HBM Partners (2012), “HBM Pharma/Biotech Report 2012”, http://hbmpartners.com/en/about-hbm-partners/report-hbm-partners.php

³ Cf. Schenk (1996), Fauli-Oller (2000), Dietrich & Schenk (1995)

⁴ Lifestyle drugs treat non-life threatening and non-painful diseases such as baldness or erectile dysfunction.

⁵ New York Times, http://opinionator.blogs.nytimes.com/2013/06/30/the-orphan-jackpot/, July 2013