1. Introduction
Developing new drugs in the pharmaceutical industry requires astronomical investment and a multidisciplinary team of experts for periods exceeding a decade. In spite of these long-term and high-cost investments, the risk of successful commercialization of blockbuster products is limited to less than 10% [
1,
2,
3].
As the pharmaceutical industry is characterized by (1) high-levels of knowledge [
4], (2) high-costs and long-term research and development (R&D), and (3) high-risk and high-returns. The survival of the industry depends on its R&D efficiency [
5]. However, R&D efficiency in this industry is gradually declining [
6,
7] although companies have made various efforts to increase efficiency by reducing investment cost and increasing profit.
Because it is difficult for one company to sustain the cost and duration of developing a new drug, several companies share the product development stage similarly to a relay race [
8]. Over the past two decades, large international pharmaceutical companies have been increasingly involved in the transition to an open innovation R&D system from the traditionally closed R&D, through trading and collaboration with external research institutes and companies [
9,
10]. Each company has a business model that creates value to be exchanged for revenue in a specific disease field, new drug development stage, or solution [
11]. As a result, technology deals, mergers and acquisitions (M&A), collaborative networking, division of labor, and specialization are increasing in the pharmaceutical industry. Some studies confirm the activities of pharmaceutical companies of exchanging technology and information using the transaction histories of companies [
12,
13].
In the early 2000s, studies focused on increasing R&D costs [
1,
2,
14]. Specifically, efficiency measurement studies argued that (1) there is a positive correlation between stock price and individual firm’s efficiency [
15], (2) there is still inefficiency in the industry [
16], and (3) there is the need to reward technology [
17] to achieve innovation increases efficiency. After 2010, as R&D productivity decline becomes an issue in this industry, there have been studies to identify the cause [
5,
18], confirming whether M&A have increased productivity [
19,
20,
21]. Industry experts affirm that The implementation of open innovation (OI) methods in pharmaceutical R&D is no longer a philosophical problem of ‘whether to do it’ instead of a logistical problem of ‘how to do it’ [
22].
Nevertheless, studies on R&D efficiency that target pharmaceutical companies are beset by limitations as follows. (1) There has been no research on how to measure open innovation (OI) and spreading within the industry. (2) Most studies use small data samples and output variables are limited to the number of patents or academic publications rather than based on corporate financial information data. The academic results reflect the outcomes of basic research, so it is hard to see it as a company’s achievement. (3) Because of the small data size, in terms of methodology, previous studies used data envelopment analysis (DEA) [
23,
24], a nonparametric methodology, not a statistical significance method. Because of the limitations of the methodology, studies were unable to compare efficiency by groups. When the efficiency of two or more groups is analyzed, the efficiency results of the two groups cannot be compared because each group uses each different production functions with different weights. The technical efficiency, result of DEA, of more than two groups under different production functions cannot be compared. A limitation of the methodologies is that a meta-frontier analysis has emerged and can be solved by comparing the TGR (technology gap ratio) values with each other [
19]. Extant research is currently limited to understanding global pharmaceutical companies or only small- and medium-sized enterprises (SMEs) in specific countries.
In this study, we are going to measure and compare the efficiency values of pharmaceutical companies’ R&D given different OI types. We use the meta-frontier analysis (MFA)—a more advanced efficiency analysis method. We cover more than 700 U.S. pharmaceutical companies to analyze the period from 2001 to 2016. The United States, along with Europe, is considered a major axis in the pharmaceutical industry, accounting for one-third of the total market. Moreover, most biopharmaceutical companies are located in the United States.
Analyzing and comparing the production efficiency of OI type pharmaceutical company groups serves as a reference material for companies formulating strategies. This is because OI activities reflect the management strategies of individual pharmaceutical companies.
We identified OI activities through transaction information data that show a close relevance between corporate R&D activities and OI activities. Our research thus enables readers to understand whether statistically significant data are being developed and if OI strategies are being used to achieve efficiency. We thus show how the efficiency of pharmaceutical companies by OI type changes over time flow, from the time OI was introduced to the present. Identifying each type of OI would thus help readers understand the bigger picture from a broader perspective.
5. Results
The probabilistic SFA model is effective in comparing the efficiencies of companies with in a group. A company’s TE cannot be used to compare the efficiencies between groups with different technical characteristics. However, this value can overcome these limitations if we calculate the additional decomposition. We thus propose a more transparent analysis of the technology gap of different groups, as well as their efficiency level. The ratio of averaged technical differences, random error, and TE provides us with an additional explanation compared to the analysis based only on the stochastic frontier function of the different groups. The TGR plays an important role in explaining the ability of a group of companies to compete with other companies in different groups within the industry. This ratio estimates the technical gap between the entire group and the industry as a whole.
As previously mentioned, we categorized pharmaceutical companies operating in the United States into four groups based on OI strategy. We evaluated the production function of each company from the SFA operation, and then calculated the meta-frontier production function.
Table 4 and
Table 5 show the SFA and MFA calculation results, respectively. First, the first column in each of the four groups provides the value of the production functions (TE). The two columns on the right display the TGR derived from the MFA operation. The estimated results show the highest average TE value (0.587) for the closed group (see
Table 5), followed by the coupled (0.519); the outside-in (0.487); and the inside-out group (that is, TE: coupled > closed > outside-in > inside-out). The TGR, which compares the efficiency of a group by measuring the distance between the discovered meta-frontier and the group frontier production function, has the opposite result (that is, TGR: inside-out > coupled > outside-in > closed).
The question is, when we consider the industry as a whole, why does the group with the highest TE value record a relatively low TGR? Conversely, why does the inside-out group, which has significantly lower TE values, record higher TGR values?
If a company has significantly higher TE value in a group, the TE value of other companies are relatively low. Subsequently, the resulting average TE value of this group would be lower. In this case, companies that have achieved exceptionally high performance may have achieved innovation as well. Conversely, leveled groups will generally register small differences between TE values. In this case, most companies would achieve relatively high TE values.
The TE values are calculated only within the group. Therefore, the comparison between groups in the TE graph is meaningless. However, we use the graphs to observe time-series changes more easily. The TE and TGR values are both 1 when the efficiency shows perfect efficiency, and the relative comparison of the productivity of each company shows similar values. The difference is that the TGR compares one group at a time, which allows us to see the bigger picture.
The number of units represent statistical significance even when divided by year and group. Hence, we consider the changes by year (
Figure 2) in the average values of TE and TGR. TE values in four graphs, each group has a large efficiency change year after year, which also drops sharply. However, TGR values, generally, all groups operate more reliably in higher numerical ranges.
The inside-out group (A) stays at has the lowest score in the TE, but stably shows a high and stable figure in the TGR graph. It is the closest group to the meta-frontier line because it keeps the highest TGR score compared to the other three groups. As previously mentioned, a small number of innovative companies may raise efficiency standards, which may mean that the average efficiency within a group is significantly different. The analysis shows that the group is constantly innovating or pursuing high-risk, high-return. As mentioned above, this group with a low average TE value and a high average TGR value. If we were to analyze only the average value of the TE is analyzed, this group can would be misunderstood as a group one with low production efficiency. The reason for this is because difference is that the value of the frontier is set at a high standard, with a small number of outstanding companies. Although the majority of the companies have achieved satisfactory results, they can be considered to for as have low technical efficiency TEs because the frontier value is too high.
In outside-in group (B) in 2001, the highest TE value among the four groups showed a high performance, but then plummeted, and since 2011, it has continued to decline year by year. The TGR value is a weak decline, but it has risen in 2006–2008 and in 2011–2013. The change in TE value is assumed to be due to deterioration of management conditions, innovation and the spread of innovation in this group. During both periods of 2001–2003 and 2012–2013, the TE value declined and the TGR value rose. In the previous period, the productivity crisis was a concern due to a decline in the number of FDA-approved medicinal substances. In the following period, profits from large-scale patent expiration deteriorated. On the other hand, in both periods, some companies started to sell biopharmaceuticals and achieved differential growth [
6].
The coupled group (C) has the most noticeable feature—it has the highest average value in TE. The TE graph shows a steep improvement in the index between 2007 and 2011, but declines without change after peaking. In the TGR graph, this group steadily falls from the highest value to the lowest in recent years. TE prices rose until 2009 and then turned downward. Improvements such as cost reduction have been made continuously, but they have not achieved innovative results. Or it could have achieved high efficiency by succeeding in innovation in other groups in the industry. The TGR value is steadily declining, which means that this group is moving away from the highest efficiency levels in the industry. The continuing decline in TE and TGR since 2009 suggests that this group's revenue mode has deteriorated or that its business environment has changed. The net income is gradually falling, causes could be inferred that the barriers to entry were lowered or the competition among companies became more intense.
The coupled group includes outsourcing and solution providers at a high rate. While the TE values are high, the TGR is continuously declining. Although it is highly efficient in terms of revenue versus investment, it is gradual in the U.S. pharmaceutical industry as a whole, but continues to be inefficient. This can be seen as equalization among firms, or less dynamism and innovation. The exact cause needs to be understood through the transaction information or a company interview.
Closed group (D) as opposed to inside-out groups, are close to the average of TE and TGR and will rise or fall in the same direction until 2009. During this period, it seems that no innovation has taken place within this group. However, after 2009, TE values have fallen and TGR values are moving upside down. This seems to be the result of some companies achieving innovation. There were several blockbuster patents expired at this time, and sales of generic drugs increased.
Table 6 presents the estimation results. This shows the production function estimates for all groups.
Figure 3 shows the TGR values of companies for each year as translucent dots. This illustration helps us understand deviation, density, and so on, which are otherwise difficult to understand when simply observing the average value. The dark area is where many TGR values are concentrated, as the TGR value either records the median score or the point, at which both the high score and low score appear together. For the early 2000s, the TGR values are clustered in the high score range. However, over time, they gradually expanded, which is common to all groups.
The TGR values of each individual company are verified to clearly understand the characteristics of each group without falling into the error of the average value (see
Figure 3). In all groups, we find the TGR values near the meta-frontier line to be close to 1. However, the graph shows different aspects for each group. First, the inside-out group has a significantly low TGR value of 0.5, while most companies are in the 0.7 to 0.9 range. The distribution within this group has widened in the past five years, but there are a considerably larger number of companies for the frontier line than in the other groups. In the outside-in group, the pharmaceutical giant that recorded an earnings boost in 2012 and 2014 hit the frontier line, but recorded a variety of distributions. In the closed groups with the largest number of companies, we find the highest annual values, but enterprise TGR values are widely distributed from 0.4 to 0.95. As the distribution spreads more widely each year, the overall average declines. The closed groups are gradually divided into upper and lower groups after 2005. The top-tier group grew more toward the frontier line over the past two years, with better performance. The subgroups show a TGR below 0.5.
Companies with the highest average TGR values from the inside-out group (
Table 4 and
Table 5) are mostly well-established companies cutting-edge medical technology. Many of them have advanced medical technologies that emerged as new alternatives beyond traditional therapeutic approaches, such as cell therapy, regenerative therapy using stem cells, and cancer therapy using immune cells. These technologies are based on biotechnology, and have been recently subject to experimental treatment.
Pathfinder Cell Therapy, which has the highest mean TGR value, has developed a novel cell-based therapy to treat diabetes, renal disease, myocardial infarction, and other diseases. Other companies are either those that specialize in chemical products and materials technology, or in information technology (IT). The most efficient companies in the outside-in group are those that exclusively supply products in specific areas with specific expertise. The products sold by this group include cryogenic cooling technology, computerized solutions, knee joint implants, ophthalmic medical devices, and medical devices that measure and transmit patient’s heart rate defibrillation. These companies have mature technologies that can already be commercialized. Moreover, they often created goods or services based on their technologies, which are also better known to the public than those made by the inside-out group. Such goods and services have proven safety and effectiveness, and sometimes bypass existing technologies. Analyzing the homepages of these companies shows they offer the most exclusive technology in the industry and they are wary of competitors that sell similar products.
It was difficult to find information that targets a specific disease in the introduction information of the coupled group of companies. These companies have technologies or solutions that could be applied to a variety of fields, rather than products based on specific diseases. There are numerous types of technology-oriented companies that involve operations such as facility manufacturing, platform construction, gene diagnosis, critical care emergency treatment, computerized tomography system, injection and inhalation technology, and technologies that optimize solutions according to disease types. These companies emphasize their unique technologies and consider them as concrete solutions for other companies. This indicates that such companies may operate in the Business-to-Business (B2B) environment, which targets other professionals. However, the terminology used makes them inaccessible to ordinary consumers.
6. Discussion
This study used financial information from 2001 to 2016 to determine how effective have U.S. pharmaceutical companies been. Based on negotiation data from the same period, companies were designated an “OI type” and grouped accordingly. We analyzed as many U.S. companies involved in the pharmaceutical industry as possible.
According to the analysis, the efficiency of the outside-in group, which comprises many large pharmaceutical companies, has been steadily declining. In fact, researchers [
5,
7,
19,
20] have been concerned with the ongoing crisis in this industry and the lack of efficiency improvement, which have yet to be solved.
Our empirical data provide evidence to show that efforts have been made to determine change in this group. The data show that large-scale investments have been made, corporate efficiency has fluctuated, and that the success or failure of the investment performance has resulted in a decline in the efficiency figures for each individual company. Next, the efficiency of the inside-out group was consistently high, with increased TGR value. The companies in this group were small in scale, with short histories. Among the companies with high efficiency scores are those that achieved success in biotechnology. These companies performed well in technologies that are expected to change the future medical paradigm, such as cell therapy, stem cells, genetic information analysis, and data analysis. Based on this technology, they can either attract investments or choose to license, that is, sell their solutions to other pharmaceutical companies. This is in line with the view that pharmaceutical R&D takes place within the enterprise but is also procured from outside, such is the case for small businesses and public institutions [
55,
56,
57]. Small biotech companies and biopharmaceutical companies maximize the success rates of their research by focusing on highly specialized technologies or specific diseases. These companies tend to pursue exit strategies to either sell their technologies or be acquired by large companies. This is because the later steps in technology or drug development are not financially feasible to them, owing to the high costs of clinical trials and commercialization. Similarly, in an environment of rapidly changing technology flows, large companies do not find it feasible to conduct research internally. Consequently, finding promising new drug candidates with the intent to buy them saves time and lowers the probability of failure. [
58,
59]. Effective R&D management is thus crucial to a company’s ongoing survival because it increases R&D efficiency, innovation, and financial performance. An accurate understanding of R&D projects contributes to a company’s strategic decision-making for sustainable management based on empirical analysis and evaluation [
60].
The OI activities of U.S. pharmaceutical companies confirm the mechanisms that help improve their economic efficiencies. Although previous studies reported the economic benefits of OI [
30], we note that OI influences business activity through a variety of factors besides financial variables. We previously defined various variables that do not translate into amounts. For example, saving time through OI is possible for both the inside-out and outside-in groups. Incorporating an innovation process, and, then taking full advantage of an innovation, can save time. Conversely, using the innovation process externally can hasten direct merchandising and profitability, resulting in the creation of direct cash flows. One innovation process, combined with various ideas, can bring products and services to the market, and, thus increase the value of R&D. Consequently, various OI activities can increase R&D efficiency in many ways [
61].
Finally, there are many variables that are difficult to quantify, such as experts’ abilities, contracting power, corporate culture, collective intelligence, accumulated learning, absorption capability, individual benefit, and intrinsic motivations [
62,
63,
64,
65,
66,
67]. Organizational learning capabilities, collective intelligence, and corporate culture also have a significant effect on the success and failure of OI activities [
68]. However, companies need appropriate research personnel and capabilities as they require extensive internal knowledge to adopt the knowledge and skills created by external organizations. Despite these reasons, the economic benefits of OI activities arise from ambiguous standards and measurement difficulties. Financial variables are often used when measuring R&D efficiency, although other variables could be used for this purpose. Nevertheless, the calculations are generally based on the ratio of expected revenue to R&D cost.
6.1. Implications
R&D efficiency in the pharmaceutical industry is a sensitive issue because R&D projects require many resources, and the success and failure of these projects determine the survival of a company. We confirmed that the prevalent concerns in this industry do not necessarily determine its downfall, but indicate a change of technology and its structure. We also examined the groups that have recently improved their efficiencies. In short, the industry has been working more collaboratively and innovatively since the 2010s. We thus identified frontier companies that have achieved higher standards each year through diverse efforts in a bid to respond to change dynamically.
Dahlander and Gann (2010) [
69] argued that openness and efficiency are inversely correlated. Thus, each company should make strategic decisions based on a balance of costs and benefits. Openness and cost increase together because of the increased external communication and complexity of internal processes. Exploring new opportunities would thus require increased costs for learning and communication, and the risk of corporate secrets being leaked. The complexity of such processes further increases when information and technology are traded with different external organizations during collaboration. For example, a company with a high degree of openness may rely on open revenue in a joint project, lower the barriers to entering the business, increase competition, and eventually lower its profits even at high growth rates. Businesses can benefit from negotiations when they have exclusive core competencies [
70]. This can be understood as risk management. Particularly, pharmaceutical R&D requires risk management because of the large size of R&D expenditures. Therefore, an effective R&D risk management enhances corporate sustainability [
60]. Additionally, OI is a positive and flexible substitute for environmental change or risk. Connecting and communicating using the knowledge systems of each organization should be based on mutual respect, responsibility, and trust, and a community with these virtues can develop robust, resilient, and agile management responsiveness to change [
71].
Our study showed that the efficiency of a company varies according to the OI strategy it pursues. Considering the worsening business environment post-2000s, openness and efficiency have differed according to the strategies U.S. pharmaceutical companies pursued. Each company has sought different strategies according to situation and the internal resources it holds. Nevertheless, the strategies resulting in higher efficiency in the 2000s were found to be inside-out strategies.
Conversely, the efficiency value of the outside-in group fluctuates or decreases. Consistently, Companies in this group should actively experiment with innovations to escape their current unstable orbit. They should also consider developing new business models that increase their inside-out transactions to recover their investment or generate new revenue. As such, it is also necessary to share their risk with other players because volatility is a business risk factor. It is dangerous to adhere to the strategy of pursuing blockbuster drugs in a changing environment. Thus, implementing a business model that generates stable margins, but also stable profits through generics or improved drugs, could be a viable option.
Companies can grow by abandoning past rules and accepting new ones. The traditional pharmaceutical business model, which was effective until the 1980s, is now losing its importance. One proof of this paradigm shift is the continued underperformance of R&D efficiency. This will eventually make it increasingly difficult to export new drugs to the market using previously successful strategies. Success will be elusive if companies continue to focus on old rules, prestige, and reputation. Our study confirmed that the companies that achieve high efficiency and reliability are also the ones that have accepted new disciplines, leveraged new technologies, and experimented with new business models. The performance of the outside-in group shows that it is not enough to accept external research results, but companies must accumulate core competencies as well.
However, we suggest that companies avoid doing this alone. It is possible to seize opportunities by abandoning exclusive closures such as “not invented here” and communicating with the outside world. Innovation takes place in knowledge-rich environments, characterized by intense fusion of knowledge from the outside as well as from within the company. Even in the closed group, we found that companies actively accepted new opportunities for generic drugs and made use of them. Thus, efficiency increases when companies increase profits or reduce costs. As such pharmaceutical companies can pursue high added value, but can also increase efficiency through cost savings. Further, they must seek creative and collaborative options to the costly processes of clinical trials, marketing, and sales. Companies in the coupled group that can borrow resources from firms and minimize inefficiencies can thus achieve high efficiencies. Yesterday’s competitors can be today’s partners. Essentially, allowing all possible stakeholders to participate in the idea pool could help achieve higher efficiency.
6.2. Limitations
However, our study is also beset by certain limitations. First, there was a limitation in dividing the OI group. An analysis of 16-year contracts according to the proportion of OI types was conducted. However, we could not reflect how the change in management strategies reflected the changes in our analysis. It would thus be interesting to compare the historical changes in contract trends until the present day.
Furthermore, we note that a single analytical method can provide only one viewpoint. When diverse methods of analyses are used, the scope of a study can be understood with greater precision and insight. Thus, despite the abundant interpretations and insights provided in our study, we must conduct further analyses to better understand the industry. Efficiency and inefficiency are good references for building corporate management strategies, but they cannot be absolute. The efficiency of a company is measured only by the presence of a better performing company because efficiency is inevitably a relative concept.
Finally, we analyzed efficiency by each company and fiscal year. Our focus was on how management strategy and openness affect corporate efficiency. However, the point in time of the investment and when revenue was generated do not coincide, which can cause the efficiency value to fluctuate. However, we urge scholars to analyze efficiency at the project le vel within the enterprise, as we did not have access to the information from inside the company. Therefore, in-house information could help measure efficiency by project and product more accurately.