Abstract
Some industries deviate from their behaviour during the later stages of their lifetime, compared to the behaviour predicted by the life cycle theory. This deviating behaviour can be caused by changes in the innovative activities of the industry’s firms. In order to test the robustness of the life cycle theory, a closer look is taken at an industry in the later stages of its lifetime. The industry chosen in this research is the athletic shoe industry and this industry will be analysed from 2003 to 2013. The patents that were issued in this time period to the different firms in this industry are used to gain insight about the innovativeness of the industry. Various analytical tests were conducted to test for significant changes in the innovativeness over time, the value of the patents and the presence of life cycle transformations over time. These results are then to give the industry a Schumpeterian classification. Differences between the predicted Schumpeterian characteristics of the industry and those found in this research will provide a deeper understanding of the industry’s innovative evolution. These findings are then combined to test the fitting of the industry’s current characteristics to that of the maturity stage as described in the life cycle theories.
Table of contents
1. Introduction ... 3
2. Theory ... 4
3. The athletic shoe industry ... 8
4. Methodology ... 9 5. Results ... 15 6. Discussion ... 18 7. Limitations ... 21 8. Conclusion ... 23 References ... 25 Appendix ... 27 2
1. Introduction
Do all industries move through similar evolutionary stages? Or do some industries follow an unpredictable path? According to research done by Klepper (1997) based on entry, exit, firm survival, innovation and firm structure in new industries, the product life cycle is an accurate description of the evolving pattern of most industries. He does note however, that the product life cycle model is only able to accurately describe the earlier stages of the industry. When some industries age and reach their later stages they tend to deviate from their predicted behaviour. In order to take a closer look at this phenomenon this paper will research one of the characteristics of a relatively older industry.
The industry to be studied in this paper is the athletic shoe industry. This industry has existed for a long time and is not expected to disappear any time soon. The product itself is used by all kinds of people across the globe and its usage ranges from leisurely walks to competitive sporting events. Whereas the casual athlete is mostly interested in the fit and look of the shoe, the competitive athlete prioritizes the performance of the product. This broad range of product usage has resulted in a product that is continuously being improved upon and specialized in order to better meet the needs of the different users. The market itself is still growing and becoming increasingly profitable for the major firms in the industry. Examples of this increase in market profitability can be found by looking at the major players such as Nike and Reebok. They have reported an increase in their earnings over 2013 of 9% (Nike, 2013a) and 5% (Adidas, 2013) respectively and are still expecting growth, especially in developing countries such as China. Whilst both companies are now multi-product companies that do not solely focus on footwear anymore, the athletic shoe is what started their entrepreneurial journey and remains one of their core products.
Most firms in the industry actively openly talk about their desire to continuously increase their product’s performance in order to enhance the customer’s experience and performance. A prime example of a firm with such an innovative drive is Nike. Even though they are the firm with the largest market share they are still heavily investing in innovation. Their research and development (R&D) team has access to laboratories filled with equipment such as motion capture devices and environmental chambers. In these labs scientists cooperate with top-tier athletes in order to not only design new products but also test and evaluate their existing products and prototypes (Nike, 2014a). Whereas the larger firms in the industry tend to focus more on incremental innovation and less on radical innovation, there have been a considerable number of (smaller) firms that successfully created their own market niche. Anatomic Research Inc. is one of these smaller, niche-focussed firms. By focussing on creating shoe soles that prevent ankle injuries they have secured their share of the industry and have even managed to partner up with Adidas (Ellis, 2013). This interesting combination of large, multi-product firms and smaller, niche-targeting firms, the incentive for both firm types to stay innovative and the age of the industry is what makes the athletic shoe industry a fitting industry for this research.
This thesis will begin with a discussion of theories and models used, and the research hypotheses to be tested. Following these parts is a brief overview of the most notable events in the athletic shoe industry that happened to date. After this overview the research methods will be explained, followed by the research results. The results from the tests and their limitations will be discussed in order to answer the research hypotheses and draw the final conclusions.
2. Theory
Usage of patents has enabled inventors and firms to protect their intellectual properties from imitators. Without these property rights on patents, technological knowledge would become public property. This would enable competitors to easily and quickly imitate products without penalty. By using patents the inventor gains a, temporary, monopoly position in the market with his innovation. This monopoly position provides encouragement for private inventors to come up with new ideas, which in turn favours technological advance and economic growth. Aside from enabling the inventor to protect its property, patents have a number of different functions. The patent-system stimulates innovation by spreading the invention information across the industry. This way patents can be used by people other than the inventor to gain insight into the progress of technological knowledge. By spreading this information, the patent-system also reduces the amount of needless duplication of the R&D activities in the industry. The time and costs saved by this reduction enables firms to focus on new products and is believed accelerate the technological process of the industry as a whole. The publication volume of patents enables insight in the R&D activity of firms in the industry and it has been shown that there is a strong, significant relationship between R&D expenditures and patent counts. This is to be expected since a patent application logically flows from a successful R&D activity (DeBackere et al. 2002). These characteristics make patents one of the best indicators of the innovativeness of an industry. Or, as stated by Grupp in 1998: “No other innovation indicator can be traced back over comparatively long periods of time, may at the same time be disaggregated at a very low level allocable to individual economic units, and is also precise and accurate insofar as identification of the timing of the innovation event is concerned.”
Even though most researchers agree that patents are one of the best methods for analysing innovativeness, there is still a considerable amount of disagreement about the valuation method of patents. While most of the economic literature use the amount of forward citations, the length of the patent renewal or the number of countries where the patent has been granted as measurements for the value of the patent; it has been shown that not all of these variables have a linear relationship with the patent value and that there are also other variables such as the patenting strategy that influence the patent value (Guellec & Potterie, 2000). Further research into the usefulness of patent citations as a measure of their importance showed that when the patents granted to a firm receive, on average, one additional forward citation their market value will increase by 3%. There is also a difference in market value between internal forward citations and external forward citations. Internal forward citations
seem to have a far greater impact on the market value than the external forward citations. Reasons for the extra value gained from internal forward citations could be the internalizing of the knowledge spill-overs or it could be an indicator of the R&D strength of the firm compared to the other firms in the industry. However, this positive effect of the internal forward citations on the market value of the patent seems to decrease with the size of the firm’s patent portfolio. This positive effect is strong for firms that have a relatively small or average sized patent portfolio (<200 patents) but for larger firms with a larger portfolio (1000 patents) this effect becomes negligible and it can even have a negative impact for firms with an even larger patent portfolio. This reduced effect size can be explained by chances of the firm citing one of its own patents increasing with the size of their overall patent portfolio (Hall, Jaffe, & Trajtenberg, 2005).
When analysing the innovativeness of industries a clear distinction must be made between the different levels of industry analysis. In the empirical and theoretical articles about the evolution of industries, three different analysis levels have been used. The analysis levels have been characterized as: the specific dimensions of industry dynamics, the structural dynamics or the structural evolution of the industry. Empirical research undertaken at the industry dynamics level has uncovered some of its defining characteristics. It has shown that every firm in an industry has its own unique bundle of capabilities, strategies, structure and performance. These lead to differences in costs and productivity, profitability, output and innovative activities. The differences between firms that make them unique seem to be persistent over time. Therefore if a firm is particularly innovative at time t it is reasonable to expect the firm to be highly innovative at time t+1 as well. This relation also applies to the profitability of firms. Firms that have been highly profitable in the past will be more likely to also be more profitable in the future. These findings lead to the first stylized fact of industrial dynamics: There exists an ongoing and notable diversity within and across industrial sectors when comparing firm behaviour, organizations and performances (Malerba & Orsenigo, 1996a).
Due to entry, exit and changes in market share every industry has its ever changing list of participants. Whereas most entrants are new small firms, there are also larger firms looking to enter the industry through usage of a diversification strategy. Research has shown that a large percentage of these new entrants exit the industry within a few years and their survival chance grows with their age and initial size. Surviving firms also appear to have either a higher initial size or higher growth rate compared to failing firms. These findings result in the second stylized fact of industrial dynamics: There is high amount of turbulence in the industry (Malerba & Orsenigo, 1996a).
Industry specific characteristics such as R&D intensity and the intensity of advertising appear to be correlated with the persistence of higher profits. The amount of turbulence also seems to differ drastically across sectors. It tends to be negatively correlated with the rate of innovation, the amount of advertising and its capital intensity. Concentration of firms and their growth seem to be positively correlated with the amount of turbulence in the sector. These sector specific variables appear to have a
significant effect on the firms’ growth and, in turn, its survival rate. This shows that: Every sector has its own specific characteristics (Malerba & Orsenigo, 1996a).
Categorizing industries based on analysis at the industrial dynamics level is done based on the patterns of organization of innovative activities and features of technological change. In his 1984 paper “Sectoral patterns of technical change: towards a taxonomy and a theory” Pavitt divides the industries in four categories. There are science-based industries, in which R&D and learning by searching are the main elements in the innovative activities. The second type of industry is the scale-intensive industry, in which economies of scale and productive features play the leading role in competition between firms. In the specialized supplier industries, firms create products for a specific task. In this type of industry interaction with the (end-) users is an integrative part of the innovative process. The last type of industry is the supplier R&D dominated industry. Here innovative activities are small and based on incremental improvements and the integration of new technologies.
The second analysis level, structural dynamics, refers to the industry dynamics over time such as entry, exit, firms’ size and concentration, product- and process innovation. When analysing an industry at this level most, if not all, researchers apply the industry life cycle model. This model is the only stylization of the dynamics of industry structure currently supported with empirical evidence. It consists of four stages that every industry seems to move through. In the first stage of the industry life cycle, called the introduction stage, fundamental scientific and technological problems of the new technology have to be solved. This results in a time of radical innovations where the number of patent applications is still low but slowly increasing. Due to the high R&D risk only a few pioneering firms will apply for patents which leads to a high number of patents applications per firm, also known as a high patent concentration ratio. Towards the end of this introduction stage the development of patent applications stagnates or declines. Possible explanations for this are: innovative products are still too expensive, the acceptance of the customer is not high enough, the range of application possibilities is not clear yet or the absence of the dominant design.
In the next stage, called the growth stage, the basic technological and market uncertainties have disappeared, and the development of the market applications of the technology starts. Due to the innovations becoming less radical in nature, the R&D risk decreases. This leads to an increase in the number of patent applications. The increasing number of new competitors results in a smaller patent concentration ratio. Following the growth stage is the maturity stage. In this stage the number of patent applications remains constant and consists mostly of incremental innovations. And finally, in the decline stage, when the potential for new product innovations on the basis of the technology decreases the number of annual patent applications decreases as well (Haupt et al. 2007).
When categorizing the innovative stages of an industry most researchers use patents as an innovativeness indicator. Several of the empirical studies have shown that the number of patent applications in an industry follows an S-shaped or even a double S-shaped growth path. This growth path is also known as a Sigmoid or S-curve (Anderson, 1998; Haupt et al. et al. 2007). These growth
paths show the cyclical motion of the industries’ innovativeness across the different industry life cycle stages. While most industries follow this S-shaped innovation curve, the type of innovative activities and distribution of innovative activities between firms in the industry differs per industry. Joseph Schumpeter was one of the first researchers to address this difference and has proposed two different categories to classify these innovative patterns. The first innovative pattern for an industry is dubbed the Schumpeter Mark I. Proposed in his 1934 article “The theory of economic development”, this pattern is best described as ‘creative destruction’. The Mark I pattern is described as an industry with technological ease of entry and, a major role played by entrepreneurs and new firms in innovative activities. These new entrants add innovative ideas and products to the continuously growing innovative base of the industry. Examples of Schumpeter Mark I industries are toys, agriculture, industrial automation, shoes and clothing (Malerba & Orsenigo, 1996b). The other type of pattern, dubbed the Schumpeter Mark II, was proposed in: “Capitalism, socialism and democracy” (Schumpeter, 1942). Described by ‘creative accumulation’ this pattern is characterized by the prevalence of large established firms and the presence of relevant barriers to entry for innovators. These large firms dominate the market and are innovative though their own accumulation of knowledge and capabilities. Industries that fit the Schumpeter Mark II category are for example the gas, aircraft and laser industries (Malerba & Orsenigo, 1996b).
Whereas the Schumpeter Mark type of analysis seems to put a very static label on the industry, Klepper (1997) showed that during the evolution of an industry, changes may occur in these Schumpeterian patterns of innovation. When analysing industries and comparing the Schumpeterian innovation patterns with the industry life cycle, Klepper noted that the Schumpeter Mark I patterns may transform into Schumpeter Mark II patterns. When looking at the early stages of the industry life cycle model, when technology is changing very rapidly, uncertainty is high and barriers to entry are still very low, new firms are the major innovators. During the life time of the industry, technological changes start to become less radical and follow well-defined trajectories. This is where economies of scale, learning curves, barriers to entry and financial resources become increasingly important to maintain a competitive advantage in the industry. This results in the industries larger firms to become the leading innovators, resulting in an industry that fits the Schumpeter Mark II pattern. Transitioning from a Schumpeter Mark II to a Schumpeter Mark II industry is also possible. During major technological and market discontinuities, a stable industry characterized by incumbents with monopolistic power is taken over by a more turbulent industry with new entrants that are using the new technology or focussing on the new demand (Klepper, 1997; Malerba & Orsenigo, 1997).
The final level of analysis used is the structural evolution level. Using this level of analysis shows a broader view of the industrial structure and its evolution over time. It looks at the emergence of new industries, the generation and transformation of technologies and products within the industry, the changing boundaries of the firms, the vertical integration, diversification, development of networks
between firms and the role of the institutions in the industry. Malerba & Orsenigo (1996a) showed that common industrial patterns found at this analysis level are:
• Technological discontinuities that are followed by periods of incremental technological change, which are followed by new discontinuities.
• Entry to the industry tends to occur in specific evolutionary periods and does not exclusively happen after technological discontinuities.
• Following a process of industrial concentration and competence enhancing a technical change is a period of competence destroying technological change.
• The vertical integration, diversification and specialization processes differ in their forms and intensity over the different stages of industry evolution.
• A major role is to be played by the governmental policies and public institutions in the industry.
• A great number of institutional and organizational differences across countries.
By analysing these patterns and other taxonomies of the clothing and shoe industry, Malerba & Orsenigo (1997) showed that this industry is best defined as a Schumpeter Mark I industry. The athletic shoe, however, might differ from his categorisation of the entire clothing and shoes industry. Athletic shoes are a very specialized product and innovation of the product requires a considerable amount of R&D dedication from the firm. These intense R&D requirements increase the entry barriers for new firm. These high barriers to entry and the majority of the innovations done by the incumbent firms may result in the athletic shoe industry being a Schumpeter Mark II industry. In order to test the fitting of the Schumpeter model and the industry life cycle, the innovativeness of the athletic shoe industry over the last 10 years (1/1/2003-31/12/2013) will be researched. This leads to our main research question: “How can the last 10 years of the athletic shoe industry be classified as a Schumpeterian industry?” In addition to the Schumpeterian characteristics tests, three research hypotheses will be used to answer the main research question:
• H1: The number of issued patents in the athletic shoe industry is stable during the last 10 years.
• H2: The impact of the innovations patented in the last 10 years of the athletic shoe industry is the same for the innovations made by the established firms and those made by the independent innovators.
• H3: The last 10 years of the athletic shoe industry fit the characteristics of the maturity stage of the industry life cycle model.
3. The athletic shoe industry
Archaeologists have found shoes that date back as far as 10.000 years, making them one of the earliest inventions made by mankind (Connolly, n.d.). Not much attention was given to the design of the
earliest shoes, since they were mostly seen as tools to enable exploration of the harsher parts of the world. During the ancient Olympic Games the first race sandals were created. In contrast to the barefooted Greek athletes, the competing nations gave their athlete sandals to increase traction. The Romans then applied this knowledge to the sandals of their military and kept investigating ways to increase the performance of man through sandals (Werd & Knight, 2010, pp. 3-5).
The first shoes solely produced for athletic purposes, were created in 19th century Britain to aid the cricket players. It featured leather housing with spikes under the sole to increase the player’s grip. Whereas most athletic shoes were sold as a by-product of the local shoemaker, the first firm dedicated solely to athletic shoes was J.W. Foster & Sons, later known as Reebok. Observing the earlier created cricket shoe and the increasing amount of runners, J.W. Foster saw a gap in the market he wanted to fill. And so, in 1890, he invented the first spiked running shoe (Reebok, 2014a). This market entry was followed by Spalding in the late 19th century and Adidas in 1925 (Adidas, 2013).The founding of Blue Ribbon Sports, later known as Nike, in 1964 created a new surge of innovation in the industry. Their usage of rubble-waffle soles and breathable nylon was highly innovative at the time and secured their position in the market (Nike 2013b). Nike has since then taken the athletic apparel industry by storm and remains one of the biggest players in the industry to date. The dominance of Nike has been cited as one of the main reasons for the acquisition of Reebok by Adidas for $3.8 billion in 2005 (NY Times, 2005).
Whilst most of these firms started out small and produced only shoes they have grown into multi-product, multinational firms that are producing everything related to sports and are still working hard to maintain their innovative status. Reebok is heavily investing in the new Crossfit hype with an increasingly amount of products specifically made (and branded) for the Crossfit activities. With the support of the Reebok franchise this spin-off brand has even spawned its own fitness games where athletes compete for the right to be called “The fittest in the world” (Reebok, 2014b). Whereas Nike is convincing the soccer players of the world to buy their newest Magista soccer shoes that should improve their ball handling and fit like a glove (Nike, 2014b).
Even though the industry is dominated by multibillion dollar companies, there is still room for independent innovators and private contractors. A prime example of a smaller, but highly innovative firm in the athletic shoe industry is Anatomic Research. Focussing on a specific niche, ankle injuries, they are still in the industry after nearly 30 years of service (Ellis, 2013).
4. Methodology
In order to test the hypotheses, the patent data found in the database of the United States Patent and Trademark Office will be used. This database holds the information of every patent issued from 1790 till the present day. Using this database to find the earliest patents that were assigned to the athletic shoe industry yielded a trio of patents from 1977. Observing the amount of patents issued every year after this will give a general overview of the innovativeness of this industry. The resulting graph may
show patterns that relate to the different stages of the industry life cycle model. This graph can then be used to, intuitively, show the current stage of the industry.
The remaining research requires a more focussed lens. Firstly, only the patents that are issued in the last ten years (01/01/2003 – 31/12/2013) are used for testing. This will enable the research to be up to date while still maintaining a reasonable scope of patents to test the hypotheses on. Secondly, this research will exclude patents that are only related to product design. While the firms in the clothing and apparel industry do place a heavy importance on design in order to make their products more recognizable and appealable, design in itself is not the type of product innovation that is relevant for this research. Using these two parameters (ccl/36/114 and isd/1/1/2003->31/12/2013) to search for patents in the United States Patent and Trademark Office yields 115 patents. The data of these selected patents and their coding scheme is available upon request.
By analysing the data of these 115 patents extracted from the database of the United States Patent and Trademark Office, the firms will be ranked based on their amount of issued patents. This ranking will show the distribution of the innovativeness of the firms in the industry and identifies the major innovative firms. A comparison of patents issued to firms and patents issued to independent innovators in the same year will show the distribution of innovative activities between the established firms and the independent innovators in the industry.
Measuring the impact of the patents will be done by using their internal and external number of forward citations. The study done by Hall et al. (2005) showed that the market value of a patent increases, on average, by 3% for every forward citation it received and that this market value premium is increased to 10% if the forward citation was made by the same firm that was assigned the patent (internal forward citation). Since the patents that are older have had more time to accumulate these forward citations, this research will then divide this value by the difference in years between the patent issue date and the amount of years in which they could accumulate citations. The patents that are used for this analysis have issue dates between 01/01/2003 and 31/12/2013, therefore the denominator in the equation becomes: (2014 – the issue year of the patent). Any forwardly cited patents that have been issued after 31/12/2013 are not taken into account for this analysis. This results in the patent value being calculated as:
𝑃𝑃𝑃𝑃 = 1.03 ∗ 𝐸𝐸𝐸𝐸𝐸𝐸 + 1.1 ∗ 𝐼𝐼𝐸𝐸𝐸𝐸2014 − 𝐼𝐼𝐼𝐼
The issued patents will then be divided over the three groups of innovators in the industry; the established firms, the fringe firms and the independent innovators. Usage of a One-Way ANOVA Scheffé-test will then test for significant differences in the patent value between these groups.
Equation 1
PV is the patent value, EFC represents the number of external forward citations, IFC represents the number of internal forward citations and ID being the issued year of the patent.
To test the current life cycle stage and the changes in innovativeness in the industry this research will use a selection of the patent indices tested by Haupt et al. (2007). In their research, Haupt et al. stated that in order to form a general idea of the innovativeness of an industry, only the patent data of the two main innovative companies need to be analysed. This means that, in this research, the two companies that have issued the most patents in the last 10 years will be used for the analysis. Haupt et al. tested seven different patent indices, but the patent data used in this research only enables usage of four of these indices. Therefore this research will use forward patent citations, backward patent citations, literature citations and the duration of the examination process as the patent indices. The mean value of these indices will be calculated and a One-Way ANOVA Scheffé-test will be used to test for significant differences in the indices between the different life cycle stages. The changes in these patent indices should show one of eight patterns (fig 1). These change patterns are indicators of the life cycle transitions and they are described by Haupt et al. (2007) as:
1. An increase in the mean value of the indicator when the technology goes from the introduction to the growth stage. And a subsequent drop in the mean value of the indicator when the technology moves from the growth to the maturity stage.
2. A decrease in the mean value of the indicator when the technology goes from the introduction to the growth stage. And a subsequent rise in the mean value of the indicator when the technology moves from the growth to the maturity stage.
3. A stagnation or non-significant change of the mean value of the indicator when transitioning from the introduction to the growth stage followed by an increase of the mean value when the technology moves from the growth to the maturity stage.
4. A stagnation or non-significant change of the mean value of the indicator when transitioning from the introduction to the growth stage followed by a decrease of the mean value when the technology moves from the growth to the maturity stage.
5. An increase in the mean value of the indicator when moving from the introduction to the growth stage. This is then followed by a stagnation or non-significant change when moving from the growth to the maturity stage.
6. A decrease in the mean value of the indicator when moving from the introduction to the growth stage. This is then followed by a stagnation or non-significant change when moving from the growth to the maturity stage.
7. An increase of the mean value of the indicator across all three stages. 8. A decrease of the mean value of the indicator across all three stages.
By testing a number of different indicators in his research, Haupt et al. (2007) successfully showed which patent indicator fits which graph. When using forwards citations as an indicator, Haupt et al. (2007) found that the mean value drops significantly when comparing the introduction stage to the growth stage, and dropping even further in the maturity stage. They also argued that this was to be expected considering the nature of this indicator. Older patents have a higher change of being forwardly cited compared to newer patents based on the cumulative character of the index, this results in a graph that shows a type 7 pattern. The number of backwards citations should increase significantly in each stage. Intuitively this makes sense as well, since an older industry has a larger amount of patents to build new research upon compared to a younger industry. This means that, when using backwards citations as an indicator a type 7 pattern would be expected. The number of literature citations will differ greatly between the introduction and the growth stage; it will not have a significant difference between the growth and the maturity stage. This is due to the fact that in the later stages of the industry most information will be found in other patents and not published literature, resulting in a type 5 pattern. The final index is the duration of the examination process. This is the only indicator
Figure 1
The eight different patterns of life cycle stage transitions (Haupt et al., 2007)
found by Haupt et al. (2007) that shows significant differences between both stages since the duration of the examination process is longer in the introduction and growth stages than in the maturity stage. Therefore the examination process index results in a type 2 pattern.
Classifying the industry as a Schumpeter Mark I or Mark II industry is done by analysing the characteristics of the innovative activities. This research will use the same characteristics described in the research by Malerba & Orsenigo (1996b) to be able to compare the results. In their research the different industry characteristics are described as:
• The concentration and asymmetries of innovative activities between firms in the industry; • The size of the innovating firms;
• The changes over time in the hierarchy of the innovators;
• The relevance of new innovators compared to the established innovators.
The concentration and size of the firms in the industry are two characteristics that have been used in empirical research for a longer period of time whereas the changes of the hierarchy over time and the relevance of new innovators are two characteristics that have been added more recently. The first two characteristics are able to give a more general measurement of the industry, and the latter two are more focussed on the degree of stability and creative accumulation in the industry.
To test these characteristics and to be able to use previous empirical results as a comparison, the same methods will be used as those that were used by Malerba & Orsenigo (1996b). They divided the number of patents issued to the top four innovative firms by the total number of issued patents to calculate the concentration ratio of innovative activities in the industry. To test for asymmetries in the innovative activities between the firms a Herfindahl-index will be used. For this index the share of each firm’s issued patents compared to the total amount of patents issued by firms is calculated, squared and then summed as shown in equation 2. The size of the innovating firms is calculated as the share of the amount of patents issued by firms with more than 500 employees compared to the total amount of patents issued for in the industry.
� �𝑇𝑇𝐼𝐼𝑃𝑃�𝐼𝐼𝑃𝑃𝑖𝑖 2
𝑛𝑛
𝑖𝑖=0
To measure the latter two characteristics, the period of industry analysis needs to be split into two to compare differences over time. Since the data available for this research is spread out over a time period of 10 years, 2003-2013, an even split can be made. This results in the first time period being from 01/01/2003 to 30/06/2008 and the second time period being between 01/07/2008 and 31/12/2013. To measure stability of the industry, or the changes over time in the hierarchy of the innovators, a Spearman-rank correlation test is used on these two time periods. Testing the relevance
Equation 2
IP is the number of issued patents, TIP is the total number of issued patents, i is the number of the firm
of the new innovators requires a base line to compare the new entrants too. To provide this baseline firms that are innovating in the first time period are classified as established firms. The relevance of new innovators will then be measured by the share of patents issued by firms that apply for the first time in the second stage over the total number of patents issued in the second stage. After analysing the characteristics of the industry it will be classified as a Schumpeter Mark I industry if there is low amount of concentration and asymmetry in the innovative activities, if there is low amount of stability in the ranking of the innovators, if there is a high amount of entry of new innovators and the majority of the innovating firms have a small size. However if the analyses show that there is a high amount of concentration and asymmetry in innovative activities, a high amount of stability of the ranking of innovators, a low amount of entry by new innovators and a large size of the innovators then the industry will be classified as a Schumpeter Mark II industry.
5. Results
The amount of patents issued per year in this industry is shown in figure 2. Looking at this bar chart it becomes clear that the amount of innovation fluctuates over the years. The low amount of patents issued in the first five years is a typical characteristic of the introduction stage in the life cycle model.
In 1982 the amount of issued patents quadruples from 3 to 12 and, with the exception of a decrease in 1984, keeps increasing until 1986. This increase in issued patents makes the time period between 1982 and 1986 identifiable as the growth stage of the industry. The years following 1986 show a stagnation of issued patents until 1991. The amount of patents issued is almost halved when it drops from 21 in 1990 to 13 in 1991. This event shows the first of the S-curves in the industry. After this drop the innovativeness rises again to previous levels until falling in 1996. This is then followed by a short increase in innovativeness and, in turn, followed by a decrease of the issued patents in 2003. The slowly but continuously decreasing amount of issued patents show that this time period can intuitively be classified as the maturity stage or even the start of the decline stage of the life cycle model. The last 10 years of the industry however show a new resurgence of the innovativeness of the industry. Following 2003, patents were issued in an increasing amount with a small drop in 2009. The amount of patents issued in the last three years show levels comparable to the all-time high of the earlier years of the industry. This can be caused by a radical new product or the entrance of a new and highly innovative firm.
In order to answer this question a more focussed lens is required for the analysis. By analysing the patent data over the last 10 years, a ranking of the firms based on their number of patents issued is 0 5 10 15 20 25 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 Year
Number of patents issued per year
Figure 2
created. Firms that have issued only 1 or 2 patents are considered to be fringe firms and are grouped as such. All patents issued to innovators without an assignee firm are grouped as being independent innovators. The number of patents issued by the non-fringe firms in this industry can be found in table A1 of the appendix. The number of patents issued per year and their distribution
between these groups is shown in figure 3.
This data shows that, when it comes to number of patents issued per firm, Nike is the biggest innovator in the industry. Nike has issued more than three times the amount of the second-most innovative firm, Anatomic Research, and there seems to be an upward trend in their innovativeness over the years. Interestingly, the next 5 most innovative firms have issued 25 patents in total, whereas the firms outside this top 6 have issued a combined total of 30 patents and the independent innovators combined have issued 32 patents. This results in the total amount of issued patents in the industry being almost evenly divided between Nike, the other firms in the athletic shoe industry, fringe firms and the independent innovators.
This shows that the innovations made by independent innovators make up a substantial amount of the total innovations, but are their innovations just as valuable as those made by the established firms? In order to answer this question the issued patents were valuated based on their internal and external forward citation count as shown in equation 1. For this analysis the innovators have been divided into three groups. They were classified as an established firm if they have issued more than 2 patents in the time period (type 1), were classified as a fringe firm if they issued one or two patents (type 2) and have been classified as an independent innovator if there was no assignee firm mentioned on the patent data (type 3). The results of the One-Way ANOVA Scheffé-test can be found in table A4 of the appendix. The test revealed that there were no significant differences in patent value in the patents issued to either firm type (p > 0.05). However, the patent values of the established firms and the independent innovators did show a significant difference (p = 0.005). It seems that patents issued to an established firm have an, on average, higher value compared to those of the
Figure 3 0 5 10 15 20 25 Year
Number of patents issued per year
(2003-2013)
Fringe Reebok Athletic Propulsion Labs Adidas Ringstar Anatomic Research Nike Independent 16independent innovators. The patent values of the fringe firms compared to the independent innovators showed no significant differences as well (p > 0.05).
Based on the volume of the issued patents shown in figure 2 and figure 3, there might have been a life cycle transition in the last 10 years of the industry. Intuitively three different patterns can be distinguished. First the amount of patents issued between 2003 and 2005 appears to rise, then from 2005 until 2009 the amount appears to drop and finally, between 2010 and 2013, the amount seems to rise even faster than before. To test if any of these stages are significantly different from each other and therefore indicate a life cycle transition, four of the patent indices as proposed by Haupt et al. (2007) are used in a One-Way ANOVA Scheffé-test with the two most innovative firms of the industry. For the athletic shoe industry these two firms are Nike, with 28 patents and Anatomic Research, with 9 patents. Of these firms the number of forward patent citations, backward patent citations, literature citations and duration of the examination process per year can be found in table A2 of the appendix. The results of the One-Way ANOVA Scheffé-test between these three stages are shown in table A3 of the appendix.
No significant differences (p > 0.05) have been found in the number of issued patents, average forward citations or the duration of the examination process when comparing two adjacent stages to each other (i.e. stage 1 & stage 2 or stage 2 & stage 3). The amount of average backwards citations did show a significant drop when the industry moved from stage 1 to stage 2 (p = 0.03), but not when the industry moved from stage 2 to stage 3. There was also a significant drop in the average literature citations when moving from stage 1 to stage 2 (p = 0.002), but there is no significant difference between stage 2 and stage 3.
The first step to classify the athletic shoe industry as a Schumpeter Mark I or Mark II industry requires an analysis of the concentration of innovative activities in the top 4 firms of the industry. The four firms in the athletic shoe industry with the most patents issued are: Nike, Anatomic Research, Ringstar and Adidas. Together they have issued 47 of the 115 patents. This leads to a concentration ratio of 0.41. By using the Herfindahl-index to calculate the amount of asymmetry between the innovative activities of the firms in the industry, an index value of 0.14 is found. In order to test share of innovations done by larger firms the firms were grouped based on their number of employees. Firms that have more than 500 employees are considered to be large firms and firms with less than 500 employees are considered to be small firms. Division of the amount of patents issued to large firms with the total amount of patents issued resulted in a size-ratio of 0.65.
In order to measure changes over time in the hierarchy of the innovators, the firms are ranked based on their amount of patents issued in two time periods (01/01/2003 – 31/06/2008 and 01/07/2009 – 31/12/2013). Of these two rankings a Spearman’s rho is calculated. This resulted a Spearman’s rho of -0.402 (p > 0.031) as shown in table A5 of the appendix. The final step in the classification of the industry requires a measurement of the relevance of new innovators compared to the established innovators. To compare the innovative relevance of the firms over time, the time period is split into
two (01/01/2003 – 31/06/2008 and 01/07/2009 – 31/12/2013). Firms that have innovated in the first time period are classified as being established firms and firms that innovate for the first time in the second time period are classified as new innovators. Applying this categorisation to the data results in 15 out of the 29 firms being established firms and the remaining 14 being new innovators. These new innovators have issued 22 out of the 66 total issued patents of the second time period resulting in an entry-ratio of 0.33.
Based on these innovative characteristics Malerba & Orsenigo (1996b, 1997) have classified 49 different technological sectors as either a Schumpeter Mark I or a Schumpeter Mark II industry. Their interpretation of the results from the five different analyses is found in table 1. Applying these interpretations of Malerba &
Orsenigo to the results from the athletic shoe industry enables the classification of the industry. The concentration ratio of the top 4 innovative firms is 0.41, which would indicate a high amount of concentration in the industry, although only barely so. A Herfindahl-index of 0.14 was found, this would suggest that there is a large amount of asymmetry between the innovative activities of the firms in the athletic shoe industry. The industry’s size-ratio of 0.65 shows that most innovative activities in this industry are completed by larger sized firms (more than 500 employees). Since the analysis of the changes over time in the hierarchy of the innovators via a Spearman-rank correlation test yielded a Spearman’s rho of -0,402, there seems to be a large amount of instability in the rankings of the innovative firms. The entry-ratio of the industry was found to be 0.33, which shows a low amount of new firms engaging in innovative activities.
6. Discussion
Combining all the results and data enables the three research hypotheses and the research question to be answered. First the total innovativeness of the most innovative companies in the industry was examined. Based on the total amount of patents issued in the selected time period of 10 years, the two most innovative firms where Nike and Anatomic Research. The patents issued by these two firms where combined to check for any changes in the innovativeness of the industry over time. Data used in this analysis included: the number of forward patent citations, the number of backward patent citations, the number of literature citations in the issued patents and the average duration of the patent examination process. The analysis showed that there were no significant differences in the number of issued patents, average forward citations or the duration of the examination process when comparing two adjacent stages to each other. There were significant drops in the number of backwards citations and the number of literature citations found when comparing stage 1 and stage 2. The drop in
Table 1
Interpretation of the findings by Malerba & Orsenigo (1996b, 1997)
Low High
Top 4 concentration ratio x<0,4 x>0,4 Herfindahl-index x<0,1 x>0,1
Size-ratio x<0,5 x>0,5
Spearman-rank -0,2<x<0,2 x<-0,2 or x>0,2
Entry-ratio x<0,4 x>0,4
backwards citation was unexpected, since the research done by Haupt et al. (2007) showed that the number of backwards citations should show a significant increase over time, due to the growth of the patent base over time.
This anomaly can be explained by taking a closer look at the data of the individual firms. The patents issued by Anatomic Research where all issued before 2007 whereas Nike had a significant share of its patents issued after 2007 (75%). The patents of Anatomic Research have, on average, 172 backwards citations whereas the patents issued to Nike average 33 backwards citations. Therefore this drop in average number of citations over time can be explained by the difference in patenting style of the firms. The difference in literature citations can also be contributed to differences between the firms. The patents issued to Nike cite, on average, less than 1 scientific article whereas the patents issued to Anatomic cite, on average, 9 scientific articles. This difference may be explained by their innovative focus. Whilst both firms are innovating in the same industry, they are focussing their innovative efforts on different products. Most of Nike’s patents were issued for an innovation of a complete shoe whereas the patents concerning the inventions of Anatomic Research where issued for shoe soles to prevent ankle injuries. The larger amount of literature citations is in line with the patenting of a radically new innovation, which fits the background of Anatomic Research. They were the firm to pioneer a footwear sole design based on the barefoot running style and its natural biomechanical properties. Whereas the innovations made by Nike can be classified as incremental innovations of an older product, the athletic shoe. Since both firms appear to have a different product focus within the athletic shoe industry and their products appear to be in different technology life cycles, the total innovativeness of the industry is best measured by the total amount of patents issued. Since there are no significant differences in the total amount of issued patents when comparing the different stages support for H1 is found.
When looking at the total amounts of patent issued over the years in the athletic shoe industry, a near even division in patents between the innovative leader (Nike), firms that have issued more than two patents, firms that have issued only one or two patents and the combined issued patents of the independent innovators was found. This showed that, when comparing patent volumes, there is a considerable amount of patents issued to independent innovators in this industry. However, one cannot conclude the impact of their innovative efforts based on patent volumes alone. Therefore all the issued patents were weighed based on their amount of internal and external forward citations. Their value was then adjusted for their age since older patents have a higher probability of being cited compared to the newer patents. The patent assignee’s where then divided into three groups. The first group consisted of innovating firms that had issued more than two patents in the time period, named the established firms. The firms in the second group had issued only one or two patents in the same time period and where named the fringe firms. The final group consisted of patents that where assigned to a person instead of a firm and thus named the independent innovators group. The One-Way ANOVA Scheffé-test checked for significant differences between the patent values of these groups. There was
only one significant difference found, the value of the patent between the established firms and the independent innovators. It seems that the patent value of the established firms is significantly higher when compared the patent value of the independent innovators. This could also be an effect of the portfolio size of the established firms compared to those of independent innovators or the fact that multiple innovative people in the same firm are working in closely related projects. Previous empirical research did recognize the diminishing values of a new patent compared to the existing patent portfolio, but this effect has not taken into account in this analysis since the chosen patent database did not allow the collection of such data. Based on these results H2 has been rejected since the impact of innovations made by the established firms is significantly greater than the impact of the innovations made by the independent innovators.
In order to classify the current life cycle stage of the athletic shoe industry, a broad overview of the total innovativeness of the industry was given. Although this overview only used the total amounts of patents issued every year as an innovative indicator, it did enable intuitive identification of the different stages and S-curves. The first stage of the life cycle model, the introduction stage, is characterized by a low amount of issued patents. The low patent count is due to the numerous fundamental scientific and technological problems of a new technology that still have to be solved. By applying this characteristic to the patent data from the athletic shoe industry, the years before 1982 seem to be the best fit for the introduction stage. After the introduction stage the industry moves into the growth stage. Most of the innovations done in this stage are incremental innovations compared to the radical innovations in the introduction stage. Since the basic technology problems have been solved and most of the market uncertainties have disappeared the industry is becoming more attractive for new innovators. These events lead to an explosive increase in the number of patents. In the patent data of the athletic shoe industry an explosive growth in patents is seen starting in 1982 and, with the exception of 1984, continuing until 1986, resulting in this time period being the growth stage. Until 1995 the issued amount of patents seems to not differentiate that much between the years. A period where the number of patents remains constant is described as the maturity stage of the industry. The time periods following 1995 show a sequential increase and decrease in the amount of patents issued. These sequential humps can be seen as the many S-curves of the industry. At the start of each of these curves a new radical innovation has been made. This product has then been incrementally improved and implemented until the firms no longer recognized any innovation opportunities of the product. The first S-curve innovation pattern in the industry that is recognizable in the last 10 years of the industry follows, mostly, out of the R&D effort of Anatomic Research. In the first four years of the industry they were the innovative leader. By focussing their effort on a market niche, ankle injuries, the firm went through all three S-curve stages until their final issued patent in 2006. While number of patents issued by Anatomic Research kept declining, a new innovative industry leader was found in Nike. They have issued small amounts of patents in the earlier years of the industry and have seen a
significant increase in their issued patents from 2011 onwards. This can be seen as the start of a new S-curve pattern.
Combining these two innovative firms in order to test for any significant changes in innovativeness over the time period yielded no significant results in three out of the five patent indices used. There were differences found in the number of backwards citations and the number of literature citations when comparing the first two stages of this 10 year time period. Both dropped a significant amount but this difference can mostly be contributed to the difference in patents between Anatomic Research and Nike, since these firms exchange their position as the innovative leader between the two analysis periods. When taking these firm specific patent details and the other three patent indices into account, it is shown that the innovativeness of the athletic shoe industry has remained constant over the last 10 years. These years can therefore be classified as the same industry life cycle stage, and all the evidence points to it being the maturity stage of the life cycle model, supporting H3. This shows that the athletic shoe industry, in terms of innovativeness, does not deviate from the characteristics predicted by the life cycle model and reinforces the correctness of this analysis method.
The five characteristics of the Schumpeterian innovation patterns where tested and showed that a large amount of the innovative activities are undertaken by the top 4 innovative companies, that there is a large amount of asymmetry between the innovative activities of the firms in the industry and that the majority of the innovation is done by large firms. Results also indicated a high amount of instability in the innovative ranking of the firms and a small amount of innovative activities undertaken by newly entering firms. The high amount of innovative concentration combined with the large amount of asymmetry, the large share of the innovative activities being completed by the large firms and the low amount of new firms being innovative are characteristic of a Schumpeter Mark II industry. Contrary to this, the large amount of instability in the rankings of the innovative firms is a characteristic of a Schumpeter Mark I industry. However, the measurements of the concentration ratio, the asymmetry and the size of new entrants are all relatively close to their respective division points. This could suggest that when using a different or larger time frame of the industry these results could be reversed. These results would suggest that the athletic shoe industry is an industry that combines the characteristics of the apparel and clothing industry, a Schumpeter Mark I industry, with those of the more technologically advanced industries. However, when taking the previous results into account and considering that four out of the five industry defining characteristics point towards the Schumpeter Mark II class, we can conclude that the last 10 years of the athletic shoe industry should be classified as a Schumpeter Mark II industry. This is in contrast with the Schumpeter Mark I classification often given to the clothing and apparel industries.
7. Limitations
The patents used in this study where based on their issue date. Using the issue date of patent instead of the application date may result in a skewed analysis. Since there is a significant amount of time
between the application date of the patent and the patent issue date, using the patent issue dates results in lag in the innovativeness. The amount of days between the application and the issue date is also not constant across all the patents, it varied between 407 days and 3757 days. Using the patents applied for instead of the issued patents would have resolved this issue. The database used for this research however, did not support cataloguing the patents based on their application date. Another notable difference between the patents applied for and the patents issued is the fact that while all issued patents have been applied patents in the past, not all patents that are applied for are issued. This could result in there being a significantly greater amount of applied patents compared to those used in this research. Scholars looking to further delve into this industry, or any other comparable type of research, have to take these differences between patent types into account and would be advised to sort the patents based on their application date. It was also assumed that every firm a separate entity and every independent innovator is not related to a firm. This may not be the case as some firms may outsource R&D activities and these innovations will then be assigned to the other firm. Firms in the industry may also be working closely together or firms could even be part of the same group. Uncovering these relationships between the firms in the industry may lead to a more complete image of the innovativeness of the industry.
Using only the last 10 years of the athletic shoe industry for the more extensive analyses restricts the generalizability of the research. There may very well have been events prior to this time period that are more influential for the evolution of the industry. By using the most recent period of events the number of forward citations arguably loses some of its viability as an innovative indicator. Patents that are issued later in the later stages of the industry have a smaller chance of being used for further research due to their age. Therefore it is hard to indicate the impact or usefulness of these newer patents when using the number of forward citations as an indicator. While this problem can never be fully resolved if the analysis of the industry includes the most recent period, this negative side of the indicator can be partially negated by testing the differences in the indicator over longer time periods instead of the relatively small time periods tested in this research.
Another problem that arises from the usage of this shorter time frame became apparent when having to classify the firms as either established firms or new entrants. This was done in the same way that previous empirical research handled this problem, by splitting the total available time period into two and classifying firms that innovated in the firms time period as an established firm. However, the two time periods in previous research consisted of at least 10 years and were often even longer. The short, 5-year time frames used in this research may have resulted in a flawed categorization of the firms. There may have been firms that have been issued patents before 2003 and after 2009, and therefore should have been categorised as an established firm. This limitation is only resolvable by looking at the entire industry’s life instead of only considering a period out of the industry’s life for the analysis. Only then can the firms be correctly classified based on their first issued patent.
However, the question remains if the firm is innovative the moment it has applied for a patent that is related to the industry or if the firm is innovative the moment its patent has been issued and approved.
8. Conclusion
This study tried to create a better understanding of the changes in innovativeness of an industry that has reached the maturity stages of the industry life cycle model since previous research has shown that not all industries follow the same predicted patterns. The industry chosen for this research was the athletic shoe industry due to its age and continuously evolving product. In order to create this deeper understanding of the innovative changes and to see if the industry life cycle model properly predicts the events in this industry various analytical tests were used on patents issued in the last 10 years of the industry (2003-2013).
Insight in the total innovativeness of the industry was gained by analysing changes in the patent data over time. In addition to the total amount of patent issued per year, the patents were compared on their average number of forward patent citations, their average number of backward patent citations, their average number of literature citations and their average duration of the patent examination process. There were only significant changes over time in the number of backwards citations and the number of literature citations. However, a closer look at these patent indices revealed that these differences are a result of the patenting style between the two firms that have issued the most patents in the industry. Therefore it has been shown that the innovativeness of the industry as a whole is constant over this time period.
The impact of the innovations made by the established firms and the independent innovators were compared based on their amount of internal and external citations. The One-Way ANOVA Scheffé-test revealed that there was a significant difference between the patent values of these groups. The results from this test show that patents issued to an established firm have, on average, a higher impact than those issued to an independent innovator.
The different analyses combined with the graphing of the patent volume showed that, in the time period chosen for this research, the total amount of innovativeness in the industry remains constant. Combining this result with the stages identified in the industry overview, shows that the innovativeness of the athletic shoe industry between 01/01/2003 and 31/12/2013 fits the characteristics of the industry life cycle’s maturity stage. This finding further solidifies the robustness of the life cycle theory and shows that the life cycle theory is able to accurately predict the evolution of the athletic shoe industry.
In order to classify the innovative patterns in the athletic shoe industry the Schumpeterian industry classifications were used. This required testing of the concentration and asymmetries of the innovative activities between firms in the industry, the size of the innovating firms, the changes over time in the hierarchy of the innovators and the relevance of new innovators compared to the established innovators. The results from these tests showed that the athletic shoe industry has a high
amount of innovative concentration combined with the large amount of asymmetry, a large share of the innovative activities being completed by the larger firms and a small amount of new firms being innovative. These are fitting with the definition of a Schumpeter Mark II industry. Contrary to this, a large amount of instability in the rankings of the innovative firms was found, which is a characteristic of a Schumpeter Mark I industry. This shows that the innovative activities in the athletic shoe industry are a combination both Schumpeterian innovation patterns but the industry is best classified as a Schumpeter Mark II industry.
References
Adidas (2013). adidas Group - History. Retrieved May 23, 2014, from http://www.adidas-group.com/en/group/history/
Adidas (2014). Nine months 2013 results. Retrieved April 17, 2014, from http://www.adidas-group.com/media/filer_public/2013/11/07/adidasgroup_2013q3results-en.pdf
Andersen, B. (1999). The hunt for S -shaped growth paths in technological innovation: a patent study. Journal of Evolutionary Economics, 9(4), 487-526.
Connolly, T. (n.d.). Fort Rock Sandals. Retrieved May 23, 2014, from http://pages.uoregon.edu/connolly/FRsandals.htm
Debackere, K., Luwel, M., & Zimmermann, E. (2002). Measuring progress and evolution in science and technology ? II: The multiple uses of technometric indicators. International Journal of Management Reviews, 4(3), 213-231.
Ellis, F. (2013). Anatomic Research Inc. | Barefoot Shoe Sole Design. Retrieved May 21, 2014, from http://anatomicresearch.com/
Grupp, H. (1998). Foundations of the economics of innovation: Theory, measurement, and practice. Cheltenham, England: E. Elgar.
Guellec, D., & Potterie, B. V. (2000). Applications, grants and the value of patent. Economics Letters, 69, 109–114.
Hall, B. H., Jaffe, A., & Trajtenberg, M. (2005). Market Value and Patent Citations. The RAND Journal of Economics, 36(1), 16-38.
Haupt, R., Kloyer, M., & Lange, M. (2007). Patent indicators for the technology life cycle development. Research Policy, 36(3), 387-398.
Klepper, S. (1997). Industry Life Cycles. Industrial and corporate change, 6(1), 145-182. Malerba, F., & Orsenigo, L. (1996a). The dynamics and evolution of industries. Industrial and
Corporate Change, 5(1), 51-87.
Malerba, F., & Orsenigo, L. (1996b). Schumpeterian patterns of innovation are technology-specific. Research Policy, 25(3), 451-478.
Malerba, F., & Orsenigo, L. (1997). Technological regimes and sectoral patterns of innovative activities. Industrial and corporate change, 6(1), 83-118.
NIKE, Inc. (2013a). Nike Sport Research Lab Incubates Innovation. Retrieved May 17, 2014, from
http://nikeinc.com/news/nike-sport-research-lab-incubates-innovation
NIKE, Inc. (2013b). History & Heritage. Retrieved May 23, 2014, from http://nikeinc.com/pages/history-heritage
NIKE, Inc (2014a). NIKE, Inc. Reports FY2013 Q4 and Full Year Results. Retrieved April 17, 2014, from http://nikeinc.com/news/nike-inc-reports-fy2013-q4-and-full-year-results
NIKE, Inc (2014b). Nike Magista Football Boot. Retrieved from www.nike.com/nl/en_gb/c/football/magista
NY Times (2005, August 3). Adidas Agrees to Acquire Reebok in $3.8 Billion Deal - New York Times. Retrieved May 23, 2014, from
http://www.nytimes.com/2005/08/03/business/03cnd-shoe.html?_r=0
Pavitt, K. (1984). Sectoral patterns of technical change: Towards a taxonomy and a theory. Research Policy, 13(6), 343-373.
Reebok (2014a). Reebok History - adidas Group. Retrieved May 23, 2014, from http://careers.adidas-group.com/reebok/brand-history.aspx
Reebok (2014b). CrossFit Games | The Fittest on Earth. Retrieved May 23, 2014, from
http://games.crossfit.com/
Werd, M. B., & Knight, E. L. (2010). Evolution of Athletic Footwear. In Athletic footwear and orthoses in sports medicine (pp. 3-5). New York: Springer.
Appendix
Year Patents total Independent Nike Anatomic Research Ringstar Adidas Athletic Propulsion Labs Reebok Misc.
2003 6 2 0 3 0 1 0 0 0 2004 11 5 0 4 0 0 0 0 2 2005 10 2 4 1 1 1 0 0 1 2006 8 1 3 1 0 0 0 0 3 2007 11 5 1 0 0 1 0 0 4 2008 9 2 1 0 1 1 0 0 4 2009 4 0 2 0 0 0 0 0 2 2010 8 2 1 0 1 0 0 0 4 2011 16 5 7 0 1 0 0 0 3 2012 12 3 3 0 2 0 1 2 1 2013 20 5 6 0 0 0 2 1 6 Total 115 32 28 9 6 4 3 3 30 Table A1
Number of issued patents per firm and independent innovator
Table A2
Mean values of the patent indices Nike
Year Issued patents Average forward citations Average backwards citations Average literature citations Average duration of examination process
2003 0 0 0 0 0 2004 0 0 0 0 0 2005 4 38,5 65,75 1,75 980 2006 3 12,33333333 30,33333333 0 985,6666667 2007 1 16 24 0 1106 2008 1 12 41 0 1527 2009 2 3,5 28,5 0 1014 2010 1 2 31 0 1299 2011 7 1,428571429 20,28571429 0 1298 2012 3 0,333333333 32,33333333 0 2013,333333 2013 6 0,166666667 30,5 0 970,5 Anatomic Research
Year Issued patents Average forward citations Average backwards citations Average literature citations Average duration of examination process
2003 3 16,66666667 96 6,666666667 2509 2004 4 18,75 203,25 10,25 2200 2005 1 5 219 15 881 2006 1 16 227 8 1383 2007 0 0 0 0 0 2008 0 0 0 0 0 2009 0 0 0 0 0 2010 0 0 0 0 0 2011 0 0 0 0 0 2012 0 0 0 0 0 2013 0 0 0 0 0 Combined
Year Issued patents Average forward citations Average backwards citations Average literature citations Average duration of examination process
2003 3 16,66666667 96 6,666666667 2509 2004 4 18,75 203,25 10,25 2200 2005 5 31,8 96,4 4,4 960,2 2006 4 13,25 79,5 2 1085 2007 1 16 24 0 1106 2008 1 12 41 0 1527 2009 2 3,5 28,5 0 1014 2010 1 2 31 0 1299 2011 7 1,428571429 20,28571429 0 1298 2012 3 0,333333333 32,33333333 0 2013,333333 2013 6 0,166666667 30,5 0 970,5 28
Table A3
Scheffé-test results between stages
Table A5
Spearman correlation results Table A4
Scheffé-test results between firm types
