Technological footprint similarity and firm performance

Abstract

The existing literature has documented significant levels of technological similarity among firms in the same industry. However, there is no theory of how technological similarity affects firm performance. This study builds and tests a theory of performance consequences of technological similarity of a firm to its industry (“technological footprint similarity”). It argues that invention performance and market performance may be affected in different ways and that specific firm characteristics moderate these relationships. This study also argues that technological footprint similarity is different from the established concepts of technological diversification and strategic similarity. The contribution of the article centers on developing the concept of technological footprint similarity and developing and testing theory of its performance consequences in the presence of firm-specific moderators.

Keywords

financial performance firm size invention performance R&D slack technological diversification technological footprint similarity

Introduction

Is it beneficial for a firm to be similar to its rivals and/or its industry as a whole? This question has been extensively studied in the literature on strategic similarity (e.g. Deephouse, 1996, 1999; Miller et al., 2013; Norman et al., 2007). While researchers have discussed competing pressures for similarity and distinctiveness in terms of a firm’s strategic positioning (e.g. Deephouse, 1999), less is known about consequences of firms’ technological positioning. Firms in the same industry tend to resemble one another in their technological competencies (e.g. Patel and Pavitt, 1997), yet we know little about the performance effects of such resemblance. Many firms are technologically diversified, and the patterns of technological diversification tend to be similar across firms (Granstrand et al., 1997) and stable across time (Brusoni and Geuna, 2003). Firms tend to know more than they make (Patel and Pavitt, 1997, 2000), suggesting that diverse knowledge may be necessary to initiate, monitor, and control outsourced manufacturing.

In addition to diversification, a firm’s technology portfolio is also characterized by technological footprint (dis)similarity (adapting the concept in Aharonson and Schilling, 2016). This concept indicates the degree to which a firm’s technology portfolio is similar to (dissimilar from) those of other firms in the industry or the industry as a whole (minus the focal firm). If we combine technology portfolios for all firms in the industry, the result is the total technological footprint of the entire industry. Each individual firm has its own technological footprint that may overlap to a certain extent with that of the rest of the industry. The greater this overlap, the more similar a firm is to the rest of the industry in its technological footprint. This technological footprint similarity (TFS) is clearly a different concept from technological diversification. A firm may be highly diversified technologically (meaning that its technological competencies are distributed across a wide range of technology areas—see Granstrand et al., 1997), yet its technological footprint may have little resemblance to that of the rest of the industry if the firm’s competencies are distributed over sparsely populated areas of technology. Such a firm’s TFS will be low. Conversely, a firm may have relatively low technological diversification, yet its few areas of technological competence may be among the most densely populated in the industry. Such a firm will have a relatively high TFS.

TFS is related to the concept of strategic similarity which is defined as “the extent to which a firm’s strategic position resembles the strategic positions of other firms competing in its market at a particular point in time” (Deephouse, 1999: 148). The difference is that strategic similarity refers to being close to industry averages in terms of strategic investments and competitive positioning, not technological portfolios. Strategic (dis)similarity has been measured as being similar to others in terms of the structure of a bank’s loan and asset portfolio (Deephouse, 1999), the distribution of routes in the airline industry (Norman et al., 2007), or distance from industry averages on various aspects of strategy and investments (Deephouse, 1996; Miller et al., 2013). None of these measures tap into a firm’s technological competencies or their relationship with those of the rest of the industry. TFS fills this conceptual void.

While previous authors noted a significant degree of technological similarity within many industries (e.g. Patel and Pavitt, 1997), none of them developed a theory of how TFS may affect various aspects of firm performance. TFS exposes a firm to two opposing forces. On the positive side, a high TFS means that the firm participates in densely populated technology areas, which may provide ample opportunities for learning and building on the inventions of others. A technology area may be “popular” due to the commercial and/or technical attractiveness of the underlying technologies. This argument is based upon the ecological view of competition: firms doing the “right” things will tend to prosper and grow while firms doing the “wrong” things will tend to shrink and/or disappear (Teece et al., 1994). TFS may also bestow legitimacy on a firm, which is seen as a positive factor in institutional theory (Deephouse, 1996).

On the negative side, densely populated technology domains are likely to see greater competition. Participation in densely populated technology domains exposes firms to strong competitive forces when rivals vie for the necessary resources (e.g. scientists with relevant knowledge and experience) and for superior positions in the markets (e.g. because many other firms introduce products with similar characteristics). Sparsely populated technology domains may be less competitive and thereby offer a better chance of distinguishing a firm from its rivals. Since competition is often based on strategic differentiation (Kennedy, 2002; McKnight and Zietsma, 2018), firms may benefit from distinct technological portfolios.

The tradeoffs between the positive and negative consequences of being similar to others are well known (Deephouse, 1999; Miller et al., 2013; Norman et al., 2007). Firms may benefit from moderate strategic distinctiveness: being too similar or too dissimilar from the industry may be detrimental to a firm’s differentiation, legitimacy, and performance (Deephouse, 1996; Kennedy, 2002). These arguments were made in the context of strategic (dis)similarity between a firm and its industry environment. Little is known about the possible effects of technological or knowledge-based similarity between a firm and the rest of its industry, even though this similarity is well known (Patel and Pavitt, 1997). Based on the arguments above, a positive, negative, or nonlinear performance effect of TFS is possible. While many scholars have studied technological similarity among firms, even those that make different products within the same industry (e.g. Orsenigo et al., 2001; Patel and Pavitt, 1997; Pavitt, 1998), no research has examined the effect of TFS on firm performance.

This article contributes to the literature by building and testing a theory of performance consequences of TFS between a firm and the rest of its industry. Earlier studies concentrated on technological diversification (e.g. Miller, 2006), relatedness of technologies in a firm’s portfolio (e.g. Teece et al., 1994), similarities among technological footprints of firms in the same industry (Patel and Pavitt, 1997), and differences among industries in terms of their technological footprints (e.g. Patel and Pavitt, 1997, 2000). This article builds on and extends this literature. It enriches the literature by theoretically exploring the various mechanisms that may affect the relationship between TFS and firm performance. This article goes beyond looking at the degree of technological diversification (e.g. Chen et al., 2013; Granstrand et al., 1997; Miller, 2006). I view the entire technological footprint of the industry as an expression of all technologies that might be relevant in this specific industry. Earlier research argued that technological diversity may be beneficial to firm performance; I add to this literature by showing that technological diversification is insufficient to explain the pattern of a firm’s technology strategy because it ignores its relationship with the industry. I consider both a firm’s invention performance (patenting) and financial performance (market valuation). In addition, I investigate firm-level factors that may affect this relationship. I invoke absorptive capacity as the likely mechanism that enables firms to benefit from closely matching the industry’s technological footprint. I also show the moderating effects of firm size and slack resources.

Literature review

Observers have long noted that organizations tend to resemble one another. DiMaggio and Powell (1983) in their discussion of organizational isomorphism wrote: “We ask, instead, why there is such startling homogeneity of organizational forms and practices; and we seek to explain homogeneity, not variation” (p. 148). TFS is a special case of organizational isomorphism. Firms within the same industry tend to have similar patterns of technological diversification (Granstrand et al., 1997; Patel and Pavitt, 1997, 2000; Pavitt, 1998). Technological diversification may have a positive effect on firm performance because a wider knowledge base allows a firm to utilize different technologies in making the same product (e.g. Miller, 2006; Patel and Pavitt, 2000). Technological diversification may also allow firms to be at the forefront of innovation and avoid technological “core rigidities” (Leonard-Barton, 1992).

Several studies supported the idea of a positive link between technological diversification and firm performance. For example, Miller (2006) showed a positive effect of technological diversification on firms’ market values. Nesta and Saviotti (2005) reported a positive effect of the scope of a firm’s knowledge base (a measure of technological diversification) on the patent output of US pharmaceutical firms. In a later paper, Nesta and Saviotti (2006) found that integration of knowledge across technology domains was a significant predictor of a firm’s market value. Further evidence for a positive effect of technological diversification on firm performance was provided by Garcia-Vega (2006) and Lin and Chang (2015). Other researchers (Caner et al., 2018; Chen et al., 2013) found a negative effect of technological diversification on firm performance. A third group of studies showed an inverse-U effect of technological diversification on firm performance (Dindaroğlu, 2018; Leten et al., 2007).

A link between technological diversification and TFS is seen in large firms. Pavitt (1998) stated that

[t]here is low diversity in the level and mix of technological competencies amongst large firms producing similar products. What is more, the degree of technological diversity¹ is lowest in the product fields with the highest rate of technical change: computers and pharmaceuticals. (p. 440)

The largest and most successful firms in highly innovative industries possess very broad technological competencies that essentially match the overall distribution of key competencies in the industry. This idea found support in the results of Arora et al. (2009), Brusoni et al. (2005), Gambardella and Torrisi (1998), and Garcia-Vega (2006).

Thus, firms tend to have broad technological competencies which are more diverse than the products that they make. “Their patent mix by technological field depends on their principal product, and changes only slowly” (Granstrand et al., 1997: 13). Overall, Granstrand et al. (1997) argue that technological diversification is a necessary consequence of increasing product complexity that requires the knowledge of different technologies. Similar ideas were developed by Patel and Pavitt (1997), Pavitt (1998), Brusoni et al. (2001) and others. Empirical findings tend to support the idea that broad technological diversification improves a firm’s innovation performance (Brusoni et al., 2005; Brusoni and Geuna, 2003; Nesta and Saviotti, 2005, 2006). However, the existing research is silent on the following important questions: (1) What is the effect of TFS on firm performance? (2) What firm-level factors amplify or mitigate the effect of TFS on firm performance? As I argued above, the fact of TFS is well established in the literature (e.g. Patel and Pavitt, 1997) yet its performance consequences are poorly understood. Furthermore, there is little understanding what companies might benefit from TFS. While we know that large firms may be more innovative (Arora et al., 2009; Bottazzi et al., 2001), this effect may be caused by large firms’ greater commercialization capabilities, not their inventive prowess.

Theory and hypotheses

Patel and Pavitt (2000) argued that firms’ technological competencies are highly diversified, stable over time, and heavily differentiated across industries. While firms in different industries showed very different sets of technological competencies (e.g. there were sharp differences between chemical and computer companies), firms in the same industry were much more homogeneous in terms of their technological competencies and patterns of technological diversification. This similarity suggests that some combinations of technological areas may complement one another for competing in a specific industry.

Different theories may be used to make different predictions regarding the effect of TFS on invention performance. According to organizational ecology (Hannan and Freeman, 1977, 1984) and the idea of competitive isomorphism (DiMaggio and Powell, 1983), firms in the same industry face similar competitive pressures and are likely to (1) respond to strategic and technological challenges in similar ways and (2) benefit from adopting similar strategies and technologies. The presence of densely populated technology domains may signify the existence of competitive advantages accruing to firms working in these domains. Summing up this argument, the first explanatory mechanism is the superior potential of densely populated technology domains for supporting invention. This mechanism helps explain why participation in densely populated technology domains may result in higher invention rates.

The second explanatory mechanism invokes the concept of related diversification (Miller, 2006; Palich et al., 2000). The industry provides the upper limit to related technological diversification. The technological footprint of the entire industry is wider than that of any specific firm participating in this industry. Bottazzi et al. (2001) argued that firms in the “oligopolistic core” undertake the majority of “pioneering R&D” in the pharmaceutical industry. These firms are usually large and diversified, often participating in a significant portion of the industry’s technological areas. Technological diversification may be beneficial when firms closely match the pattern of technological diversification set by the entire industry. For example, Gambardella and Torrisi (1998) showed that for firms in the electronics industry, their technological diversification was much more similar than their product market diversification. The existence of common patterns of technological diversification suggests (Teece et al., 1994) that some combinations of technological areas may be more beneficial to firms than others.

Different industries are likely to differ in the related bundles of knowledge required to compete in those industries so that firms within one industry will be likely to patent in a specific range of related technology domains. For example, biopharmaceutical companies are likely to innovate in domains that deal with chemical compounds, biological substances, and treating diseases; they will be less likely to innovate in domains such as software, mechanical devices, or energy. Firms that closely match their industry’s technological footprint may enjoy economies of scope and knowledge transfer between the related technology domains that they participate in.

The third explanatory mechanism invokes the concept of absorptive capacity (Cohen and Levinthal, 1990). Firms participating in densely populated technology domains may have access to a greater body of knowledge that exists in those domains compared to sparsely populated domains. This outside knowledge may enable firms to be more inventive by building on the inventions of others. The closer the match between the firm’s technological footprint and that of the entire industry, the more outside knowledge the firm is potentially exposed to. While firms differ in their absorptive capacity, densely populated technology domains will offer greater opportunities to learn from others and apply this knowledge to a firm’s own innovation, ceteris paribus.

Therefore, TFS may benefit firms’ invention performance because it will be closely matching the industry’s “distributed competencies” (Granstrand et al., 1997), which help firms compete in this specific industry. Since firms compete within the industry’s technological ecosystem,² the industry as a whole is likely to provide a snapshot of technological diversification that may improve invention performance in this industry. This may stem from the following observations: (1) the survivor principle suggests that the collection of the existing firms is likely to be well adjusted to the competitive conditions prevailing in the industry and (2) the combined areas of technological expertise of firms competing in the same industry suggest the boundaries of useful knowledge and differentiates this industry from other industries (e.g. pharmaceuticals from computers). Therefore:

Hypothesis 1: TFS is positively associated with the firm’s invention performance.

Arguments in favor of an inverse-U-shaped effect of TFS

A high-TFS firm is likely to be a technological generalist, active in many areas of technology that are useful in this specific industry. The opposite of being a generalist is specializing in a few technological areas while not participating in most of the other areas. According to Swaminathan (1995), generalists tend to be larger while specialists are often smaller. Almost any market contains niches that are too small for large companies to serve: if there are no economies of scale, firm size stops being an advantage and may become a disadvantage. “With increasing concentration, generalists tend to compete vigorously for the center of the market, thus allowing specialists to thrive in the periphery” (Swaminathan, 1995: 663). Thus, the presence of small niches along with mainstream areas dictates peaceful coexistence of strategies of different degrees of specialization. Too much technological similarity to the industry may result in intense competition in densely populated areas while too little technological similarity may result in a lack of learning opportunities in the sparsely populated areas. Firms may have to hit a “sweet spot” of moderate TFS to have both moderate competition and opportunities for external learning.

TFS is analogous to organizational conformity, which has received ample treatment in the literature. Miller et al. (2013) argue that there are two competing approaches to organizational conformity: institutional theory with its emphasis on legitimation (Deephouse, 1996) and strategic theory with its emphasis on differentiation (e.g. McKnight and Zietsma, 2018). Both legitimation and differentiation may improve firm performance; however, legitimation is achieved by conforming to norms (i.e. greater TFS) while differentiation is achieved by lower conformity (i.e. lower TFS). The presence of opposite forces affecting outcomes of TFS is likely to result in an inverse-U effect of TFS on firm performance (Haans et al., 2016). As an example of such an effect, Deephouse (1999) found that moderate amounts of strategic similarity increased firm performance while too much similarity or too much differentiation hurt firm performance. Financial market performance is likely to be an inverse-U function of TFS because perceived benefits of conformity and differentiation affect investors’ decisions but not the invention performance of firms. Investors’ decisions affect market valuations of firms but not the invention process. Investors will likely reward moderate levels of TFS while punishing extremes (Deephouse, 1999; Norman et al., 2007). Therefore:

Hypothesis 2a: TFS has an inverse U-shaped effect on the firm’s market performance.

Combining the logic in Hypotheses 1 and 2a, we can expect the following: if (1) the effect of TFS on market-based performance is inverse-U-shaped and (2) the effect of TFS on invention performance is positive, then (3) we should expect an overall positive slope of the inverse-U curve for the market performance. We can expect the market to reward firms for higher invention performance, which (if Hypothesis 1 is correct) is achieved at higher levels of TFS. Therefore:

Hypothesis 2b: Market performance of firms with extremely low TFS is lower than market performance of firms with extremely high TFS, ceteris paribus.

The moderating effect of a firm’s size

Previous research has shown that technological diversification tends to be beneficial to firms and that larger firms are the main recipients of those benefits. The positive effect of technological diversification was theoretically argued by Granstrand et al. (1997), Bottazzi et al. (2001) and others and empirically demonstrated by Miller (2006), Garcia-Vega (2006), Quintana-García and Benavides-Velasco (2008) and others. Furthermore, benefits of technological diversification do not accrue to all firms equally: large firms benefit from technological diversification more (Bottazzi et al., 2001; Granstrand et al., 1997; Lin and Chang, 2015; Pavitt, 1998). Large firms possess the necessary resources to diversify and they also have the need to diversify because each technological niche is limited in its capacity.

One important aspect of technological diversification has escaped the attention of researchers. Does a large firm benefit from diversifying into the same areas as many other firms? Or are there benefits to a large firm diversifying into “areas less traveled?” In terms of the discussion in this article, do large firms benefit from greater TFS?

This question has not been explored theoretically or empirically. Existing research suggests ambivalent answers to this question. One school of thought concentrates on great similarity among large firms’ technological footprints (e.g. Bottazzi et al., 2001; Brusoni et al., 2001; Granstrand et al., 1997; Patel and Pavitt, 1997, 2000). The larger the firm, the more likely it is to be similar to the rest of the industry in terms of its technological footprint. This similarity can be viewed as a sign of desirability of TFS if we follow the competitive isomorphism logic (DiMaggio and Powell, 1983). A similar argument can be made based on the survivor principle which indicates the most desirable combinations of technologies (Teece et al., 1994). If large firms tend to follow similar patterns of technological diversification, there may be advantages to doing that: large firms closely matching the industry’s technological diversification pattern will have a higher invention performance. Finally, Swaminathan (1995) found that large firms tended to be successful when competing in the densely populated areas while smaller firms often found success on the market periphery.

The alternative argument is that large firms may be able to explore new frontiers and learn from newcomers occupying niche technology areas if the large firms are themselves diversified into niche areas away from the densely populated technology domains. For a large firm, patenting in sparsely populated domains presents an opportunity to exploit economies of scope by cross-utilizing their distributed competencies in multiple technological fields. The larger the firm, the more resources it will have that may be cross-utilized in different knowledge domains. Furthermore, large firms may be able to benefit from inventions in sparsely populated areas of knowledge more than small firms because large firms have ample resources and highly diverse technological portfolios, which help them reduce risk. A small firm’s failures in a sparsely populated technological area may be doomed while a large firm may withstand repeated failures waiting for a breakthrough. This higher tolerance for risk and an ability to withstand setbacks may make large firms more successful when their invention takes them off the “beaten path.” Thus, large firms may be more successful than small firms when their technological portfolios deviate from industry norms.

Thus, there are two theoretical arguments predicting different interactions of TFS and firm size. Competitive isomorphism theory and the survivor principle (ultimately based on ecological reasoning) predict more positive effects of TFS for large firms’ invention performance. Tolerance for risk and resource endowments predict less positive effects of TFS for large firms’ invention performance. Both predictions appear to be valid based on the underlying theory. Furthermore, it is difficult to argue that one of these theories is correct. Therefore, I will formulate two competing hypotheses testing these alternative theoretical predictions:

Hypothesis 3a: Larger firms’ invention performance benefits from greater TFS than smaller firms’ invention performance.

Hypothesis 3b: Larger firms’ invention performance benefits from lower TFS than smaller firms’ invention performance.

The moderating effect of slack

Slack was defined as “the difference between total resources and total necessary payments” (Cyert and March, 1963: 42). Different authors conceptualized slack in different ways. Daniel et al. (2004) suggested that there are three kinds of slack: available, recoverable, and potential slack. Available slack is conceptually close to the original definition by Cyert and March (1963): it is liquid resources that are not needed to pay off obligations of the firm. Recoverable slack is usually conceptualized as the investments that the firm can control, such as selling, general, and administrative expenses (SGA) or R&D expenses. Finally, potential slack is usually understood as absence of heavy debt in the structure of the firm’s capital because the less debt the firm has, the more it can borrow, ceteris paribus.

The most important roles of slack are protection from risk via buffering of a firm’s technical core (Bourgeois, 1981) and creation of opportunities (Kim et al., 2008). Firms that have a lot of slack face less downside risk from poor performance of their investments. Such firms also have the resources to make new investments even when their current performance is relatively poor. As a result, slack may help firms perform better in risky situations by protecting their technical core from cuts and developing this core through further investments.

The downside of available slack is the forgone investments. The more slack the firm has that is not invested in any productive activities, the more opportunities the firm may be missing. Therefore, too much available slack may have a negative effect on firm performance. While some empirical studies reported a positive effect of available slack on firm performance (e.g. Ahuja, 2000; Bromiley, 1991; Daily and Dalton, 1994; George, 2005), other studies found a negative effect of available slack on firm performance (e.g. Bergh, 1997; Palmer and Wiseman, 1999). According to the meta-analysis by Daniel et al. (2004), available slack tends to have an overall positive association with firm performance. A similar effect was observed for potential slack. The association between recoverable slack and firm performance was mixed.

If slack helps firms cope with downside risk and exploit opportunities, we should expect greater benefits from slack accruing to firms exposed to risk and/or multiple opportunities. In this study, I am interested in the performance implications of TFS. High TFS may be a safe strategy due to lower risks of being similar to the industry (Deephouse, 1996). Low TFS may expose a firm to higher risk. This higher risk stems from several factors such as (1) the perceived “maverick status” of firms with technological portfolios far removed from the industry’s mainstream, (2) lower availability of resources in sparsely populated technology domains, and (3) higher stake on each individual invention in sparsely populated technology domains. Based on these considerations, firms with low TFS may benefit from higher slack due to its risk-mitigation properties while firms with high TFS may benefit from lower slack because they can safely invest more into developing new inventions.

Hypothesis 4: The positive effect of TFS on invention performance becomes weaker or non-existent when the firm has high slack.

The moderating effect of TFS on the impact of R&D investments

As I argued above, one potential benefit of participating in densely populated technology domains is the opportunity to learn from other participants in the same domain (similar to knowledge spillovers—see Nieto and Quevedo, 2005). It is likely that multiple companies patent in densely populated technology domains. A firm with high TFS is thereby exposed to a lot of information about the activities of other industry participants that are closely related to what the focal firm is doing. However, a firm with low TFS tends to participate in sparsely populated domains with relatively little available external information.

Mere exposure to knowledge is insufficient. A firm must be able to learn from other industry participants. The theory of absorptive capacity (Cohen and Levinthal, 1990) posits that firms differ in their abilities to learn from outside sources. High absorptive capacity allows a firm to acquire and assimilate outside knowledge and transform and exploit it for internal innovation purposes (Zahra and George, 2002). Using R&D investments as a proxy for absorptive capacity has been justifiably criticized in the literature because absorptive capacity is a rich, multidimensional construct that cannot be reduced to R&D investments alone (Chang et al., 2014; Flatten et al., 2011). However, R&D investments are an important factor that affects a firm’s absorptive capacity (Cohen and Levinthal, 1990; Tsai, 2001).

Previous researchers showed that R&D investments are beneficial for invention and innovation. For example, Tsai (2001) found that R&D intensity had a positive effect on business unit innovation; furthermore, R&D intensity was even more important for business units that occupied central network positions. Network centrality exposes a business unit to a potentially wide array of knowledge, which in turn requires a greater absorptive capacity to assimilate and use this knowledge. Fosfuri and Tribo (2008) found that the amount of internal R&D investments had a positive effect on innovation and that this effect was even stronger in the presence of socialization mechanisms, once again supporting the idea that R&D investments may help a firm learn from others. Escribano et al. (2009) reported a positive effect of R&D investments on sales from new products and product/process innovations. They also found that this effect was even stronger in the presence of external knowledge flows. Lin et al. (2012) demonstrated that R&D investments were particularly important for firms that had multiple alliance ties with other companies. All these studies underscore the fact that R&D investments are vital to learning when a firm is exposed to rich outside knowledge.

The results reported above suggest a twofold mechanism by which R&D investments may affect invention performance. First, R&D investments are likely to have a positive direct effect on invention (Fosfuri and Tribo, 2008; Kostopoulos et al., 2011) by enabling a firm to transform and exploit its existing knowledge and create new knowledge as well as by creating a costly-to-imitate intangible capability for invention. Second, R&D investments are especially impactful in the presence of exposure to outside knowledge (Lin et al., 2012; Tsai, 2001) because they may help the firm acquire and assimilate outside knowledge. R&D is likely to be positively related to invention performance; this effect should be particularly strong for firms that have ample exposure to relevant outside knowledge. TFS is one such factor. Firms with high TFS that invest heavily in R&D are likely to be able to learn from this outside knowledge and apply it to generate their own inventions.

Firms with low TFS are less likely to benefit significantly from outside knowledge. The relative poverty of the available knowledge that they face in sparsely populated technology domains means that such firms must primarily rely on their own knowledge in their quest for inventions. While R&D investments will benefit such firms as well, firms with low TFS will not benefit from outside knowledge as much, and the overall effect of R&D investments on their invention performance will be lower. Therefore, the effect of R&D on invention performance will be smaller for firms with low TFS.

One of the main advantages of higher TFS is greater learning opportunities. However, as R&D investments increase, the firm’s absorptive capacity increases at a decreasing rate (Stock et al., 2001). Because of these declining increases in marginal absorptive capacity, firms with high TFS will likely experience declining marginal returns to R&D investments. Therefore, high TFS will result in a positive-sloping, inverse-U-shaped effect of R&D investments on invention performance (Haans et al., 2016).

If a firm’s TFS is low, it gains little from higher absorptive capacity. R&D for such firms is likely to be directed toward internal invention because of a limited opportunity for benefiting from external knowledge spillovers. Low-TFS firms may have to build up their R&D investments before they can reap economies of scale or scope in R&D (Cockburn and Henderson, 2001; Henderson and Cockburn, 1994; Klette, 1996). If this is true, then firms with low TFS will benefit from R&D at an increasing rate, so that their invention performance will have a positive-sloping, U-shaped relationship with R&D investments.

Thus, we can expect that the quadratic effect of R&D on invention performance will change its sign depending on a firm’s TFS. Combining the changes in the quadratic effect with the overall positive effect of R&D on invention performance, I can formulate the following moderating hypotheses:

Hypothesis 5a: R&D investments will have a positive-sloping, concave (inverse U-shaped) effect on invention performance in the presence of high TFS.

Hypothesis 5b: R&D investments will result in a positive-sloping, convex (U-shaped) effect on invention performance in the presence of low TFS.

Method

To test the hypotheses, I collected a dataset on all publicly traded US-based pharmaceutical and biotechnology companies (SIC codes 2834-2836) for which financial data were available. The choice of these companies was dictated by the following factors. Biopharm companies are by their nature innovative and routinely patent their inventions. Publicly traded companies offer a wealth of financial data which may be difficult to collect from private companies. Also, limiting the sample to one country (the United States) serves to diminish institutional differences among companies. The final sample was 327 companies and 1475 firm-year observations covering the period 2000–2009. The source of financial data was Compustat. Patent data were collected from the USPTO website.

Dependent variables

The first dependent variable in this study is invention performance. I used raw patent counts and citation-weighted patent counts to measure invention performance (Lin et al., 2012; Trajtenberg, 1990; Wang and Hagedoorn, 2014). I counted all patents in the year of application rather than granting. There may be a considerable time lag between application and granting and the length of this lag is often beyond the firm’s control. The application date is closer to the actual invention date than the date of patent granting. I did not consider applications that were not eventually granted because these indicate lack of originality of these inventions.

Values of individual patents differ greatly (Trajtenberg, 1990). Some patents sink into relative obscurity while others become foundations of new technological breakthroughs. Scholars have shown forward citations that a patent receives to be a good measure of a patent’s importance and value (Bessen, 2008; Hall et al., 2005). However, raw patent counts may also be a good measure of a firm’s inventiveness, particularly in less popular technology domains or for patents that were recently applied for and have not received many forward citations yet. I use both raw patent counts and citation-weighted patent counts for robustness purposes. I calculated the citation-weighted patents (CWPs) using the formula in Hall et al. (2005) as the total of all forward citations received by all of a firm’s patents plus the raw number of patents (to correct for the fact that some patents did not receive any forward citations)

CWPs = \sum_{i = 1}^{N} (1 + F o r w a r d c i t a t i o n s_{i})

To measure a firm’s financial market performance, I used the natural log of a firm’s end-of-year market-to-book value measured 1 year into the future. To aid with interpretation of the results, I measured all dependent variables 1 year after the independent variables.

Independent variables

TFS was measured as the correlation between the distribution of the firm’s patents across main patent classes (the firm’s patenting footprint) and that of the whole industry minus the focal firm (the patenting footprint of the overall industry). I counted all patents that the focal firm applied for (and was eventually granted) in a given year. If a patent was assigned to two or more classes, it was counted under each of those classes. Formally, I used the following formula for TFS:

TFS = \frac{F_{i} * F_{n - 1}^{'}}{\sqrt{(F_{i} * F_{i}^{'}) (F_{n - i} * F_{n - i}^{'})}}

where F_i is the patenting footprint for firm i in the current period (the vector of all patents assigned to specific patent classes); F_n−i is the patenting footprint for the entire industry without the focal firm’s patents (the vector of all patents assigned to all firms in the industry minus the focal firm in the current year). Sampson (2007) used one minus the correlation computed above as a measure of difference between technological portfolios. I am interested in a measure of proximity or similarity between a firm and the rest of its industry; thus, I used the correlation itself. I computed this number for each firm in the sample in every year that this firm was active and had data available.

When computing TFS, I only considered those technology domains (patent classes) in which the focal firm had at least one patent. If a firm has patents in just a handful of knowledge domains (patent classes), it will have no patents in most patent classes. Computing a correlation between this firm’s patent portfolio and that for the rest of the industry by considering all classes would be misleading since most patent classes are irrelevant to the focal firm (Aharonson and Schilling, 2016). I also computed TFS using the entire range of patent classes for all firms as a robustness check and found the results to be very similar (available from the author upon request).

I measured a firm’s degree of diversification using the natural log of entropy computed using the formula in Palepu (1985)

Diversification = \ln (\sum_{i = 1}^{N} (P_{i} \times l n (1 / P_{i})))

where P_i is the share of patents in class i in the total patent portfolio for the focal firm in the specific year; N is the total number of patent classes that the firm patented in.

I measured the firm’s size as the natural log of sales in a given year. Using the number of employees and total assets (with the natural log transformation) as alternative measures of firm size produced very similar results (available upon request).

The firm’s slack is a measure of resources immediately available for investment which the firm does not have to use to satisfy demands of creditors, suppliers, and so on. Some previous researchers (e.g. George, 2005) used cash as a measure of high-discretion slack. While cash is the most liquid asset that can be invested immediately in various projects, it may not be the best measure of slack as it is often understood in the literature. Cyert and March (1963) defined slack as “the disparity between the resources available to the organization and the payments required to maintain the coalition” (p. 36). Cohen et al. (1972) defined slack as “the difference between the resources of the organization and the combination of demands made on it” (p. 12). Dimick and Murray (1978) stated that slack is “[t]hose resources which an organization has acquired which are not committed to a necessary expenditure” (p. 616). These definitions of slack underscore the fact that the mere amount of cash a firm has may not be sufficient to measure slack. While cash gives a firm some freedom in the very short term, a better measure of slack must consider the firm’s current liabilities as well. I measured slack by cash plus other short-term assets minus short-term liabilities, which reflects the conceptual definition of slack by Cohen et al. (1972). To reflect differences between available, recoverable, and potential slack Daniel et al., 2004), I also used SGA intensity (selling, general, and administrative expenses divided by sales) as a measure of recoverable slack (Bergh and Lawless, 1998) and the debt-to-equity ratio as a measure of potential slack (Stuart, 1998).

I measured R&D expenditures using the natural log transformation of the R&D investments that the firm made during a specific year due to the highly skewed nature of R&D expenditures.

Dealing with endogeneity

In econometrics, endogeneity is correlation between an independent variable and the error term (Wooldridge, 2010: 54). “[T]he field of strategic management is fundamentally predicated on the idea that management’s decisions are endogenous to their expected performance outcomes” (Hamilton and Nickerson, 2003: 51). Since decisions are not randomly distributed across firms, observing true performance effects of managerial decisions may be difficult. In my sample, TFS, Diversification, and R&D investments were all endogenous in the econometric sense because they were correlated with the error term in models that did not control for endogeneity. It is also obvious that managers tend to make decisions regarding the amount and direction of R&D investments based on the firm’s present technological portfolio and expected returns from future investments. Interest lies in deviations from the expected values of those variables because those deviations are more likely to reflect decision making by the firm’s managers. In order to correct for endogeneity and capture managerial decision making better, I adopted a technique called “residual centering” (Lance, 1988; Little et al., 2006). Applied to my needs, this technique calls for regressing each endogenous variable on other variables that may affect its value and using the residual of this regression instead of the original variable. This technique allows researchers to capture the unique variance in each independent variable that is not explained by other variables in the model. Using the residuals instead of the original values resulted in elimination of endogeneity and a better glimpse at the effect of deviation from what is “expected” of the firm. I used this method to compute the residual-centered values of TFS, technological diversification, and R&D investments.

Control variables

I controlled for other factors that may affect firm performance. Acquisitions of other companies can change the firm’s invention rate due to changes in the firm’s size, acquiring additional R&D capacity, and potential diversion of resources from R&D to acquisitions (Ahuja and Katila, 2001; De Man and Duysters, 2005); therefore, I controlled for the natural log of one plus the dollar amount of acquisitions that the firm made in a given year (to include firms that made no acquisitions).

Alliances may affect a firm’s invention performance and financial performance. The source of data for this variable was SDC Platinum by Thompson-Reuters (Cui and O’Connor, 2012; Martynov, 2017; Tafti et al., 2013). I counted all alliances that the firm had entered in the 3-year period ending with the current year (years t − 2, t − 1, and t). Since the average duration of an alliance is about 3 years (Pangarkar, 2003), this 3-year moving window allowed me to measure the size of a firm’s alliance portfolio that may have affected the firm’s invention performance. Considering the highly skewed nature of alliance participation, I used the natural log transformation of the alliance portfolio sizes over the past 3 years.

Since different types of alliances may differ in their contribution to the firm’s invention performance, I counted all alliances over the past 3 years and R&D alliances only over the past 3 years as two separate measures of alliancing activity (using the natural log transformation) to be used in separate models. The results were similar whether I used the count of all alliances or R&D alliances only.

There may be time trends in inventions. In particular, the number of forward citations is likely to be smaller for more recent patents since they have been out for a shorter period of time. Therefore, I controlled for the year effects in the model.

Another important factor to control for is the presence of inertia in inventions (technological inertia). Organizational ecology theory (Amburgey et al., 1993; Hannan and Freeman, 1977, 1984) argues that a certain amount of inertia may be positive for organizations since too much change may disrupt organizational processes and worsen organizational performance. Furthermore, Patel and Pavitt, (1997, 2000) and Brusoni and Geuna (2003) noted that firm’s technological portfolios tend to be stable over time, which may reflect the positive effect of stability (inertia) on performance. Therefore, I controlled for Technological Inertia, measured as the correlation between the firm’s patenting footprints in subsequent years

Technological Inertia = \frac{F_{i t} \times F_{i, t - 1}^{'}}{\sqrt{(F_{i t} \times F_{i, t}^{'}) (F_{i, t - 1} \times F_{i, t - 1}^{'})}}

where F_it is the vector of the firm’s distribution of patents over patent classes in year t (its current technological footprint); F_{i, t−1} is the vector of the firm’s distribution of patents over patent classes in year t − 1 (its previous technological footprint). The larger this correlation, the greater the firm’s Technological Inertia.

Statistical technique

The main dependent variables (patents and citation-weighted patents) are non-negative and highly skewed. They only take integer values (count variables). There are two possible ways to analyze these data. The Poisson model requires that the mean be equal to the variance and estimates one parameter. Otherwise, the Poisson model will produce biased results. Since the variance for the patent counts and citation-weighted patent counts was much higher than the mean, the data were characterized by overdispersion. The proper model to analyze these data is the negative binomial model (namely, its panel version due to the panel nature of my data). I used the command xtnbreg, re in Stata. I used the random effects model rather than the fixed effects because the negative binomial fixed effects model is not a true fixed effects model like linear fixed effects models (Allison and Waterman, 2002). I am interested in between-subjects comparisons as well as within-subjects comparisons, and the random effects model allows researchers to combine between-subjects and within-subjects variance in the same model. I also ran the xtnbreg, fe model as a robustness check and found the result to be similar (available upon request).

The natural log of a firm’s market-to-book value (my second dependent variable) was distributed close to the normal distribution. This fact justified using panel regression with fixed effects for each firm (command xtreg, fe in Stata).

Results

Tables 1 and 2 show the descriptive statistics and correlations for the sample.

Table 1.

Descriptive statistics.

Variable	Mean	SD	Min	Max
Patents	3.711	19.411	0	431
Citation-weighted patents	46.806	280.3	0	6603
Log (sales)	2.331	2.338	−0.348	11.216
Log (Market-to-book)	1.046	1.213	−4.761	10.959
TFS	−0.006	0.219	−0.742	0.481
Log (entropy)	0	0.57	−1.27	1.841
Log (acquisitions)	0.285	1.133	−4.711	10.672
Log (R&D)	−0.213	1.522	−3.193	5.672
Log (alliances)	0.555	0.773	0	4.111
Log (R&D alliances)	0.363	0.628	0	3.526
Slack	213.491	1450.405	−6076.633	36071
SGA intensity	10.908	107.442	−20.603	3094.455
Debt-to-equity	−0.265	19.967	−1642.857	195.069
Inertia	0.644	0.317	−0.015	1

SD: standard deviation; TFS: technological footprint similarity; R&D: research and development; SGA: selling, general, and administrative expenses.

Table 2.

Matrix of correlations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)
(1) Patents	1.000
(2) Cit.-weighted patents	0.675	1.000
(3) Log (sales)	0.564	0.411	1.000
(4) Log (Market-to-book)	0.070	0.145	−0.034	1.000
(5) TFS	0.127	0.089	0.152	0.011	1.000
(6) Log (entropy)	0.501	0.394	0.438	0.102	−0.110	1.000
(7) Log (acquisitions)	0.310	0.222	0.560	−0.123	0.073	0.228	1.000
(8) Log (R&D)	0.564	0.410	0.839	0.077	0.339	0.488	0.526	1.000
(9) Log (alliances)	0.487	0.414	0.642	0.049	0.168	0.345	0.447	0.663	1.000
(10) Log (R&D alliances)	0.500	0.418	0.634	0.047	0.215	0.410	0.464	0.676	0.905	1.000
(11) Slack	0.429	0.217	0.617	−0.033	0.150	0.276	0.507	0.607	0.537	0.553	1.000
(12) SGA intensity	−0.030	−0.027	−0.134	0.163	−0.022	−0.013	−0.035	−0.071	−0.061	−0.055	−0.026	1.000
(13) Debt-to-equity	0.022	0.022	0.000	0.074	−0.019	0.123	0.001	0.023	0.064	0.090	−0.012	−0.005	1.000
(14) Inertia	0.295	0.234	0.254	0.102	0.302	0.445	0.135	0.337	0.195	0.264	0.206	0.015	0.083

TFS: technological footprint similarity; R&D: research and development; SGA: selling, general, and administrative expenses.

Table 3 shows the results of panel negative binomial estimation of raw patent counts in year t + 1 while all dependent variables were measured in year t.

Table 3.

Results of random effects, panel negative binomial regression. Dependent variable: total number of patent applications that were eventually granted for each company-year in year t + 1.

Variables	Panel negative binomial regression with random effects.Dependent variable: Patent applications in year t + 1 that were eventually granted).Standard errors are in parentheses.
	Main effect	Quadratic effect	Size interaction	R&D interaction	Slack interaction	SGA interaction	Debt-to-equity interaction
Year and firm effects	included	included	included	included	included	included	included
Log (acquisitions)	−0.020**	−0.019**	−0.017**	−0.007	−0.014*	−0.024***	−0.028***
	(0.008)	(0.008)	(0.008)	(0.00956)	(0.00811)	(0.00850)	(0.00771)
Log (sales)	0.0917***	0.092***	0.097***	0.093***	0.094***	0.077	0.087***
	(0.015)	(0.016)	(0.015)	(0.016)	(0.0157)	(0.048)	(0.016)
Log (alliances)	0.092***	0.092***	0.100***	0.095***	0.106***	0.109***	0.088***
	(0.031)	(0.031)	(0.031)	(0.031)	(0.030)	(0.042)	(0.031)
Inertia	0.122	0.123	0.113	0.124	0.109	0.158	0.109
	(0.101)	(0.101)	(0.101)	(0.101)	(0.100)	(0.164)	(0.101)
TFS	0.632***	0.634***	1.110***	0.666***	0.778***	0.0913	0.595***
	(0.137)	0.138	(0.215)	(0.164)	(0.140)	(0.283)	(0.136)
TFS²		0.049
		0.473
Log (R&D)	0.268***	0.268*** (0.029)	0.268***	0.314***	0.256***	0.280***	0.271***
	(0.029)		(0.030)	(0.038)	(0.029)	(0.055)	(0.030)
TFS×Log (R&D)				0.281**
				(0.121)
Log (R&D)²				−0.013
				(0.010)
TFS×Log (R&D)²				−0.094***
				(0.030)
Log (entropy)	0.683***	0.684***	0.675***	0.667***	0.697***	0.845***	0.668***
	(0.059)	(0.060)	(0.060)	(0.060)	(0.059)	(0.101)	(0.059)
Slack	−0.000***	−0.000***	−0.000**	−0.000**	0.000
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
TFS×Log (sales)			−0.115***
			(0.039)
TFS×Slack					−0.0002***
					(0.000)
SGA intensity						−0.126
						(0.104)
TFS × SGA intensity						0.518
						(0.329)
Debt-to-equity							0.001
							(0.002)
TFS × Debt-to-equity							−0.003
							(0.015)
Constant	0.850***	0.848***	0.834***	0.828***	0.842***	0.667**	0.891***
	(0.131)	(0.133)	(0.131)	(0.130)	(0.130)	(0.327)	(0.131)
Observations	1666	1666	1666	1666	1666	685	1,664
Number of unique firms	347	347	347	347	347	159	347

: p<0.05; **: p<0.01; ***: p<0.001.

TFS: technological footprint similarity; R&D: research and development; SGA: selling, general, and administrative expenses.

Because some independent variables are correlated with one another, collinearity may be an issue. I ran collinearity diagnostics and found that the largest VIFs, 2.36 for Log (sales) and 2.43 for Log (R&D), are safely below levels which may potentially signify multicollinearity problems (O’Brien, 2007).

Hypothesis 1 predicted that TFS would benefit a firm’s invention performance. The coefficient for TFS in Model 1 is positive (0.632) and highly significant. An increase of 0.5 in TFS was associated with an e^0.316 = 1.371 multiplier for the number of patents (i.e. a 37.1% increase in the number of patents). This effect is conceptually important because I control for the degree of technological diversification, thereby adding to the work of Granstrand et al. (1997) Miller (2006), and Chen et al. (2013) I show that greater TFS has a positive effect on invention performance above and beyond the effect of technological diversification.

Hypothesis 2a predicted that firms would have an inverse U-shaped relationship with market performance. I found that the coefficient for TFS squared is not significantly different from zero when market performance is the dependent variable while the linear effect of TFS on market performance is positive (results available from the author). Therefore, Hypothesis 1 is supported while Hypothesis 2a is not supported. Overall, the results of testing Hypotheses 1 and 2a suggest that TFS has a positive, linear effect on invention performance and on market valuations while there is no quadratic effect of TFS on either dependent variable. Hypothesis 2b is supported: firms with very high TFS have higher market valuations than firms with very low TFS.

Hypotheses 3a and 3b presented alternative theoretical arguments for the moderation effect of firm size. According to Table 3, the interaction between firm size and TFS is negative, lending support to Hypothesis 3b. I find that TFS has a more positive effect on smaller firm’s invention performance. This effect is even more interesting since the interaction between firm size and technological diversification is positive, showing that large firms (as can be expected) benefit more from technological diversification than small firms (the result is available upon request). This result shows a clear difference between the effects of technological diversification and TFS. Figure 1 presents an interaction plot between firm size and TFS. We see that small firms enjoy greater gains in invention rates from higher TFS than large firms. Additional checks showed no significant interaction effect between firms size and TFS when market-to-book valuations were the dependent variable (available upon request).

Figure 1.

Interaction of firm size with TFS.

Hypothesis 4 predicted that firms with low TFS would benefit from higher levels of slack. According to Table 3, Hypothesis 4 is supported only for available slack (short-term assets minus short-term liabilities) but not for absorbed slack (SGA expenses) or potential slack (debt to equity). No effect on market-to-book valuations was found. Figure 2 plots this interaction:

Figure 2.

Interaction of available slack with TFS.

Finally, Hypotheses 5a and 5b argued that TFS would moderate the effect of R&D investments on invention performance. Furthermore, the expected quadratic effect of R&D would have different signs at different levels of TFS. These hypotheses are supported. According to Table 3, there is a significant interaction between TFS and simple and squared term for R&D. Graphs in Figure 3 show this interaction.

Figure 3.

Interaction of R&D Investments with TFS.

According to Figure 2, firms with high TFS experience much higher marginal benefits from R&D than firms with low TFS, especially when R&D increases from low values. The shape of the curve flips when firms transition between low and high TFS, supporting Hypotheses 5a and 5b. When market-to-book was the dependent variable, there was no interaction effect between TFS and R&D investments (available upon request).

The results for CWPs are very similar to those for raw patent counts (available upon request), which lends further support to my findings.

Discussion

The goal of this study was to investigate the effects of TFS on firm performance and firm-level factors that may either amplify or weaken these effects. The results indicate that (1) the effect of TFS on firm performance is positive: ceteris paribus, firms with technological footprints most closely mirroring that of the rest of the industry tend to have the highest invention and market performance; (2) there is no quadratic effect of TFS on firm performance; (3) firm size weakens the effect of TFS on invention performance so that large firms do not experience a significant positive effect of TFS on their invention performance; (4) available slack is a negative moderator: firms with high TFS have lower invention performance if they have ample slack; (5) while the effect of R&D on invention performance is positive overall, its shape depends on the firm’s TFS: it is positive and U-shaped for firms with low TFS; it is positive and inverse-U-shaped for firms with high TFS.

The findings reported in this article help extend and qualify the previous theory on the performance effects of similarity of the firm to its environment (Deephouse, 1999; Oh and Barker, 2018). TFS has a positive effect on invention and financial market performance, supporting the arguments of population ecology and competitive isomorphism theory. Lack of a quadratic effect of TFS suggests that performance of bio-pharm companies is highest in the most densely populated areas of knowledge, ceteris paribus. Firms with high TFS may have high capabilities for monitoring their rivals’ R&D and engaging in the most promising areas of technology.

The interactions between TFS and firm size suggests that firms may use their resources to overcome the straitjacket of the industry’s technological footprint. I found that larger firms tend to be nearly as inventive while deviating from the industry technological footprint as their peers with high TFS. I also found that firms with high TFS experience positive marginal benefits from additional R&D investments at all levels of R&D investments while firms with low TFS need to build up their R&D before they start experiencing positive marginal benefits from additional R&D investments.

Lack of support for Hypothesis 2a is interesting. I did not find a quadratic effect of TFS on financial market performance: the effect was positive and linear. A possible explanation is that investors in biopharm companies reward safe inventions that have the highest chance of being approved by the Food and Drug Administration (FDA). Perhaps patenting in densely populated areas is seen as less risky and is therefore rewarded by investors. Additional research in other industries is needed.

TFS captures an important aspect technological positioning of firms: being similar to the industry in terms of their technological portfolios. I show that the degree of technological diversification (e.g. Chen et al., 2013; Granstrand et al., 1997; Miller, 2006) is an insufficient characteristic of a firm’s technological portfolio without referring to the degree of similarity between the portfolio and the rest of the industry. Firms can certainly change their technological portfolios by entering or exiting specific technological areas. Active strategizing by a firm’s management can take two forms: (1) adjusting the fit between the firm’s technological portfolio and that of the rest of the industry and (2) changing the firm’s slack resources, R&D investments, and growing larger. An interesting extension of this study would examine the interaction between the firm’s TFS and strategic similarity as it was conceptualized by Deephouse (1999) and Oh and Barker (2018): the extent to which a firm’s overall strategy matches that of the entire industry.

Low TFS does not necessarily mean outsourced R&D. It might mean that the focal firm is a specialist in a narrow technology domain that is poorly served by the rest of the industry, perhaps due to highly specialized assets necessary to compete in this domain. Conversely, high TFS is a sign of a generalist (a firm with broad technological competencies) or a conformist (a firm that is active only in the most popular technology domains). A generalist is likely to benefit from economies of scale and scope while a conformist is likely to benefit from outside learning and improving upon the inventions of others. Both of these technological strategies have their benefits. Distinguishing between them and learning their relative benefits should be an important direction of future research.

The moderating effect of firm size and slack resources on TFS—invention performance suggests an interesting avenue of further research. Since larger firms and those with greater slack are relatively better buffered against risk of immediate failure, they may successfully tackle risky technological strategies that require deviating from industry norms. It is possible that such firms may use their ample resources to develop inventions in sparsely populated areas of technology or produce novel combinations of technologies in their portfolios that deviate from industry norms. More research is needed in this intriguing area.

I also show that firms with high R&D investments have higher invention performance when they had high TFS. The shapes of the curves in Figure 3 suggest that R&D investments may play a different role depending on the degree of a firm’s TFS. Firms with high TFS experienced high invention performance gains with increases in R&D investments from low to medium, but then their marginal increases in invention rates become lower as R&D investments continue to grow. This result suggests that densely populated technology domains may reward even small initial increases in R&D investments because they allow firms to benefit from the rich knowledge dispersed in the industry. Conversely, firms with low TFS need to build up their R&D before they experience an increase in their invention performance. This result suggests that firms with low TFS may have to rely on their own R&D and that R&D investments may play somewhat different roles for firms with different degrees of TFS: significant external learning at high levels of TFS versus mostly internal development at low levels of TFS.

An interesting result that was consistent across all models was a negative effect of mergers and acquisitions on invention performance. While similar results have been reported in the literature (e.g. Cloodt et al., 2006), no previous studies have addressed a possible interaction of acquisitions with TFS. A post hoc analysis revealed that acquisitions were especially detrimental to invention performance in the presence of high TFS. It is possible that firms patenting in densely populated technological areas tend to acquire firms with similar technological portfolios while paying top dollar for similar invention capabilities. Conversely, firms with low TFS might make acquisitions of companies working in sparsely populated and therefore “less popular” areas of technology. Such companies may be relatively less expensive while providing the acquirer with access to breakthrough inventions. My dataset does not allow me to distinguish between these possibilities. Further research is needed to conduct finer-grained analyses of mergers and acquisitions and their interactions with TFS.

Based on the results of this study, bio-pharmaceutical firms tend to have higher invention performance when their technological footprints match those of the industry as a whole. One may wonder why not all firms migrate toward densely populated technological areas. Several factors may prevent firms from doing that. One is inertia. Firms cannot easily change their technological portfolios. Investing in certain technologies creates competencies that may be hard to transfer to other areas of knowledge. In fact, as Patel and Pavitt (1997) showed, large firms tend to be more alike in their technological capabilities than smaller firms. This fact may be the outcome of an evolutionary process (firms slowly drifting toward each other and thereby toward industry averages over time) and of the competitive selection process (firms whose technological portfolios are too far from the industry averages tend to be selected out). Another factor may be that TFS may have a positive effect on invention performance but a U-shaped effect on innovation performance (commercialization of technologies). Being too close to the industry mainstream may mean greater difficulties in commercializing one’s inventions since they will have to compete with many other inventions. Some difference between the firm’s technology portfolio and those of other firms may be desirable. This dichotomy suggests two equally viable strategies: conformity or differentiation. While the former leads to greater invention performance in terms of the number of patents, the latter may lead to greater commercialized innovation.

TFS means patenting in “popular” areas of technology. One side effect of this may be greater impact of inventions as measured by forward citation rates because there are many other companies that may be interested in the focal firm’s patents. I ran a post hoc analysis and found that greater TFS was indeed a predictor of higher forward citation rates. Forward citations are an imperfect measure of patent value (compare the findings of Hall et al. (2005) and De Carolis (2003)); more research is needed connecting the effect of TFS and the impact of firms’ inventions.

Limitations and future directions

This study was done using two closely related industries (pharmaceuticals and biotechnology). The results of this study may not be transferable to another industry, especially if patenting is not a good measure of invention performance in that industry. More research is needed to test the main propositions of this article in other industries and using other measures of invention and innovation performance.

The measure of TFS used in this study relies on correlations between the firm’s patenting footprint and that of the rest of the industry. There could be other measures of TFS between the firm and its industry, such as those using common patent citations (Gambardella et al., 2008) or cross-citations (Mowery et al., 1996). The results of this study could be replicated and extended using other measures of TFS.

I used patent classes to represent technology domains. While each patent class reflects a specific area of application (e.g. pharmaceutical preparations), there could be other ways to classify technologies and construct technology domains (e.g. groups of patent classes that are closely related to one another or subclasses for finer-grained analysis). Future research may be necessary to replicate and extend the results in this study using other definitions of technology domains.

An important area of future research is the effect of TFS on innovation outcomes instead of pure invention outcomes. Invention is defined as “a commercially promising product or service idea” while innovation is defined as “the successful entry of a new science or technology-based product into a particular market” (Branscomb and Auerswald, 2002: 1). A company with more inventions may not be more innovative if it fails to commercialize its technologies. Firms with similar technological capabilities may have very different integration/architectural capabilities that are needed to commercialize a product.³ While we may expect firms with more patents to introduce more new products, this connection is by no means clear. Further research is needed to examine the effects of TFS on introductions of new products, especially because firms competing in densely populated technology areas may have to face severe competition. In particular, strategic alliances and other partnerships in marketing and manufacturing may help firms overcome their limited capabilities to commercialize their inventions successfully. Considering potential instability in vertical relationships, it may be important to study interactions between stability of technological competencies in the industry and stability of vertical relationships, especially in the bio-pharm industry.

Conclusion

This study examined the effect of TFS on firm performance and the moderating effects of firm size, slack resources, and R&D investments. The results indicate that TFS is positively associated with invention and financial market performance. The size of the firm and the amount of its available slack resources negatively moderate this relationship so that large firms and firms with ample slack do not suffer negative consequences of low TFS. Finally, the overall positive effect of R&D investments on invention performance of firms changes shape (from convex to concave) as TFS increases. These findings contribute to our understanding of the interplay between firm-level and industry-level factors as they jointly affect a firm’s invention and market performance.

Footnotes

Acknowledgements

The author expresses his gratitude to Lakshmi Sameera Peyyalamitta, Krutika Parab, and Anusha Samudrala for their help with data collection.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was partially funded by an FRSF grant from the University of Houston–Clear Lake.

ORCID iD

Aleksey Martynov

Notes

Author biography

Aleksey (Alex) Martynov is an associate professor of Management at the University of Houston–Clear Lake. He obtained his PhD from the University of Kansas. His research interests lie in the areas of invention, innovation, patenting strategies, the impact of innovation and institutions on resilience to climate change, and the interplay between competitive and institutional forces that shape both markets and institutions. Alex has published in the Journal of Business Ethics, Journal of Management, Long Range Planning, Managerial and Decision Economics, the Journal of Strategy and Management, Team Performance Management, and Research in International Business and Finance.

References

Aharonson

Schilling

(2016) Mapping the technological landscape: Measuring technology distance, technological footprints, and technology evolution. Research Policy 45: 81–96.

Ahuja

(2000) Collaboration networks, structural holes, and innovation: A longitudinal study. Administrative Science Quarterly 45: 425–455.

Ahuja

Katila

(2001) Technological acquisitions and the innovation performance of acquiring firms: A longitudinal study. Strategic Management Journal 22: 197–220.

Allison

Waterman

(2002) 7. Fixed-effects negative binomial regression models. Sociological Methodology 32: 247–265.

Amburgey

Kelly

Barnett

(1993) Resetting the clock: The dynamics of organizational change and failure. Administrative Science Quarterly 38: 51–73.

Arora

Gambardella

Magazzini

, et al. (2009) A breath of fresh air? Firm type, scale, scope, and selection effects in drug development. Management Science 55: 1638–1653.

Bergh

(1997) Predicting divestiture of unrelated acquisitions: An integrative model of ex ante conditions. Strategic Management Journal 18: 715–731.

Bergh

Lawless

(1998) Portfolio restructuring and limits to hierarchical governance: The effects of environmental uncertainty and diversification strategy. Organization Science 9: 87–102.

Bessen

(2008) The value of US patents by owner and patent characteristics. Research Policy 37: 932–945.

10.

Bottazzi

Dosi

Lippi

, et al. (2001) Innovation and corporate growth in the evolution of the drug industry. International Journal of Industrial Organization 19: 1161–1187.

11.

Bourgeois

(1981) On the Measurement of Organizational Slack. The Academy of Management Review 6: 29–39.

12.

Branscomb

Auerswald

(2002) Between invention and innovation an analysis of funding for early-stage technology development. NIST GCR 02-841. Available at: https://www.nist.gov/system/files/documents/2017/05/09/gcr02-841.pdf

13.

Bromiley

(1991) Testing a causal model of corporate risk taking and performance. Academy of Management Journal 34: 37–59.

14.

Brusoni

Geuna

(2003) An international comparison of sectoral knowledge bases: Persistence and integration in the pharmaceutical industry. Research Policy 32: 1897–1912.

15.

Brusoni

Criscuolo

Geuna

(2005) The knowledge bases of the world’s largest pharmaceutical groups: What do patent citations to non-patent literature reveal? Economics of Innovation and New Technology 14: 395–415.

16.

Brusoni

Prencipe

Pavitt

(2001) Knowledge specialization, organizational coupling, and the boundaries of the firm: Why do firms know more than they make? Administrative Science Quarterly 46: 597–621.

17.

Caner

Bruyaka

Prescott

(2018) Flow signals: Evidence from patent and alliance portfolios in the U.S. biopharmaceutical industry. Journal of Management Studies 55: 232–264.

18.

Chang

Chen

Lin

MJJ

(2014) Determinants of absorptive capacity: Contrasting manufacturing vs services enterprises. R&D Management 44: 466–483.

19.

Chen

Y-M

Yang

D-H

Lin

F-J

(2013) Does technological diversification matter to firm performance? The moderating role of organizational slack. Journal of Business Research 66: 1970–1975.

20.

Cloodt

Hagedoorn

Van Kranenburg

(2006) Mergers and acquisitions: Their effect on the innovative performance of companies in high-tech industries. Research Policy 35: 642–654.

21.

Cockburn

Henderson

(2001) Scale and scope in drug development: Unpacking the advantages of size in pharmaceutical research. Journal of Health Economics 20: 1033–1057.

22.

Cohen

March

Olsen

(1972) A garbage can model of organizational choice. Administrative Science Quarterly 17: 1–25.

23.

Cohen

Levinthal

(1990) Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly 35: 128–152.

24.

Cui

O’Connor

(2012) Alliance portfolio resource diversity and firm innovation. Journal of Marketing 76: 24–43.

25.

Cyert

March

(1963) A Behavioral Theory of the Firm:. Englewood Cliffs, NJ: Prentice Hall.

26.

Daily

Dalton

(1994) Corporate governance and the bankrupt firm: An empirical assessment. Strategic Management Journal 15: 643–654.

27.

Daniel

Lohrke

Fornaciari

, et al. (2004) Slack resources and firm performance: A meta-analysis. Journal of Business Research 57: 565–574.

28.

De Carolis

(2003) Competencies and imitability in the pharmaceutical industry: An analysis of their relationship with firm performance. Journal of Management 29: 27–50.

29.

De Man

Duysters

(2005) Collaboration and innovation: A review of the effects of mergers, acquisitions and alliances on innovation. Technovation 25: 1377–1387.

30.

Deephouse

(1996) Does isomorphism legitimate? Academy of Management Journal 39: 1024–1039.

31.

Deephouse

(1999) To be different, or to be the same? It’s a question (and theory) of strategic balance. Strategic Management Journal 20: 147–166.

32.

DiMaggio

Powell

(1983) The iron cage revisited: Collective rationality and institutional isomorphism in organizational fields. American Sociological Review 48: 147–160.

33.

Dimick

Murray

(1978) Correlates of substantive policy decisions in organizations: The case of human resource management. Academy of Management Journal 21: 611–623.

34.

Dindaroğlu

(2018) Determinants of patent quality in US manufacturing: Technological diversity, appropriability, and firm size. The Journal of Technology Transfer 43: 1083–1106.

35.

Escribano

Fosfuri

Tribo

(2009) Managing external knowledge flows: The moderating role of absorptive capacity. Research Policy 38: 96–105.

36.

Flatten

Engelen

Zahra

, et al. (2011) A measure of absorptive capacity: Scale development and validation. European Management Journal 29: 98–116.

37.

Fosfuri

Tribo

(2008) Exploring the antecedents of potential absorptive capacity and its impact on innovation performance. Omega-international Journal of Management Science 36: 173–187.

38.

Gambardella

Torrisi

(1998) Does technological convergence imply convergence in markets? Evidence from the electronics industry. Research Policy 27: 445–463.

39.

Gambardella

Harhoff

Verspagen

(2008) The value of European patents. European Management Review 5: 69–84.

40.

Garcia-Vega

(2006) Does technological diversification promote innovation? An empirical analysis for European firms. Research Policy 35: 230–246.

41.

George

(2005) Slack resources and the performance of privately held firms. Academy of Management Journal 48: 661–676.

42.

Granstrand

Patel

Pavitt

(1997) Multi-technology corporations: Why they have “distributed” rather than “distinctive core” competencies. California Management Review 39: 8–25.

43.

Haans

Pieters

(2016) Thinking about U: Theorizing and testing U- and inverted U-shaped relationships in strategy research. Strategic Management Journal 37: 1177–1195.

44.

Hall

Jaffe

Trajtenberg

(2005) Market value and patent citations. The RAND Journal of Economics 36: 16–38.

45.

Hamilton

Nickerson

(2003) Correcting for endogeneity in strategic management research. Strategic Organization 1: 51–78.

46.

Hannan

Freeman

(1977) The population ecology of organizations. American Journal of Sociology 82: 929–964.

47.

Hannan

Freeman

(1984) Structural inertia and organizational change. American Sociological Review 49: 149–164.

48.

Henderson

Cockburn

(1994) Measuring competence? Exploring firm effects in pharmaceutical research. Strategic Management Journal 15: 63–84.

49.

Kennedy

(2002) Strategy fads and competitive convergence: An empirical test for herd behavior in prime-time television programming. The Journal of Industrial Economics 50: 57–84.

50.

Kim

Lee

(2008) Ownership structure and the relationship between financial slack and R&D investments: Evidence from Korean firms. Organization Science 19: 404–418.

51.

Klette

(1996) R&D, scope economies, and plant performance. The RAND Journal of Economics 27: 502–522.

52.

Kostopoulos

Papalexandris

Papachroni

, et al. (2011) Absorptive capacity, innovation, and financial performance. Journal of Business Research 64: 1335–1343.

53.

Lance

(1988) Residual centering, exploratory and confirmatory moderator analysis, and decomposition of effects in path models containing interactions. Applied Psychological Measurement 12: 163–175.

54.

Leonard-Barton

(1992) Core capabilities and core rigidities: A paradox in managing new product development. Strategic Management Journal 13: 111–125.

55.

Leten

Belderbos

Van Looy

(2007) Technological diversification, coherence, and performance of firms. Journal of Product Innovation Management 24: 567–579.

56.

Lin

Chang

C-C

(2015) The effect of technological diversification on organizational performance: An empirical study of S&P 500 manufacturing firms. Technological Forecasting and Social Change 90: 575–586.

57.

Lin

Y-J

Chang

, et al. (2012) The alliance innovation performance of R&D alliances—The absorptive capacity perspective. Technovation 32: 282–292.

58.

Little

Bovaird

Widaman

(2006) On the merits of orthogonalizing powered and product terms: Implications for modeling interactions among latent variables. Structural Equation Modeling 13: 497–519.

59.

McKnight

Zietsma

(2018) Finding the threshold: A configurational approach to optimal distinctiveness. Journal of Business Venturing 33: 493–512.

60.

Martynov

(2017) Alliance portfolios and firm performance: The moderating role of firms’ strategic positioning. Journal of Strategy and Management 10: 206–226.

61.

Miller

Breton-Miller

Lester

(2013) Family firm governance, strategic conformity, and performance: Institutional vs. strategic perspectives. Organization Science 24: 189–209.

62.

Miller

(2006) Technological diversity, related diversification, and firm performance. Strategic Management Journal 27: 601–619.

63.

Mowery

Oxley

Silverman

(1996) Strategic alliances and interfirm knowledge transfer. Strategic Management Journal 17: 77–91.

64.

Nesta

Saviotti

(2005) Coherence of the knowledge base and the firm’s innovative performance: Evidence from the US pharmaceutical industry. The Journal of Industrial Economics 53: 123–142.

65.

Nesta

Saviotti

(2006) Firm knowledge and market value in biotechnology. Industrial and Corporate Change 15: 625–652.

66.

Nieto

Quevedo

(2005) Absorptive capacity, technological opportunity, knowledge spillovers, and innovative effort. Technovation 25: 1141–1157.

67.

Norman

Artz

Martinez

(2007) Does it pay to be different? Competitive non-conformity under different regulatory regimes. Journal of Business Research 60: 1135–1143.

68.

O’Brien

(2007) A caution regarding rules of thumb for variance inflation factors. Quality & Quantity 41: 673–690.

69.

W-Y

Barker

III (2018) Not all ties are equal: CEO outside directorships and strategic imitation in R&D investment. Journal of Management 44: 1312–1337.

70.

Orsenigo

Pammolli

Riccaboni

(2001) Technological change and network dynamics: Lessons from the pharmaceutical industry. Research Policy 30: 485–508.

71.

Palepu

(1985) Diversification strategy, profit performance and the entropy measure. Strategic Management Journal 6: 239–255.

72.

Palich

Cardinal

Miller

(2000) Curvilinearity in the diversification-performance linkage: An examination of over three decades of research. Strategic Management Journal 21: 155–174.

73.

Palmer

Wiseman

(1999) Decoupling risk taking from income stream uncertainty: A holistic model of risk. Strategic Management Journal 20: 1037–1062.

74.

Pangarkar

(2003) Determinants of alliance duration in uncertain environments: The case of the biotechnology sector. Long Range Planning 36: 269–284.

75.

Patel

Pavitt

(1997) The technological competencies of the world’s largest firms: Complex and path-dependent, but not much variety. Research Policy 26: 141–156.

76.

Patel

Pavitt

(2000) How technological competencies help define the core (not the boundaries) of the firm. In: Dosi

Nelson

Winter

(eds) The Nature and Dynamics of Organizational Capabilities. Oxford: Oxford University Press, pp. 313–333.

77.

Pavitt

(1998) Technologies, products and organization in the innovating firm: What Adam Smith tells us and Joseph Schumpeter doesn’t. Industrial and Corporate Change 7: 433–452.

78.

Quintana-García

Benavides-Velasco

(2008) Innovative competence, exploration and exploitation: The influence of technological diversification. Research Policy 37: 492–507.

79.

Sampson

(2007) R&D alliances and firm performance: The impact of technological diversity and alliance organization on innovation. Academy of Management Journal 50: 364–386.

80.

Stock

Greis

Fischer

(2001) Absorptive capacity and new product development. The Journal of High Technology Management Research 12: 77–91.

81.

Stuart

(1998) Network positions and propensities to collaborate: An investigation of strategic alliance formation in a high-technology industry. Administrative Science Quarterly 43: 668–698.

82.

Swaminathan

(1995) The proliferation of specialist organizations in the American wine industry, 1941-1990. Administrative Science Quarterly 40: 653–680.

83.

Tafti

Mithas

Krishnan

(2013) The effect of information technology-enabled flexibility on formation and market value of alliances. Management Science 59: 207–225.

84.

Teece

Rumelt

Dosi

, et al. (1994) Understanding corporate coherence: Theory and evidence. Journal of Economic Behavior and Organization 23: 1–30.

85.

Trajtenberg

(1990) A penny for your quotes: Patent citations and the value of innovations. The RAND Journal of Economics 21: 172–187.

86.

Tsai

(2001) Knowledge transfer in intraorganizational networks: Effects of network position and absorptive capacity on business unit innovation and performance. Academy of Management Journal 44: 996–1004.

87.

Wang

Hagedoorn

(2014) The lag structure of the relationship between patenting and internal R&D revisited. Research Policy 43: 1275–1285.

88.

Wooldridge

(2010) Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.

89.

Zahra

George

(2002) Absorptive capacity: A review, reconceptualization, and extension. The Academy of Management Review 27: 185–203.