Growth - intentions, perceptions and reality: Evidence based directions for innovation researchers Martie-Louise Verreynne UQ Business School, University of Queensland Brisbane, Australia
While measurement of innovation and growth is critical for academics, policy-makers and practitioners alike, the field is characterized by contradictory approaches, which have hampered the development of theory and methods. The relationship between innovation and firm growth also remains inconclusive. It is argued here that this can be at least partially attributed to how we conceptualize the measures that form part of hypothesized relationships. This paper develops a set of recommendations regarding the measurement of innovation and firm growth in research that studies the relationships between these constructs. The focus is on how innovation and growth is measured in Community Innovation Survey type studies. It uses data gathered during 2010/1 in Australia to illustrate different ways to measure innovation and growth and suggests which variables are to be used in different research questions.
Key words: Innovation measures, growth measures, CIS studies
INTRODUCTION Innovation is seen as the cornerstone of economic and firm growth (Baldwin & Gellatly 2003). Firm survival and growth has broad economic and social implications. Growing firms contribute to increased employment and exports, higher GDP and other flow on effects such as improved living standards. Both innovation and growth are constructs that have been defined from many different perspectives. Penrose (1959) argues that growth is not only a change in firm size, but in fact a process which poses many challenges to the growing firm. Growth occurs in many guises. It can be organic, or can be based on mergers or acquisitions, integration, diversification, or licencing. It can be one-dimensional, based on growth in employment numbers, or multifaceted (Delmar, Davidsson & Gartner, 2003). Similarly, innovation has been conceptualized broadly as technical and non-technical (Damanpour, 1991; Rothwell, 1992). Within these broad categories many forms present; including product, service, process, marketing, and managerial (Schumpeter, 1934) as well as Research and Development (R&D) intensity (Freel, 2007).
A review of the literature reveals that a myriad of approaches have been used to measure organization growth (Weinzimmer, Nystrom & Freeman, 1998) and innovation (Adams, Bessant, & Phelps, 2006). It is therefore not surprising that not noting the effects of the use of different measures has led to conflicting findings (Adams, et al. 2006). The importance of noting and understanding these conflicts when choosing measures can therefore not be understated. This paper attempts to remedy this by drawing on theories of Industrial Organization (IO) (Geroski, 1998) and firm resources (Resource Based Theory - RBT) (Barney, 1996) to ask: What are the best approaches to measure growth for research models of different forms of innovation?
To this end, methods to measure growth are evaluated against innovation types and innovation processes such as R&D and sales from innovation. The paper therefore investigates the use of alternative innovation and firm growth measures, and their effect on the findings from statistical analysis. It proceeds as follows: First, an overview of innovation and growth measurement types and issues is provided. Second, a framework for analysis, which includes a number of innovation variables, is developed. Third, this framework is with different operationalizations of growth as the dependent variable. Based on this analysis, the paper concludes with a number of recommendations regarding the measurement of
1
innovation and growth; identifying the limitations and future research directions that emanate from the findings.
INNOVATION MEASUREMENT The issues with measurement can be traced back to the definitional issues facing the field. Adams, et al. (2006) suggest that defining it as ‘the successful exploitation of new ideas’ (p. 22) allows researchers to view innovation as innovation types, as suggested by the OSLO Manual (OECD, 2005). We use the broader definition from the Oslo Manual to fit with the purpose of this paper, namely to investigate the use of innovation in CIS studies. Therefore, innovation is defined as “the implementation of any new or significantly improved product (goods or services), operational processes (methods of production and service delivery), any new marketing methods (packaging, sales and distribution methods), or new organisational or managerial methods or processes in business practices, workplace organisation or external relations” (OECD 2005, p. 46). This definition links well with Schumpeter’s (1934) five types of innovation.
Innovation has developed into a substantial field of enquiry (Adams, et al. 2006) since the early works of Schumpeter (1934). The view that innovation is the basis of economic (Porter & Ketels, 2003) and firm growth (Damanpour & Evan, 1984; Gopalakrishnan, 2000; Lööf & Heshmati, 2006; Wong, Page, Abello, & Pang, 2007) has gained wide acceptance. Measuring the impact of innovation on growth accurately is important to a number of audiences (Adams, et al. 2006). To firms, the challenge is to understand how innovation can be optimally managed (Cordero, 1990). To academics, the importance of an approach that is commonly accepted will improve comparability of results, limit conflicting findings and advance the development of the field. To policy-makers, improved regional and international comparison can help to guide policies and investment.
Yet the treatment of innovation has been fragmented. Innovation is typically measured as an input and output, and the processes that go between are often neglected (Adams, et al. 2006). Freel (2007) voices his disappointment with the lack of a unified measure for innovation. He identifies a number of inputs (e.g. R&D, qualified scientists and engineers) and output measures (e.g. patents or new product/process introductions). He also, in his own research based on the Centre for Business Research studies conducted out of Cambridge University,
2
calculates a number of other variables such as R&D expenditure (as a proportion of turnover), innovation in turnover, and innovation in profits.
In this paper, the focus is on the conceptualization of relationships, rather than measures. Therefore, for innovation measures the conceptualization proposed by Adams, et al. (2006) is used as a basis for the proposed hypotheses. These authors identify the measures that have been used to account for the disaggregation of the concept which has been observed in the literature, and which has left several idiosyncratic operationalizations of innovation. The resultant seven-factor framework includes; inputs management, knowledge management, innovation strategy, organizational culture and structure, portfolio management, project management and commercialization.
In this article, three of these areas are investigated for their relationships with different forms of growth. The first, inputs management, is frequently used in studies. In particular, the construct research and development (R&D) intensity has been investigated (e.g. Deeds 2001; Greve, 2003; Parthasarthy & Hammond 2002). R&D intensity is calculated as the ratio between expenditure (Parthasarthy & Hammond 2002) or numbers employed in R&D roles (Kivimäki, Lansisalmi, Elovainio, Heikkila, Lindstrom, Harisalo, Sipila & Puolimatka, 2000) and an expression of output, such as sales. Findings regarding the relationship between R&D intensity and performance or growth have varied greatly, with a linear, non-linear and no relationship emanating from different studies (Bougrain & Haudeville, 2002; Stock, Greis, & Fischer, 2001).
The second factor is knowledge management. Again, several scale measures have been developed to assess different aspects of knowledge management, including knowledge absorption or absorptive capacity (Cohen & Levinthal 1990; Zahra & George 2002) and patent counts (Coad & Rao, 2008). However, these authors argue that problems with this measure may be why they did not find a strong correlation with growth. For this reason, this variable is not investigated in this paper. Of greater importance to studies that use CIS approaches is idea generation and information flows. These measures are essentially focused on collaboration and networking between different firms and institutes in the knowledge network or supply chain. These measures are predicated on the premise that idea generation equals the raw materials for innovation (Cooper 1988). Researchers either measure the number of ideas generated in a particular period (e.g. Chiesa, Coughlan & Voss, 1996; Lee, 3
Son & Lee, 1996) or assess the breadth of tools and techniques (e.g. Cebon & Newton 1999; Thompson 2003). In CIS studies, this latter measure is represented by the linkages with other firms or sources of ideas as well as internal sources. These measures are typically dichotomous in nature.
The third innovation factor is commercialization. Commercialization is the implementation of an innovation into the market (Chakravorti, 2004), and include for example market investigation, market testing, and promotion (Verhaeghe & Kfir, 2002). Measures include numbers of products launched in a given period (e.g. Yoon & Lilien, 1985) and launch proficiency (Parry, 1996). Adams et al. (2006) note that this area is the least developed of all innovation measurement areas. In this article ‘sales from innovation’ is used to measure commercialisation. However, introduction of innovation into the market, even without sales, is also measured. In this paper it is referred to as innovation type. In CIS studies, this factor has played a central role, and is based on Schumpeter’s (1934) five types of innovation, identified earlier. Eight variables were initially developed, but later whittled down to five to test innovation type.
Adams et al. (2006) also identify a number of other measures, which will not be covered in this paper. For example, innovation strategy is very popular in the strategy and entrepreneurship areas alike. Here measures such as entrepreneurial orientation (Covin & Slevin, 1989; Lumpkin & Dess, 1996) and prospector strategies (Miles & Snow, 1978) have gained acceptance. Organizational culture and structure are often used in contingency approaches, either as antecedents or moderating variables. Informal structures (Burns & Stalker, 1961) and innovative cultures (Amabile, Conti, Coon, Lazenby & Herron, 1996) are widely used in the literature. Portfolio Management looks at the effectiveness with which firms manage their innovation projects portfolio (Brenner 1994; Cooper, Edgett, & Kleinschmidt, 2001) or how well innovation management is integrated in other managerial practices (Cebon & Newton, 1999). Like the next measure, project management, which is typically measured with efficiency and speed measures these measures are about the environment that is created within a firm to facilitate innovation, and therefore not the focus of this study.
4
GROWTH MEASUREMENT Performance has been understood to mean effectiveness and efficiency, lean production competitiveness, cost reduction, value creation, growth, survival and job creation (Lebas & Euske, 2002). Lebas and Euske discuss various definitions of performance and then define it as ‘the sum of all processes that will lead managers to taking appropriate actions in the present that will create a performing [firm] in the future’ in other words, ‘doing today what will lead to measured value outcome tomorrow’ (2002, p. 68).
Growth, although often used as a proxy for performance, is a unique construct. Growth is defined as firm expansion (Penrose, 1959) and operationalized as growth in output (e.g. exports, sales), increase in size, process of development (e.g. biological) (Penrose, 1959). While it is often assumed that there is an optimal, or most profitable, firm size to which a firm can grow, the measurement of expansion is conceptually difficult (Penrose, 1959). This is because monitory measurements suffer from price, production and technological differences.
Theoretically firm growth is typically viewed as a reaction to external or internal conditions. Industrial organization (IO) theories are concerned with firm growth at an industry level. They often seek to understand the optimal size of a firm, which is seen to be determined by competition, market power, economies of scale, degree to which costs are sunk and internal organizational factors (Geroski, 1998). Gibrat’s (1962) hypothesis (the law of proportionate effect) on size has received mixed results. Mansfield (1962, p. 1030) summarizes this hypothesis as follows; ‘the probability of a given proportionate change in size during a specified period is the same for all firms in a given industry – regardless of their size at the beginning of the period’. The results from these studies vary, and some indicate that size and growth vary independently (e.g. Prais, 1974) while others find a negative relationship (e.g. Audretsch, 1991; Mansfield, 1962), albeit only for smaller firms. One explanation is that Gibrat’s law only holds for smaller firms, and that beyond a certain point it becomes obsolete.
Other IO theories have also been used to explain growth differences among firms. For example, Jovanovic’s (1982) theory holds that age and growth are negatively correlated, and that age leads to growth, rather than the other way around (Coad, 2007; Evans, 1987). IO
5
theorists are often fascinated with industry differences, which they see as including stylized ideas of competition (Bottazzi & Secchi, 2003; Nelson & Winter, 1978; Simon, 1977). Many small firms enter markets, but only a few are able to survive and grow (Mata, 1994) because industry effects are numerous and varied. Industry effects are numerous and varied. For example, Audretch (1995, p. 454) finds that ‘the ability of new entrants to differentiate their products through innovative activity and otherwise is clearly a key strategic instrument deployed by entrants to offset scale disadvantages and to occupy small-scale product niches’. Over time these effects vanish. IO theories focus on stochastic models of firm growth dynamics (Botlazzi & Secchi, 2006).
Resource based theorists are concerned with internal differences in resources among firms. Most works in this field build on Penrose’s (1959) research which was concerned with growth in already established firms. Penrose had a broader view of growth, espousing that ‘market opportunities of the firm and the productive services available from its own resources’ (1960, p. 14) form the bases of the growth process. RBT (resource based theory) as applied to small firms views the owner/manager as the key decision maker in small firms (Storey, 1997; Verreynne, 2006) who profoundly influences the growth propensity of the firm.
One approach that stems from RBT is the investigation of the internal and external factors that underpin growth. Many studies have been undertaken to investigate the factors that are present in high growth firms, perceived as the reasons why these firms grow while others do not (e.g. Chan Bhargava, & Street, 2006), or, on the flip side, fail (e.g. Hartarska & Gonzalez-Vega, 2006). Generally these are summarized as environmental, strategic, entrepreneurial and organizational; including factors such as industry concentration, capital investment, market share, innovation and capabilities (Capon, Farley & Hoenig, 1990).
Depending on the theoretical perspective, different growth measures seem to be popular. IO theorists typically measure growth by growth in firm size as proxy for growth (e.g. Brown, Earle, Lup, 2005). The latter authors include a working owner as an employee for measurement purposes. Garnsey, Stam and Heffernan (2006), from a RBT perspective, consider the different methods available to measure firm growth, including funds invested at different stages, sales figures, profits, employment figures, valuation of firm assets and composite measures such as the Birch Employment Growth index. Each of these methods 6
provides problems with comparability, but Garnsey et al. (2006) agree that changes in employment size are a conservative measure, and that it has been used to construct growth episodes and operational definitions of turning points.
Methods to measure growth are divided into three groups in this paper, namely growth intentions, objective growth and perceived growth. Growth intentions focus on the motive of a firm to grow. Researchers use the terms growth intentions (LeBrasseur, Blanco & Dodge, 2006) aspirations (Autio & Acs, 2010) or motivations (Davidsson, 1991) to indicate that owners/managers aim to grow their business. This is measured with multiple approaches, varying from a direct question to measuring the difference between current and expected employee and/or sales numbers.
Not all firms wish to grow, and in small firms, that is more likely to be the norm (e.g. Penrose, 1959) mainly because it is argued that the small firm owner does not want his/her control over the firm reduced. Wiklund, Davidsson and Delmar (2003) find that owners of young, small firms do not see that a doubling of firms’ size will lead to an increase in personal income, thereby removing one of the main motivators to grow. These authors use theories of planned behaviour to underpin their arguments. The theory of planned behaviour was developed by Icek Ajzen (1988, 1991). It incorporates reasoned actions and individual ability. These authors classify intentions as an aspect of aspirations, using Ajzen’s theory. They further stress that intentions only lead to outcomes in the presence of resources and opportunities.
Objective growth is typically measured as growth in employment numbers or sales/turnover; however, here growth in exports and productivity is also examined. Garnsey, et al. (2006) argue that change in employment size is a conservative measure and ease of access makes it prevalent (Hubbard & Bromiley, 1995). Sales growth has been widely used as an indicator of firm growth (e.g. Coad & Rao, 2008; Roper, 1999; Weinzimmer et al. 1998), while export growth features in internationalisation studies. These are seen as appropriate measures for entrepreneurial constructs.
Productivity growth has mostly been studied at the industry/sectoral level. However, some studies focus on its relationship with firm size (Pagano & Schivardi, 2003) and show that larger firms are more likely to have productivity growth, likely as a result of their ability to 7
take advantage of increasing returns associated with R&D and/or innovation. Last, indices such as The Birch (Employee Growth) Index multiply absolute job growth by relative job growth to determine the employment-creation power of differently sized enterprizes. Others have also created their own index to test their hypotheses on growth by summing either four items (Wiklund & Shepherd, 2003) - employment growth, sales growth, and comparative sales and employment numbers against competitors or six items (Delmar, Davidsson & Gartner, 2003) to measure growth. In general, results using objective measures are inconsistent, to illustrate, measures of innovation that include product and process innovation are problematic, seeing that process innovation does not seek sales growth.
When calculating objective growth, Weinzimmer et al. (1998) show that in articles using objective measures, such as sales, employment and asset growth, three formulae are used most often to calculate growth. In the first, a beta estimate from an OLS model is used by regressing size (sales, employee numbers and assets) on time. While these authors found this method to be superior, it does require a minimum of 15 observations to fit the regression line properly, a condition that is out of reach to most researchers. The second is termed ‘real growth’ and is calculated by subtracting size (sales, employment OR assets) in Period 1 from Period 2 and to then divide it by the number of periods represented to provide an annual number. While easy to calculate, the resulting dollar value is highly variable among industries, firms in different life cycle stages and sizes. The last method, used in this paper, subtracts size in Period 1 from Period 2 and then divide it by Period 1 ([t2-t1]/t1). This number can and should also be adjusted for inflation if used with other change sensitive variables.
Perceptions of growth are subjective measures of performance (Dess & Robinson, 1984). These measures typically come in the form of indices which consist of satisfaction data that have been weighted by importance data (Covin & Slevin, 1989). Perceptions of different types of performance, including growth, are used over objective data because of the inability and unwillingness of private and small firms to share the archival data. However, this approach is also better at explaining goal attainment (Glancey, 1998) and improves comparison across industries, firm sizes and age.
The use of perceptual measures of strategic constructs, including performance, has been supported by various studies (Clark, 2002; Dess, Lumpkin & Covin, 1997; Lebas & Euske, 8
2002; Lyon, Lumpkin, & Dess, 2000). Many reasons are given for this, mostly to do with the nature of small firms, which constitutes the majority of firms. These include heavy investment in development, the interrelationship between the firm and the owner, the delayed nature of performance testing, and the goal of the owner which may not be growth, but lifestyle (Cooper, 1979). Not all entrepreneurs will pursue profit maximisation and growth (Glancey, 1998). Owners of new ventures and small businesses can choose to either grow their business or to maintain it. This becomes a lifestyle choice and a decision to maintain control for the entrepreneur, and firm performance may be judged by basic financial criteria such as cash flow, or even survival (Lumpkin & Dess, 1996). Furthermore, some businessorientated entrepreneurs may also choose not to expand their businesses beyond some level that they can control without delegation of key functions (Glancey, 1998).
Although these advantages are important, the most compelling reason for using perceptions of performance in a small firm study is the vast diversity in how small firm manager-owners identify good performance specific to their firm. This study argues that small firm managerowners may have varying goals for their firms, ranging from growth goals to ensuring enough profit to sustain their lifestyle. Firm performance can only be measured accurately when compared to the intentions of the firm, namely its goals. It is therefore impossible for a researcher to make a judgement call on whether a small firm is successful based on objective performance measures. Unfortunately the use of perceptions to measure performance is not without disadvantages. Lyon et al. (2000) identify some of the limitations of this approach, including difficulty in identifying sources of variation in responses; perceptions may be subjected to ‘retrospective bias and other attributional phenomena’ (Clark, 2002, p. 33). Most of these limitations have been accounted for in the research design, for example questionnaires were used instead of interviews and more than one measure of performance was included to compare for issues such as variation in responses, bias and single respondent error. Furthermore, Dess and Robinson (1984) find a strong correlation between the use of subjective and objective measures of firm improvement and decline over a five year period. This is supported by Hart and Banbury (1994). This study will therefore employ subjective measures (perceptions) of firm performance.
That said; there are several issues with objective growth measurement that must be noted. First is that not all firms, especially privately held firms, will divulge their financial information. Second, objective growth, whether based on sales, profit or employee numbers 9
can vary widely. For example, erratic growth rates at the onset of firms may skew results. Further, staff numbers are distorted between different types of firms such as service versus manufacturing whereas sales do not account for the degree of integration. Third, and most important, growth in different aspects of the firm, that is, different growth measures, will have different results, for example, measures of innovation that include product and process innovation are problematic, seeing that process innovation does not focus on sales growth.
THE CONSEQUENCES OF USING ALTERNATIVE CONCEPTS AND FORMULAE To examine the relationship between innovation and growth, and to provide evidence for the guidelines presented at the end of this paper, a base model that regressed different forms of innovation on each of the proposed growth measures was developed and analysed. Potential predictors were taken from the discussion above, and include various measures of innovation, collaboration (proxy for knowledge management), age and size. To avoid the danger of using overlapping forms of innovation in one regression, which might have caused multicollinearity, more than one baseline model was developed after investigating the correlations among innovation variables (see Table 1). This led to the deletion of a number of variables from the analysis. Next the question: How should growth and innovation be matched? was attempted. To answer the research question posed earlier, two steps were taken. First, using an IO approach the relationships between growth intentions, perceptions of growth, and objective growth measures (see Table 2) and firm size and age were investigated. Second, the use of different growth measures as dependent variables in innovation related models was examined (see Table 3) using a RBT lens, as explained next.
METHODS Survey The Centre for Business Research (CBR) at Cambridge has been conducting small firm panel studies since 1991. Their survey instrument, on which the Australian survey is based, examines general business characteristics; workforce and training; innovation activity; commercial activity; competitive situation; factors affecting expansion and efficiency; capital expenditure and finance structure; and acquisition activity; providing around 400 variables. The Cambridge studies have become a widespread methodology (e.g. Cosh, Fu & Hughes, 2010; Freel, 2000; 2007) and builds on the Community Innovation Survey (CIS) 10
methodology. This paper uses data from a survey conducted during the first half of 2011. Questionnaires were mailed to 28,300 Australian firms, followed by two reminders. A stratified sampling strategy was used to ensure representativeness across firm sizes, industries and states. The response rate was 7.5 per cent with a total of 2,107 responses, following a number of natural disasters that occurred during data collection.
Data analyses Several tests for reliability and validity were conducted. The survey instrument included questions about the owner/manager, and organisational demographics, practices and performance. For the purpose of this paper, responses to questions about the age, size, industry and innovation practices of the firm were supplemented with data on growth intentions/perceptions and financial data. Data were analysed using chi-square tests of differences, Spearman’s and Pearson’s correlations, multiple and logistic regression.
FINDINGS Overview The responding firms had a median size of six FTEs and the median firm age of 19 years. Firms represented a broad range of industries, firm types and start-up motives. The demographics of the innovators identified in this study are very similar to other studies, for example, innovating firms tended to be older, larger and from industries such as manufacturing, mining and IT/professional services. R&D engagement was reported by 34 per cent of all firms. Approximately 40 per cent of all firms engaged in formal or informal collaboration or partnership arrangements with another organisation. Most Australian firms had moderate growth intentions (60 per cent), while approximately 19 per cent of firms indicated that they wanted to grow substantially.
Correlations As explained earlier, correlations among innovation variables were investigated first. Table 1 indicates that high correlations existed between: Process innovation and measures of innovation breadth; Product innovation and measures of innovation breadth; and Novel innovation breadth and significantly improved/new products or services introduced. High correlations among other variables were ignored, because these variables were conceptually different. These included product and process innovation, also with commercialisation (sales
11
from innovation). Therefore innovation breadth measures were not included in subsequent analysis.
Table 1: Correlations among innovation variables (1) (2) (3) (4) (5) R&D Intensity (1) 1 -.001 .062 .120* -.037 Process Innovation .037 1 .647** .049 .628** binary (2) Product/service .054 .647** 1 .262** .491** innovation binary (3) Sales from novel .010 .045 .212** 1 .033 innovation (4) Non-novel Innovation -.009 .550** .480** .055 1 breadth (5) Novel Innovation .111* .445** .522** .176** -.205** breadth (6) Significantly improved -.013 .064 .197** .832** .095** products/services (7) New products or .027 .010 .109** .794** -.026 services introduced (8) ** ** ** Commercialisation -.011 .585 .641 .371 .372** binary (9) Innovation binary (10) .000 .059* .049 -.047 .015 *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). Pearson’s correlations in bottom half, Spearman’s correlations in top half
(6) .066 .505**
(7) .063 .057
(8) .127* .026
(9) .019 .585**
(10) .030 .059*
.597**
.260**
.208**
.641**
.049
.193**
.862**
.809**
.490**
-.058
-.128**
.071*
-.013
.403**
.040
1
.156**
.185**
.432**
.057**
.114**
1
.496**
.240**
-.042
.138**
.324**
1
.653**
-.054
1
.026
.026
1
.348 .034
**
.075
*
-.035
.400
**
-.010
Next, correlations among all the variables were investigated (see Table 2 at end of paper). The correlations identified a number of interesting patterns. First, considering age and size, it seems that innovation variables are generally positively related to (larger) sized and (older) aged firms. There are, however, two notable exceptions. R&D intensity and sales from novel innovations and new products/services introduced, both proxies for novelty or industry level innovation, seem to be negatively related with these variables. It therefore seems that higher levels of innovation in older, larger firms are mostly driven by firm level innovation or imitation.
Second, the only input management variable, namely R&D intensity, presents very few significant correlations, and none with growth measures. As stated before, it seems to be more prevalent in smaller, younger firms, which may just be an artefact of it being a larger proportion of smaller turnover numbers. But it may also be as a result of the start-up phase that some of these firms may find themselves. Without further analysis only conjecture can be presented. The fact that there are no correlations between R&D intensity and any of the growth measures, may be a result of a time lag in the data, but it may also be evidence that not all R&D investment lead to commercializable products or services.
12
Third, the only knowledge management measure, collaboration, is only correlated to commercialisation, firm size and growth intentions. The latter correlation may again indicate that a time lag may exist in the data. This is further supported with the correlation with commercialisation, which indicates that the relationship between collaboration and growth is potentially mediated by innovation. A similar argument has been made by Gronum, et al. (2012) in their investigation of data provided by the Australian Bureau of Statistics.
Fourth, perceived growth seems to be the form of growth that exhibits the strongest correlations with innovation variables, while export growth is not related to any form of innovation and productivity growth only related to commercialisation. Employment growth is positively related to most forms of innovation, which is counter intuitive at least on the count of process innovation. Sales growth exhibits strong positive correlations with most innovation variables. These patterns were further investigated through the regression models described next.
Comparisons of alternative concepts of growth and forms of innovation Table 3: Regression results for alternative formulae of growth Dependent variables Independent variables Forms of innovation Commercialisation R&D Intensity Process innovation Product/service innovation Significantly improved products/services New products or services introduced
Collaboration Controls Size Age Constant R² Adjusted R² F
Sales growth
Employment growth
Models Productivity Export growth growth
Perceived growth
Growth intentions
-.117 .010 -0.010 -.167* .047 .004 .081
.003 -.016 .066 -.006 .077 -.102 .034
-.074 -.001 .004 -.130 -.004 .001 .035
.057 -.048 .115 .001 .197 -.009 .080
-.025 -.082 .169** .023 .131 .132 -.050
1.106* .013 .467 1.711** .098 .199 .576
-.030 .206**
-.034 -.074
-.023 .113
.056 -.436
-.030 .066
.000 .010
615.901 .101** .059** 2.412**
34.961 .017 -.032 .342
446.169 .052 .006 1.120
-190.519 .071 -.059 .547
2.534** .092** .053** 2.367**
2.398** .199**
*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). Note: Growth intentions were examined using logistic regression. The R² reported is Nagelkerke
Tables 3 reveals that only the sales growth, perceived growth, and growth intentions models are significant. Interestingly, process innovation seems to be significant to perceived growth, while product/service innovation is important to growth intentions. Further, product/service innovation is significant in the sales growth model. However, the beta is negative, indicating 13
a negative contribution. In this model, age is also significant, indicating that older firms are more likely to have higher levels of sales growth. This model also had the highest amount of variance explained.
DISCUSSION First the relationships between different measures of growth were explored. While most measures were related, productivity growth was not related to growth intentions, and negatively correlated with employment growth. This indicates that the purpose of improved productivity is not to grow the firm, but rather to do the same or more with less. Some correlations were also weak at .05, for example between export growth and employment/perceived growth. From an IO perspective, Gibrat’s (1962) law was only supported for export growth, which had no relationship with size. Size was positively correlated with growth intentions, employment, perceived and sales growth, and negatively with productivity growth. Jovanovic’s (1982) law was supported with age negatively correlated with growth intentions and sales growth, but not correlated to any other form of growth. From a RBT perspective, the relationships with innovation measures exhibited interesting patterns. R&D and innovation (planned but unrealized innovation), showed stronger relationships with growth intentions than any other form of growth. Innovations that were new to firm only, were related to most growth measures. Industry level innovation was correlated with growth intentions and perceived growth, but not with objective measures. Just being innovation active was not correlated with any form of growth. However, actual commercialisation of innovation was correlated with all forms of growth.
Guidelines From the analysis a number of guidelines emerge. First, one innovation input measure, R&D intensity, was investigated. Inputs are the furthest removed from actual firm growth in terms of time, and this was evident in the absence of any effect on growth. Therefore, it is suggested that researchers ensure that there is an appropriate time lag in data when measuring the effect of R&D intensity on growth. Further, as typical with time lagged variables, it is possible that a number of mediating variables may be important to fully explain this relationship. Additional research into the type of mediators is therefore warranted. However, Adams et al. (2006) argue that R&D intensity is just on input into the process, and therefore not ideal as a single proxy. In particular, they question its usefulness for small firms and
14
service industries. Instead, they argue for the use of funding committed to the innovation process as a better measure. Such expenditure on innovation can then be expressed as a proportion of sales. In similar vein, the knowledge management measure, collaboration, also seem to have a time lag effect with growth, with the potential that other forms of innovation may be acting as mediators in this relationship (Gronum, et al. 2012).
Second, a number of commercialisation type variables were investigated. Process innovation was highly correlated with growth intentions and perceived growth, which seem to illustrate that managers view this as an effective way to score quick runs. In contrast, product/service innovation is used for the purpose of growing the firm. It also takes longer to show a strong effect on sales growth, indicating its importance as a longer term strategy. Innovation breadth variables, and in particular novel innovation breadth which is negatively correlated with export, are not very important to growth. This may be attributed to the inability of resource strapped small firms to deal with innovation across a broad range of areas. From these results it is evident that different innovation and growth measures are better matched. It is the purpose of this paper, in a more developed form, to make definitive suggestions regarding such matches.
Third, when comparing the suitability of different growth variables, it seems that perceived growth and sales growth are most important, both in terms of variance explained as well as number of variables with which they have significant correlations. For employment growth, process innovation seemed to be most important, which is counter intuitive. From the regression analysis it became clear that process innovation seems to be significant to perceived growth, while product/service innovation is important to growth intentions.
These findings must be viewed against a number of limitations. First, growth can only be measured for firms that survive (Chan, 1998; Mata, 1994) and with slow growing firms more likely to disappear in time (Dunne & Hughes, 1990), thus research is hampered by the use of surviving firms. Second, the data presented here include measurement at two points in time. However, at least three points will be ideal for future research (Ployhart & Vandenberg, 2010). Longitudinal research designs emphasize change in focal variables over a period of time. Two aspects of change are important, namely within-unit change or growth trajectories, and changes in the differences between units. While appropriate spacing and management of attrition needs further discussion, data collection to this effect is currently in progress. Last, 15
and most important, this research is still very exploratory and further analysis is needed to fully understand the relationships presented here.
CONCLUSION As noted, the inconsistent availability and use of measures for growth and innovation has hampered theoretical development in these fields. This paper highlights a number of these issues and provides some guidelines for researchers when designing models containing innovation and growth measures. Several implications were noted. Theoretically, the importance for small firm researchers to carefully consider their choice of growth measures is highlighted. Independent variables (such as different types of innovation constructs) should be matched with a growth dependent variable that is appropriate in terms of temporality, nature and practicality. To illustrate, there is no point to argue that R&D has no effect on sales growth if it has not generated sales yet. Some practical insights are also identified. From a strategy point of view, it is important to also set growth goals and to follow through on those. The results also indicate that imitative innovations have a shorter path to market, but that the expectation to grow from new-to-industry innovations are higher. The tensions between short and long term expectations are further explored.
REFERENCES Adams, R., Bessant, J. & Phelps, R. (2006). Innovation management measurement: A review. International Journal of Management Reviews, 8(1), 21-47. Ajzen, I. (1988). Attitudes, personality, and behavior. Chicago, IL: Dorsey Press. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211. Amabile, T.M., Conti, R., Coon, H., Lazenby, J. & Herron, M. (1996). Assessing the work environment for creativity. Academy of Management Journal, 39, 1154–1184. Audretsch, D.B. (1991). New firm survival and the technological regime. Review of Economics and Statistics, 60(3), 441-450. Audretsch, D.B. (1995). Firm profitability, growth, and innovation. Review of Industrial Organization, 10(5), 579-588. Autio, E. & Acs, Z. (2010). Intellectual property protection and the formation of entrepreneurial growth aspirations. Strategic Entrepreneurship Journal, 4, 234-251. Barney J. (1996). The resource-based theory of the firm. Organizational Science, 7, 469481. Baldwin, J. & Gellatly, G. (2003). Innovation strategies and performance in small firms, Edward Elgar, Cheltenham, UK. Bottazzi, G. & Secchi, A. (2003). Why are distributions of firm growth rates tent-shaped? Economic Letters, 80, 415-420. Botlazzi, G. & Secchi, A. (2006). Explaining the distribution of firm growth rates. RAND Journal of Economics, 37(2), 235-256. 16
Bougrain, F. & Haudeville, B. (2002). Innovation collaboration and SMEs internal research capacities. Research Policy, 31, 735–747. Brenner, M.S. (1994). Practical R&D project prioritization. Research Technology Management, 37, 38–43. Burns, T.R. & Stalker, G.M. (1961). The Managementof Innovation. London: Tavistock. Capon, N., Farley, J.U., & Hoenig, S. (1990). Determinants of financial performance: a meta-analysis. Management Science, 36(10), 1143-1159. Cebon, P. & Newton, P. (1999). Innovation in firms: towards a framework for indicator development. Melbourne Business School Working Paper 99-9. Chakravorti, B. (2004). The new rules for bringing innovations to market. Harvard Business Review, 82, 58–67. Chan, D. (1998). The conceptualization and analysis of change over time: An integrative approach incorporating longitudinal mean and covariance structures analysis (LMACS) and multiple indicator latent growth modelling (MLGM). Organizational Research Methods, 1, 421-483. Chan, Y.E., Bhargava, N. & Street, C.T. (2006). Having arrived: the homogeneity of highgrowth small firms. Journal of Small Business Management, 44(3), 426-440. Chiesa, V., Coughlan, P. & Voss, A. (1996). Development of a technical innovation audit. Journal of Product Innovation Management, 13, 105–136. Clark, B. (2002). Measuring performance: The marketing perspective. In A. Neely (Ed.), Business performance measurement: Theory and practice. Cambridge: Cambridge University Press. Coad, A. (2007). Testing the principle of ‘growth of the fitter’: The relationship between profits and firm growth. Structural Change and Economic Dynamics, 18, 370-386. Coad, A. & Rao, R. (2008) Innovation and firm growth in high-tech sectors: A quantile regression approach. Research Policy, 37, 633-648. Cohen, W.M. & Levinthal, D.A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35, 128–152. Cooper, A. C. (1979). Strategic management: new ventures and small business. In D. E. Schendel & C. W. Hofer (Eds.), Strategic management: a new view of business policy and planning (pp. 316-327). Boston: Little Brown. Cooper, R.G. (1988). Predevelopment activities determine new product success. Industrial Marketing Management, 17(3), 237-247. Cooper, R.G., Edgett, S.J. & Kleinschmidt, E.J. (2001). Portfolio Management for New Products, 2nd edition. Cambridge, MA: Perseus. Cosh, A., Fu, X. and Hughes, A. (2010). Organisation structure and innovation performance in different environments. Small Business Economics, DOI 10.1007/s11187010-9304-5, 17 pages. Covin, J. G. & Slevin, D. P. (1989). Strategic management of small firms in hostile and benign environments. Strategic Management Journal, 10, 75-87. Damanpour, F. (1991). Organizational innovation: a meta-analysis of effects of determinants and moderators. Academy of Management Journal, 34, 555–590. Damanpour, F. & Evan, W.M. (1984). Organizational innovation and performance: the problem of "organizational lag’. Administrative Science Quarterly, 29(3), 392-409. Davidsson, P. (1991). Continued entrepreneurship: Ability, need, and opportunity as determinants of small firm growth. Journal of Business Venturing, 6, 405-429. Deeds, D.L. (2001). The role of R&D intensity, technical development and absorptive capacity in creating entrepreneurial wealth in high technology start-ups. Journal of Engineering and Technology Management, 18, 29–47.
17
Delmar, F., Davidsson, P. & Gartner, W.B. (2003). Arriving at the high-growth firm. Journal of Business Venturing, 18, 189-216. Dess, G. G., Lumpkin, G. T. & Covin, J. G. (1997). Entrepreneurial strategy making and firm performance: Tests of contingency and configurational models. Strategic Management Journal, 18(9), 677-695. Dess, G.G. & Robinson, R.B. Jr. (1984). Measuring organizational performance in the absence of objective measures: The case of the privately-held firm and conglomerate business unit. Strategic Management Journal, 5, 265-273. Dobbs, M. & Hamilton, R.T. (2007). Small business growth: Recent evidence and new directions. International Journal of Entrepreneurial Behaviour & Research, 15(5), 296-322. Dunne, P. & Hughes, A. (1990). Small firms, age, growth and survival in the 1990s, Paper presented at the 17th EARIE Conference, Lisbon. Evans, D.S. (1987). Tests of alternative theories of firm growth. Journal of Political Economy, 95(4), 657-674. Freel, M.S. (2000). Do small innovating firms outperform non-innovators? Small Business Economics, 14, 195-210. Freel, M.S. (2007). Are small innovators credit rationed? Small Business Economics, 28, 23-35. Garnsey, E., Stam, E. & Heffernan, P. (2006). New firm growth: exploring processes and paths. Industry and Innovation, 13(1), 1-20. Geroski, P.A. (1998). The growth of firms in theory and practice. DRUID conference on Competencies, Governance and Entrepreneurship. 14 pages. Gilbrat, R. (1930). Les inégalités économiques. Paris: Librairie du Receuil Sirey. Glancey, K. (1998). Determinants of growth and profitability in small entrepreneurial firms. International Journal of Entrepreneurial Behaviour and Research, 4(1), 18-27. Gopalakrishnan, S. (2000). Unraveling the links between dimensions of innovation and organizational performance. The Journal of High Technology Management Research, 11(1), 137-153. Greve, H.R. (2003). A behavioral theory of R&D expenditures and innovations: Evidence from shipbuilding. Academy of Management Journal, 46, 685–702. Hartarska, V. & Gonzalez-Vega, C. (2006). What affects new and established firms’ expansion? Evidence from small firms in Russia. Small Business Economics, 27, 195-206. Hubbard, G. & Bromiley, P. (1995). Researchers and top managers: How do they measure firm performance? Jovanovic, B. (1982). Selection and evolution of industry. Econometrica, 50(May), 649670. Kivimäki, M., Lansisalmi, H., Elovainio, M., Heikkila, A., Lindstrom, K., Harisalo, R., Sipila, K. & Puolimatka, L. (2000). Communication as a determinant of organizational innovation. R&D Management, 30, 33–42. Lebas, M. & Euske, K. (2002). A conceptual and operational delineation of performance. In A. Neely (Ed.), Business performance measurement: Theory and practice. Cambridge: Cambridge University Press. LeBrasseur, R., Blanco, H. & Dodge, J. (2006). Growth intentions of owner-managers of young microfirms. New England Journal of Entrepreneurship, 9(1), 9-20. Lee, M., Son, B. & Lee, H. (1996). Measuring R&D effectiveness in Korean companies. Research-Technology Management, 39, 28–31. Lööf, H. & Heshmati, A. (2006). On the relationship between innovation and performance: A sensitivity analysis’, Economics of Innovation and New Technology, 15(4/5), 317-344.
18
Lumpkin, G. T. & Dess, G. G. (1996). Clarifying the entrepreneurial orientation construct and linking it to performance. The Academy of Management Review, 21(1), 135-154. Lyon, D. W., Lumpkin, G. T., & Dess, G. G. (2000). Enhancing entrepreneurial orientation research: operationalizing and measuring a key strategic decision making process. Journal of Management, 26(5), 1055-1085. Mansfield, E. (1962). Entry, Gibrat’s law, and the growth of firms. American Economic Review, 52(5), 1023-1051. Mata, J. (1994). Firm growth during infancy. Small Business Economics, 6, 27-39. Miles, R.E. & Snow, C.C. (1978). Organization Strategy, Structure and Process. New York: McGraw-Hill. Miller, D. & Friesen, P.H. (1982). Innovation in conservative and entrepreneurial firms: two models of strategic momentum. Strategic Management Journal, 3, 1–24. Nelson, R.R. & Winter, S.G. (1978). Forces generating and limiting concentration under Schumpeterian competition. The Bell Journal of Economics, 9, 524-548. OECD 2005, Oslo Manual: Guidelines for Collecting and Interpreting Innovation Data, 3rd ed., Joint Publication of the OECD and the Statistical Office of the European Communities, Paris. Pagano, P. & Schivardi, F. (2003). Firm size distribution and growth. Scandinavian Journal of Economics, 105(2), 255-274. Parry, S.B. (1996). Measuring training’s ROI. Training and Development, 50(5), 72-77. Parthasarthy, R. & Hammond, J. (2002). Product innovation input and outcome: moderating effects of the innovation process. Journal of Engineering and Technology Management, 19, 75–91. Penrose, E.T. (1959). The theory of the growth of the firm. New York: Wiley. Penrose, E. (1960). The growth of the firm – a case study: The Hercules Power Company. Business History review, XXXIV, 1-23. Ployart, R.E. & Vandenberg, R.J. (2010). Longitudinal research: The theory, design, and analysis of change. Journal of Management, 36(1), 94-120. Porter, M.E. & Ketels, C.H.M. (2003). UK Competitiveness: Moving to the Next Stage. DTI Economics Paper No 3, URN 03/899. Prais, S. (1974). A new look at the growth of industrial concentration. Oxford Economics Papers, 26(2), 273-288. Roper, S. (1999). Modelling small business growth and profitability. Small Business Economics, 13(3), 235-252. Rothwell, R. (1992). Successful industrial innovation: critical factors for the 1990s. R&D Management, 22, 221–239. Schumpeter, J.A. (1934). The Theory of Economic Development: An Inquiry into Profits, Capital, Credit, Interest, and the Business Cycles, Harvard University Press, Cambridge. Schumpeter, J.A. (1950). Capitalism, Socialism and Democracy, 3rd ed., Harper Torchbooks, New York. Simon, H.A. (1968). On judging the plausibility of theories. In B. Van Rootsellar and J.F. Staal, eds. Logic, Methodology and Philosophy of Sciences Vol. III. Amsterdam: North Holland. Stock, G.N., Greis, N.P. & Fischer, W.A. (2001). Absorptive capacity and new product development. Journal of High Technology Management Research, 12, 77–91. Storey, D. (1997). The ten percenters, fast growing SMEs in Great Britain. Deloitte, Touche, Tohmatzu International. Thompson, L. (2003). Improving the creativity of organizational work groups. Academy of Management Executive, 17, 96–109.
19
Verhaeghe, A. & Kfir, R. (2002). Managing innovation in a knowledge intensive technology organisation (KITO). R&D Management, 32, 409–417. Verreynne, M. (2006). Strategy-making process and firm performance in small firms. Journal of Management and Organisation. 12(3), 209-222. Weinzimmer, L.G., Nystrom, P.D. & Freeman, S.J. (1998). Measuring organizational growth: Issues, consequences and guidelines. Journal of Management, 24(2), 235-262. Wiklund, J. & Shephard, D. (2003). Aspiring for, and achieving growth: the moderating role of resources and opportunities. Journal of Management Studies, 40(8), 1919-1941. Wiklund, J., Davidsson, P. & Delmar, F. (2003). What do they think and feel about growth? An expectancy-value approach to small business managers’ attitude toward growth. Entrepreneurship Theory and Practice, 27(Spring), 247-269. Wong, M.H., Page, D., Abello, R. & Pang, K.P. (2007). Explorations of innovation and business performance using linked firm-level data, Research Paper, Cat. no. 1351.0.55.020. Australian Bureau of Statistics: Canberra. Yoon, E. & Lilien, G.L. (1985). A new product launch-time decision model. Journal of Product Innovation Management, 3, 134–144. Zahra, S.A. & George, G. (2002). Absorptive capacity: a review, reconceptualization, and extension. Academy of Management Review, 27, 185-203.
20
Table 2: Correlations R&D Intensity (1) Process Innovation (bin) (2) Product/service innovation (bin) (3) Sales from novel innovation (4) Non-novel Innovation breadth (5) Novel Innovation breadth (6) Significantly improved pdt/serv (7) New products/services(8) Commercialisation (bin) (9) Innovation (bin) (10) Collaboration (11) Firm Size (12) Firm age (13) Growth intention (bin) (14) Perception of growth (15) Sales growth (16) Export growth (17) Productivity growth (18) Employment growth (19)
Means 13.14 .6006
SD 88.38 .49
n 377 1540
(1) 1 .037
(2) -.001 1
(3) .062 .647**
(4) .120* .049
(5) -.037 .628**
(6) .066 .505**
(7) .063 .057
(8) .127* .026
(9) .019 .585**
(10) .030 .059*
(11) .017 .049
(12) -.226** .261**
(13) -.150** .072**
(14) .049 .189**
(15) .028 .154**
(16) .097 .120**
(17) .052 .133
(18) .039 .022
(19) .014 .113**
.5336
.499
1578
.054
.647**
1
.262**
.491**
.597**
.260**
.208**
.641**
.049
.028
.265**
.066**
.199**
.148**
.101**
-.009
.034
.092**
36.53
31.41
1058
.010
.045
.212**
1
.033
.193**
.862**
.809**
.490**
-.058
-.073*
-.050
-.065*
.070*
.158**
.132**
.106
.051
.091*
1.12
1.51
2107
-.009
.550**
.480**
.055
1
-.128**
.071*
-.013
.403**
.040
.057**
.205**
.008
.146**
.082**
.086**
.090
.029
.072**
.96
1.56
2107
.111*
.445**
.522**
.176**
-.205**
1
.156**
.185**
.432**
.057**
.039
.183**
.097**
.132**
.119**
.052
-.028
.025
.047
20.91
21.41
1116
-.013
.064
.197**
.832**
.095**
.114**
1
.496**
.240**
-.042
-.061*
-.025
-.042
.039
.167**
.130**
.116
.054
.090*
17.20
19.286
1142
.027
.010
.109**
.794**
-.026
.138**
.324**
1
.653**
-.054
-.063*
-.063*
-.061*
.074*
.131**
.081*
.016
.051
**
**
**
**
**
.641
.371
.372
**
.348
**
.075
*
.45
.498
2107
-.011
.585
.6744 1.41 131.40 25.38 1.796
.46870 .491 1553.1 23.94 .40
2107 2037 2107 2079 2061
.000 -.008 -.014 -.021 .042
.059* .049 .027 .104** .189**
.049 .028 .032 .068** .199**
-.047 -.051 -.010 -.078* .056
.015 .046* .013 .035 .156**
.034 .026 .036 .084** .103**
-.035 -.027 .016 -.070* .035
3.81
1.462
1910
.009
.159**
.146**
.155**
.084**
.134**
74.87 62.18 58.64
736.74 224.68 546.04
1414 185 1307
-.009 -.023 -.012
.037 -.007 .040
-.006 -.044 .012
-.025 .087 -.030
.015 -.008 .015
13.86
54.48
1421
-.001
.069*
.071*
.052
.084**
.400
**
1
.026
.087
-.010 -.034 -.040 -.068* .069*
.026 .087** .015 .047* .170**
1 .006 .032 .057** .002
.006 1 .002 .033 .097**
.167**
.106**
.082**
.026
.016 -.003 .019
-.009 .139 -.016
-.034 -.016 -.033
.018 -.052 .035
.007
.053
.018
.075**
.228
**
.100
**
.125
**
.051
-.059
.070
*
.101**
.030
.170
.051* .094** 1 .111** .033
.063** -.016 .351** 1 .003
.002 .097** .244** -.079** 1
.024 -.006 .152** -.021 .234**
.001 .019 .103** -.090** .154**
-.105 -.103 .070 .061 .188*
-.033 .036 -.028 -.037 .045
.021 .003 .205** -.040 .136**
-.020
.000
.017
.209**
1
.324**
.205**
.122**
.274**
.026 -.109 .028
.055* -.051 .043
.072** -.006 .081**
.061* -.055 .045
.011 .120 .008
.029 .178* .011
1 .021 .879**
.466** 1 .014
.576** .262** 1
.501** .183* -.289**
.006
.012
-.016
-.077**
.069**
.223**
.103**
.188*
-.010
1
*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). Pearson’s correlations in bottom half, Spearman’s correlations in top half
21
