The responding companies consist of manufacturers of paper, suppliers of raw material or equipments, and those which may be categorized as both supplier as well as manufacturer. The sample companies have been labeled as small or large based on the number of their full-time permanent employees. Companies with 100 or less employees are called small, and the ones with more than 100 are referred to as large.
Compatible with the Oslo Manual of OECD (2005)and based on the evidence from the literature (Wan et al. 2005; De Jong and Vermeulen 2006), a set of six questions was developed to measure the magnitude and novelty of innovation commercialized by the firms. The questions ask about the number of technologically new or improved processes or products (raw material in case of suppliers) implemented or introduced by a firm first time to itself or to the industry during the last three years (April 2006 to March 2009).
For computation convenience, an equal weightage has been assigned to each item, and hence, the simple average of the scores of a company on these six items represents its technological innovativeness (INNO). Such scores have ranged between 0.33 and 5.83 with a mean of 3.78. For further analysis, the companies with average score of 3.78 or less are designated as poor innovators (denoted by numeral 1); those with higher values, as good innovators (denoted by numeral 2). As shown in Figure 1, around 62% of the manufacturers have been assessed as good innovators, whereas there are more poor innovators (55%) in the others category. (Those who are suppliers-cum-manufacturers have been grouped with suppliers to form a single category, others). It could be understood that due to their nature of operations, the manufacturers might have scored higher than the suppliers on product and process innovation. This appears as one strong reason for them to outperform the others on the overall innovativeness.
Technological innovation and performance
Another set of questions measures the effect of innovations on company sales (SALE) and on other important determinants of company performance such as production time (TIME), production cost (COST), production flexibility (FLEX), and production quality (QUAL) (Neely et al. 1995; Kaplan and Norton 1992). Since the study deals with the paper industry, environmental factors such as emission of hazardous fumes (EMSN) and disposal of solid waste (DISP) are also included in the study as another two important (dependent) variables. This is evidenced from the literature that researchers have examined innovation as a dependent (Wan et al. 2005) as well as an independent variable (Lin and Chen 2007; Mansury and Love 2008). The present paper studies innovation as an independent variable.
Respondents were requested to report the effect of technological innovations, which they have introduced during the mentioned period of three years, on these performance parameters. A five-point Likert scale (Sawang 2006) ranging from improved significantly through improved moderately, no effect, worsened moderately to worsened significantly was proposed to them for this purpose. Bivariate correlation analysis is then run on the data obtained on company innovativeness.
Table 1 shows that whether performance is in the form of outputs (such as sales), production factors (such as production time, cost, flexibility, and quality) or environmental hazards (such as emissions and disposal of wastes), innovativeness is likely to have its impact on it. Except for production flexibility, all performance indicators have a significant correlation with the company’s innovativeness. Sales, the most commonly used measure of performance, shows a highly significant positive relationship with innovative-ness. The ability of a company to quickly incorporate the changes required in volume and deign (flexibility) and improve the quality has been found positively linked with how innovative the company is. However, the strength of relationship between innovativeness and flexibility is not significant. Since greater flexibility is more a factor of process innovation, this insignificant correlation may be explained by examining the companies’ scores on process and product innovations separately. This argument is also supported by the literature, which reports that considering product and process innovation together or separately does influence the effect of innovation on performance (Michie and Sheehan 2003). The correlation coefficients also indicate that the greater the company innovativeness, the lower is the production time, cost, emission, and waste. This may be inferred that innovative companies not only perform better in competitive terms, but also are less harmful to the environment.
To further investigate the relationship between innovation and performance, MANOVA is carried out with company innovativeness as the predictor variable, whereas changes occurred due to innovation in the sales, time, cost, quality, flexibility, emissions, and waste production as dependent variables. The results are shown in Tables 2 and 3.
Descriptive statistics indicate that the mean scores on the factors such as sales, quality, and flexibility are higher for the good innovators than those for the poor innovators. Conversely, the factors which are likely to be reduced as a result of innovations, such as time, cost, emissions, and disposal of waste, have shown higher means for the poor innovators. This can also be observed from the standard deviation column that good innovators have been more consistent than their poor counterparts, except in cases of quality and flexibility, where wider dispersions are reported for the good innovators. The overall situation signals that innovation is likely to affect all variables positively.
The Box’s test of equality of covariance matrices is conducted to validate the assumption of homogeneity before proceeding further. The result (Box’s M = 36.229; p = .530) suggests that the assumption is valid, and hence, the multivariate tests (Table 3) are reliable (Field 2005). The significance column and INNO row of Table 3 indicates that innovations have a significant effect on performance. However, the results do not tell anything in detail. To investigate this effect with reference to the individual variables and their combinations, DFA is applied. Since there are only two groups (good innovators and poor innovators) involved in this analysis, there has to be a single discriminant function variate. The initial outcome of the DFA reveals that this variate is significant (Wilk’s lamda = .361; p = .000). Finally, the structure matrix (Table 4) is obtained, which explains the relationship between the dependent variables and the variate. The values (canonical variate correlation coefficients) in this matrix indicate the relative contribution of each variable (and its direction, positive or negative) in differentiating the two groups, poor innovators and good innovators, from each other.