The multifaceted role of information and communication technology in innovation: evidence from Chinese manufacturing firms

Abstract

Using a sample of 1397 manufacturing firms in China, this study explores how information and communication technology (ICT) stimulates innovation performance through interaction with investment in internal and external knowledge. The results of this study reveal that the adoption of ICT not only contributes to innovation by directly enhancing its output, but also affects the creation of knowledge along two different paths in Chinese manufacturing firms. Specifically, ICT mediates research and development to foster innovative sales and it is an effective moderator in leveraging externally acquired knowledge.

Keywords:

• ICT
• innovation
• moderating effect
• mediating effect
• Chinese manufacturing firms

View correction statement:

Erratum

ORCID

Jizhen Li http://orcid.org/0000-0003-1940-5633

Notes

1. Data source: http://www.itu.int/net/pressoffice/press_releases/2013/41.aspx#.Vz0YL-x7LX4.

2. Innovation-oriented ICT refers to investments in ICT that are specifically deployed for the purpose of supporting knowledge-creation processes such as hardware and software, technical equipment, business components, and networks for knowledge acquisition, integration, learning, and development.

3. Llorca (2002) analyses the impact of process innovations on productivity growth using number of product and process innovations. However, the exact number of process innovations is not available and is proxied by a categorical variable. Another study was conducted by Peters (2008) in which the output or success of process innovation is measured as the share of cost reduction in unit costs due to process innovations, which can therefore be interpreted as cost-weighted innovation counts. This indicator and its corresponding definition were given in the CIS Germany.

4. The mediating effect of ICT here refers to partial mediation as discussed by Baron and Kenny (1986). It is different from complete mediation that defines the mediator as a must-be-satisfied condition to form the relationship between the explanatory variable and the dependent variable. When the complete mediator is controlled, the explanatory variable will no longer affect the dependent variable when the complete mediator is controlled. The partial mediation is the case in which the path from the explanatory variable to the dependent variable is reduced in an absolute size but is still different from zero when the mediator is controlled. In the current setting, ICT partially mediates R&D by increasing its innovation effects while R&D itself also produces directly innovation impacts.

5. Policy Studies Institute, University of Westminster. ‘Small Firms Innovation’. Retrieved 27 May 2010 from http://en.wikipedia.org/wiki/Product_innovation.

6. Enterprise Surveys (http://www.enterprisesurveys.org), The World Bank.

7. Firms with fewer than five employees (3.6%) have been deleted from our sample. Firms that do not aim at innovating are also filtered out from sample (Mohnen and Roller, 2005; D’Este, Iammarino, Savona, and von Tunzelmann, 2008; Savignac, 2008). Specifically, firms not aiming at innovating do not carry out innovation activities at all and more likely to ignore obstacles to innovation. We assume that non-innovative firms which experienced failure of innovation would encounter at least one of four types of obstacles. Hence, we drop firms which have no new or updated products/services and, at the same time, report none of the chosen obstacles is important. It accounts for 6.2% of the total sample.

8. The main drawback of the innovation input variable is that it is a flow variable and was observed only in the year 2012 – in other words, the same year in which we observed the innovation output. This means that the lag between investment in research and actual innovation is ignored, along with the lag between product innovation and market acceptance. However, Griliches (1998) reports of some scattered evidence from questionnaire studies that such lags are rather short in the manufacturing since most research expenditure are related to development and applied topics.

9. See Baron and Kenny (1986) for more details on the detailed computation of moderate effects.

10. The detailed calculation involved in the estimating of some equations not presented in this manuscript, such as a simple equation without ICT adoption variables. The estimation results of this simple equation will be available upon request.

Additional information

Funding

The authors would like to thank the generous funding from the National Natural Science Foundation of China awarded to Zhongjuan Sun (Project No. 71602127) and Jizhen Li (Project No. 71573148). This paper is also supported by the China MOE Project of Key Research Institute of Humanities and Social Sciences at Universities (Project No. 15JJD630003) and the project of Capital University of Economics and Business (Project No. 00191654490317).