Procedural Justice, Not Absorptive Capacity, Matters in Multinational Enterprise ICT Transfers

Birkinshaw (1996) already noted more than 15 years ago that no definitive list exists of foreign subsidiaries in Canada, and this was still the case in 2008.

The website addresses with linkages for each of the mentioned lists are:

Items used in our questionnaire are available upon request.

For brevity, the measure of organizational factors and conditions ‘organizational structure’ (formalization-centralization) was omitted, as this measure had no significant direct or interaction effects.

Titles of positions with the number of respondents having that title in parentheses include President (3), CIO (8), Vice President (5), Director (12), General Manager (4), Manager (34), other (11), and not identified (9). The parent companies involved, represented eleven home countries: US (53), Germany (6), Japan (6), UK (6), Switzerland (3), Hong Kong (2), France (2), Australia (1), Netherlands (1), Spain (1), Norway (1) and not reported (4). This resulting distribution is largely consistent with the regional distribution of foreign direct investment stocks in Canada, with the United States still largely dominating these stocks, and the various European countries together representing a distant, second investment region. We also coded each company’s main industry (two-digit SIC) where available: Resources (13), manufacturing (36), services (32), or unknown (5). This distribution is also largely consistent with the broader population of 746 foreign-owned subsidiaries in Canada.

The factor analysis for the absorptive capacity variable identified one element that did not contribute. This element was omitted from the final absorptive capacity variable, with eleven elements remaining, resulting in a Cronbach’s alpha of 0.847. The factor analysis for the procedural justice variable indicates that all five elements contributed, with a Cronbach’s alpha of 0.83. The factor loading for one element was slightly below the established cut-off value but its continued inclusion was considered to be necessary for consistency with the foundational theoretical concepts. The factor analysis for the knowledge tacitness variable identified eight elements that did not contribute and four elements below the factor loading cut-off value. The resultant knowledge tacitness variable was reduced to the ten remaining items with a resulting Cronbach’s alpha of 0.766. The factor analysis for the performance outcome variable identified one element that did not contribute. The resulting performance outcome variable was reduced by this element with ten remaining elements and a Cronbach’s alpha of 0.874.

Mean-centering of all variables was considered as the literature shows that mean-centering can reduce the covariance between the linear and the interaction terms, thereby suggesting that applying such a statistical technique reduces collinearity. However, Echambadi and Hess (2007) analytically demonstrate that mean-centering neither changes the computational precision of parameters, the sampling accuracy of main effects, simple effects, interaction effects, nor the R². Therefore mean-centering was not applied.

The factor loadings of each scale item from the factor analysis are available upon request.

Only 11 of the respondents were from countries with a Hofstede power distance score more than one point higher than Canada’s.