Strategic action in network industries: an empirical analysis of the European mobile, страница 13

Notes:

Dependent variable: Log [Prob(cooperative move)/Prob(competitive move)].

Standardized coefficients reported.

N=199 observations (strategic actions).

Results control for 12 country dummy variables (coefficients not shown).

Represent significant results at 0.10, 0.05 and 0.01 confidence level respectively.

customers of other networks. Finally, once the effect of penetration has been discounted, more strategic actions are taken in collaboration with other firms in the industry in the early stages, because of the need to set up standards and obtain licenses at that time, although as time passes operators are eventually forced to compete head-to-head.

The empirical results confirm the important role of strategic agreements and alliances in network industries. Over 35% of the strategic actions in the sample were cooperative moves. As expected, the likelihood of a strategic action being cooperative is determined to a large extent by the stage in the evolution the industry, well as the degree of concentration and the level of market penetration. However, to get a better understanding of the conduct of firms in the mobile phone industry, we can look beyond the cooperation/competition dichotomy. Mobile phone operators engage in infrastructure operation, in promotion and contract management, and in service provisioning activities. The 199 actions were also categorized according to these three types of primary activity, as shown in Exhibit 1. Cooperative moves are frequent when it comes to creating the network infrastructure and layering new services over and above an established network, but are not usual when it comes to actions related to promotion, which are more likely to be taken by firms independently.

These three types of strategic action are analyzed in greater depth in Exhibit 3. Again, a multinomial logit model is used to investigate the circumstances under which each type of strategic action is more likely to occur relative to the other two. In this case, there are three independent comparisons among the three types of strategic actions concerned. However, to make it easier to interpret these results, six logit regressions are reported as well as the corresponding hypotheses.

In the first set of two regressions, the log odds of a service-related move (versus a promotion-related move in the first column as denominator and an infrastructurerelated move in the second column) is regressed over Market Penetration,

Exhibit 3

Likelihood of Strategic Actions: Infrastructure, Promotion, and Services Logit Regression Analysis