5.3.1 The distinct levels of compliance

As clarified above in the description of the previous studies performed on code compliance, different levels of compliance (or in other words different ways of rating code compliance) must be distinguished so proper research can be conducted. Therefore, table 5.3.1 shows and explains the defined levels of compliance before the hypotheses are formulated further in this section.

In this chapter and in table 5.3.1 the different levels of compliance are numbered (I to III) and named (applied, applied and explained, applied and explained sufficiently) for the purposes of the underlying empirical research. The names have been chosen to provide a clear understanding of what is measured and, for the time being, a possible terminology discussion is set aside. In short, in the case of I.a and I.b all the code provisions applied, as declared by the company, are measured. II.a and II.b concern the provisions applied as well as the deviations mentioned by the company with any kind of explanation, meaning that only the deviations, for which no explanation at all is provided, have been excluded from the level of compliance (see table 5.3.1a category no explanation). So a higher level of compliance is expected than for I, since it concerns applied and explained. In fact the remaining percentage up to one hundred per cent is of importance, since this concerns the non-compliance without explanation. For the underlying study, level III is most important since the quality of the explanations is taken into account (see chapter 6). One hundred per cent 'appliance' of the code (level of compliance I) is not desired since the 'one size does not fit all' and 'made to measure' approach are believed in. Deviations from the code are allowed as long as the explanation is sufficient, hence one must aim for one hundred per cent in level of compliance III (applied and explained sufficiently). The explanations are distinguished into six categories based on the subdivision of Arcot and Bruno as used in their research (Arcot and Bruno 2007, p. 12). They distinguish the following taxonomy: no explanation (type 1), general (type 2), inline (type 3), limited (type 4), transitional (type 5) and genuine (type 6). Table 5.3.1a elucidates this subdivision further (see also Annex II).

Arcot and Bruno give no judgment regarding the validity of the explanations provided, but construct a corporate governance score based on their taxonomy (Arcot and Bruno 2007, p. 13) (0 points for no explanation to 5 points for a genuine explanation). The underlying study considers the type 1, type 2 and type 3 explanations insufficient. Although type 4 and type 5 need further improvement, together with type 6 they are considered 'sufficient'. Therefore, levels of compliance III.a and III.b consist of the provisions complied with, as well as the deviations with a sufficient explanation (types 4 to 6, categories limited, transitional or genuine)(applied and explained sufficiently). To sum up, level of compliance III measures the code provisions complied with together with the sufficient explanations in the case of non-compliance. Levels I to III give a comprehensive overview of the compliance with corporate governance codes as realised so far in the countries and time period under research. An explanation of the difference between the a-levels and b-levels remains; serial numbers I.a, II.a and III.a are for the level of compliance of the countries under review per se and serial numbers I.b, II.b and III.b compare the levels across all the five countries under review simultaneously. In order to be able to do so, only the code provisions comparable for all the countries are included in the level of compliance, which implies that a certain number of code provisions are excluded from this part of the research as explained further in section 5.5 below.

The reason for the distinction in these levels of compliance (applied, applied and explained, applied and sufficiently explained) - besides the fact that, above all, this provides a more in-depth empirical research - is to be able to compare the levels of compliance between the countries. The levels must be qualified and defined before the empirical research and formulation of the hypotheses given below. The research - as will be explained further in section 5.4 -consists of a univariate analysis (descriptive statistics: a description of the characteristics of the observation results (Van Dalen and De Leede 2002, p. 32)), a bivariate analysis and regression models (multivariate analysis). The hypotheses as formulated below and issues to be researched are placed in the same order (univariate, bivariate and multivariate analysis).