No se

--decreases the risk of unexpected losses and the risk of damage to the bank's reputation.

4.2 Empirical comparative approach of the Romanian control environment framework by reference to international COSO and CoCo models

In order to achieve our aim--to identify on which of the international internal control models (COSO or CoCo) is based the Romanian internal control system framework, weconducted an empirical study based on an analysis with character of comparison between the three sets of regulations (the two international models an the Romanian one).

We have started from the components of the control environment as these are settled by COSO and the criteria of control as these are defined by CoCo, trying to establish the link between them. Thus, we have identified a seriesof issues regarding control environment, which we organized within seven main topics as follows: (1) integrity and ethical values/standards, (2) commitment to competence, (3) board of directors and audit committee, (4) management's philosophy and operating style, (5) organizational structure, (6) assignment of authority and responsibility, and (7) human resource policies and practices. Startingfrom these topics, we proceeded to compare the control environment as it appears within the three frameworks. Thus, we have allocated the 1 or 0 value for each possible and/or existent requirement withit at least one of the considered regulation, where the 1 value shows that the requirement exists within that framework, and 0 value is given for the situation when the requirement is not found withinthe considered framework.

In the above table (Table I.) there are presented the 1 and 0 values that have been allocated to each requirement of the seven components analyzed.

In order to achieve the proposed comparison, we have considered that the best analysis, in case of this type of approach, is represented by the nonparametric correlation and the association degree between two or more thantwo considered variables. Thus, we have used for our research the Jaccards' association coefficients, which have been used before in studies focused on comparisons between different sets of regulations. On the other hand, the two Jaccard's coefficients offer the possibility of quantifying both the association degree and the dissimilarity degree between different sets of requirements for controlenvironment taken into consideration for analysis.

So, in order to dimension the compatibility degree or, in other words, the association between two or more internal control systems, the calculation formula for the Jaccards' coefficient shows as follows:

[S.sub.ij] = a / (a + b + c)

and

[D.sub.ij] = (b + c) / (a + b + c)

where:

--[S.sub.ij]--represents the similarity degreebetween the two sets of analyzed control environment frameworks;

--[D.sub.ij]--represents the degree of dissimilitude or diversity between the two sets of analyzed control environment frameworks;

--a--represents the number of elements which take the 1 value for both sets of control environment frameworks;

--b--represents the number of elements which take the 1 value within the j set offrameworks and the 0 value for the i set of frameworks;

--c--represents the number of elements which take the 0 value within the j set of frameworks and the 1 value for the i set of frameworks.

The control environment elements analyzed in this empirical study are there fore given the 1 value for containing a certain requirement and the 0 value for non-containing that considered requirement. As aresult of the effective measurement of the comparability degree between the Romanian control environment framework and the international models COSO and CoCo, based on Jaccard's coefficients, we have reached to the conclusion that there is a high degree of similarity between the Romanian regulation and the COSO model on the approached areas, as presented in the following table (table II.)

5....