Independence

advantage of being dependent therefore does not seem to reduce the type-1 agency problem or detect the moral hazard problem (that top management proposes sub-optimal projects) as the alignment with top management seems to outweigh said advantage. In other words, it could be argued that shareholders do not gain from having dependent directors on the boards, because their assumed informative advantage is not greater than their alignment with the executive management.

On the other hand, if the CEO’s risk preference is shifted from the risk averse agency prediction to the risk-seeking alternative prediction outlined in section 4.7.1, the discussion on the value of independent directors is turned to another angle: if Danish bank CEOs prefer risky projects, then it can be argued that independent directors are merely rubberstamping his proposals. However, this does not present an increase in the type-1 agency problem. While this may be counterintuitive – section 1.3 argues that the problem in Danish banks has been excessive loan-giving – the finding of this study is that in respect to shareholders, independent directors do what shareholders require even if that means unconditionally approving projects that are relatively risky, as the shareholders, being risk-neutral, can diversify their portfolio and thus seek risk to maximize the value of their call option held on the bank’s assets (see section 4.5). In other words, when the bankrupt state is not considered (and it can safely be concluded that it has not, since none of the actors prefer the bankrupt state), independent directors are better aligned with shareholders than dependent directors as they increase the bank’s risk-taking. The causality in this relation is less clear, however, because it depends on the CEO’s inherent risk preference, which is not the research focus of this paper.

10.1.2 The sample

When discussing the impact of having independent directors on the board, it is imperative to note that nearly all outside board members have been deemed independent. Thus, the variation in risk-taking comes from two subsamples that differ considerably in size.

Also, the literature and empirical research on dependent board members mostly stem from one-tier management systems in which the board is composed of executive employees.

In Denmark, virtually all inside directors are not part of top management – and only few are middle-managers. The dependent outside directors present some sample issues as well, because they come nearly all from the same banks, which are parts of larger financial corporations – Nordea Bank Denmark, Alm. Brand Bank and Nykredit Bank. This may mean that the external validity of this particular variable is somewhat compromised in relation to the possible generalization of the finding, because other researchers in different corporate

governance areas than the Danish may yield significantly different results using the same sampling technique.

In determining the degree of dependence of the individual, outside director, some sampling issues present themselves as well: unfortunately, due to banking secrecy and privacy in some business transactions as well as limitations on the investigate nature of this study, some links between the banks and the director may have gone unnoticed and will do so to the researcher contemplating to copy this research method on the particular field.

For example, the management may be a distant relative to a director without this being common knowledge or surfacing in the online directories. Also, big borrowers (businesses, farmers, individuals) in banks are dependent, but will not immediately be recognized as such, as the banks’ individual outstanding loans are not available to the public⁴².

Cross-directorships have been detected in some cases, but not between management and directors and vice-versa, but only in the form of directors taking seat on each other’s (non-bank) firms’ boards. These directors have been deemed independent of the bank in question, though they may owe their place on other boards to the place on the bank’s board, thus making them dependent on having the bank board seat. These types of dependencies are nearly impossible to detect in a research design like the one deployed in this thesis.

It should be noted that the high percentage of independent directors is not (only) attributable to the research design. Danish companies have a strong tradition of separating ownership and control when the companies grow bigger and certainly when they reach the size of most of the banks in this sample in this study (Thomsen 2008)

10.1.3 Implications

The Danish Recommendations on Corporate Governance recommend that a majority of directors be independent to ensure that special interests do not influence the running of the company – in this case, the bank. Insofar as the “special interests” refers to that of the shareholders at the expense of other stakeholders of a bank as a financial intermediary (see section 3.2), the findings of the data analysis in the previous section do not unconditionally support that the board should be as independent as possible. While perhaps controversial, having a completely independent board would – ceteris paribus – increase risk, which for other stakeholders is not necessarily desirable. The implication of this is that the bank governance system benefits the shareholders – and the shareholders alone – when following

42 Although some details of the board members’ combined loans are available in some of the annual reports, the nature of the loans is not disclosed.

the recommendation at the expense of the remaining stakeholders it was originally established to (also) shield.

It should be noted, however, that ‘dependence’ comes in many other forms than the ones surveyed and accounted for in the analysis at hand as described just above. For example, a considerable share of directors hold just one board seat and are possibly dependent on the income derived from this.