Chapter 4 Internal Audit: Theoretical Foundations
4.3 Agency Theory
Agency Theory is based on two essential actors: a principal and his agent. In its simplest form the principal wants to know whether the agent is acting in the best interests of the principal.
Or, and not at all necessarily the same, the best interests of the agent. As a somewhat free and self interested individual, however, the agent will almost by definition have another agenda.
He can shirk (free riding) or misstate what he is doing and the results of that behaviour. In either case, the principal needs to "monitor" the (reported) outcome (hidden information) or the behaviour (hidden action) of the agent. Or both. Surely, if the agent was able to do that and satisfy everyone's needs, it wouldn't be necessary to have someone else do the job.
Unfortunately - or ç'est la vie - that is generally not the case. Especially in a large organization. Therefore, the principal needs a third party for the monitoring task.
Agency Theory operates at level 4 - 3rd order economizing (see Figure 5 above) - and deals with the problems of marginal conditions. In this section the focus will be on Agency Theory's contributions in determining incentives to influence the behaviour of agents in such a way that their performance is as efficient as possible and meets to the max the interests of the principal.
Agency Theory deals with situations that are quite common when principals and agents sign contracts with each other. The principal cannot observe the actions of the agent (hidden action; Arrow, 1985b, 38), and/or the agent has information the principal doesn't (hidden information). The terms "hidden action" and "hidden information" originate in the concept of
"moral hazard" (insurance theory). Moral hazard can be described as an ex post situation of Information Asymmetry, because the agent knows more than the principal. In cases where results are visible the amount of Information Asymmetry can be assessed. But there are huge areas where assessment is not an option.
Stockholders are principals, who certainly cannot observe in detail whether the management, their agent, is making appropriate decisions. The principal-agent theory provides an instrument to discuss the rationale of the "separation of ownership and control" problem which Adam Smith (1776) focused on and which Berle and Means (1932) popularized 157 years later. (Jensen and Meckling, 1976, 327)
Thus in principal-agent relationships a need for monitoring arises. In theory, of course, the principal himself can do the monitoring. But there are some fairly obvious reasons why that is not done. First, it would reduce the distinction between the principal and agent and, consequently, the need for an agent. Two, many principals do not wish to spend their time monitoring agents. Three, many principals are either partially or totally incapable of monitoring their agents. Thus, in many, if not all, cases, principals decide to hire a specialized monitor: an auditor. The scheme is complicated further because the monitor himself is an
"economic agent" (Antle, 1982) who has a contract with the principal and may also be subject to shirking or misstating the results or his behaviour. Or both.
There are two variants in the Agency Theory (Jensen, 1983). First, a normative variant, which uses the achievements of micro economics and tries to create insights in optimizing problems via utility functions (constrained maximization). This theory is by nature predominantly mathematical and rarely empirically tested. Second, the positive variant, which focuses
mainly on monitoring and bonding of contracts and organizations. This variant is predominantly non-mathematical and empirically tested. It tries to model the effects of the context in which contracts are designed and detect what organizational form has the most chances of survival.
Although the separation between the normative and the positive variant are not always that clear (Eisenhardt, 1989a)64, attention will be paid to the positive variant, because of its wider applicability for theorizing about internal auditing.
Positive Agency Theory
Jensen and Meckling (1976) are the most important contributors to positive Agency Theory.
According to them, the concept of agency costs is very important.
The principal can limit divergences from his interest by establishing appropriate incentives for the agent and by incurring monitoring costs designed to limit the aberrant activities of the agent. In addition in some situations it will pay the agent to expend resources (bonding costs) to guarantee that he will not take certain actions, which would harm the principal or insure that the principal will be compensated if he does take such actions…. In most agency relationships the principal and the agent will incur positive monitoring and bonding costs (non-pecuniary as well as pecuniary).
And in all there will be some divergence between the agent's decisions and those decisions which would maximize the welfare of the principal. (1976, 308)
According to Jensen and Meckling, agency costs are the sum of:
Monitoring costs (e.g., costs of external auditor)
Bonding costs (costs agents incur to convince principals of their efforts; e.g., by having these efforts measured by internal auditors)
So-called residual loss (total other costs that need to be incurred)
Even if it were cheaper to have the agent himself collect and verify information necessary for monitoring, it is still better to have a third party do it. This is driven by the need for independence and objectivity (see Chapter 2). According to Agency Theory, this is one of the reasons audit has a license to operate.
Agency problems within the organization are predominantly the consequence of a classical division of capital and labor, and ownership and control. Fama (1980) and Fama and Jensen (1983, 303) have questioned how decisions are made within organizations and distinguished four steps:
1. Initiation: ideas and initiatives are generated 2. Ratification: choices are made
3. Implementation
4. Monitoring: measuring results and acknowledging rewards
The first two steps are called decision management and the last two decision control. It is important to achieve the most efficient distribution between the first two and the latter two
64 Eisenhardt (1989a, 60) stated that "the important point is that the two streams are complementary".
and the residual risks between the agents. This makes the organizational structure relevant.
The question is whether these first three functions should remain in the hands of one agent. In
"simple" organizations - law or consultancy firms - it is economically advantageous (more efficient) to combine them in the hands of one or a limited number of agents. Why? Because transferring the relevant and often extremely detailed information to others can be very expensive.
The question of separation of ownership and control is important for large organizations and, therefore, internal auditors. Fama and Jensen (1983) have said that in large organizations decision management and decision control should be separated. Especially when specific knowledge is necessary to reach a good decision it is wise to delegate those decisions to the level where the needed information is available. This will reduce information and agency costs. Decision control can then be delegated to those who also evaluate performance and may decide on appropriate remuneration. From an efficiency point of view, the number of residual claimants (residual risk bearers) - those authorized to decide what to do after all costs have been paid - should be limited. The residual claimant is the one who can dispose of what's left after all costs have been subtracted from revenues.
Alchian and Demsetz (1972), who studied incentive structures and residual claimants and analyzed team production, concluded that teams can do much more than individuals. But the downside of that is free riding, or shirking. That is, not all the members of the team carry their fair share of the weight. This then requires the appointment of a specialized monitor, who has the right to change the contract and distribute incentives (decision control). High performing team members can expect bonuses. Since opportunism on the side of the monitor might lead to biased numbers, this reported information might also be subjected to scrutiny by an auditor.
With regard to the capitalistic entrepreneurial form, Alchian and Demsetz (1972) also conclude that there should be one residual claimant, who is the central party to all contracts with inputs, entitled to observe behaviour and, if necessary, alter team membership. The latter is needed to ensure that when noticed shirking will be punished, and, therefore, increase team production. He also has the option to sell these rights.
Some Criticism of Positive Agency Theory
Criticisms of Positive Agency Theory are predominantly focused on its assumptions about human behaviour. Baker, Jensen and Murphy (1988) demonstrated that the relationship between performance and reward is not always straight forward and suggested that "a hierarchy that can induce the right kind of cooperation – defined as voluntary deviation from self-interest behaviour – will have an important competitive edge over other firms" (1988, 12). This leads us into the area of trust, which will be elaborated on later in this chapter.
Miller (1992) criticized the attitude of incentives being a cure all. According to him, using rewards and punishments to discipline individual behaviour is too mechanical. Using the work of Baker et al, he based his criticisms on the "prisoners' dilemma". That is, there are situations when following your limited self interests is not totally, or even partially, in your best interests. For example, in a leadership position in a modern organization, where conflicts are inevitable. It is in the best interest of the leader to put the interests of the group above his own. It is his task:
to shape expectations among subordinates about cooperation among employees, and between employees and their hierarchical superiors. This is done through a set of activities that have been in the realm of politics rather than economics:
communication, exhortation, and symbolic position taking. Most important perhaps, the leader has a central role in committing the organization to what is in effect the constitution of the hierarchy – the allocation of generally accepted responsibilities, rules of the game, and property rights that provide the long-run incentives for investment in the firm. (Miller, 1992, 217)
Miller (1992, 2) suggests that there are also other possible approaches, such as "leadership that is inspiring a willingness to cooperate and take risks, to innovate, to go beyond the level of effort that a narrow, self-interested analysis of incentives would summon". He is convinced that conflicts between individuals and group interests are the central characteristic of large modern organizations (1992, 236-7).
Positive Agency Theory has had some impact on the debate of mitigating agency problems via incentives and rewards. But there is also a serious debate about its adverse effects. What might be called "the negatives of positive Agency Theory". Jensen and Murphy (2004) thoroughly researched the remuneration issue, and their conclusion is worth pausing over.
In their 1990 study of CEO compensation Jensen and Murphy (1990) had this to say:
"Are current levels of CEO compensation high enough to attract the best and brightest individuals to careers in corporate management? The answer is, probably not." … As the reader of this report has undoubtedly surmised, Jensen and Murphy would not give that answer today. Indeed, we have emphasized here that while executive compensation can be a powerful tool for reducing the agency conflicts between managers and the firm, compensation can also be a substantial source of agency costs if it is not managed properly. And as we've summarized, there is substantial evidence that we can do much better in the future. (2004, 98)
The issue is of course that there may always be dysfunctional behaviour. One of which - called "earnings management" - has surfaced in spades inside the corporate scandals of Enron, WorldCom, Tyco, Adelphia, Ahold, Parmalat, and many others we know so well from reading the newspapers. Incentives, in the form of stock options, induced managers to "cook the books" to such high temperatures that a number of them boiled over and melted down. ,A number of studies (Bruns and Merchant, 1990; Degeorge, Patel, and Zeckhauser, 1999) showed that "earnings management" has been around for a while. One (Bruns and Merchant, 1990) concluded that "we have no doubt that short-term earnings are being manipulated in many, if not all, companies".
4.4 Transaction Cost Economics