Perfect trio or lone fighter: Contracting authority, monitoring, and incentives in hierarchies

Incentive compensation, performance monitoring, and the delegation of authority to managers – these three management control measures are often considered inextricably linked. In many cases, they are considered complementary. A common line of reasoning is that, for example, high-quality performance measures allow for stronger incentives, and stronger incentives are essential when managers have a lot of decision-making authority. But does this always hold true? Do these three measures always act as complements? How do they interact in hierarchies? TRR 266 researcher Christian Hofmann has addressed these questions together with Raffi J. Indjejikian. They identified some interesting empirical implications.


What happens when a three-legged stool loses one of its legs? It becomes unstable and can carry less weight than before. A very similar logic – at least according to conventional predictions – applies to these three management control measures: incentive compensation, performance monitoring, and decision authority. According to this analogy, optimal firm performance requires all three control measures. Take away one of them and firm performance falters. In the economics and accounting literature, this trio is therefore also known as the three-legged stool. We examined these three controls closely and investigated how they interact directly and indirectly: In what situations do they act as complements – increasing each other’s returns? And under what circumstances do they substitute each other – and thus contribute more successfully to a better performance of the firm as a single measure than they would as a duo or a trio?

What is complementarity?

Unlike previous studies, we have not only examined these measures with regard to their pairwise associations, we also identified circumstances where they exhibit three-way complementarity. We define complementarity as follows:

1) We define two control measures to be pairwise complements when their comparative statics with respect to an exogenous variable have the same sign. In other words, the decision to increase one control measure increases the return of the other measure.

2) We speak of three-way complementarity when any environmental change that is conducive for a rise in one measure will let respond the other measures in a consistent way (by increasing or decreasing their return) – and thus form a coherent pattern. For example: strong incentives, broad contracting authority, and low monitoring quality.

Our model

To identify the circumstances where the control measures are complements or substitutes, we based our study on a hierarchical principal-agent model. The model consists of three actors, an owner-principal, an agent-manager, and several agent-workers. Using a retail company as example, I will explain the key features of the model. The owner of the company contracts directly with some salespeople – and with the sales manager to whom she delegates some of her contracting authority. The sales manager, in turn, contracts with some salespeople herself. In the model, the owner makes three interrelated decisions:

1) Contracting authority:
The owner decides about the sales manager’s contracting authority or span of control, i.e., the number of workers she assigns to the manager.
2) Performance measurement:
The owner invests resources to monitor the sales manager’s performance.
3) Incentive compensation:
The owner also designs compensation contracts for all salespersons for whom she retains contracting authority and for the sales manager. The manager’s contract ties the sales manager’s compensation to her own performance and to the performance of her salespeople. In doing so, the owner ensures that the sales manager is ‘motivated to motivate’ her salespeople to provide effort. The sales manager, in turn, designs the contracts for the salespeople assigned to her.






Based on the model, we identified some important direct and indirect interactions between the control measures:

Higher performance without intensive monitoring

Conventional predictions hold that high quality measures promote steep effort incentives. In contrast, our results show that this may not always be the case. For example, this does not apply in hierarchical organisations, when monitoring quality is determined endogenously (i.e., a conscious decision made by the principal), effort incentives and monitoring quality may be substitutes. More specifically, a manager who has contracting authority has an implicit incentive to put in more effort. Simply because the incentive compensations of her workers are tied to the manager’s performance. As leading indicators of the manager’s performance increases, so do employees’ incentive compensation expectations, resulting in lower regular salary demands.

This implicates that a manager in a hierarchy provides more effort even though the principal monitors the manager less intensively. This also suggests that firms spend less time and fewer resources in monitoring managers entrusted with contracting authority than otherwise comparable managers. Since contracting authority resides with higher-ranking managers in most organisations, this means that performance monitoring practices at lower management ranks are far more comprehensive than similar practices in higher ranks.

Complementarity among all three measures

Taken together, we find that when all three control measures are endogenously determined (i.e., a conscious decision made by the principal), they may appear as pairwise complements or substitutes. Moreover, we can identify certain situations where the control measures form a coherent pattern of practices and thus can be understood as three-way complements. This is the case, for example, when contracting authority is highly sensitive to changes in firm fundamentals. Let us illustrate this with an example: the owner’s decision to delegate contracting authority. This decision reflects a benefit-cost trade-off. The owner benefits by increasing the manager’s efforts as a result of the implicit incentives associated with contracting authority. Potential costs arise from the owner’s control loss. This control loss manifests in the manager’s workers contributing less to firm profitability than the principal’s workers.

Now, if the incentive problem of lower-level employees is modest, the owner’s control loss via delegation is modest as well. Then, the decision about the number of workers the manager has under contract is highly sensitive to changes in firm fundamentals – since control loss plays only a tangential role for the owner’s decision. In this case, incentive pay and contracting authority substitute for monitoring quality, while incentive pay complements contracting authority. Taken together, these three pairwise associations imply a coherent pattern of strong incentives, broad contracting authority, and low monitoring quality.

In contrast, if the incentive problem of lower-level employees is severe, then the delegation decision is rather insensitive to changes in firm fundamentals and the control measures may be pairwise complements or substitutes. Their constellation varies. Thus, complementarity among all three measures is unlikely.

Added value for firms

Identifying coherent patterns of control measures is important because such patterns significantly reduce a firm’s design problem. In contrast, “mix and match” of seemingly “best practices” will likely not generate the desired results. However, with coherent patterns, it may well be that changing just one control measure would worsen firm performance, whereas changing all control measures together would increase firm performance substantially. Thus, to respond to changes in firm fundamentals, the presence of coherent patterns requires a more comprehensive “turn around” rather than small step changes.



To cite this blog:

Hofmann, C. (2021, April 7). Perfect trio or lone fighter: Contracting authority, monitoring, and incentives in hierarchies, TRR 266 Accounting for Transparency Blog.



Your email address is required to be able to post a comment. Only your name will be visible to others, your email address will NOT be published. Please note that the comment section is moderated – this is needed to keep spammers and net abusers away.

Related content