‘No artificial incentive can ever match the power of intrinsic motivation. People who do exceptional work may be glad to be paid and even more glad to be paid well, but they do not work to collect a pay cheque. They work because they love what they do’.
Although this may be true to some extent in that ‘intrinsically motivating jobs require less compensation than intrinsically less motivating jobs’1, people nonetheless require a wage.
Nursing is arguably one of the most intrinsically motivating professions yet nursing in the UK is in crisis due to a shortage of qualified personnel who have left the profession due to their poor earnings capability (nursing fails to satisfy the individual rationality constraint). However, the result of pay increases in the sector has been to increase the number of nurses prepared to rejoin the profession. Thus it appears that interest in a job is simply not enough to attain good performance or even participation in the profession if the individual’s rationality constraint is not satisfied.
However, not everyone gets to do their dream job, so what about the many individuals pursuing less agreeable careers. It seems that extrinsic motivation in the form of rewards can be crucial to ensuring that optimal performance is provided.
There is an argument for designing jobs so that people are given more responsibility, less monotony and the opportunity to express creativity, thereby ensuring greater enjoyment of the task.
There have been numerous empirical studies that have investigated employee and/ or firm behaviour pre- and post-adoption of a compensation plan and many have illustrated considerable evidence of positive behaviour in response to the introduction of incentives. For example, Lazear (1996) found that the output of workers installing automobile windshields increased after a switch from hourly wages to piece rates. Similarly, Banker et al (1996) found that sales increased after a sales-based performance plan was implemented at a retail establishment and the effect persisted over time. This seems to put a dent into Kohn’s argument of the temporary nature of the compliance that results from extrinsic motivation offered by incentives. Kohn argues that ‘once the rewards run out, people revert to their old behaviours’ but this is not at odds with agency theory.
If we view the agency relationship as a contractual relationship among individuals who are rational, evaluative utility maximisers, we would expect the contractual parties to take actions to maximise on their own behalf. Thus the basic agency model seeks to tie the agent’s behaviour to what would be good for the principal by the use of a reward which induces the agent to behave as if he were maximising the principal’s welfare (the incentive compatibility constraint). The agency model expects the agent to act rationally and behave less than optimally from the principal’s perspective on the removal of the reward (i.e. at the end of the contract). At the end of the contract, we would expect a new contract to be issued which would again seek to ensure that the agent maximise on the principal’s behalf by offering a contract that rewards the agent taking into account what the agent could receive elsewhere as well as satisfying the incentive compatibility constraint.
Kohn may be making a fair point where he argues that if income depends on productivity or performance rating, then individuals will focus on the numbers and quality may be sacrificed to achieve quantitative targets. There may be even more serious consequences such as earnings management (bordering on fraud as in the case of Enron). This is a problem exacerbated by the use of single (one dimensional) performance measures which are not likely to reflect all dimensions of managerial activity perfectly thus leading to managers paying more attention to those aspects that are included in their performance appraisal. For example, increased managerial attention on new investments may come at the expense of the efficient management of existing assets. This seems to indicate that managerial performance measures need to take into account the ‘myriad of decisions and responsibilities’ that managers may view as complements or substitutes in order to understand the overall benefit to the shareholders.
It is in the light of such evidence that the multi tasking framework has developed especially due to growing concern that neither risk aversion nor effort aversion are as important as the traditional agency model suggests and that the problem increasingly is not in ensuring that managers work harder but in getting them to choose the right combination of actions and decisions that increase shareholder value.
This may be why there is a reduced role for accounting based measures which do not capture the full consequence of all current actions although there may still be a role where we seek only to motivate certain tasks (e.g. measures of profitability may be very important if a new CEO is hired to turn a loss making company around).
Performance evaluation is rarely linked to a single performance measure (as in the traditional agency approach) on the appearance of which the reward is issued but rather the result of a ‘complex and time consuming process involving a variety of informal and qualitative assessments and judgements.
Another critical observation of the traditional agency model may be in its assumption of making explicit all terms to the contract although in reality ‘many employer/ employee relationships involve informal arrangements and understandings about individuals duties and responsibilities, an organisation’s evaluation and reward practices’. This is something Kohn fails to recognise – Kohn argues of the negative consequences of having explicit rewards yet rewards are an intrinsic part of the work place. Every work place has rewards in place although not necessarily explicitly contracted for. Repeated high performance may lead to recognition and promotion but this is accepted as an essential part of the work environment so why are financial rewards any different.
The informal contract is often a better motivating device since it does not specify terms but rather takes a broad view of performance removing the incentive for negative actions that result from multi-tasking. It is not limited to any specific information source such as the annual report and so can take a variety of performance measures into account in management evaluation.
As discussed earlier, a large part of Kohn’s argument revolves around the temporary nature of compliance brought about by extrinsic rewards. Yet in practice we see that CEOs that are significantly rewarded and we can only assume that it has motivated individuals such as Jack Welch of General Electric to the high levels of performance that he has have achieved during his long tenure. Managers will often start their careers wanting to learn the trade and develop their reputation thereby not requiring huge incentives in the hope of attaining success early in their careers, which will be rewarded in later years. Managers are not simply interested to maximise over one period – just as the firm expects to remain a going concern over the foreseeable future, so the manager expects to remain in position beyond the current period. This is where Kohn may be correct in his critique and it is a fault of the traditional agency model recognised by many – that the model may be best served in a one-period setting where both principal and agent are concerned only to maximise their wealth (and thus their utility) with no other concerns.
Once we introduce the multi-period extension of the model we obtain an almost enlightened principal and agent relationship which is not merely concerned with the ‘here and now’. The agent has long term goals just as the firm does - to maximise their lifetime utility.
This is where many of Kohn’s critiques fall through the floor. Both the agent and the principal will need to align their short-term goals with the long term. The principal will seek to provide a suitable work environment and appropriate support as well as ensuring that the individual rationality constraint and the incentive compatibility constraint are satisfied (to
ensure that agents participate while maximising on the principal’s behalf). The company will thus enhance its reputation as a good employer and attract the best of the future talent pool, thereby ensuring its future success.
In a similar fashion, the agent will be concerned as to his reputation and perception by potential employers (and even the current employer) and will therefore strive to display favourable characteristics. The agent will not seek to maximise in the current period if it hampers his future prospects. Thus an agent who reports truthfully and attempts to maximise on the principals behalf will be recognised and be even more greatly rewarded by promotion or obtaining a new appointment. This is why agents will seek to perform highly in all areas and not only those on which they are evaluated. Whereas Kohn argues the number one casualty of rewards to be creativity, it appears that managers would be more keen to display their creativity since it is highly sought after.
In conclusion, it appears that most employer-employee relationships and firms’ compensation practices are far more complicated than standard principal-agent theory allows. Although Kohn’s arguments do reflect many of the dysfunctions associated with introducing incentive schemes, in terms of agency theory, they appear to be a critique of the standard agency model with its rather narrow outlook. Kohn has failed to appreciate the extensions to the agency model in terms of dealing with multi-tasking and multi-period issues. The intuition behind the agency model is rather robust which with slight adaptation can very well explain the practical insights which we observe in practice and which Kohn uses to illustrate his arguments. Many of Kohn’s illustrations simply appear to revolve around bad practice and organisations which have not thought very much about their incentive implementation process and job design processes. Most crucially, incentives need to be developed for the specific situation and need to be aligned to ensure that it is based on broad measures of the organisational and managerial performance. Incentives are not a ‘quick-fix’ and they cannot simply be implemented to make up for organisational lackings and bad practice.