Informed discretion in performance evaluations

by Marcel van Rinsum

A study into managerial behaviour relative to compensation decisions for individual performances reveals that most managers are driven by powerful, non-selfish motives that include a strong preference for fairness.

Altruism, fairness and self-sacrifice are not words that typically spring to mind when we think of the managerial motives behind incentive compensation decisions. Yet, this is precisely what our recent paper In Search of Informed Discretion: An Experimental Investigation of Fairness and Trust Reciprocity reveals.

This result has impor t ant implications. The vast majority of companies today rely solely on financial instruments as “objective” metrics for structuring their incentive compensation contracts – a norm that has been blamed for creating harmful incentives and asymmetrical, inaccurate pay-offs. So, could the subjective input of managers in fact create fairer compensation allocations more reflective of performance?

Performance evaluations

Standard accounting reports provide the basis for the majority of incentive compensation contracts today. These data-driven measurements provide a supposedly “objective” evaluation of an employee’s performance and thus share of the bonus pool.

Yet, the inherent flaws of financial tools as a measure of performance and reward have been widely demonstrated in both research and practice. At best, these tools offer inaccurate and incomplete measures of employee performance that fail to take into account the influence of external factors such as exchange rates or the performance of other business units. At worst, they encourage a myopic view of profit-making and a strong incentive for employees to “cook the books”.

However, what’s the alternative? One option is to bring more subjective, discretionary input from managers into the information flow. Discretionary bonus pools are good examples of how companies are using subjective input from managers to counter incomplete reports.

The theory? That managers can subsidise financial reports with additional information garnered from variance investigations, discussions, examination of work documents, and other investigative processes, thus reaching a compensation decision more proportionate to the employee’s performance than a standardised report could alone.

Moreover, herein lies the problem. Empirical studies from accounting literature reveal that managers often fail to use their discretion to seek out this extra information.

Related research has also indicated the presence of managerial bias towards understating differences in employee performance.

Yet, is this the whole picture? In our study, we investigated how managers behave when making compensation decisions for individual performances in a team setting. Specifically, we wanted to know how willing managers were to obtain the additional information that would enable them to make a more accurate assessment of an individual’s contribution, when this information came at a personal cost.

“When employees chose to invest effort in their work, it conveyed trust to managers that this effort would be recognised and rewarded.”

Potential unfairness

Our experiment involved analysing the behaviour of managers making compensation decisions on behalf of individual employees operating in teams of two. Individuals acting as managers were given the results of a team’s aggregate performance and asked to split the bonus between the two team members.

Each manager was given a sum of money and, before making their decision, the opportunity to use it on learning more about the effort levels of each team member (this money representing the managers’ time spent on a costly information search), or to simply make a decision based on the team’s aggregate performance and keep the endowment for themselves.

How interested were the supervisors in disentangling these performance levels to make more precise evaluations? At the  extreme  ends of performance – where the teams performed either extremely badly or extremely well, most managers felt no need to investigate further. This is because one can reasonably assume that when a team performs extremely well, both team members did a good job and thus an equal split of the bonus a fair distribution of the reward. Likewise, if a team performs terribly, the chances are high  that  both team members performed poorly.

But when a team performs at an average level, it is impossible to ascertain whether both members performed equally well, or one very badly and one very well. In this context, the aggregate performance becomes a very “noisy” measure of individual output.

A need for fairness

In our research, we found that it was precisely in these instances that most managers elected to find out how much each individual contributed and thus allocate the bonuses proportionately. In other words: the greater the uncertainty as to the potential for unfairness, the greater the managers’ willingness to incur personal cost to avoid it.

Most managers revealed an inherent interest in fairness and trust reciprocity. What was their motive? These managers received no material pay-off or other benefit for trying to unravel the noisy metric of aggregate performance. Quite the opposite: acquiring additional information came at a personal price. How can we explain this?

Numerous empirical studies have revealed a strong preference for fairness in humans and a willingness to incur personal cost to achieve it. Our study confirms that this preference also plays an important role in the decision- making process of managers. In fact, even when there is only a potential for unfairness (and not certain unfairness), these managers were willing to sacrifice personal wealth in order to bring about a fairer result.

Like fairness, trust reciprocity has been identified in studies as a driving human concern. In our study, we found that managers were more willing to obtain information on individual effort levels when the teams performed relatively well (but still within the “noisy” average range). Why? When employees chose to  invest  effort in their work, it conveyed trust to managers that this effor t  would be recognised and rewarded. This perceived trust was reciprocated by managers in the form of greater time and money spent ensuring these expectations were met, thus revealing trust reciprocity as a motivating factor in their behaviour.

Clearly, managers have strong preferences other than that of wealth maximisation, self-interest and cognitive bias influencing the decisions they make. In this instance, we could view their decisions as a trade-off between the potential risk of unfairness, the degree of trust reciprocity present, and the personal cost of acquiring the extra information.

Companies could do with integrating subjective input into the evaluation process. But how can this information be of use to companies?

One important implication concerns how companies can achieve a more optimal process for determining the appropriate incentive compensation plan for a par ticular employee. More research  needs to  be  done into the interplay of managerial preferences. But the evidence of strong, non-selfish motives provides a powerful basis for companies to consider including more subjective input from managers into the decision- making process as an effective means of remedying the very real shortcomings of purely data-driven performance measures.

Another implication relates to efficiency. Many managers with strong social preferences for fairness are likely to invest considerable time and energy in acquiring additional info r mat i o n to subsi dise the perceived inadequacy of standardised performance reports – regardless of the policy of the company.

Companies could thus do well to facilitate this investigation by making data more readily available, such as via more flexible accounting systems that cater for user-driven ad hoc exploration of data beyond that routinely produced by accounting reports. This would improve the efficiency in which managers’ can acquire their information and thus reduce the cost of time spent on these activities.

These changes could be immensely beneficial if they serve to structure incentives and rewards more appropriately. Indeed, what could be more important in a market economy than the per formance incentives we provide for employees, managers and investors?

Marcel van Rinsum is Associate Professor of Management Accounting and Control, Department of Accounting and Control, Rotterdam School of Management, Erasmus University.

This article is based on the paper In Search of Informed Discretion: An Experimental Investigation of Fairness and Trust Reciprocity, which was written by Victor S.Maas, Marcel van Rinsum and Kristy L.Towry and published in The Accounting Review Vol. 87, No. 2, 2012, pp. 617-644.

This article was published in RSM Insight 14. More information about and back copies of RSM Insight can be found here.

Download Original Article (.pdf)

Share this article: