Some interesting insights into what can go seriously wrong in a system of reward-linked performance appraisal is found in the work of Deets & Tyler (1986).
The Xerox Experience
The Xerox system included all the common features of rating scale systems. The appraisal interviews were held annually and conducted by the employee's immediate supervisor.
Accomplishments of the preceding year were recorded and performance levels were judged according to various predefined criteria. The system included some elements of essay appraisal, since appraisers were required to write brief supporting statements for each rated criterion.
The Xerox system also called for a summary rating; an ultimate digit, from a low of 1 (for unsatisfactory) to a high of 5 (exceptional). The summary rating attempted to encapsulate the whole year's performance in a single number.
The significance of that final number was immense. It literally determined the size of the employee's annual pay rise. The higher the rating, the bigger the rise. For Xerox employees, the thing that really mattered about appraisal was getting the biggest possible final number.
Analysis of ratings over time showed that more than 95 percent of employees were either 3s, 4s or 5s; that is, the spread of ratings heavily favored the higher end of the scale. Almost every employee, according to the appraisal system, was performing at or above the average.
The lower end of the scale, the ratings of "below average" and "unsatisfactory" were very rarely used. The effect of this distortion was that any employee who scored less than a 4 ("exceeds expected performance level") began to feel like a failure!
The appraisal process became a sort of ratings lottery; the aim of the game was to get the highest possible score and win the jackpot. The process became fixated on that all-important final digit.
This situation placed tremendous pressure on appraisees and appraisers alike. The appraisers had the unenviable task of deciding the winners from the losers. No wonder most of them preferred to hand out an abundance of overly-generous ratings!
Xerox eventually replaced this system with an MBO/essay form of appraisal. They abandoned rating scale methods completely.
That may have been an over-reaction, since the fault did not lie with the method itself so much as with its intimate - and ultimately inflexible linkage - to the annual pay rise. When reward outcomes are so closely linked to the size of a rating on a five point scale, the difference of one point either way can become very important and provocative.
The Xerox rating system might have worked if the direct
causal relationship between the summary rating and merit pay outcomes had been
eliminated or at least softened.
One of the main reasons for separating appraisal
results from reward decisions is the belief that a too-close link would create
an overly-threatening and potentially punitive system.
There is a deep irony in the fact that many organizations, while having excellent systems of appraisal, allow their merit pay and promotion decisions to be made by inferior means. Often the matter is left to the discretion of one or two supervisors or managers, with a cursory review being made by the HR department.
There is also the work of Bannister & Balkin (1990), which has reported that "discussions of pay at the time of performance appraisal" increases employee acceptance of appraisal and their satisfaction with the process. This undermines the arguments for separation.
As well, there is evidence that incongruity between appraisal results and later pay and promotion outcomes is a source of employee discontent and de-motivation.
Pay increases and promotions send powerful messages to
employees. If these messages don't match up with the appraisal results,
employees are quick to dismiss the whole process as a farce. Efforts have been
made to convince employees otherwise, but the "bottom line" for many
is who got the extra money or who got the new job.
Nor is the practice of putting a six-month buffer between appraisals and pay reviews an effective method of avoiding the issue. Far better to define and clarify the relationship between appraisal, performance and reward outcomes.
The view of the separatists, which insists that appraisal results and reward outcomes should be insulated from each other, may be an over- reaction to the potential abuses.
There is evidence that appraisees appreciate the existence of a link between appraisal and reward results. To many, the existence of such a link is intuitively sensible.
From the perspective of the organization, the inclusion of carefully collected appraisal data in pay and promotion matters may contribute to better quality decisions. It should also help ensure a greater degree of congruity between appraisal results and subsequent reward outcomes.
Even so, many advocates of separation will be reluctant to concede the possibility of any form of constructive linkage between appraisal and rewards.
This is a shame, because the potential of performance appraisal encompasses more than employee development. Admittedly there are risks in linking reward outcomes; but there are also risks, and a potential for harm, in contriving to deny that any linkage exists.
At the very least, an organization wishing to form the mildest of reward links might consider a frank discussion of reward criteria during the appraisal interview.
Those organizations that are determined to keep their appraisal and reward issues separated might ask themselves whether performance appraisal is really the tool they need. Perhaps what they actually desire is some form of developmental appraisal.