Pay-for-performance is the current holy grail in compensation, and while many companies have implemented some sort of strategy, it seems that no one has been able to figure it out quite yet. More than three-quarters of the companies in i4cp’s Pay-for-Performance Pulse Survey said they tie pay to performance. It makes sense on paper. When people perform better, pay them more. How hard could that be, right? OK, perform better than what? Better than whom? How much more do I pay them? And what exactly is it that these employees are supposed to be performing?

Many experts believe that the merit raise has breathed its last breath. It has become so misused and mishandled that employees expect to receive one every year, whether it was merited or not. High performers are turned off because their extra effort is rewarded no differently than the slackers, and the slackers keep slacking because they keep getting raises anyway. The new push is to have greater differentiation between high and low performers. In the i4cp survey, the most common raise range for high performers was from 4% to 4.9%. For average performers it was 3% to 3.9% and for low performers it was no raise. While this definitely separates out the low performers, the high performers are not really being set apart from the average Joe.

So let’s say you’ve maximized your merit budget by allocating 80% to the top 20%, and your performance review process isn’t a joke, issuing 4s and 5s all around. Productivity should be through the roof. Your high performers are performing highly, your average workers are striving to achieve and your slackers are looking for another job (or maybe they’ve stopped slacking). For some reason, however, this doesn’t seem to be how it plays out in the real world.

A study published in Industrial and Labor Relations Review looked at data from 5,736 Danish firms from 1992 to 1997 and compared wage increase dispersion to net value added (revenue less purchases) per employee. As the spread between raises for low performers and high performers widened, the value added went down. That’s right. The perceived unfairness of wage allocations among employees was enough to hurt productivity and bring down company performance. What’s happening? There are studies showing that a majority of employees believe they are top performers, but in reality that’s not the case. So when someone who believes they are a top performer sees someone else get a much larger raise than they do, morale drops, taking productivity with it. In the i4cp survey, 70% of respondents rated their companies’ plans as either average or poor when it came to being perceived as fair by employees.

Jeffrey Pfeffer, professor of organizational behavior at Stanford University’s Graduate School of Business, says that high levels of pay differentiation can kill engagement, trust and teamwork. And to top it off, it’s not so great for the high performers, either. Pfeffer says the increase in pay is often not enough to sustain the performance it rewarded and that many plans reward the wrong things and encourage the wrong behaviors.

There is no single “Eureka!” solution to fix the pay-for-performance issue. Companies need to get better at setting goals and determining the criteria on which to base compensation. Performance management needs to improve. It needs to be unbiased and true. Communication also needs to ramp up to help alleviate the problem of perceived unfairness. And lastly, organizations need to get better at capturing and analyzing their human capital data. This is the only way to determine if a pay for performance plan is actually working the way it’s supposed to.