Debates about pay for performance have been part of our landscape for many years and will, I suspect, continue on long after you and I have ceased to care. All the same, the debate today seems particularly acute - likely fueled by a number of forces, not the least of which is the handy platform that social media serves up to any of us with an opinion and a keyboard.
One thing strikes me, though. The performance pay debate is hobbled by a lack of clarity and agreement as to what pay for performance actually is. Some seem to speak exclusively of base salary increases that are linked to performance appraisal results. Others are focused on incentives. Still others aren't clear at all as to the type of program they are referencing.
In order to have a productive and ultimately useful debate about pay for performance, I think we need to define our topic more clearly. And I will submit that we would all be best served with a definition of performance pay that is broad enough to cover the reasonable assumptions and possibilities. Otherwise, we aren't really debating reward or motivational strategy at all; we are merely arguing design choices and tactics.
With this, my suggestion is to define pay for performance as employee rewards which are differentiated by performance. Recent (2010) research released by Aon Hewitt, used a similar definition in producing the information in the chart below.
In addition to spanning multiple vehicles, performance pay can also cover different levels of performance: individual, team/work group, business unit, organizational, etc.
So, in asking whether pay for performance works, what we should really be considering is whether differentiating employee rewards (any vehicle) by some level of performance will produce appropriate and desirable outcomes. Which is, of course, a little more complex and a little more challenging to answer with an easy and absolute "yes" or "no".
Which is, by the way, exactly my point.