Thank you for commenting.
What Holmström in particular addressed is how *not* to address the cobra problem. He points out the risks of too narrowly looking at the layoff as a proxy for the effort, and the definition of the measure, and he argues for using more information about the effort in evaluating performance.
What you seem to do here is to explore an alternative, show it is absurd, and hence suggest that the original position is mistaken.
There are many other challenges in resolving cobra or black widow problems than introducing inappropriate incentives. That doesn’t diminish Holmström’s contribution.