So his take on executive compensation is noteworthy ...
... Tom Peters, a management expert, suggests capping the pay of C.E.O.’s receiving bailouts at that of a four-star general. As for the concern that the executives would quit, who cares? Mr. Peters writes that if all the top executives of the Fortune 500 companies were exiled to Elba, “performance of their companies would not on average deteriorate.”...
CEO compensation is the inverse of the zero-bound problem that confounds fiscal policy and produces perverse (high institutional risk) but logical (highest expected return) behaviors in financial managers and bank traders (Paul Wilmot, NYT). CEO compensation has reached the upper-bound limit of what used to be called "f-you money" -- the $30 million plus personal wealth that makes employment optional.
So CEOs don't need their jobs. They work extremely long hours because they adore their work - for reasons that are good and not so good. If you reduce their compensation, so they're now falling down the power curve, many of them will quit.
Then what happens?
I'm betting Peters is right, in part because I don't think CEOs are "selected" on the basis of excellence in leadership or judgment. I suspect that if a large majority of CEOs of publicly traded companies were to take early retirement that corporate performance would be unchanged or improved.
That may be what we have to do to eliminate the upper bound problem.
 Update: Ok, confession time. I added Sun Tzu because I was influenced by the opinion of others. My bad. I've since read it; Machiavelli stands alone. In modern times I still like 'Reengineering the Corporation' even though it got quite a bad rep later, and I like 'The Innovator's Dilemma' (but not later, much weaker, "Innovator's Solutions" book). and 'Getting to Yes'.