I think Larry Summers is wrong, on a point of theory. He commits a fallacy of composition. (That's a brave way to start the day.)
He says (HT John Cochrane):
"Businesses will raise wages to a point where the cost is just balanced by the reduced bill for recruiting and motivating workers. At that point, a further increase in wages does not appreciably change their total costs but higher wages certainly makes their workers better off. So there is a strong case for robust minimum wages."
Suppose wages were set by an omniscient social planner, who took into account the effect of higher wages on recruitment and motivation. Now consider the point of view of an individual employer, who contemplates what he would want to do, if he could set his wages higher or lower than the social planner allows. By raising his wages relative to other employers' wages, he could increase recruitment and motivation at his firm, but would reduce recruitment and motivation at other firms. The individual employer would ignore the negative externality he imposes on other firms' recruitment and motivation of workers. But the social planner would not ignore it. The social planner would prevent the individual employer raising wages at his firm.
Nope. That makes a strong case for robust maximum wages.
We don't call it "the efficiency wage theory of unemployment" for nothing.
I look forward to those who like government intervention to eagerly adopt the efficiency wages case for maximum wage laws, to help eliminate the scourge of unemployment that results from individual firms' profit-maximising behaviour in setting their relative wages too high.