While the metrics behind his assertions are short-sighted, specifically ( years in workforce / number of jobs ) = job-hopping likelyhood, I think he left out the most absolutely critical reason why I and other hiring managers look at the duration of employ at previous establishments.
I've reviewed hundreds of resumes, hired great people, and hired some not great people. Sometimes they leave because of you and your company, sometimes they don't. In my experience, a history of short employment can mean one of two things:
1. This candidate is extremely talented, independent, and a self-starter. Hire these people. Now.
2. This candidate leaves before he or she is truly accountable for the code they write.
2 is very dangerous. You don't want to hire an engineer who will write a subscription billing platform and quit the week after launch. It's often tough to determine the difference between 1 and 2, and it takes many pointed questions to figure it out.