This is a good point and has been used to show that ivy league schools are not worth the money. The alumni make more money but they would make more money anyway because they are in the top few percent of all students.
So I can agree in one respect that grad school may just be another gate keeper statistic like number of years of experience. It may or may not have relevance to a particular job.
I guess one could say that one should strive to get into grad school but once accepted it may not matter whether they go? I have some problems with this. First, it's like saying as long as you have good grades through your junior year you may or may not finish your senior year. You can extend this logic back to going to school at all, right? Certainly if a bachelor's degree has value then a master's degree has value.
The second problem I have is that, as a person who has been in continuing education since my master's degree, I know that the corporate training and education are totally different. In corporate you go for a couple days, learn one tool or management practice, you're never tested and you may or may not retain the information for more than a week. In school I have to learn topics/material that I might never encounter in my direct job but still teach me useful engineering principles and forces me to be current on the latest tools.
In all, I think every engineer should be in continuing education. You don't know what you don't know. I think a lot of the folks who claim to know it all based on their work history would be quite surprised to enroll in a modern graduate class on a topic even in their own field.