First off my obligatory declaration that I'm an individual contributor. I offer up my own opinion from decades as a mainframe professional in real commercial mainframe applications. I have nothing to sell, just voicing my original take.
I've been in the industry since the late 80's, which actually makes me a young'un in comparison to many mainframers. I started just as universities began pumping out computer graduates while large corporations still had internal trainee programs to ensure a pipeline of developers. My first job as a uni grad was with a large bank. I went through an intensive 6 weeks training to skill me up as a PL/1 mainframe developer. I don't recall when that trainee program ended but I had plenty of highly skilled mentors that had entered IT from the mailroom, or were admins, or changed tapes, or started as tellers.
I worked solidly as a PL/1 developer until 1995. A company in the US brought me in on a H1B for a PL/1 contract they had but by the time I got to Detroit (sorry, meant Farmington Hills) the position was filled so I sat in on the tail end of their internal three month mainframe Cobol training program. Pre Y2K this was all the rage, intensely training uni grads and then contracting them out at incredulous amounts.
Without understanding my situation at the time I look back now and realize I was in the precursor of a body shop. Complete Business Solutions Inc started in 1985. It grew on the back of many companies trying to cut their IT costs by bringing in consultants as staff supplementation. Plenty of other consulting companies were doing the same thing except CBSI seemed to specialize in H1B workers and had opened up a branch in India in 1991 (paying wages much lower than in the US). This was the first time I saw work being sent off shore.
As the Y2K deadline crept closer desperation saw applications packaged up and sent offshore to be analysed/modified/tested. This was a lightbulb moment for the bean counters at large corporations of how to cut IT costs even further. Work could be sent overseas to take advantage of cheaper labour rates.
As I look back I'd have to say that following the successful implementation of the world's largest IT effort, preparing all mainframe code bases for the year 2000, this was the pivotal point that has led us to our current predicament - namely the shortage of adequately trained mainframe aware professionals.
Stay tuned, part 2 dropping soon.