Or I suppose they can just do without if they want to be stubborn.
You're correct that big companies have positions for people at all levels. However, they usually aren't willing to pay more to fill a junior position than they'd need to pay to fill a more senior role.
And I'm confident it's possible to develop aptitude interview processes that avoid 10:1 false positives.
The bottom line though, is this- If there's a genuine shortage, on the job training is a solution with centuries of history behind it. If companies aren't willing to train, they can't be that desperate.
Here's another example to illustrate the point with more realistic numbers. Using the same training program described above and an improved filter a company can get one qualified graduate for every two people it sends through the program. On the first day of work he company has invested $50 000 in this new jr programmer. Alternatively, the company could offer a generous, $30 000-40 000, signing bonus to new qualified jr programmers and pocket the difference. Hey can also structure the terms of this bonus so that it is spread over one to two years to incentivize retention.
There are lots of real world training programs with qualified, motivated students that have high failure/drop out ratios so I'd say 2:1 is reasonable here.
I'm not saying that paid training is a bad idea. I am saying that it doesn't make sense for most companies currently and they will continue to use cheaper alternatives while they exist.
This actually happens in India with companies like Infosys, TCS hire people from all different domains and then give them basic programming training.
Investing in employees only makes sense if the company thinks the investment will pay off. Let's say a company views the expected value of an employee at $100 000 a year. Instead of trying to hire someone at that rate they decide to train someone. Training costs money and doesn't always result in someone that can fill the role. Let's say it takes $20 000 to successfully train someone. To break even on their investment the company would need to pay below market rates for some period of time. If they pay $90 000 a year for the next two years we plans to bump the person's salary to the full $100 000 afterwards what are the odds the person will stick around for the full two years? In a hot job market why wouldn't they leave and start getting the full $100 000 earlier?
I'm not saying that the current system works well but I'm not convinced this approach makes sense for most companies.
Neither of the companies you mentioned focuses on tech. Their management and financial consultants. I'm not convinced that you can apply the same plan to programming and get the same results.