Re: We're not in Kansas anymore, Tony.
The banks are, on the whole, paying the price for moving fastest into IT, like, 30-40 years ago, then upgrading continuously.
They invested heavily in mainframes early on, and built new on that (remarkably solid) base, then new on that, then new on that... and now they're stuck with seven layers of legacy hell because a full replacement is expensive and risks looking exactly like this news item. They've traded a system that mostly worked for one that doesn't. There's a chance they could have done it right, but few people are willing to take that risk, given how often IT change programs fail.
It's the same story all through finance as a whole - new shiny companies have great responsive public-facing IT but no understanding of what will go wrong at scale, and no built-in pile of fixes for problems most people have entirely forgotten about.
Old companies have rock-solid mainframes buried under layers of cruft and middleware before approaching the customer-facing applications, which inevitably run like cold treacle and fragment at the merest thought of a change request.
That's your choice, really: do you want your hard problems addressed with organically developed spaghetti-code in five languages, or with clean modern code that's not been battletested in the same depth?