Because some fraction of people are cheaters. Just think of how many people you know who cheated in school. I'd say the number is 10% or higher. Some of those people became CEOs who continued to cheat, but many became low-level bank employees who continued to cheat.

And if you're being measured on how many accounts you create, and the system doesn't have checks to keep people from opening accounts without consent, it's not exactly hard for all the cheaters to figure out what to do. Again, think back to all the ways people cheated in school. You think they had to coordinate to figure out how to cheat?

The management in one branch likely socializes with the management in near-by branches. It spreads through drinks and "Yeah, it's dumb; we just do <sketchy behaviour> so we can get back to work".

Have you ever had a job? Have you ever had management insist upon some stupid metric or some deadline that just could not be met and they refused to hear about it?

People had to open new accounts or lose their jobs. Some people opened fake accounts to accomplish this. When they were found out, they were fired. Sounds like a horrible place to work, but the firings were proof that management didn't want this to happen, not some bizarro proof that this is what they wanted all along.

You don't need coordination to figure out how to game a system.

This is not quite right. They were pressured by managers to open fake accounts, fired for not making quotas if not, fired if they reported the practice, and then all fired once the government finds out and it blows up. Of course the senior management plays dumb and keeps their jobs.

They were not "all fired, once the government found out."

Wells Fargo confirmed to CNNMoney that it had fired 5,300 employees over the last few years related to the shady behavior. Employees went so far as to create phony PIN numbers and fake email addresses to enroll customers in online banking services, the CFPB said.

http://money.cnn.com/2016/09/08/investing/wells-fargo-create...

Making fake accounts is a pretty obvious solution to increasing cross-selling numbers. I'm sure almost all of the branch employees thought about it, but only some actually did it.

You get what you measure. You measure account creation, you'll get account creation. You measure legitimate account creation, and it's a different ballgame.