Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

What kinds of systems did you have in mind, then? It's hard to see how a CEO could keep personal tabs on all the bank's activities without automated surveillance - it's not like she can individually review every deal the bank makes.


What forces do you think are capable of causing leaders at financial institutions to internalize the cost of criminal banking activity such that they adequately prevent it?


I don't agree with the premise. The way to fight criminal banking activity is to write clear, straightforward, and consistently enforced laws. No system of incentives is going to meaningfully reduce financial crime in a world where grey areas are gigantic and almost all cases end with a settlement.


You can try to write all the laws you want, eventually it boils down to manually reviewing a ton of transactions and everyone loses. The bank needs to pay for growing compliance departments, customers are scrutinized on a growing number of transactions.

Small businesses who are in higher risk industries, however that is chosen to be defined, become inundated with KYC tasks. Every transaction needs an accompanying stack of documents showing the entire trail.

Finance is becoming a bureaucratic hellhole which disproportionately affects lower income people and small businesses. If you don't have a hired compliance agent you will spend your time as a business owner complying to endless demands for onboarding/KYC/transaction reporting.


No but she puts the organisation and processes in place so that the deals are reviewed with a chain of responsibility


That's how it works today. The problem is that any sequence of links in the chain can decide to lie. That's how most major banking scandals happen; some guy, his boss, and maybe his boss's boss conspire against the rest of the chain to be dishonest together. I feel like I end up circling to 1MDB very often here, but it's just such a good example; the 1MDB bankers had to lie extensively to the chain of responsibility at Goldman Sachs, since the compliance department had repeatedly instructed people not to do business with their major point of contact.


An investigation will reveal a subset of the people in the chain involved in the misdeed and/or coverup. Could the following process work?

1. Identify the subset of people involved and color them red.

2. For each of the red-colored nodes, color their immediate superiors yellow if not colored.

3. Until there are only uncolored or blue nodes left in the company, do:

3a. Put all red and yellow nodes in front of a court (presumably in parallel)

3b. When a verdict is reached for a given node, color it blue.

3c. If a yellow node is revealed to be involved, color it red, and goto step 2.

3d. If an investigation reveals further people possibly involved, color them yellow, goto step 2.

This is based on a principle of "your subordinate, your responsibility", and is meant to give everyone involved and their superiors a fair chance of landing jail time. I would hope the courts are already doing something like this.


Random auditing by machine learning algorithms looking for suspect transactions.


That is already done.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: