When Banks Become Cops

The argument is a fairly obvious extension of the rationalization for in rem asset forfeiture, to “take the profit out of crime.” It’s a great slogan, given that crime is bad and profit is its motive. When the money seized was the cash in a traveler’s pocket, seized not because there was any particular basis to believe that a crime was committed or that money snatched was either derived from or used in crime, but just cash that cops could grab, however, the slogan rang hollow.

But that was cash, and there was more money to be had. And much like drug dealers might carry suitcases filled with cash, bad dudes might use banks to hold, launder and pay for bad purposes. Something must be done, activists cried. And so something was done.

These situations are what banks refer to as “exiting” or “de-risking.” This isn’t your standard boot for people who have bounced too many checks. Instead, a vast security apparatus has kicked into gear, starting with regulators in Washington and trickling down to bank security managers and branch staff eyeballing customers. The goal is to crack down on fraud, terrorism, money laundering, human trafficking and other crimes.

Banking law put the onus on banks to identify “suspicious” transactions and file a “suspicious activity report” or the banks are held culpable to the federal government. And the last thing in the world banks want is to have the feds going after them for failing to snag a customer, whether individual or business, engaged in bad acts. So bank compliance departments created algos to red flag activity that strayed from what banks consider the usual.

Bank customers get a letter in the mail saying their institution is closing all of their checking and savings accounts. Their debit and credit cards are shuttered, too. The explanation, if there is one, usually lacks any useful detail.

Or maybe the customers don’t see the letter, or never get one at all. Instead, they discover that their accounts no longer work while they’re at the grocery store, rental car counter or A.T.M. When they call their bank, frantic, representatives show concern at first. “Oh, no, so sorry,” they say. “We’ll do whatever we can to fix this.”

But then comes the telltale pause and shift in tone. “Per your account agreement, we can close your account for any reason at any time,” the script often goes.

The point isn’t that the customer service rep doesn’t want to tell you, but that they have no clue why the algos kicked your account back, but it did and that’s that. Far better to lose a few thousand questionable customers than fail to throw out a fraudster or money launderer and get slammed by the feds, with the ensuing fines if not worse, for complicity in crime.

In the process, banks are evicting what appear to be an increasing number of individuals, families and small-business owners. Often, they don’t have the faintest idea why their banks turned against them.

But there are almost always red flags — transactions that appear out of character, for example — that lead to the eviction. The algorithmically generated alerts are reviewed every day by human employees.

Algos are bludgeons, and easily pick up on activity outside the “norm” of banking. The problem is that there are a great many perfectly lawful and, indeed, entirely normal transactions that are “out of character” unless you ask why. Algos do not ask questions. Buying a used car from someone on Craig’s List? You’ll need cash to complete the transaction. There’s nothing unusual about buying a used car. People do so all the time. But they don’t do so everyday, and so the algo raises a red flag over an unexplained cash transaction and you’re suddenly a potential criminal. Banks won’t take that chance.

At the outset, the filing of SARs was intended to alert the government so it could investigate suspicious banking activity, but the sheer number of SARs filed made that an impossibility. While banks weren’t in a position to seize money, they were in a position to protect themselves from being targets that were easier for the feds to nail than individuals and businesses whose accounts were flagged. After all, when some bad dude makes the front page or 60 Minutes, the feds can then search their SAR databank to see if the bank flagged the account. It’s backwards investigation to find out who gets blamed, and banks do not want to get that call asking for comment or what follows.

So what’s the problem? There are regular people and businesses burned in the banks’ efforts to make sure their own butts are well and truly covered.

But there are almost always red flags — transactions that appear out of character, for example — that lead to the eviction. The algorithmically generated alerts are reviewed every day by human employees.

Individuals can’t pay their bills on time. Banks often take weeks to send them their balances. When the institutions close their credit cards, their credit scores can suffer.

Upon cancellation, small businesses often struggle to make payroll — and must explain to vendors and partners that they don’t have a bank account for the time being.

Imagine you, a law-abiding person with bills to pay and mouths to feed, suddenly cut off from your savings and credit for an unknown extended period of time. Your kids or employees don’t want to hear about the banks’, or the federal government’s, problems when they’re hungry. The banks, and the feds, really don’t care.

To former bank employees, the bloodless data belie the havoc that banks wreak. “There is no humanization to any of this, and it’s all just numbers on a screen,” said Aaron Ansari, who used to program the algorithms that flag suspicious activity. “It’s not ‘No, that is a single mom running a babysitting business.’ “It’s ‘Hey, you’ve checked these boxes for a red flag — you’re out.’”

In a world driven by algos, explanations don’t matter. But that’s the only way to make sure that no bad dude launders money, and so what if a few good people go hungry?