Mind Matters Natural and Artificial Intelligence News and Analysis

# How Do We Know Financial Transactions Are Honest?

Let’s look at the steps we can take to find out

Recently, we’ve been asking readers to think about Alice and Bob, the famous pair in physics used to demonstrate propositions, for example, as if as if they were running for office. At Expensivity, a blog about that expensive and unpleasant subject, money, Bernard Fickser asks about better ways of preventing financial fraud:

We focus on financial Alice (the situation with financial Bob is parallel). Alice wants the record of deposits and disbursements in her ledger to reflect deposits that she has knowingly and willingly received as well as disbursements that she has authorized to go to the intended parties.

If this is the case, the record of deposits and disbursements in her ledger as well as the running totals will correspond to the true state of affairs of what her transaction history and running totals ought to be.

In trying to secure Alice against financial fraud, we therefore need to consider what could go wrong. In fact, three things could go wrong:

Alice finds a disbursement that she did not authorize.

Alice finds a disbursement that she did authorize but that went to an unintended party.

Alice fails to receive a deposit that she was supposed to receive.

All three cases represent a failure. With fraud, we consider them a failure of security. With unintended error, we consider them a failure of numerical input, calculation, or technology.

But given that unintended errors could always be intentional (though the intentionality may be implausible, as with certain types of common math errors), and given that the damage done by unintentional errors can be as profound, and sometimes more so, than done by fraud, it is convenient to treat all such failures, at least potentially, as cases of fraud and to strive to protect against them. (For an unintentional math error that, in 2012, caused J.P. Morgan Chase to lose $6B dollars, see Chapter 3 of Matt Parker’s Humble Pi.) We therefore consider a bad actor, the fraudster Frank. Frank is going to try to divert money from Alice’s account so that disbursements intended for Bob end up on Frank’s financial ledger, or deposits intended to go on Alice’s ledger likewise end up on Frank’s financial ledger. How could this happen? Let’s ignore microthefts, in which pennies or even fractional pennies are skimmed off an account at every transaction, making the thefts numerous but almost unnoticeable. Instead, let’s focus on large discrete events where the fraud is palpable. Thus, it could happen that Alice authorizes a disbursement intended for Bob, but it never ends up in Bob’s account. Or Alice might be informed that a deposit into her account is on its way, say from Bob, but it never arrives. Or it might happen that funds simply disappear from Alice’s account without her authorization. To commit fraud in such cases, what might Frank do and what safeguards might hinder him from committing the fraud in the first place? As noted earlier, money transferred among bank accounts (such as the financial ledgers of Alice and Bob) always has a provenance. Money can’t just magically materialize. There’s always a history. For money to be deposited in one account, it must be withdrawn from another account, and there has to be a record of the transaction. The only exception is the fiat creation of money by the central banks, in which they create a deposit without transferring existing funds. But even here there has to be a record of the money being created. Let’s therefore start with fraudulent disbursements. Prior to a given transaction, Alice is, let us say, satisfied that all earlier transactions are legitimate. We can assume that data integrity methods are in play so that Alice and impartial third parties can all agree that up to the problematic transaction, all the prior transactions are legitimate. So let’s say$10,000 was transferred out of Alice’s account without her authorization. How could this have happened? Let’s run through the possibilities … in the next instalment!

You may also enjoy our earlier stories in this series:

How can we prevent financial or election fraud? Both contexts come down to an accounting problem, keeping track of money or votes over time. Let’s take two people, the famous Alice and Bob, used to demonstrate many propositions in math and science and think of them as candidates running for office.

and

How can ballots be both secret and fair? The secrecy of ballots would not be compromised if voters used some markers of their identity known only to themselves. Fickser: If you cast a ballot, it is your ballot. If the ballot is cast by someone else in your name, you deserve to challenge it and get it changed.