The Emperor’s New (Cryptographic) Clothes

As far as investing goes, two words are more frightening than anything else:

“Trust Me”.

There are many reasons this is frightening, but first, the more someone brags of his honesty, the shadier he seems. As Mark Twain once wrote, “If a man speaks of his honor, make him pay cash”. Secondly, because trust is… complicated.

Entrusting one’s money to an investment manager, broker or third party requires one to believe that the party will act in a trustworthy manner, i.e. won’t embezzle it. Trust is usually established by three different ways.
First, through personal bonds. One usually trusts her siblings more than her cousins, and her cousins more than total strangers. Second, through reputation. Someone ‘in the business’ for several decades that has demonstrated excellent results is expected to keep acting in faithful manners. In both cases, trust comes from expecting that acting in bad faith would cost more than what it would bring. For instance, stealing several thousand dollars from a family member may have dire personal consequences that probably outweigh the immediate gains. But such expectation is not always true. For instance, the need to satisfy an addiction may cloud any long-term consequences. Or perhaps that decades-old reputation of good behavior is completely bogus.

When personal connections or reputation are non-existent, trust can be built through the assessment of a third party, himself trustworthy enough, that steps in and says ’I can vouch for this one“. Such ”adult supervision" is usually fulfilled by government agencies. Since those agencies cannot assess morality itself, they put in place a number of handrails to make defrauding harder and punishable. In theory, the cost of breaking the law times the probability of getting caught is greater than the gain from criminal activities. Unfortunately (as demonstrated by the ever-growing number of prison inmates) criminals do not always behave according to economic theory. As a consequence, the trust provided through regulation remains limited and endorsement by official agencies is only a stopgap measure.

As superbly illustrated by the Bernard Madoff scandal, none of those trust mechanisms is truly fool-proof.

But technology has created an opportunity to establish working relationships without any trust needed. Both by limiting the gains from fraud and by making such fraud nearly impossible.

For instance, the incentive for fraud can be considerably reduced by splitting transactions into much lower amounts and proceeding through a continuous flow of money. When a payer wants to transfer \$1M to a payee, she must trust the intermediary not to run away with the money. The higher the amount, the bigger the risk. But imagine if the transfer could happen as a series of 1,000,000 successive payments of only \$1.00. Each time the payee receives one dollar, the payer gets a confirmation, and then sends another dollar to the intermediary. This way, even if the intermediary is fraudulent, he won’t be able to steal more than one dollar. If completely automated, the whole process would take less than a few seconds.

Fraud can be made virtually impossible if the parties in custody of money are completely transparent. Bitcoin was precisely designed to ensure trustworthy transactions between people who don’t trust each others. Its genius is that the history of all the transactions are publicly available and unforgeable, through an underlying technology called ‘Blockchain’.

At its core is the ‘Hash Function’. This function converts a document, of any size, into a cryptographic code. Creating this code is extremely fast, but the inverse function (re-generating the input from the hash code) is impossible, and modifying the input while keeping the hash code identical is almost impossible. This allows to create a kind of digital signature that can be publicly disclosed while keeping the input private. Auditing is greatly simplified, because one has only to verify that verified data generates the same hash code than the one publicly disclosed. For instance, imagine a game where you choose a pangram (a sentence that includes every letter in the alphabet). You want to give a proof of the sentence you chose, but without disclosing it. You could say that its MD5 hash code is ‘896145144ce306f507839e6cc5db5343’. It would be near-impossible for the other players to infer what is your sentence from that code. Yet, later in the game, you could prove you chose “The quick brown fox jumps over the lazy dog” by letting others verify the MD5 hash code from that sentence is identical to the one you communicated previously. Change a single character in the sentence, and the hash code would be wildly different.

Now imagine that you play that game once every day, and that you want to be able to prove today’s choice along with all your previous choices. To do this, you would add the previous hash code to your new choice. Day 1, we saw that you chose “The quick brown fox jumps over the lazy dog”, so you write down the corresponding hash code, which is still ‘896145144ce306f507839e6cc5db5343’. The second day, you choose the French pangram “Portez ce vieux whisky au juge blond qui fume”. This time you generate a hash code for a text that combines both the previous code ‘89614…’ and ‘Portez ce vieux…’, which gives “b95d0b0aefc090f9e13b4f982249d24a”. When you communicate this code, you will be able to prove both that you chose ‘The quick brown fox…’ yesterday, and ‘Portez ce vieux…’ today. Because only ‘The quick brown fox…’ would have given ’89614…’, and only ’89614…’ and ‘Portez ce vieux…’ combined could give ‘b95d0b…’.

This method, calculating hash codes and adding them to subsequent records, creates a tamper-proof chain of documents. This is called the Blockchain.

What if a fund (pure hypothesis, of course) publishes a ledger with the properties of all the notes its hold, down to the last cent of payments and a loan’s properties at issuance, and ‘fingerprints’ this ledger issue with a hash code that is integrated into each successive ledger issues?

Essentially, this hypothetical fun has created its own Blockchain. Now, one could argue that it could lie about those hash codes by retroactively creating each record and pretending that they were untouched. This is why it is important that it also ‘notarize’ those hash codes in Bitcoin’s Blockchain itself, effectively time-stamping the record. Every week, it would add its tiny hash code to the giant ledger of all Bitcoin transactions, making it possible for anybody to verify which hash code were generated, and when.

Investors wouldn’t have to trust a fund like that. Because it could never lie. Maybe one day all advisors will be this transparent, and the industry will follow Heidi Klum’s excellent advice on how to maintain a successful relationship: “The trick is that you have to stay naked.”