Will Blockchain Change Wall Street? The More Things Change, the More They Stay the Same
In a recent blog post, I discussed whether the blockchain is likely to be sustaining or disruptive to the financial industry and summarised the current consensus around crypto-currencies as “Bitcoin bad, blockchain good”. Every bulge-bracket firm has ‘innovation labs’ working on blockchains (as discussed in this video by the Financial Times’ Phil Stafford), and the days of sneering at this weird hobby of ‘cypherpunks’ are definitely over.
However, the devil — as always — is in the details. I recently found some excellent explainer articles written by Tim Swanson (@ofnumbers) on why people awaiting the demise of the modern financial system may be disappointed. The issues are quite wonky, but I have tried to summarize and make them understandable for someone with a basic knowledge of blockchains (view our policy brief on the topic).
The underlying motivation for interest in financial applications of the blockchain is easy enough to understand. Everyone agrees that the existing systems of settling financial transactions are very slow. Blockchain solutions promise to decentralise (good, right?) the recording of transactions, making settlement more efficient (sorry back office) and, most importantly, faster.
This is great but unfortunately goes against the raison d’etre of the blockchain’s original design. To understand why, let’s look at three issues that any ledger, such as a database, has to consider (thanks to Robert Sams’ (@codelogic) ‘Blockchain Finance’ presentation notes):
- Sin of Commission — forgery of a transaction
- Sin of Omission — censorship (i.e., not recording) of a transaction
- Sin of Deletion — reversal of transaction
Contrary to the popular impression that Bitcoin is designed to be ‘unhackable’ (i.e., not vulnerable to forgery or corruption), its first design priority is to be censorship-proof. Here, censorship refers to the ability to double-spend a crypto-currency by not recording one or more transactions on the ledger.
To avoid this, the Bitcoin protocol is designed to operate in a similar way to cash. Cash works on the principle that whoever physically holds the note is the owner; this means that the note cannot be double-spent since it can only physically be in one place at a time. Bitcoin also operates on a possession equals ownership basis — whoever is recorded as the owner of the bitcoins on the blockchain is the owner — there is no other recourse. This means that bitcoins can only digitally be in one place at a time and therefore cannot be double-spent.
As a result of this design priority, forgery and deletion are actually technically perfectly possible, if currently economically unattractive. This is a problem for applying the blockchain to finance as we know it. It is reckoned that only about 30% of total network computing power is necessary for a cooperating group of anonymous individual “miners” (who verify transactions on the blockchain) to act as one and reverse history for its own benefit (known as a ‘Sybil’ or 51% attack).
Existing collectives of miners have been known to exceed this threshold. The investment to do this is not large by the standards of large corporations or nation states, but it is currently not profitable because the assets up for grabs (i.e., bitcoins) are worth less than the investment required. Further, an attack of this scale would destroy the value of the coins being stolen.
However, if large numbers of off-chain assets (i.e., assets that exist outside the blockchain, such as stocks, real estate) start to be recorded and traded on the blockchain (this could be done by representing them as ‘coloured coins’, which are bitcoins that are ‘tagged’ as representing some off-chain asset), the lure of forging history and launching a Sybil attack will begin to satisfy the basic marginal benefit/ marginal cost trade-off. This is unacceptable in any jurisdiction with property rights.
Legal systems in the ‘wet’ (i.e., real) world require irrevocability of ownership, beyond some identifiable point in time, in order to avoid potentially limitless arbitration over property rights. The blockchain cannot guarantee this because it is theoretically possible for a collective of miners to forge ownership history by hijacking the blockchain.
Luckily, this finality problem can be solved with a consensus protocol (i.e., blockchain) that authenticates validators (i.e., the miners are not anonymous). This is known as a permissioned blockchain (as opposed to the unpermissioned ‘free-to-participate’ Bitcoin blockchain) and is the type of protocol being developed by incumbent banks.
In a permissioned blockchain it would be impossible to launch a Sybil attack because the validators are known and legal action could be taken against those committing sins of commission or deletion, in a manner similar to attempted forgery of any other database of records. This means finality of ownership can be achieved and the blockchain can be made compatible with legal systems.
However, if we implement this permissioned blockchain design then what are we left with? Looks a lot like a database run by trusted third parties.
Plus a lot can change.
If you liked this post, consider subscribing to Market Integrity Insights.
Image Credit: iStockphoto.com/sorbetto