Blockchains and Banks
As we worked with Eris Industries in the Anthemis Foundry, we came across the conundrum concerning blockchains: Bitcoin bundled together various existing technologies in a unique fashion to create something genuinely new — an almost unhackable, replicated database with no master server, via updates which are based on quickly verifiable effort rather than permission. Blockchains (permissioned chains) remove the genuinely new bit, but they are the current focus of activity in the Financial Services sector. Why?
[ tl;dr: Bitcoin created an awareness of a mechanism that could conceivably disrupt both banks and existing banking infrastructure providers via a financial network without either middle men or trusted entities. This created the incentive for strategic investment to look into blockchain applications. These replace banking infrastructure providers with software that doesn’t need to be run by a separate entity (e.g. SWIFT), and lower internal costs but still keep the requirement for trusted entities (banks themselves).
What blockchains achieve could have been done before, albeit in a less elegant way, but there wasn’t the alignment of incentives to produce the applications that will create a de facto reality.
Even without the advantages of the Bitcoin approach, this will stimulate the creation of Internet era banking infrastructure. With them, we could see the web of money and, more generally, any kind of contract. ]
If I have a digital copy of a song, I can duplicate it millions of times for almost zero cost. I can’t do that with a physical product like vinyl, I’d have to stamp a lot of plastic.
This is how file sharing threatened to disrupt the music industry, and how Napster was born the last time there was buzz around decentralized systems in 2001. This was the period that produced Skype and JXTA and BitTorrent and when O’Reilly created the Peer-to-Peer conference that morphed into Web 2.0.
15 years later there is renewed buzz around decentralized systems, but this time it is not about duplicating information but preventing it being duplicated.
Duplicates of a Unique Version
When people duplicated a song on Napster, each version of that song was actually identical. Ironically, if you could make it so that the file could not be changed without permission or costly effort, you could create a system where the identical copies meant there was only one version of a song, but duplicates of that identical version that everyone could listen to. That single version of a song could equally be a file that said who (or what address) was associated with an account balance. Everyone could ‘listen’ to that account balance, but nobody could alter it, thus creating a unique version of the truth, widely duplicated. This would be a shared account balance database — a shared ledger.
This kind of system would not work for creating a Music Industry version of Napster, because reading the database — listening to the song — is where the value is (nb. a modified version could). But for contractual exchange of value (i.e. buying and selling) and more generally agreement of any contract, where the value is where the file says it is (e.g. someone else has $5000) not in the act of reading it, it works. And a system like this works well for a lot of banking.
If you create a network where every user uses the same database, you can prevent people from copying or changing data within that network, without permission. If you tie the update of that data to finding the unique way to lock it up and that requires doing something physically costly like doing millions of calculations that inevitably cost money in electricity spend — then you have inextricably linked the digital world back to the physical one and you prevent anyone from changing the data with or without permission unless they are prepared to spend that money.
This means the data can represent money, much like a piece of paper can. But because the data is in a ‘network’ it can move around geographically, much more easily that a physical piece of paper.
Ironically, by using a similar technology that allowed people to remove scarcity for things like music by copying files, Bitcoin solved the problem of making digital things scarce by preventing people changing their contents.
Blockchains take much of this approach and apply it to more general use cases that don’t necessarily involve the creation of currency. This is a necessary component of the Bitcoin model, to reward and incentivise the ‘untrusted’ people who update the database and so blockchains require trusted entities as nodes, even if the network itself doesn’t require a separate trusted third party to run it.
Banking’s Not Becoming Digital, it Became Digital Too Early, Pre Web.
By using blockchain technology to replace banking infrastructure the banking industry is not suddenly becoming digital, it has been for decades. But it became so too early, before the web, therefore much of its back end infrastructure is based on a less networked world. The system runs in batches, stopping and starting, and databases are not linked together sufficiently. Banking infrastructure is less like an Internet platform with a packaged end user service that combines data and communication of that data, than a telecoms one which is merely concerned with connecting the data (the pipes not the whole end user application).
A more modern, Internet era banking system would consist of realtime transactions settling on synced, duplicated ledgers. This could have been achieved years ago, in the days of Napster, but like many regulated industries, the ecosystem had self-configured so there wasn’t the incentive to innovate. Blockchains have created the market awareness and incentives to innovate as much as the technological means.
Virtual (Software Only) Consortia
Current banking databases are disparate and therefore require reconciliation between each other, something which doesn’t happen in the realtime manner we have become accustomed to with the advent of the web. The network connections between these databases, the ‘rails’, are separate and run by 3rd parties and many of the rails that banking systems operate on are actually standardised networks operated by middle men, created as a consortia operations by the banks themselves.
It’s the existence of these consortia that has prevented banking systems from evolving, because there was no incentive for a single bank to create a better system outside of the consortia. They would be in control of a non-interoperable standard that nobody used. All transactions happen between multiple parties, that is their nature, so some form of synchronized system and innovation which is simultaneously adopted by multiple parties is required.
To replace legacy systems with Internet era ones, there are other models than that of multiple entities tied together by mutually owned consortia: (a) the winner-takes-all platform monopoly model that works in many industries and created Googles, Facebooks and Ubers and (b) the utility one, where the entire banking service, not just the rails becomes a utility much like roads.
(a) will not happen because regulatory requirements mean that even if there are global Financial Service brands, these sit on different local systems that are tied to local jurisdictions. There are certain Internet platforms that stubbornly remain regional, one of these is real estate listings, another is banking. There will be no Facebook of banking.
(b) could happen, where there is no Facebook of banking but both enterprise and consumer Internet platforms from social media to SaaS based accounting own corporate and retail customers. i.e. there will be no Facebook of banking but Facebook etc. could own the banks’ customers. Precisely because this would be an undesirable outcome for the banks, it will naturally drive movement towards a new version of the consortium model.
This updated model can be done much more efficiently with blockchains, which could replace 3rd party consortium entities with software alone. By removing a middle man from doing this, the constraint (that the consortium cannot threaten the banks) that ringfences the consortia from not encroaching outside of their designated territories disappears and would allow blockchain based systems software to fulfil their potential to innovate and do more than existing cooperatives like SWIFT can.
As a side note, the Napster or Bitcoin style model of truly decentralized systems with no entity in overall control of any portion will probably not happen in Financial Services unless there is a fundamental shift in how societies work, triggered by them.
Who knows, maybe this could happen, but not in the short term. The reasons for this are that decentralized systems can’t be governed by anyone, so they end up being either compromised or outlawed and secondly, they decentralize revenue. It’s difficult (not impossible) to imagine a decentralized Facebook usurping them because Facebook centralizes revenue and therefore may be able to compete more effectively. Both of these reasons are why the last wave of decentralization faded.
In summary, the ‘multiple entities tied together by software only consortia’ model will be where banking innovation for back end processes and infrastructure happens, and blockchain technology will certainly be the initial focus.
How Blockchain Innovation could Play Out
How might the Blockchain based virtual consortia (consortia that are purely in software) model evolve?
Until today, banking innovation through Internet startups, ‘Fintech’, has been concerned one of three things:
1. better interfaces to existing technology through fancy APIs (e.g. Stripe).
2. better design, simplicity, brand and positioning (e.g. Transferwise).
3. regulatory arbitrage (new kinds of marketplaces with different rules e.g. LendingClub)
All of this has happened against a backdrop of unusually low Interest rates which have subdued existing revenue streams and created an opportunity for startups (e.g. an unregulated pre-paid credit card startup, offering accounts with no interest and with a mobile phone interface looks and feels much like a real bank’s current account when Interest rates are low).
None of the three types of Fintech innovation have tackled the underlying technology, because until now, nothing has threatened the consortia businesses to untangle the connections that bind everything together. And because these consortia take fees which are less than their members do, you can only disentangle these systems if there is a simultaneous threat to both the consortia and their members.
Bitcoin does theoretically threaten both — it gets rid of trusted middlemen and trusted parties and counterparties altogether with a new kind of software architecture for transactions, comprised of existing technologies put together in a truly novel way.
What Bitcoin does is provide a perceived threat to both the consortia model and the banks, and in doing has stimulated banks to break rank and abstract the blockchain architecture from Bitcoin to innovate on the back end technology.
Blockchain interest within banks hasn’t been driven by panic, but by intellectual curiosity, and the ability to authorise R&D expense to look into strategically threatening technology. This money will create applications and the applications will be useful. The crack has been opened for a feedback loop of innovation that delivers results that in turn justifies the expense of new innovation.
Internet era banking technology has found its ‘ecosystem fit’ in the form of blockchains.
A New Kind of Database
Ironically it may not matter if blockchains do anything new that couldn’t have been achieved with relational databases, when they are stripped of the principal innovation that Bitcoin offered (no requirement for trusted entities). If enough code is written and enough people use it banking rails will provide realtime execution of any type of financial contract, anywhere and blockchain systems will evolve into the most elegant systems to do that. Given that their attributes are what is needed for the job and that configuring existing systems to achieve what is needed is complex and messy, blockchains will become the most elegant de facto solution even if they were not, a priori.
With blockchains added to the mix, the database universe could be: (1) full-text search systems for fast retrieval, slow updating, and low accuracy (useful for search engines, social networks and document systems). (2) relational and object databases, used for complex structured data and balanced retrieval and update speeds (useful for business processes and general purpose applications). (3) blockchain systems, useful for highly accurate transactions between multiple parties with structured data that is simple enough (as most financial contracts are) that it won’t change too much over time (new fields added etc.) these systems are possibly slower and the data much less compact.
A New Kind of Web
But the outcome may be more than this. Public chains, with their genuine innovation could play a role in making the promise of the blockchain not being a new database, but a new protocol or network built on that protocol. This would create not just a web of money, like Bitcoin has, but the web of money and would possibly go beyond that, even, to create a universal ‘transactional web’ for any kind of contract.
Nobody knows how this will play out, whether there will be lots of private chains, like disconnected Intranets, or there will be one web, one public chain to rule them all. In the interim, what is more likely is that the separate strands of public and private chains will later join up if the backbone evolves. The slower, but more secure Bitcoin style public chain could become the backbone for an Internet of transactions, where private chains among groups of corporations checkpoint against that chain at periodic intervals.
If this results in the web of transactions it’s a very big deal.