Filed Under: Finance and Banking, Money

You can’t rob a glass bank, even if you work for it

Leave a Comment

At the Imperial College (packed) discussion on “Distributed Ledgers – Future Research Challenges“, chaired by Professor Bernard Silverman FRS, the Home Office Chief Scientific Adviser (and a mathematician), a series of speakers (including yours truly) sparked a valuable and fascinating series of discussions around the topic and, in my case at least, left me feeling as if I’d actually learned something.

In the morning, Iain Stewart from Imperial College introduced us to his “Nonsense Watch”. It turned out that his nonsense watch only had two things on it:

  • We hate Bitcoin but we love the blockchain.
  • The blockchain is efficient.

In a memorable presentation, elaborating on these topics, he told the assembled group that the a good way to think about the blockchain is to compare it to somebody swallowing condoms full of heroin and carrying them through customs in your stomach. It’s a really inefficient way to transport heroin around but you have to do it because “powerful forces” (as Iain called them) are trying to stop you from doing it!

I will never forget that example! Anyway, just to explain the background. Consult Hyperion were asked to become part of a consortium bidding to examine the potential for Bitcoin, the blockchain and suchlike across a variety of sectors in response to the Treasury’s decision to allocate £10m in funding for the topic. In this context I (along with a couple of my colleagues) took part in discussion at Imperial that brought together academics, technologists, government and a number of different businesses (including banks), which is why we were listening to Iain.

I thought it would be helpful, with such a mixed group, to use a narrative that would help people to communicate effectively and share ideas. This is why I used the “glass bank” example that I’ve used before and built on the presentation that I gave to the Dutch National Bitcoin Congress in June. As it turned out, it worked very well on the day and after discussing it with a couple of other people I’ve decided to expand it as clients might find it a helpful way to think about the new technology (as they get a bit bogged down in Bitcoin and cryptography). I have to say that it worked largely because Richard Brown from IBM had set things up so nicely for me with his discussion about “Creation Myths and Shared Ledgers” that immediately preceded my talk.

The actual purpose of my talk, narrative aside, was to put forward three solid ideas for research threads that could form part of the project. I’ll blog about this, but I was looking for examples of areas for genuine research, areas where the answers aren’t known, that could complement shared ledger technology in some way to deliver something special or different groundbreaking.

In the end the three examples I settled on were:

Homomorphic encryption. Although I wouldn’t say I was absolutely up to speed on the state-of-the-art in this field I do understand the rudiments and it strikes me as an area where any small improvements could lead to pretty significant benefits. This is an area where pure mathematics is needed and I would’ve thought that most businesses and even technology companies just do not have that kind of research going on.

Publicly-private records. This builds on the idea of “translucent” databases to use homomorphic encryption encryption to put data on public blockchains that can be audited in necessary ways but remain private. I don’t think it’s enough just to store encrypted data on public blockchains. If we can agree on the use of the word translucent to mean data that can be audited while remaining encrypted, then I genuinely do feel that a new kind of financial services industry could be on the horizon.

Bottom-up identity. It occurs to me that if it was possible to use homomorphic encryption to store publicly private records about an individual then the cryptographic techniques that are currently used to demonstrate attributes without revealing them (e.g., interval proofs) might be transformed to help creates a shared infrastructure for identity built on very different foundations (e.g., testing that an age is >18 without decrypting the age).

As I say, these are areas for research. I don’t know what might be discovered in these fields any more than anyone else does, but I have a feeling that it might be both important and of immediate practical application. Now imagine that we bring those technologies together to create “glass institutions” in the financial services world. This would be utterly transformational, in a way that making payments cheaper and quicker (even if this were true) is simply not.

The idea of glass institutions may seem paradoxical but with the advances in technology and our evolving understanding of how replicated shared ledgers might transform a variety of different kinds of systems, I think we can begin to explore their impact. I rather like the language of translucent transactions and I think it works well with the glass bank narrative to open up sensible discussions at the business (and regulatory) level.

So where does this take us? Well, as Richard said in his talk, a replicated shared ledger in financial services is unlikely to be “permissionless” in the censorship-resistant sense that Iain was talking about at the start of the day. However, it is entirely possible and highly desirable to construct replicated shared ledgers that allow for permission and innovation in the use of the ledger even if the ability to create transactions on the ledger is permissioned. Of course, this is not to say that both permissioned and permissionless ledgers cannot co-exist. Michael Mainelli provides an excellent narrative for this perspective, talking about the “Temple of Financial Services” in comparison to the “Souk of Sharing Economies”.

While my heart is with the Souk of Sharing Economies, my head recognises that there may be room for both. A sensible union would be a few, competing, ‘blockchain-type’ services encircling the globe providing end-of-day validation and recording of transactions, while thousands of mutual distributed ledgers do the busy work of serving thousands of shared economies. In effect, the merchants of the Souk bring their ledgers up to the Temple to be validated and timestamped by whichever priests occupy the Temple of Financial Services. It may not be orthodoxy, but it’s not heresy either.

[From iGTB – Liquidity Management – The Temple & The Souk – The Future Of Mutual Distributed Ledgers]

The permission, distributed shared ledger of the Temple will mean disruptive change. I can show this by giving a couple of obvious examples: what if a company chose from a group of regulator-certified auditing applications instead of from a competing group of auditors? Auditing banks’ books would become a continual process and you might even have multiple different applications constantly auditing the same bank on behalf of regulators, shareholders, customers, pressure groups and even rival banks. Anti-money-laundering processes would shift from expensive and rather useless gatekeeping combined with floods of suspicious transaction monitoring to being a variety of different anti-money-laundering applications combing through the shared ledger entries to find transactions indicative of misbehaviour (at which point, law enforcement agencies could apply for warranted access to the unencrypted ledger entry or relevant meta data).

This is why I don’t think it is an exaggeration to say that the shift to shared ledger technologies might be one of the most important innovations of our image of our age, and I will close by making another historical analogy to support that point.

In Victorian Britain, the collapse of railway companies led to a colossal crash in 1866. It was caused (and here’s a surprise) by the banking sector, but in that case it was because they had been lending money to railways companies who couldn’t pay it back rather than American homeowners who couldn’t pay it back. The British government then, as in 2008, had to respond. It suspended the Bank Act of 1844 to allow banks to pay out in paper money rather than gold, which kept them going, but they were not too big to fail and the famous Overend & Gurney went down. When it suspended payments after a run on 10th May 1866 (as frequently noted, the last run on a British bank until the Northern Rock debacle), it not only ruined its own shareholders but caused the collapse of about 200 other companies (including other banks). The directors were, incidentally, charged with fraud but got off as the judge said that they were merely idiots, not criminals.

The reason I choose this example is that railway companies then held the same commanding position in the economy as banks do today, so the impact on UK plc was substantial. Bear in mind that the first railway service in the world started running between Liverpool and Manchester in 1830 and less than two decades later (by 1849), the London & North Western railway was already the biggest company in the world. When the Directors of these gigantic enterprises went to see the Prime Minister in 1867 to ask for the nationalisation of the railway companies to stop them from collapsing (with dread consequences for the whole of the British economy) because they couldn’t pay back their loans or attract new capital, they didn’t get the Gordon Brown, investment bank advisers, suspension of competition law and the tea and sympathy of 2008. Disreali sent them packing as he didn’t see why the public should bail out badly run businesses, no matter how big they might be.

Needless to say, the economy didn’t collapse. As you may have noticed, we still have trains and tracks. A new railway industry was born from the ruins, the services kept running and the economy kept growing. And there was another impact. Andrew Odlyzko’s paper The collapse of railway mania, the development of capital markets, and Robert Lucas Nash, a forgotten pioneer of financial analysis argues that the introduction of basic corporate accounting standards following the collapse of the railway companies was a significant benefit to Britain and aided the development of Victorian capitalism.

So, with the well-worn maxim about not letting a good crisis go to waste in mind, I would like to advance this hypothesis: the long-term impact of the financial crash of 2008 will be a shift to the replicated shared ledger as the central organising principal for financial services. An entirely new way, as Richard Brown notes, of building financial institutions based on common ledgers and APIs.

Francis Keally’s vision will be realised and to the great benefit of society as a whole. After all, you can’t rob a glass bank, even if you work for it.

One thought on “You can’t rob a glass bank, even if you work for it”

  1.' Simon Taylor says:

    Not only does the transparency / please don’t fine me again story lend itself to shared ledger use, but think about some other dimensions of compliance.

    The over head of sub accounts structures in capital markets will make many businesses un-economic using current systems. A shared ledger is not only more transparent, but it’s reconciliation through cryptography holds the promise (not proven) of greater efficiency.

    With the spectre of structural reform looming, shared ledgers are the natural evolution of accounting engines. The traditional vendors have been very slow to grasp this fact and are woefully behind on this subject.

    Couple of other use cases for you:

    Gov.Verify as a shared ledger instead of SAML based logins.

    Heck even credit card clearing as shared utility from the schemes with reconciliation through cryptography. Much better than how it works today!

Leave a Reply

Your email address will not be published. Required fields are marked *

Tags: , , , ,