For ten years Angus de Crespigny worked for EY. In the two years until he left in August, he headed up blockchain for financial services in the US. During that time de Crespigny became critical of private blockchains.
A large part of Ledger Insight’s coverage is about private permissioned blockchains. To reflect de Crespigny’s opinions, no further commentary or analysis has been added. And a transcript of the interview is appended.
What does blockchain mean?
De Crespigny asks what the differentiating factor of a blockchain is? Because the definition of blockchain is used loosely, including to describe distributed ledger technologies (DLT) that are not blockchains. GitHub is a repository where developers share and commit code. Even if the code were in blocks, we don’t view GitHub as a blockchain.
He points to two critical features: Who can arbitrarily change data and what’s the governance? He witnessed many groups developing governance processes and central organizations to establish standards for how to deploy a blockchain. In his opinion, if you have a central party doing that, then you might as well use a centralized database.
That’s because blockchain and DLT have disadvantages. Duplicating code and data across numerous nodes don’t come cheap, and upgrading involves repeating the process on every server. Plus blockchains are far slower and less scalable compared to centralized databases.
Given that code in the form of smart contracts is used to change the data, if you don’t want people making arbitrary changes to the data, then what about changes to the code? “Who is it who’s actually propagating the changes, deciding the changes to the code?” asked de Crespigny.
Hence companies need to review code and have the option to reject changes. In many “consortium” governance structures that opportunity won’t exist.
“The thing is how people get there is typically by forming a consortium to agree on all of those standards to sort of lay it all out,” said de Crespigny. “And if you agreed on that consortium, you’ve created that trusted party that defines the standards anyway. So the technology may make sense, but the process that you get to that point, that can often mean another technology is just as good or sometimes better.”
Don’t solve human problems with tech
Given the desire to stipulate how to do business and how things should operate, this could be done in a legal agreement rather than using technology.
“You see this in technology a lot, that people and certainly in this industry, that people will look to technology to solve what are human problems,” observed de Crespigny. “Which will be corruption or fraud or more broadly with the blockchain world, coordination. Now technology can make things faster can make things more efficient… but I think expecting it to solve those problems is a challenge.”
Lack of due diligence
Additionally, the technology carries risks because it’s new. IT departments don’t necessarily have the expertise.
“I don’t think people are really doing … objective due diligence on the technology and instead really going with what they’re being told by some of the providers,” said de Crespigny.
What the technology is suitable for
Bitcoin was the first “use case” and the purpose of the blockchain design. It solved the double spending problem that prevents a person sending the same Bitcoin to two people, and it does so without trusting a central party.
De Crespigny said: “Now if you are not looking to do that or if there is a central party that you can establish a legal relationship with, you’re taking on a lot of the disadvantages, a lot of the trade-offs that the technology made for a benefit that may not be applicable to your problem.”
He continued: “The challenge at the moment is financial services is centralized. Now if you’re to do anything to try to decentralize financial services, you’re not going to be able to do that unless you decentralize money, to begin with. Now if you decentralize money, then you can start creating certain decentralized assets on top of that. But decentralization in itself isn’t necessarily beneficial.”
Comparing the internet to blockchain, it took years before e-commerce, social networks, and more advanced applications appeared. First, the foundation of the network had to settle in, which was followed by mass adoption. Likewise, de Crespigny believes that the core of a blockchain is a cryptocurrency like Bitcoin and Ether. And only once those are in day-to-day use will the other applications that sit on top of them make sense.
On reconciliation and distributed databases
With a conventional distributed database every company could have a node and automate reconciliation as with a blockchain. De Crespigny acknowledged that a major difference is that a master node can overwrite previous data, which cannot happen with a blockchain.
“I think if you’re relying on the technology to ensure that people are following the law, I think that is a dangerous proposition,” said de Crespigny.
Specific use cases
Ledger Insights asked de Crespigny about quite a few specific use cases. These are a few highlights.
Many fraud and anti-counterfeit blockchains store hashes on a blockchain for later verification. “If all you’re storing is a hash, then you may as well just do that on the public blockchain because you have less infrastructure cost,” commented de Crespigny. “So you don’t need to store a node on your system. And you can leverage the sort of irreversible power of the network that’s already established.”
Some sectors that are embracing blockchain and DLT are ripe for digitization. Syndicated loans and insurance are two examples. And the halo of blockchain is helping in that process. De Crespigny said: “If we were to look at it and say, look in the end blockchain may not be the best way to do it. But this has got us to a point where we have digitized all of these things that we were holding off. And now we’ve got a long-term, a far more sustainable long-term efficient business. I think that’s a success.”
Below is a transcript of the phone interview on 26 September 2018. Minor edits were made to remove repetition.
Why de Crespigny left EY
I mean a very very quick summary of that is really, I wanted to focus specifically on Bitcoin and cryptocurrencies and dive into what I’m most passionate about.
So, that’s really the sort of the sum of that. And as a consulting firm and at least in the position I was, you need to, you should cover, more broadly the industry. You need to do that, and that’s really all it was. I just wanted to focus on what my passion was.
And so consequently what I’m doing now is well a few things. But I’m independently, so I’m advising a few sort of different projects.
The two main things that I am working on is cryptocurrency custody. So better ways how asset managers and the industry can manage the custody implications or the custody process for these assets which were specifically designed to change how they are custodied.
And the second piece which is a big interest of mine which I’m starting to help out within certain areas is: How Bitcoin and cryptocurrencies more broadly can operate as a store of value in certain developing countries around the world. Where people look for stores of value outside of the local fiat currency.
So that’s kind of the path or overview of those points.
Store of value in regions
There are two aspects. There is on one side the Venezuela bucket where you could see that people have so little faith in local currency that they’re almost looking to use this out of necessity. I think that’s an interesting bucket.
I think that’s not necessarily the only one. What you do find in other parts of the world is a big appetite just simply to store value in say gold, right. You get that in India. You get that in Turkey. You get that in Iran.
And what some of these countries see is that when you’ve for centuries, there’s very much a cultural appetite for some store of value outside of the local government. And gold is often the fallback.
There does seem to be appetite around in some of those countries for something other than gold. Particularly since gold can be rather cumbersome to carry, to hold, to secure.
Now Bitcoin and other cryptocurrencies aren’t much better at the moment with the managing of keys and wallets. But we’re still very early days with that, and I think that’s where there’s a lot of opportunity for those countries.
Skepticism over permissioned blockchains
There are a few aspects to it. One is that everyone who’s developing some sort of new technology at the moment calls it a blockchain, right? So there are challenges with defining exactly what people are talking about when they talk about a blockchain.
Is it something that because if you’re talking about something that simply just shares blocks, shares data in blocks, then by some measures GitHub could be a blockchain. But that’s not really what people mean. They’ll often talk about coming to consensus amongst a lot of parties. Well, so does Google Docs so do distributed databases.
So what’s the sort of key differentiating factor? And I think the only thing that really sort of makes sense is when you talk about technology that changes the dynamic of who can arbitrarily control the data.
So if you look at Google Docs or distributed databases, then Google in the former or whoever owns the master in the latter, can arbitrarily change data. But if you can do that with a blockchain, then you should just compare it with other distributed databases.
If you can’t arbitrarily change data, you’ve got to consider the other side of the coin of that which is: how do you govern something where there is no one party that’s governing? So what we’ve seen a lot. Or what I saw a lot was that a lot of people were developing sort of governance processes and central organizations to develop the standards around how a blockchain should be rolled out.
But when you develop that central trusted party to do all of that, then the question is why would you sacrifice what you need to from … to implement a blockchain so that you can also enforce them at the technical level.
Because implementing blockchains aren’t cheap. You’ve got data that gets duplicated across a lot of different parties. You have code that gets executed across a lot of different parties. If you want to do any upgrades, you need to get everyone to individually upgrade.
So there is a trade-off in taking away that Central coordination. On the plus side, nobody can centrally coordinate. On the negative side, nobody can centrally coordinate, right.
And in an Enterprise environment where systems go wrong, systems fail. You know coding goes wrong. That’s why we have big large IT organizations in every Enterprise. I think it’s a big ask to put in this technology and say that it’s going to solve all of the problems that people say it’s going to solve.
And I think when you have a setup that is, say called the Maersk IBM set up right then, who is it who’s actually propagating the changes? Deciding the changes to the code? All of that sort of thing. Right?
And if it’s those two, then that’s already kind of centralized. Right? So if we have a think about how a governance process might work in a permissioned blockchain world, you don’t want anyone to be able to arbitrarily make changes. So, therefore, any code that is distributed you want people to have… The counter side to that is you need people to make sure that they’re looking through the code and that they agree with the code and the opportunity to not accept those changes.
Now, you could either do that at a code level. Or you just do that like we do with anything in business. And just create a legal agreement that, here is a party that we want to do business with and this is how they’re going to operate. We do that all the time. It’s how we create financial agreements. How Banks create financial agreements between each other. How over the counter financial products are created.
So the question is. If you can do all of that at the legal level and considering that, this code and standards and all of that sort of stuff is determined by that central party that everyone has a legal agreement with anyway, why would you then use a technology that is so cumbersome to deploy? And sacrifices transaction speed. And sacrifices storage space.
And also frankly it’s just a new technology that your current IT department doesn’t necessarily have the expertise with. It’s not clear why you would make that trade-off.
And then the thing is when any technology that’s being deployed should be assessed against other technologies to make sure that it’s the right one. We don’t go down and implement a Hadoop system without having a look to see if it can already be done on our current systems.
The thing is you see a lot of going straight to blockchains just because of everything that’s been sold on. And then trying to sort of fit that into the current environment. And I don’t think people are really doing that objective due diligence on the technology and instead really going with what they’re being told by some of the providers.
What blockchains are good for
You come down to what this technology is good for and what it’s not good for. And what it’s good for was solving a problem that online digital currencies had. Which was how do you know that I have sent you something that I haven’t already sent to someone else. And that was people have been working on trying to implement digital currencies for years, and this solved that specific problem. Because previously you could only do it by trusting some sort of central party.
Now if you are not looking to do that or if there is a central party that you can establish a legal relationship with, you’re taking on a lot of the disadvantages, a lot of the trade-offs that the technology made for a benefit that may not be applicable to your problem. So if we’re looking at that specific problem, then we’re sort of looking at that problem or that solution will really sort of need to be adopted first before we can deal with that sort of stuff.
The challenge at the moment is Financial Services is centralized. Now if you’re to do anything to try to decentralize financial services, you’re not going to be able to do that unless you decentralize money to begin with.
Now if you decentralize money, then you can start creating certain decentralized assets on top of that. But decentralization in itself isn’t necessarily beneficial. The whole thing is why might people want to do that? So you kind of need to have in my view the base reason for people to adopt this technology that kind of needs to take off.
In other words censorship resistant money before the stuff on top makes sense.
So to have an analogy with say the internet, creating social networks or IoT devices on the internet in 1995 wouldn’t have really made sense. Because people weren’t using the internet to communicate. Some people were starting to get online, but really you needed people to get on to say use email or to create a website or for their business.
But until that took off, you’re not going to get people connecting to the internet to join a social network or connecting to the internet to automate their home devices. Because there were too many on ramps that you needed to build to get to that point.
So it was only once you had people having that native digital fluency and people being able to you know, people already plugged in, that okay, you can communicate with people. You can build websites. Now here is the other stuff you can do. So we went into commerce and so on.
And I think that’s how blockchains need to be looked at. For all of this to create a security on top of the blockchain, you’ve got to get everyone plugged into that blockchain first. And right now, I don’t think the economics make sense. If you have everyone or a critical mass of people using say Bitcoin or Ether as a native currency just on a day-to-day basis, creating something on top of that, on top of the network which you are already plugged into, starts to make sense. But you need that initial pick up first.
What are the issues with anti-counterfeit/fraud blockchains?
So it was spoken about using it to see where things stood along the supply chain or something like that. But you can’t ensure that someone is going to, you can’t make someone put the hash into the blockchain, right.
The second piece for that is if it’s a private blockchain or permissioned blockchain you are getting, you need to pull everyone together and decide what are going to be those standards. That if you put a hash on there, what’s actually being hashed? What’s the data that’s being hashed and how you’re going to share that information outside to validate it? So you’ve got to sort of develop standards around that, and you’ve got to pull together people to do that.
The third thing is if all you’re storing is a hash, then you may as well just do that on the public blockchain because you have less infrastructure cost. So you don’t need to store a node on your system. And you can leverage the sort of irreversible power of the network that’s already established.
So if you can with these things, if you can get everything onto the network, then it can make sense.
The thing is how people get there is typically by forming a consortium to agree on all of those standards to sort of lay it all out. And if you agreed on that consortium, you’ve created that trusted party that defines the standards anyway. So the technology may make sense, but the process that you get to that point, that can often mean another technology is just as good or sometimes better.
In other words, coordinating people. It’s a people problem, not a technology problem. And once you have coordinated everyone, it’s not always the best technology to use.
What are the issues with trade finance?
Yeah, I mean it comes down to who’s defining the standards? Right? Because if you can coordinate to define the standards, then at least on the private or permissioned blockchain side then if you’re coordinating to build the technology distributed, you may as well just build the database and distribute nodes.
If you’re not looking to build the technology and it’s simply just defining standards so that you can use it on a public blockchain, almost sort of by design it’s probably going to be simple enough to use on that. But you are really just then using the blockchain for a notary service.
So expanding outside of the value proposition of the reason why you can use a blockchain for moving value, is because you can do essentially a timestamped notary services on who had things at what time.
Now, if that’s all you want to use, then you can use public blockchain infrastructure to do that. But like with any of these things the devil’s in the details. How do you decide what is being notarized and how do you decide that? And how do you determine that people along that chain, are recording things, recording things accurately, and recording what’s recognized by the group to be the thing that everyone wants to record?
On something like Corda?
If it’s not centrally controlled the question is again who’s defining the standards for it? So if it’s totally done and all individually and each organization is somehow helping to coordinate the standards around this, and there’s no central investment, and there’s no central governance.
I’m not sure how that would work, because at some point someone needs to agree on what the standards are. At some point, there needs to be someone defining how the endpoints are coded. And unless you’re going to have every separate entity checking the code, once those are distributed from whoever’s making the changes.
I question if that’s an effective model versus just simply setting up a central sort of member organization like say the DTCC. Like it was set up a few a few decades ago to simply, to act as a central repository for all of this and centrally manage things.
Well, I think it goes to the point that you, I think you nailed it somewhat when you were saying before about insurance that the hype is pushing people to digitize and I think that’s the key thing, right.
Syndicated loans as well is a very document-heavy process. And with some of the hype and excitement around this, it’s causing people to rethink a lot of this, a lot of this stuff. And maybe we can digitize far more of this stuff as part of this process.
And if we were to look at it and say, look in the end blockchain may not be the best way to do it. But this has got us to a point where we have digitized all of these things that we were holding off. And now we’ve got a long-term, a far more sustainable long-term efficient business. I think that’s a success.
And the tokenizing is simply recording data onto a database, in this case the database that’s spread across, that’s a blockchain, right? So the question is do you want to record that data in a technology that’s set up just simply to stop people from being able to stop you from sending that data. Or is it something that once you digitize you can simply have those contracts running on an upgrade of current infrastructure.
I think with syndicated loans when it’s some of this over-the-counter stuff. There’s nothing really stopping someone from setting up a central exchange to clear all of this sort of stuff.
And that’s sort of what would end up being a case for if you created a blockchain to do it. Would there be someone who’s defining what’s the data? How do you form that syndicated loan contract? How do you record that information onto the blockchain? What are the sorts of inputs and outputs? What do we do if there’s a change in the contract or some sort of data quality issue that we didn’t expect?
There would need to be someone who’s watching and administering that. And if we got to the point where so much of that is digitized that we could put it on some market like that, I think we find that a technology that is set up to stop any one party from making changes, I don’t think that would be the most practical case. Versus setting up the system that just looks like a traditional exchange.
Blockchain reconciliation advantages?
I don’t see why that’s different from a distributed database, right?
We can have nodes of the distributed database sitting at every different company or counterparty’s machine. And they can be monitoring that like they would a blockchain.
The only difference being that if it was a blockchain, then no one could send out something that would sort of overwrite the previous pieces. If it’s a distributed database whereas the master node could.
But I think if you’re relying on the technology to ensure that people are following the law, I think that is a dangerous proposition.
The stable coins are funny, and I must admit it confuses me sometimes, the excitement around it. Because in theory it sounds great, but I don’t think anyone’s really produced a method for how it’s going to be effectively deployed. And from what I’ve seen, they fall into two buckets. One would be sort of …Firstly any sort of stable coin I think stable is a misnomer.
I think it’s really a pegged coin, right. It’s pegged to something. And but how is it pegged there? And there are two ways you can do it. Algorithmically, which is in theory the market economics around the coin will settle it to a price. Or you have it backed by something.
Now, let’s say you’ve back it by the US dollar then you have 20 million dollars that you put aside. That’s not a great capital use, right? You also have some central party that when those coins are moved around, and you get this with the regulatory one.
If they need to follow regulatory rules, then why would someone use that coin versus just the US dollar or whichever coin it’s backed to. So it’s I think the idea of stable coins sounds good for say people in the developing world that could do with something that is more stable than the current cryptocurrencies on the market, but I haven’t seen one that’s really sort of differentiating, that doesn’t just simply look like a bank account but on a blockchain.
But isn’t that the whole point that they actually want a bank account on the blockchain so that they can have finality in transactions?
Well, the thing is let’s say you get that coin and it’s kind of a claim on the asset, it’s not the dollar or the currency itself. So then it falls back onto well how do you actually get that currency into your system?
You need to actually settle it at some point so it can you know, this sort of comes into what you define as settlement. That stable coin in that sense, it’s an IOU, right?
So unless you get to a massive adoption where the asset itself, the stable coin or whichever coin it is, is recognized to have value in itself. You still have that settlement leg where you need to move the dollars.
Now I think if you get to that, if you get to that point where you’ve got a network where the asset itself is recognized, I think that does change things. But that’s also the biggest obstacle with any payments mechanisms. Payment is much more a network problem than it is a technology problem.
In that how many, Swift has value because it has so many people on board. And I think you’ve got to look at the technology level. Can this accomplish something that say SWIFT couldn’t do with a payment upgrade? And I’m not sure if I necessarily see that.
Enterprise technology assessment
I think the main thing with this is it’s not to say avoid this technology. But with any enterprise transformation, you should assess the available technologies to you. And choose the best one.
And assess that this stuff isn’t magic. It’s another technology and when you’re developing your product to assess and evaluate it against other technologies that are out there. Be they distributed databases or other forms.
I think what’s happening a lot, is people are catching on to the buzzwords and not making assumptions that this is more efficient. That this is faster. That it’s automatically reconciled.
Some of those things I think don’t necessarily hold up to scrutiny. So I’d recommend that any enterprises when looking at this have the same lens on assessing this technology that would against any other.
Is there a role for tokens to incentive ecosystem players for enterprise networks?
Look I think there could be. I think the thing that is often underestimated in this space is the complexity around decentralized networks. There was a bug discovered in Bitcoin just recently. And all Bitcoin is doing is just moving value from A to B.
There’s some other stuff, but it’s as simple a protocol as you can get. So even with that, critical bugs are coming up. So, in theory, could that happen? Yes, but I think it comes down to practice. And because any incentive model or economic model that you build by design is just going to run on the system and can’t be changed.
So in an enterprise environment, you’ve got two choices. Do you do a database which you can manage? Or do you set up a technology that will just, well rather you won’t be able to modify it. In certain cases the latter that may be useful. If you want say an industry group to basically do nothing, but to sort of time stamp to notarize data.
But I think the challenge that the public blockchain industry has seen is incentives, and economic incentives and economic models are incredibly difficult to design. Because people can often game it in ways that you just can’t anticipate when you first design the system.
Blockchain for human problems
The thing you see this and technology a lot, that people and certainly in this industry, that people will look to technology to solve what are human problems.
Which will be corruption or fraud or more broadly with the blockchain world, coordination. Now technology can make things faster can make things more efficient, but it can’t necessarily… but I think expecting it to solve those problems is a challenge.
So where you can solve that through the current legal system, I think that that’s a better system.
For now, for more complex things because law isn’t black or white and if you try to record laws or trying to record these sorts of things into something that can never be changed, I think that’s far more complex than people are expecting.
Do you characterize yourself as a Bitcoin maximalist?
I would characterize myself as a coin maximalist. I think if you read that tweet it’s not from any sort of religious belief and I think that’s somewhat of a misunderstanding often of people who classify themselves as Bitcoin maximalists.
For most, if something else comes along that solves the problems in a better way, they’re open to and would accept that.
I come to what’s probably coin maximalism and in that this technology, I haven’t seen it really being able to solve better than current technologies, anything other than that censorship resistance value transfer.
And I think considering that is a problem of networks. Then from that things will naturally coalesce around one. I think Bitcoin has certainly got the head start, has certainly got the network that a lot of others don’t.
And is being used in the developing world exactly for that. It’s not to say something else couldn’t make some changes that do that better, but I think it’s kind of like TCP/IP which is the protocol that the internet runs on.
Back in the nineties, people were very skeptical that that was going to stay or was going to be able to handle everything. But it turned out to be good enough.
And the question is if, in the end, all you need is a censorship-resistant value storage and transfer, which is critical to a significant percentage of the world’s population. If that’s all you need, is the coin good enough to do it? I think we will see. But it’s certainly in front of the moment.