Last month, HSBC announced it had issued a letter-of-credit financing over a blockchain. The deal involved supporting the shipment of soybeans between two units of Cargill, from Argentina to Malaysia, with ING serving as the seller’s bank.
Unlike a traditional L.C. deal, which takes up to a month to process because of all the paperwork involved, this transaction was completed in 24 hours, because everything was handled using the Corda distributed ledger operated by R3.
It is the latest, most concrete example of how distributed-ledger technology is poised to transform the $8 trillion world of trade finance. First, though, competing platforms are going to duke it out among corporations, banks and regulators.
This is a contest for creating a utility-like structure enabling the digitization of trade finance across many players. If it ends with no clear winner and a variety of blockchain solutions that have to connect via APIs, the industry risks creating a new generation of business silos.
On the other hand, if the industry settles on a common ledger framework, it could breathe new life into the L.C. market, and give banks new ways to both cut costs and drive new revenues.
L.C.s make up about $1.4 trillion of the total trade-finance market in terms of annual contracts. That seems like a big number, but L.C.s have been in decline as a proportion of trade financing tools, and now account for only about 11% of transactions. L.C.s are commoditized – one contract looks mostly like another – but complicated and paper-intensive in their execution, with banks coordinating documents among freight forwarders, export credit agencies, insurers and regulators – none of which have ever agreed on a standard of automation, therefore making the whole thing slow and error-prone.
Account receivables (selling invoices to banks) and open-account financing (in which exporters ship goods before getting paid) have become preferred over L.C.s, in part reflecting the greater influence of anchor corporate buyers in the process; in trade finance, banks are just another layer of service providers.
But if the L.C. industry can be automated and sped up, like the HSBC-Cargill-ING example, then perhaps L.C.s will regain their luster, and allow banks to take a more prominent role as trusted guarantors. At the very least, it would encourage more trade, and grow the pie while reducing the headcount costs to manage all that paperwork.
“Barriers to trade aren’t limited to tariffs,” said Joshua Kroeker, HSBC’s lead for distributed-ledger technology, based in Hong Kong. “It’s companies not being able to access working capital, or unable to make sales overseas due to their lack of experience with risk-mitigation products such as L.C.s. Our incentive to do this [pilot over Corda] is to have more trade: get more customers to use L.C.s, and grow the pie.”
L.C.s aren’t the first trade-related project to emerge from R3: that would be one called Marco Polo, driven by TradeIX, a technology company based in Singapore, which is aimed at the open-account segment.
R3 is a New York-based consortium of banks, insurers and other institutions, as well as regulators, with a mission to provide enterprise-level DLT to financial markets.
Its product, Corda, is not a classic blockchain like the one behind Bitcoin (for example), in which consensus is reached by broadcasting every transaction across the entire network; instead it operates permissioned, private environments in which “business network operators” – member banks, or their tech development partners – write the applications that run a node on top of Corda’s operating system, and say who gets to see what information.
Banks are involved in multiple Corda-based projects, and from Marco Polo emerged a smaller group of 11 banks interested in what became known as “Voltron”. (Blockchain engineers apparently have a thing for robot cartoons. The name is an internal moniker, not a product brand, but DigFinuses it here to differentiate it from other projects.)
Bringing in CryptoBLK
HSBC emerged to take the lead and sought developers to make Voltron happen. The finalist was selected by the bank’s London-based innovation team, which gave vendors samples of code to see how developers would improve or write it.
The mandate went to CryptoBLK, a blockchain developer in Hong Kong. CryptoBLK was set up in 2017 by Duncan Wong, a cryptographer who had spearheaded the fintech initiative at ASTRI, a government-backed research body. At ASTRI, Wong had worked with banks and financial regulatory bodies on issues around cybersecurity and blockchain proofs of concept; he then set up CryptoBLK to commercialize production-ready solutions.
With the HSBC mandate in hand, Wong says it became important to ensure Voltron could keep up with the fast pace of development of Corda. (ASTRI used Ethereum’s blockchain, but Wong says it Voltron required additional encryption mechanisms for security and access control.) In the end, this meant writing code from scratch.
“The lesson we learned was to ensure the future is scalable,” Wong told DigFin. HSBC and the other Voltron members all expect dozens of banks to join, and each one would attract hundreds of corporate clients. “So we’re designing this for hundreds of thousands of corporates and hundreds of banks,” Wong said.
CryptoBLK and the banks then had to address a number of areas before the pilot run.
One is access control. Unlike in Bitcoin, in which all nodes share the ledger, an enterprise DLT like Corda requires privacy, meaning an additional encryption layer to ensure only the relevant parties can see the L.C. details. That in turn requires complex management of private keys (the hashes, or lines of code, that let permissioned parties access information).
Second is managing the complexity of designing smart contracts to recreate the workflows associated with traditional L.C.s. Blockchain tech offers immutability: no one can change a record on their own, thus making audits easy. This is where the real cost-savings accrue, but using software (smart contracts) to ensure the integrity of documents requires a lot of development work.
That work has paid off: the Cargill soybean deal proves the system works in a live environment. But Voltron isn’t yet ready for prime time.
It still needs to provision “high availability”, meaning it can carry out transactions even when some nodes crash. The developers are also working on “multi-threading”, meaning a given node can handle a virtually unlimited number of transactions, so Voltron can scale.
And then the developers need to establish disaster recovery systems, as well as design protocols for revoking or renewing private keys in the event of cyber attacks or security breaches.
CryptoBLK’s Wong says the goal is to have a minimum viable product out and running 24/7 by the summer of 2019, but it may take longer before a comprehensive product is ready at scale.
That’s still a ways off. But the pilot presages more activity.
R3 will launch its enterprise version by the end of this month. Carl Wegner, Taipei-based managing director and head of Asia at R3, says until now Corda development has been open source, which has led to lots of shared ideas, for Voltron and for other Corda-based projects.
But open source doesn’t enable commercialization, which means deployment at a sophistication that banks require. Technical issues such as high availability, scalability and backup systems require a proprietary, enterprise-driven model, which also lets R3 begin to make money from these projects.
From there, the project becomes a race to add banks to Voltron. HSBC’s Kroeker emphasizes the consortium nature of the project. “This is an HSBC-owned platform,” he said. “This is decentralized.”
More banks will join as their big customers express more interest in blockchain. “Customers drive the world,” Kroeker said. “Banks have already been getting an earful about blockchain” and he expects the pace of adoption to quicken.
Voltron is the most prominent project for automating L.C. trade finance, but it isn’t the only one.
IBM’s Fabric, a DLT under the aegis of Hyperledger, another open-sourced platform, is working with many banks to deploy a solution.
Other banks may have their own DLT projects: in 2016, Commonwealth Bank of Australia and Wells Fargo pioneered what is probably the first blockchain-supported L.C., moving cotton from Texas to Qingdao on behalf of trading firm Brighann Cotton. (The developer of that tech, California-based Skuchain, has since pivoted from L.C.s to supply-chain invoice finance.)
In Europe, along similar lines, a consortium of banks have launched WeTrade as a DLT for trade finance.
Big corporations are driving their own solutions: Maersk, the shipping giant, has separate projects with both IBM and R3.
And governments are developing their own blockchain protocols: the Hong Kong Monetary Authority, in partnership with the Monetary Authority of Singapore, is developing the Global Trade Connectivity Network.
A HKMA spokesperson told DigFin: “When the interconnection of trade finance platforms between Hong Kong and Singapore is in place, this should…encourage other trading partners to participate.”
The GTCN is being developed to accommodate other digitized standards, so it can connect with other privately led systems, the spokesperson said.
But this leads to questions of just how many blockchain solutions are required for trade finance.
“If each consortium has its own solution, without an underlying platform that links them together, you end up with the same problem that you already have in banking today,” R3’s Wegner said. “You end up with disparate systems that have to be connected through APIs.”
For now, it seems proponents of all these different distributed ledgers say they are building to scale and to speak each other’s language, so they can one day connect. But this is untested.
Blockchain technology now offers the first serious opportunity to digitize trade finance. Consortiums, vendors, governments, banks and tech companies are building what they hope will become the industry standard. Some are doing this from a regional starting point, while others are trying to be global and then add on local partners.
Can all of these solutions actually rub along together indefinitely, or will there be a point at which the knives come out and one standard triumphs? These projects are not yet fully deployable, but they are being developed with one eye on this ultimate competition.