Blockchain: A Solution in Search of a Problem?
Countless column inches have been dedicated to distributed ledger technology (DLT) and blockchain since Bitcoin burst onto the global scene in 2009. Since then, numerous financial institutions, governments and other organizations have experimented with testing and implementing blockchain solutions. This includes development organizations such as the World Bank, IMF and the United Nations. Despite breathless enthusiasm and substantial investment, we have seen relatively few DLT applications successfully introduced at scale outside of cryptocurrencies (and these are a whole other can of worms). At CGAP, we have been thinking about which blockchain applications might make sense for financial inclusion, considering use cases such as retail payments, cross-border payments, remittances and agricultural value chains.
The question we have is not “Does blockchain work?” but "Does it work better than other technology solutions in the market?" And what are the cost-benefit trade-offs to switching to a new consensus-based technology solution? We have seen lots of proofs of concept that it can work, but not a lot of quantified analysis on why a DLT solution might be better than existing alternatives.
So, to what extent is blockchain a solution in search of a problem? There are a few important hurdles for those considering blockchain solutions in the financial inclusion space.
Governance, rules, oversight
The financial services space is highly regulated, and with good reason: There are important restrictions placed on organizations that move money to ensure they are not facilitating criminal or terrorist activities. There are important considerations around financial sector stability, as well as privacy. While blockchain can create an immutable record of two parties agreeing to a transaction, in the financial sector, we must also ensure the information entered into the ledger is legal and compliant with relevant regulations. As such, participants in a financial sector blockchain are likely to be licensed counterparties using a permissioned ledger, either in consortium or with third-party providers who set the rules that others agree to when they join the network. With the exception of a purely private scheme, the complexity of establishing governance arrangements and rules is not much different than that of setting up a payment system today, whether retail or wholesale. In this context, it is unclear that there are significant gains to be made from deploying DLT because the problem is around governance and rules, not technology.
In many financial inclusion use cases, the factor slowing things down is not technology, but regulation. Many payment systems, including cross-border systems, operate in real time. What slows things down is regulatory requirements, mainly related to anti-money laundering and combating the financing of terrorism (AML/CFT) or exchange controls. For example, in South Africa, I can send money instantaneously within South Africa using Real Time Clearing, but if I send money into the country, it can take days to get to my account. I recently used PayPal’s Xoom service to send money to bank accounts in the United Kingdom and South Africa. The transfer to the United Kingdom took five hours, end to end. Sending money to South Africa took five days (and resulted in my account being frozen for two weeks) because of the South African Reserve Bank’s exchange control policy and required processes. That is not a technology problem: It is a regulatory one. That said, from an operator’s perspective, DLT could help reduce the cost of cross-border transfers by making net settlement positions known in real time to all participants, lowering liquidity management costs and minimizing foreign exchange risks. These efficiency gains are great for banks, but it is unclear whether and how these benefits will be passed on to consumers. There is potential here, but we have not yet seen it play out in practice. And it is possible this would be manageable with existing technology solutions.
Privacy, complexity, scalability
Privacy is a major and growing challenge in the financial inclusion space. And putting proper privacy controls in place adds complexity to systems that are not centrally managed. SWIFT, the entity that handles cross-border payments on behalf of its member banks, recently completed a well-documented proof of concept using DLT among 28 of its 11,000 member banks. Although SWIFT reported that the test went extremely well, it concluded that further progress is needed on DLT before it can be ready for production-grade applications. It was clear from the test that the clearing of nostro accounts could be streamlined using DLT. But privacy was an issue: Running the test with just 28 banks required 528 subledgers to maintain the confidentiality of information. SWIFT calculated that rolling this out to its entire membership would require the creation of 100,000 subledgers, which would be a challenge to manage. This kind of complexity requires scalability of processing, and this is an area where current database solutions outperform DLT. Using today’s technology, Visa can process 24,000 transactions per second. The best I have been able to find that any DLT can do is 1,500-2,000 transactions per second. The sequential nature of DLT and the possibility of forks (incompatible versions) in the ledger because of errors or asynchronous entries introduce additional inefficiencies as solutions scale.
Intersection with cash economy
In financial inclusion, which is about bringing the poor into the financial system, there is a major challenge around the way new technologies intersect with the cash economy. No matter how money is sent around the world digitally, poor people need to be able to access and use their financial resources in cash, at least until better solutions are available for paying digitally. Mobile money has garnered a lot of attention in recent years for the important role it has played in improving financial inclusion for the poor. But mobile money is only partly a technology solution. It is mainly a cash distribution business — the cash-in/cash-out interface is critical for enabling people operating in the cash economy to use the system. There are a lot of interesting experiments underway to shift to a more digital profile. For example, using smartphones and QR codes can substantially lower the cost of delivering a payments service. But these solutions rely heavily on people having access to smartphones and bank accounts, which is not yet the case in many poorer parts of the world. And smartphones and QR codes don’t need DLT. Mobile network operators, with 2.9 million active cash-in/cash-out agents worldwide, currently reach the poor very effectively, as is clearly validated by the World Bank's Global Findex. And they are increasingly adding use cases like international remittances. GSMA estimates that the cost of sending remittances to Africa can be cut in half by sending remittances via mobile money. But this works on existing technology. It is not evident that DLT makes it work better, particularly in environments where both 3G and electricity access are sketchy.
Finally, any DLT solutions would have to incorporate mechanisms related to recourse and consumer protection, whether accepting liability for fraud or helping consumers to find payments that have gone astray or data entries that are incorrect. This is not necessarily the responsibility of the distributed ledger operator, but rules of engagement would need to be agreed to and implemented consistently among system participants. Recourse is critical to building consumer confidence and trust. Liability also enforces good provider practices: Visa and Mastercard have strong fraud management processes in place that limit fraud to a few basis points in a trillion-dollar industry. This is despite the existence of large and sophisticated crime syndicates that are focused on hacking the system.
In the financial inclusion space, we have not yet seen business models emerge that make a compelling case for DLT over existing technology solutions with trusted third-party managers. Ripple is perhaps the one exception to this, but it is basically a managed third-party solution by virtue of its business model, and there is not yet a clear connection to financial inclusion.
That is not to say solutions that incorporate DLT will not emerge. Places where we think there may be breakthroughs are mostly in areas where distributed, real-time information is important: supply chain finance, agricultural value chains, identity verification, personal data storage, clearing and settlement, collateral and land registries and maybe credit reporting. But even these use cases will require answers to some of the questions outlined above. The key point is to understand the trade-offs involved in introducing DLT and to have a clear understanding of the problem you are trying to solve. After creating a proof point to determine whether the technology can work, it then becomes important to conduct a cost-benefit analysis to see if it offers any gains over existing technology solutions. By jumping straight into DLT, you risk putting the solution before the problem.