How Algorand Can Speed Up Financial Settlements

By Pablo Azar


Any financial system is, at its core, a consensus mechanism to determine who owns which assets and which transactions are valid. This process is time consuming and costly when financial accounts are not on the same database. For example, if Alice wants to send money to Bob, their banks would have to send each other messages, sometimes through multiple intermediaries, in order to verify that the funds exist, that their accounts have been updated, and that the banks’ commercial accounts with each other reflect this change. This all must happen before the transaction may be considered complete. This costly consensus process is the reason why even simple payments such as bank transfers between individuals are expensive and take multiple business days to settle.

The current mechanism used by the financial industry—which economist Massimo Morini calls “consensus-by-reconciliation”—is drastically inefficient because it relies on every financial institution in the world hiring analysts, accountants and lawyers to reach agreement on which transactions are valid. While the execution of trades (the orders to buy and sell financial assets) may happen at the speed of the internet, the settlement of these transactions (the actual change in ownership of the assets) may take days. This delay in settlement introduces many risks, including the risk that settlement will not actually happen, or the risk that the asset being traded will decrease in value between the execution and settlement date.

If we aggregate across all deals being made across the world—between individuals, between businesses and between financial institutions—the “consensus-by-reconciliation” mechanism induces trillions of dollars per year in transaction costs, creates risk in the financial system and slows down the pace of global commerce.

If assets were instead held on a shared digital ledger, then reaching consensus on who owns which assets would be faster, cheaper and easier. This is why financial institutions have devoted a large number of resources to develop blockchain solutions that facilitate financial transactions.

However, the existing distributed ledgers introduce new risks by being either slow or permissioned. The problems with slow blockchains such as Bitcoin or Ethereum is that they cannot handle the transaction throughput that is needed to support the financial system. Furthermore, because they rely on proof-of-work, transactions cannot be considered finalized because of the risk that the blockchain will fork.

Permissioned blockchains, on the other hand, can handle fast transactions, but introduce new types of risk. One type of risk comes from the fact that the owners of the permissioned blockchain have a monopoly over transactions. They can extract rents from users under the threat of eliminating their access to the financial system. They also have an informational advantage over the users of the system and can use this to gain a strategic advantage over competitors and clients. Another type of risk is that this centralized system could be attacked by a malicious hacker, who could leak confidential financial information of the users. Finally, an unspoken risk of permissioned systems is that they stifle innovation: only institutions who are part of an approved consortium can build on a permissioned blockchain.

At Algorand, we recognize that reaching consensus on who owns which assets is one of the most important functions of the financial industry. This is why the heart of the Algorand blockchain is a novel consensus algorithm, developed by cryptography pioneer Silvio Micali, which requires minimal computation and which can settle transactions in a few seconds. By having fast settlement of transactions, the Algorand blockchain has the promise to make current institutions more efficient.

The Algorand consensus algorithm has the additional advantage that it is permissionless: anyone—from individual entrepreneurs, to small businesses, to large banks—can join without being part of an approved group. Just like open source software and the open internet led to flourishing ecosystems that allowed anyone to build on others’ work, a permissionless blockchain can also allow anyone to develop new market applications based on ownership records made public by the blockchain. This is the feature that I am most excited about, because economic growth often comes from the combination of existing ideas. By allowing anyone to build on public records of ownership, the Algorand blockchain can help unlock new markets that did not exist before.