Blockchain shows great promise for certain applications in the financial services industry, including securities clearance and settlement. Given all of the recent publicity, one has to consider the Gartner “hype cycle.” Some recent articles discuss drawbacks for the use of blockchain technology in the securities settlement business.

Scalability

Bitcoin has recently seen a slowdown in transaction processing times, indicating that the bitcoin version of blockchain may not be suitable for public markets:

The average time it takes for a bitcoin transaction to be verified is now 43 minutes, and some transactions remain unverified forever. Some of the problem stems from the fact that anyone can add a fee to every bitcoin transaction, which bumps that transaction up in the queue, meaning that those who didn’t pay such a fee — or didn’t pay a sufficiently big fee — may be waiting hours and sometimes even days for a transaction to complete.

This is how it works. When someone uses bitcoin to pay for an item in a shop, that transaction needs to be verified on the blockchain. This is done by what are known as miners, individuals or groups who use massive computing power to solve increasingly complex mathematical equations to mine new bitcoins, which come in “blocks” and are mined about every 10 minutes. These blocks are used to record all transactions made on the bitcoin network, and have a maximum size of 1 megabyte (MB), meaning they can record just seven transactions per second at most.

Confidentiality

Blockchain clearance may not permit confidential trading that some believe is necessary in capital markets that we are used to:

Capital markets expert Larry Tabb has issued a report “Blockchain Clearing and Settlement: Crossing the Chasm”, which takes a realistic — some might say pessimistic – view of hurdles in the way of blockchain-enabled real time clearing and settlement.

Tabb highlights what he sees as insurmountable problems relating to identity and transparency, and explores these in some depth from both a buy-side and sell-side perspective. One of the key issues raised in Tabb’s report is information leakage. This inevitably follows if registration changes from generic “street name”, where brokers and custodians know which securities are held by clients, to an exacting shared ledger, whereby the client name should be on the transaction record prior to entering the trade.

Large investors, hedge funds, activist investors, and especially takeover firms, do not want to let others know that they are accumulating positions, notes Tabb. Information leakage will also happen when securities are lent out by investors to cover short positions because moving to real-time clearing would require securities to be on hand before they were sold (because they would clear virtually immediately).

Netting

Larry Tabb also discusses netting, a practice common with the current DTC-based settlement system:

Whether blockchains will settle trades on a gross or a net basis is a vexing question. Some advocates of blockchain technology believe netting will be integrated into the ledger, although they concede this will not be straightforward, others say everything will be done on a gross basis. Tabb believes a blockchain model would necessarily entail gross settlement and that this would be hugely inefficient to the way markets function.

“If a blockchain is an irrevocable record then the question becomes is — how do you actually net it? It needs to be a lot more complicated than a simple ledger of who bought what, when, and a chronological list of all the transactions. You have netting transactions, so how do I figure out exactly which securities I own?

Newer, less developed blockchain systems may be subject to more security flaws:

Blockchains provide a global append-only log that is publicly writeable. Writes to the global log, called transactions, are organized as blocks and each block packages multiple transactions into a single atomic write. Writing to the global log requires a payment in the form of a transaction fee. Nodes participating in a blockchain network follow a leader-election protocol for deciding which node gets to write the next block and collect the respective transaction fees. Not all nodes in the network participate in leader election. Nodes actively competing to become the leader of the next round are called miners. At the start of each round, all miners start working on a new computation problem, derived from the last block, and the miner that is the first to solve the problem gets to write the next block.

Miners often pool their resources to form a mining pool, which is essentially a super node on the network (a lot of computational power behind a single miner node). If the amount of computational power under the control of a single miner (or pool) is more than the rest of the network, called a 51% attack, then that miner has the ability to attack the network and rewrite recent blockchain history, censor transactions (e.g., for name registrations), and steal cryptocurrency using double spend attacks. This is because it will win the leader election majority of the time, and produce a blockchain history with more proof-of-work than any disagreeing miner. The more expensive it is to control a majority of the compute power on a particular blockchain, the more secure the blockchain.