Financial Industry Looks to Automated Reasoning
Aesthetic Integration, the startup that won the UBS Future of Finance challenge this month, is working with several large financial institutions on formally verifying critical components of their trading infrastructure as algorithms and systems have become increasingly complex.
Denis Ignatovich, co-founder of Aesthetic Integration, was previously head of the central risk trading at Deutsche Bank in London and searching for a better way to test trading algorithms. He and his former college roommate, Grant Passmore, a mathematician specialising in formal verification – used to develop software in safety critical industries such as air traffic control systems – realised the deep connection between financial trading systems and autopilot algorithms.
Ignatovich told Markets Media: “Formal Verification involves use of automated mathematical techniques to ensure system specifications are consistent and designed correctly, and then using the same techniques to test whether actual production system is consistent with its design. Applying such techniques to financial algorithms was physically impossible three years ago.”
Computers can use automated reasoning to perform tasks using laws of logic and mathematics such as analysing algorithms, resulting in “formal verification” programs that test the behaviour of other programs.
In the financial industry, for example, trading venues can convert regulatory filings and trading rules into mathematically precise specifications and code which can be analysed using automatic reasoning to test fairness and whether trading behaviour matches the marketing materials.
“Formal Verification will change how the financial industry writes software,” Ignatovich added. “It is similar to an architect having to prepare a blueprint for a new building which can be checked to see if it is safe, before you call in the crews to start laying down bricks.”
For example, in August this year the Hong Kong securities regulator fined BNP Paribas HK$15m ($1.9m) and alleged that the French bank’s dark pool had failed to execute orders based on price priority for almost one and a half years i.e the trading logic was different from what the bank told clients.
Aesthetic Integration said on its blog that automated reasoning and formal verification could reduce the time taken for the regulator to perform an investigation into such cases from years to a couple of days.
“All they have to do is to ask a technical person (within the firm) to convert the regulatory submissions (like Form ATS and exchange by-laws) into a mathematically precise specifications and automatically reason about their fairness and conformance to marketing materials,” said the blog.
Other possible uses are for an investment bank which trades on hundreds of venues to quickly check how each platform operates while investors can easily test how their trades would behave on different venues.
Ignatovich added: “For the first time, financial firms can quantitatively describe their testing scenarios and analyse all possible behaviours of the system’s state space.”
At the UBS challenge Aesthetic Integration won $50,000 in cash and dedicated coaching and mentoring from UBS experts and partners. Oliver Bussmann, chief information officer of UBS, said in a blog that the bank had learnt a great deal throughout the challenge.
“We received more than 600 applications from around the world, showing that the old fashioned closed door policies of the financial services industry is now a thing of the past,” added Bussman. “If anyone in the financial services industry thinks they can close their eyes to the threat of the fintech community then they are missing the incredible depth of talent that is out there.”
UBS is one of the banks that has joined R3, the consortium launched to develop distributed ledger technologies, such as blockchain, for the financial services industry. Today R3 announced the addition of 12 new banks to the consortium bringing total membership to 42. R3 said that no more banks will be added and it will consider options for participation by non-bank institutions next year.
David Rutter, chief executive of R3, said in a statement: “R3 has long believed that distributed ledger technology has the potential to impact the financial services sector the way the Internet changed media and entertainment. Yes, that’s a big statement, but there is increasing evidence to support it.”
The Linux Foundation, the nonprofit open source organization, also announced a blockchain collaboration today that includes diverse participants such as R3, banks, payment provider Swift, exchanges, VMware and IBM.
IBM will contribute code and its corresponding intellectual property. Digital Asset is contributing the Hyperledger mark, which will be used as the project name, as well as enterprise grade code and developer resources. R3 is contributing a new financial transaction architectural framework designed to meet the requirements of financial institutions.
The blockchain distributed ledger allows virtually anything of value to be tracked and traded directly between multiple users without using intermediaries. Financial firms can create smart contracts, code versions of traditional securities, which are stored in a distributed ledger. For example, a smart contract can encode the interest payments on a bond to check a specified source for the rate and the digital currency distributions will be made automatically to all the holders of the bond recorded on the blockchain on set dates. Settlement takes place as soon as all the specified conditions in the code have been met without any manual intervention, giving huge amount time and cost savings.
Aesthetic Integration can use formal verification on blockchain technologies and smart contracts to ensure that they perform as advertised, and so firms can understand and hedge risks associated with smart contract positions.
Featured image via iStock
Asset managers seek more cost-effective and trader-friendly solutions.
It's about providing actionable insights to the buy side.
How to narrow the gap between decision price and execution price?
Buy side targets 'micro alpha' via trading >1 stock at a time.
Options traders get two new FIX-based algos.