By Dr Lewis Z. Liu, Co-founder & CEO of Eigen Technologies
Talking heads who declare that technology will transform finance are everywhere. The comment sections of financial publications are full of conversations on the topic of how AI technology will transform the sector, and every industry conference has at least one keynote speech or panel dedicated to it. The rise of chatbots or automated mortgage decisions in the consumer finance space is well known. In contrast, the potential of technology to transform the wholesale side of banking and, particularly, capital markets, has received far less attention. We recognize the potential is huge to bring greater stability to capital markets by giving both financial institutions and regulators a far better understanding of the complex instruments they trade.
Most public equities already have good data. Bloomberg, FactSet, and the like provide sufficient data on an almost live basis for traders and investors. However, the fixed income space, which has a much larger market exposure, often lacks critical data. These assets, like all assets, are underpinned by contracts and documentation. Although this is meant to be ‘standardised’, many deals result in complex, bespoke agreements, which both dramatically increases the expense of transactions and reduces liquidity.
Now we are seeing the arrival of AI technologies that can address this problem, specifically Natural Language Processing (NLP). NLP can analyse and extract information from the contracts and documents that underpin complex financial instruments. Suddenly, these previously less-liquid assets can be turned into structured information that banks can analyse and use to make decisions. I see NLP technology transforming capital markets by improving liquidity and thereby splitting the information that fuels these markets into upstream and downstream data.
Upstream data is unrefined raw data: the crude oil of capital markets, if you will. This crude oil could be anything from a facility for a non-performing loan (NPL) to a Collateralized Loan Obligation (CLO) prospectus. Downstream data is the refined petrol and fuel that powers the market. Until very recently, the highest-octane downstream data was almost entirely numerically based (and generally geared towards equities). There was simply no means of refining qualitative, contract-based data into easily usable information, which means that certain assets were far less liquid than they could have been. AI-powered NLP technology solves this problem by refining previously inaccessible qualitative data into structured, usable downstream data. Crude oil becomes jet fuel at the touch of a button.
For banks, this changes everything. A bigger, higher quality source of downstream data will provide an entirely new stream of information with which to make better decisions, thus making capital markets more liquid. Suddenly, for example, you can bring far greater transparency to CLOs. This is a meaningful breakthrough considering the justified worries about the potential of CLOs to be the source of the next financial crisis. For NPLs, which are notoriously hard to assess, there is now the potential to clean them up far faster than was previously possible.
To me, it is clear that AI technologies, particularly NLP, have immense potential to transform capital markets. It is part of a bigger revolution in our ability to process qualitative data that is still nascent. Whilst there are many clichés out there on tech, it’s no exaggeration to say that this will be transformative.
This blog was originally published by Innovate Finance for the Innovate Finance Global Summit in London.