LIBOR Remediation: Remove the Hay
An interview with the CEO of Pendo Systems, Pamela Pecs Cytron at SIBOS 2019 regarding the industry challenge of LIBOR remediation.
An interview with the CEO of Pendo Systems, Pamela Pecs Cytron at SIBOS 2019
What’s the big deal about LIBOR?
OTC: The first phase of Libor and the fall back is identifying within these what and where is the fall-back clause?
As an industry we are saying LIBOR remediation is a $400trn problem, but we don’t know if it’s that size because we need to go in and extract the information and the content out the documents to find out. Just last week we found LIBOR student loans in the United States – they are in every kind of loan: swaps and derivatives. It is a bigger problem than Y2K and makes STP look like nothing. There is a lot of work on the tail of LIBOR, but right now what we have to surface is what is the exposure: what is our risk, what do we have to do, what do we have to remediate.
How does your platform accelerate the remediation process?
We can identify data by text, by a sentence, by paragraphs and by variations of those. In our solution what we have done is looked at document types, and if you have for instance a syndicated loan – they vary in language, content and vary in terms and conditions, but the logic we have built around that is able to highlight that exact paragraph you need to find.
Where does a firm start to find the needle?
When I talk Pendo we don’t find the needle, we remove the hay. As we are removing this, the corpus is getting smaller and smaller; until its manageable. In the case of LIBOR, we partnered with a workflow tool, by us digitising the golden source of this information and we are then being able to route the information fit for purpose to the relevant end user.
If we know the replacement language, we can basically then say this is the old language, this is the new language and automatically generate the correspondence that can be delivered to the client for review and signature.
How does a firm manage litigation risk?
The biggest fear from a Libor perspective is the litigation downstream. As Pendo brings this data through and digitises it, we document every step of that and provide full cradle to grave data lineage on unstructured content.
In that case we can build on that ability, just like we are doing with the fallback terms and say you can update that system of record with the right piece of data. That reconciliation is happening. 98% of the time, the system of record is wrong because we have taken it from the golden source of the document. There is a by-product to all the work we are doing with LIBOR and that is when we have to ingest everything, and then we index everything, and it becomes searchable.
Is the LIBOR problem too big for an on-premise solution?
A major US bank had a no KYC mandate because they had to go through all of their client onboarding records (22 million customers). That was everything from client onboarding to absence and presence of signature.
We had to go through all 22 million documents, this was on premises, the bank had built the environment, they scaled it up and they were getting such good quality, to the point of running hundreds of documents a day through.
But if they were doing it on the cloud, you could scale it up for 4 days and get millions of contracts through. What happens if your BAU platform goes down? The compute power needed for BAU is limited and having access to a scalable server platform in the cloud is a huge benefit.
How long will it take for the whole industry to solve the Libor problem?
I have a philosophy about this in general, just like when we went through the GFC those banks that changed the reporting role of the risk officer to be into the CEO, they got healthier faster. Until these banks take innovation out of the technology organisation and have it reporting to the CEO, they will not make the progression at the speed that they need to. Essentially innovation needs to come from solving business people’s problems.
Pendo is fuel, we are creating fuel for machine learning, AI, but we are creating auditable fuel – premium gas. When you can trust the data and you know it’s right, that’s going to be the accelerator for the AI bots and a new wave of innovation.