Quant Labs, the research division of Quant Foundry has developed an operational risk model that supports the COO to pin point areas of process weaknesses. The model continuously learns the business operating model and enables the COO to target investment under different strategic scenarios.
Lundin Gold April 2024 Corporate Presentation v4.pdf
Optimising Business Performance with Operational Risk Modelling
1. Operational Risk Model – Optimising Business
Performance
Dr Chris Cormack and David K Kelly, Quant Foundry
Strategy, culture and policy
Understanding the complexity and interplay within an organisation that gives rise to operational risks is
something most organisations struggle with. Strategy direction provided by the Board filters down with
clarity of purpose is dispersed through policy interpretation and largely negated as front-line staff react
to poor engagement that introduces additional controls whose additional burden is expected to be
absorbed in existing budgets.
Most COOs will recall efforts in improving their business model hampered by a change-weary culture of
passive resistance. Even strong engagement, clear direction and benevolent funding face challenges as
individuals agree in principle but clearly the changes do not apply to them.
Operational risk, like other risk disciplines is for the most part not risk elimination, a typical response of
“this must never happen again” sounds laudable but is not proportionate. Zero tolerance to risk has a
place in the nuclear industry, but even in the airline and medicine industries the drive is to reduce not
eliminate risk otherwise drugs and planes would never be commercial.
At the Labs division of the Quant Foundry, we recognised that the largest impediment to improving
business performance through improvements in the operational risk environment is culture and the best
way of addressing this is to take the emotion out of dealing with operational risk issues is through
measurement. We also believe that operational risk in finance should not be something done because
there is a regulatory charge but should be seen as a critical tool for the COO.
Operational risk in finance is often viewed as the poor cousin of the risk disciplines, it is frequently under
resourced and its methodology and technology development has lagged behind. The department is seen
as reactive in the sense it is just recording losses. Many organisations do not have the processes in place
for an effective partnership across the business lines to leverage the depth of knowledge acquired from
the loss “Mea/Tua Culpa” writeups.
It is clear that finance companies will continue to not learn from their mistakes and make the targeted
investment to stop problems before they materialise – where are the warning signs?
.
COO plan of action
So, what does the COO need to do to reduce operational losses. At the Quant Foundry we have a six-
point plan: -
• Move away from a Mea/Tua Culpa mentality - the airline industry doesn’t do it so neither should
finance – and build an understanding of the process landscape and the key influencers that drive
the measurement of operational resilience
2. • Leverage of the loss data they would like to collect alongside the loss information (no matter
what the form) that is already available. Make the data sourcing relevant – details that will
improve business engagement and relevance
• Understand the causal chain – map how strings of processes interact, develop a suite of data to
collect that with provide the necessary insight to the drivers of losses
Structured Data
Operational risk always requires three categories of data: - loss data, risk factors, process taxonomy.
Sufficient amount of data is usually available, but it needs to be placed into a form that follows the
diagram below: -
This is not a straightforward exercise as it will require a degree of standardisation so that data is uniform
across organisations in particular self-assessments. Enforcing a template for RCSAs across all legacy
processes that, for example, uses binary questions, weightings and normalisations will normalise the data
and make it future proof. A key set of risk factors comes from human resources will require careful
handling - turnover, resource levels, managerial scores. Each of these will have some influence on
internally driven losses.
Quant Labs Operational Model
Quant Labs at Quant Foundry has built a machine learning model that utilises the data collected, maps
the separate components of the process taxonomy and then defines potential influencers (Headcount,
staff surveys, conduct, economic environment, BEICF, turnover). The model then works on the complex
relationships between realised losses, previous RCSAs and historic data on influencers to learn how they
can measure a baseline measure of operational resilience and that for each component.
The beauty of the Quant Labs operational risk model is that it can provide immediate insight even for
large organisations. The model can work with low granular and sporadic information so that for the COO
can address quickly some obvious pain points. With a couple of quick wins under the belt, the COO can
invest in a more expansive data normalisation and gathering exercise and a more granular process
taxonomy to feed the richer set into the model to provide much less obvious insights to pain points.
3. Thinking Forward
As the Quant Labs operational risk model continues to learn the COO can now use it as a key tool to
understand how the organisation will respond to the organisation embarking on a new strategic direction
– increase turnover, increase country coverage, reduce operational costs, clean out middle management,
introduce stricter conduct policy, address high turnover. The COO can move the dials of the influencers
and the model will pin point to the component of the process most at risk of degradation. The COO can
therefore engage with the front-line staff on the change strategy, the impact to process and gain an
understanding of the investment required to remediate. The COO can then make an informed decision
on how to proceed.
With each day passing, the Quant
Labs operational risk model
continues to learn about the
organisation and will continue to
improve the engagement
between decision-makers at all
levels and front-line staff –
everyone gets a copy of the
model - and optimise the cost of
ownership of the organisation’s
process.
4. Conclusion
The Quant Labs Operational Risk model helps the COO and all decision-makers in an organisation to
understand the complex relationships between processes, risk assessments and the influencing risk
factors that drive historical losses. The model enables each user to dial up or down the influencers to
gain an insight on where pain points will arise based on a chosen strategic path. This will enable COO to
fund pin point remediation efforts and monitor improvements. All the while, the model continues to
learn more and more about the organisation, improving its predictive powers.
The Quant Foundry
The Quant Foundry has been set up by Chris Cormack and David Kelly to provide professional consulting
services, quantitative team augmentation services, software design and development as well as mathematical,
social science and scientific research. Our research division – Quant Labs - includes the design and development
of deep-learning algorithms and Artificial Intelligence models across multiple industrial sectors.
Dr Chris Cormack Solution Partner chris.cormack@quantfoundry.com
David K Kelly Assurance Partner david.kelly@quantfoundry.com
Quant Foundry
Labs-Forge-Works