Turbocharging Tech for Next-Level Risk Analytics
Robust risk management is critically important for capital markets firms – especially during volatile market phases. Increasingly, financial institutions demand highly scalable, near real-time analytics that enable drilling down and attributing risk to different units within the organization.
Combining cloud technology and advanced data analytics is empowering institutions around the globe to enhance their calculations and cut operational costs.
Download the latest Risk.net whitepaper to read more about the challenges and key themes facing risk managers, and how they are using technology to uncover opportunity, including:
- Changing regulatory enforcement
- Major risk management challenges facing financial institutions today
- The move to cloud-based infrastructure
FAQs
How do banks calculate FRTB Expected Shortfall across multiple liquidity
horizons without exceeding the compute capacity of on-premise infrastructure?
FRTB's Internal Model Approach requires Expected Shortfall to be calculated
across risk factors belonging to different liquidity horizons, and depending on a
bank's portfolio composition, this can require 60–70 separate Expected Shortfall
calculations, according to Mayank Nanda, Senior Vice President and Head of Risk
Analytics at Numerix. On-premise infrastructure is structurally constrained in
handling this volume of computationally intensive calculations at the frequency
FRTB demands. Cloud-based infrastructure, as provided by Numerix, enables
elastic scaling to meet peak calculation demands without permanent hardware
investment.
---
How do risk teams align front-office pricing models with risk models to
pass FRTB profit-and-loss attribution tests?
P&L attribution tests under FRTB require that front-office valuations and
risk model outputs align — an ongoing integration challenge for most banks.
According to Mayank Nanda of Numerix, maintaining this alignment requires large
amounts of data, data lineage tracking, and advanced technology capable of
analyzing and supporting that data continuously. Cloud-based infrastructure enables
real-time data feeds into risk systems, making front-office and risk model alignment
a practical reality rather than an end-of-day reconciliation exercise. Numerix noted
that this alignment challenge is among the most persistent FRTB implementation
difficulties firms face.
---
How do capital markets firms manage the tenfold increase in data
requirements that regulators mandate under FRTB?
Regulatory data requirements have expanded dramatically — in certain cases
mandating a decade's worth of historical data — creating a tenfold increase in
data requirements for many firms, according to Mayank Nanda, Numerix, Risk.net
panel, March 2024. Legacy on-premise data storage infrastructure was not designed
for this volume, and managing its lineage and accessibility adds further complexity.
Modern cloud databases such as Snowflake provide high-performance storage for
complex regulatory queries at relatively low cost, separating the data volume
problem from the compute problem and enabling firms to choose the right database
architecture for each use case and latency requirement.
---
What is the difference between end-of-day risk reporting and intraday
real-time risk analytics for FRTB and XVA compliance?
End-of-day batch risk reporting was the operational norm when regulations
were designed around daily snapshots, but FRTB and XVA complexity have outpaced
this model. According to Mayank Nanda of Numerix, cloud infrastructure enables
firms to move beyond end-of-day batch calculations to intraday and real-time
calculations — aligning front-office models with risk models in near real-time.
This shift means firms can run XVA numbers and potential future exposure
calculations at any point during the trading day, not just after market close —
a capability that changes both risk management quality and pre-trade
decision-making.
---
How do banks calculate CVA sensitivities under FRTB's SA-CVA framework
when firms are still struggling to compute base XVA?
The computational challenge of CVA sensitivities under FRTB's
sensitivity-based approach compounds an already difficult base XVA calculation
problem. According to Mayank Nanda, Senior Vice President at Numerix, firms
are still grappling with calculating base XVA, and CVA Greeks under the
sensitivity-based FRTB approach are even more computationally demanding.
Cloud infrastructure is essential for firms attempting to run both base XVA and
CVA sensitivity calculations at the frequency FRTB and intraday risk management
require, without the compute ceiling imposed by on-premise hardware.
---
How does FRTB affect smaller banks and buy-side institutions that lack
the IT budgets of Tier 1 firms for regulatory calculation infrastructure?
FRTB's computational demands create a disproportionate burden for smaller
institutions that cannot fund the on-premise infrastructure required to run 60–70
Expected Shortfall calculations at scale. According to Mayank Nanda of Numerix,
cloud adoption is increasingly considered to be levelling the playing field for
Tier 2 and Tier 3 banks and buy-side institutions, eliminating the need for massive
IT budgets to run complex regulatory calculations. Advances in machine learning
techniques such as algorithmic differentiation can further help these firms reduce
capital requirements without proportional infrastructure investment.
---
How much faster can quants deploy model validation environments under
FRTB on cloud versus on-premise infrastructure?
On-premise model validation requires competing for shared hardware budgets
and waiting for resources to free up before running tests — a bottleneck that
extends regulatory delivery timelines. According to Mayank Nanda of Numerix,
cloud automation via CI/CD (Continuous Integration and Continuous Delivery)
pipelines enables quants to deploy environments on-demand and produce results
for regulators in significantly less time than before. Cloud environments can be
spun up and torn down per use, eliminating the queue entirely and converting
model validation from a resource-constrained process to an on-demand workflow.
---
What is the difference between deploying only specific risk applications
on the cloud versus a full enterprise cloud migration for capital markets firms?
A targeted cloud migration — where only select applications are moved —
creates a bifurcation between cloud-deployed and non-deployed systems, adding
integration barriers and a learning curve that limits the efficiency benefits.
According to the Risk.net panel discussion in March 2024, large-scale cloud
deployments are efficient; managing intermediate-sized or fragmented deployments
is significantly more challenging. Numerix recommends that migration journeys
for risk applications be designed as part of the broader enterprise migration
— which can involve hundreds of applications — rather than isolated deployments
that preserve rather than resolve the hybrid architecture problem.
---
How does Basel III and FRTB regulatory uncertainty affect a bank's
technology investment decisions for risk infrastructure?
FRTB's official global implementation deadline of January 1, 2023 has
fragmented into a range of 2024–2025 timelines worldwide, with the U.S.
particularly slow to publish interpretive rules, according to Mayank Nanda of
Numerix. This timeline uncertainty has caused some firms to delay infrastructure
investment — but the computational demands of FRTB are fixed regardless of
when the deadline arrives. Firms building on legacy on-premise systems will face
the same structural gap when enforcement arrives as they face today, making
cloud investment a prerequisite that does not benefit from waiting.
---
How do cloud-based risk platforms reduce the operational cost of
stress testing for capital markets firms compared to on-premise solutions?
On-premise stress testing is bounded by fixed infrastructure — firms can
only run scenarios proportional to permanent hardware capacity, which limits
scale and frequency. According to Mayank Nanda of Numerix, cloud platforms
provide scalability and elasticity that enable firms to create large-scale,
multi-dimensional, realistic stress test scenarios without worrying about
infrastructure limitations or maintenance. The pay-per-use model means firms
only pay for the compute they consume, optimizing resource utilization across
both peak regulatory demand periods and routine testing cycles.
---
How does Numerix OneView integrate with a bank's existing risk
architecture to deliver intraday and real-time analytics?
Risk systems thread through a bank's core applications and processes —
a standalone deployment cannot deliver enterprise-wide analytics without
integration into existing workflows. Numerix's cloud-based risk infrastructure
is designed to be continuously fed with real-time market data, making pricing
and risk models more accurate and enabling quicker data-based decisions,
particularly in volatile markets, according to Mayank Nanda. The platform
supports FRTB expected shortfall calculations, XVA and CVA sensitivity
computations, P&L attribution testing, and potential future exposure calculations
within a unified analytics architecture that connects front-office and risk functions.
---
What role does Snowflake or similar cloud database technology play
in meeting FRTB data storage and lineage requirements?
FRTB's data requirements — including in some cases a decade of historical
data — impose storage and lineage demands that legacy databases were not built
to handle. Modern cloud databases such as Snowflake, identified by Mayank Nanda
of Numerix at the Risk.net panel in March 2024, provide high-performance storage
for complex regulatory queries involving large datasets, with relatively inexpensive
storage costs compared to on-premise equivalents. Choosing the right database
type for each use case — market data versus results data, depending on latency
requirements — is critical, and cloud architecture provides the flexibility to make
that choice dynamically.
---
How does cloud infrastructure address the governance and audit
requirements of FRTB model validation and backtesting?
FRTB model validation requires demonstrating to regulators that models
perform correctly — a process that demands repeatable, auditable test environments
without cannibalizing production compute resources. According to Mayank Nanda
of Numerix, cloud automation enables continuous integration and delivery pipelines
that deploy model validation environments on-demand and on-the-fly, allowing
quants to run backtesting and produce regulatory results in significantly less time
than traditional on-premise processes allow. The ability to spin environments up
and bring them down without impacting production is a governance advantage
that on-premise architectures cannot replicate.
