Menghan Xiao

New York

Risk Management Reporter at Risk.net

Cover all things risk @RiskDotNet. [email protected]

Featured in: Favicon

risk.net

newsbreak.com Favicon

scmagazine.com Favicon

waterstechnology.com Favicon

centralbanking.com

Articles

Market whipsaw spurs calls to rethink buy-side stress-testing

2 weeks ago | risk.net | Menghan Xiao

Buy-side risk experts have said the tariff-driven volatility of the past few days has highlighted the need for flexible approaches to stress-testing, with a move beyond traditional assumptions and risk factor definitions. "We live in times where we see things we have never seen before [and] we have dynamics that play out differently than the past," said Mohamed El Hioum, executive director at Morgan Stanley.
Top 10 op risks: AI arms race leaves risk teams playing catch-up

2 weeks ago | risk.net | Menghan Xiao

This year's Top 10 Operational Risks survey featured a new open-ended question: "What will be the most significant trend influencing your firm's operational risk management in 2025?" The most cited response, by some margin, was artificial intelligence. That's about as far as consensus goes, though. Respondents were split down the middle on whether AI would be an influence for good for their firm, or for ill.
SEC's Peirce calls for rethink of international standards

2 weeks ago | risk.net | Menghan Xiao

A top US regulator has called on international regulatory standard-setters to rethink their agendas at a time when national governments are asserting more control over domestic policy.
DeepSeek success spurs banks to consider do-it-yourself AI

3 weeks ago | waterstechnology.com | Menghan Xiao |Anthony Malakian |Max Bowie |Rebecca Natale

Tweet Facebook LinkedIn Save this article Send to Print this page The success of DeepSeek, the Chinese artificial intelligence firm whose smash-hit large language model was built at a fraction of the cost of its big-tech rivals, is prompting some banks to take a fresh look at creating their own LLMs.The goal: technology that is better suited to core banking tasks and is easier to control.
DeepSeek success spurs banks to consider do-it-yourself AI

3 weeks ago | risk.net | Menghan Xiao

The success of DeepSeek, the Chinese artificial intelligence firm whose smash-hit large language model was built at a fraction of the cost of its big-tech rivals, is prompting some banks to take a fresh look at creating their own LLMs.The goal: technology that is better suited to core banking tasks and is easier to control. "Previously, it was rare for banks to discuss this option, because the mindset was always: "Look at OpenAI - they've spent so much to get where they are.
CCAR methodology may give first glimpse of Fed's transparency drive

4 weeks ago | risk.net | Menghan Xiao

The in-depth methodology for this year's capital analysis and review (CCAR) will shed light the US Federal Reserve's commitment to making the process more transparent, say lobbyists. They are keen for a more detailed description of the models used by the regulator, especially around the Global Market Shock (GMS) scenario, which has a significant impact on capital levels at the largest dealers.
Fed stress-testing operational readiness of discount window

1 month ago | nginx.production.bb8-centralbanking.uk3.amazee.io | Menghan Xiao |Christopher Jeffery |Daniel Hinge |Daniel Blackburn

The US Federal Reserve has been stress-testing its discount window to ensure it could provide liquidity to large numbers of banks simultaneously during a stress event, Risk.net has learned. Although the internal testing is not directly related to a request for information (RFI) on possible operational upgrades to the discount window, sources say both initiatives stem from criticism of the Fed’s emergency backstop following the US regional banking crisis in March 2023. “The RFI might be part of
Fed's NBFI scenario may be more use than CCAR

1 month ago | risk.net | Menghan Xiao

The US Federal Reserve's exploratory scenario focusing on non-bank financial institutions (NBFIs) could be more useful than the main severely adverse scenario in the annual stress-testing exercise, according to risk management experts. "The Fed has already squeezed as much as it could from the severely adverse scenarios," says Viktor Tsyrennikov, an independent stress-testing consultant.
Credit loss database reveals holes in Basel's IRB formula

1 month ago | risk.net | Menghan Xiao

When the Basel Committee on Banking Supervision first permitted banks to model their own credit risk capital requirements as part of the Basel II framework in 2006, they had to set out a new methodology without being able to calibrate it based on historical data. Fast forward almost 20 years, and banks have compiled vast databases to feed their models for Basel's internal ratings-based (IRB) approach to calculating credit risk capital.
Banks urged to track vendor AI use, before it's too late

1 month ago | risk.net | Menghan Xiao

Banks risk being caught unawares if they fail to keep an eye on which direction their vendors are taking in the fast-changing world of artificial intelligence, says veteran risk manager Jeannie Pumphrey. Since the release of US interagency guidance on third-party risk management (TPRM) in 2023, banks have continued to reassess and update their TPRM frameworks.