Virus risk, pandemic modelling and the challenge for AI

Covid brings op risk to the fore at Credit Suisse

Jim Barkley, head of non-financial risk, has added pandemics to the long list of threats on his radar

Covid-19 tumult is testing AI fund returns

Some machine learning strategies have coped well, but others began to struggle as panic mounted

Quants pitch in to improve pandemic models

The finance industry’s quants are trying their hand at modelling the virus and its economic impact

COMMENTARY: Faith in the machine

Machine learning fraud detection systems, as Risk.net discussed last week, haven’t coped well with the turbulence caused by the coronavirus pandemic. But other automated systems are having a comparatively good war. Contrary to expectations, algorithmic execution is seeing a boom, with Goldman’s Ralf Donner estimating a 50% increase in algo foreign exchange volumes since markets began to crash last month. Customers have grown comfortable with algorithmic execution during quieter times, he speculates, and are willing to shift more volume to the algos during a crisis, in order to free up more time to cover the more volatile equity and commodity markets.

And machine learning has had some notable successes in the fund management sector . Not every fund using machine learning methods has produced good returns, but an index of the sector shows decent performance, and some managers report that machine learning has made a significant – and lucrative – difference.

In particular, the machine learning advantage showed itself early on in the pandemic, spotting the first signs of pandemic risk and recommending derisking or even a stop to trading. And, as the spread of the virus gives the world a lethal demonstration of exactly what exponential growth looks like, we’re all learning that an apparently insignificant delay in responding early on can make a catastrophic difference once the curve steepens. The US introduced social distancing on March 16; had it done so on March 2 instead, 90% of those now dead of Covid-19 would still be alive, epidemiologists estimate. Similar delays of a few days in response may have made the difference between European countries like Greece and Ireland suffering hundreds of deaths, and countries like the UK and Spain suffering thousands. Taiwan’s health ministry, meanwhile, has reportedly credited part of the country’s prompt response to the pandemic to an unlikely source: a December 2019 post on the virus on the country’s unruly PTT online bulletin board.

The combination of machine learning and alt data can be a powerful one – and there’s room for it to be applied in many areas outside finance. If the pandemic serves as a stress test to cull the worst-performing machine learning applications and move the state of the art on, 2021 could be a revolutionary year for finance.

STAT OF THE WEEK

Changes to the valuation of derivatives forced on JP Morgan because of last month’s market turmoil inflicted a $951 million loss in Q1. The US dealer said a widening of the funding spread for derivatives drove the loss, captured under “credit adjustments and other”. This line item aggregates costs and benefits to trading revenues caused by funding valuation adjustment (FVA) and credit valuation adjustment (CVA). The FVA/CVA loss is JP Morgan’s biggest in at least five years.

QUOTE OF THE WEEK

“[The VAR exceptions] are causing procyclical capital buffers that banks need to set aside when capital is a scarce resource. Our Common Equity Tier 1 ratio is already under strain on its own and what is happening is that all of the mechanisms that typically work in normal market conditions are now exacerbating problems” – a head of market risk modelling at a eurozone investment bank