Risktech start-ups, market making and poor Mifid II data

Risktech start-ups struggle to clinch big-bank contracts

Light on cash, risk management fintechs face an extended gauntlet most won’t survive

Machine earning: how tech is shaking up bank market-making

As banks get serious about e-trading, humans are being asked to give up their secrets to the machines that could replace them

Poor Mifid data could condemn OTC market to the dark

Many derivatives likely to fail first full liquidity test and escape EU transparency obligations

Commentary: Culture clashes

Life can be tough for fintech start-ups. Clinching a sale takes on average two years; nine out of 10 new firms fail in their first five. Elegant programming code and neat ideas are no guarantee of success when the seed money is running out.

Risk.net spoke to 15 new players in the space and found many struggling to wade through the glop of banks’ procurement processes.

In the face of inertia, the start-ups are seeking different ways to get on. Some are switching targets, lowering their ambitions from selling to top-tier banks and favouring second-tier firms or buy-siders where access to decision-makers is easier.

Some have sought partners that can lend them prestige or help make introductions, including tying-up with the banks themselves as part of incubator programmes. Others are selling into clients from the bottom up, aiming to get individual users hooked and turn them into advocates for corporate contracts from the inside.

The game is to hold on long enough to overcome banks’ distrust of the unknown or unproven.

Inside banks, meanwhile, showing the worth of innovation can be equally trying. There, efforts to bring automation to pricing and hedging for less liquid, more complex products is a metamorphosis that’s been waiting 20 years to happen.

It’s happening now. Barclays, Goldman Sachs, JP Morgan and UBS have all created beefier cross-asset teams to extend automation into areas where traditional approaches still dominate.

Human traders may still be deciding on the price, but algorithms are advising them how long it will take to move the risk along, and finding the optimal hedge.  

Part of the reason this is happening only now is that such projects rely on rendering in computer code a trader’s experience and instinct. That means convincing the traders in question to give up their secrets – and reassuring them they’re not simply doing themselves out of a job.

There are concerns also about putting too much power in the hands of algorithms, effectively fixing the assumptions of quants into the circuitry of a bank, with the risk those assumptions prove wrong or break down in a crisis.

But change is happening all over: in sales, where banks are building an information exoskeleton around salespeople to help them churn through small-customer business more quickly; in the development of bank-constructed ‘bots’ connecting to the Symphony chat system to cut trade times from 15 minutes to seconds; in the rollout of tools to identify toxic flows.

It might be slow going, but there’s no turning back.

STAT OF THE WEEK

The basis between 10-year euro swaps cleared through London’s LCH SwapClear and the same instruments cleared at Frankfurt’s Eurex Clearing had collapsed to 0.25 basis points on August 3 from 1.35bp on May 21, according to Icap pricing.

LCH-Eurex basis falls 80% as insurers head to Frankfurt

QUOTE OF THE WEEK

“We have zero tolerance for rogue traders. … So, if we had such a person at UBS – as we had some years ago – how long would it take us to detect what’s happening, and how could we limit the amount of money we lose?” – Christian Bluhm, UBS