Machine-readable sovereign downgrades are here | FT Alphaville

Machine-readable sovereign downgrades are here

Here’s how those macro-obsessed markets are helping to reshape market technology:

New York, NY – May 16, 2011 – Selerity, a provider of real-time event data solutions for the financial services industry announced today an agreement with Standard & Poor’s (S&P) to deliver low-latency credit rating announcements in a machine-readable format. Clients using the Selerity Platform will now have access to breaking credit-related announcements from S&P, including ratings on public companies, along with developed and emerging sovereign nations globally.

This specific low-latency machine-readable technology, Standard & Poor’s Roger Fahy tells us, was developed specifically for Selerity, though a handful of other machine-readable news platforms use it.

Put simply, it involves S&P tagging its ratings events — upgrades, downgrades and outlook changes — with ‘machine readable’ coding, i.e. numerals. Selerity then takes the feed from S&P and if needed structures it further into machine-digestable format, making it available to its high-frequency and algorithmic trading clients. As for the latency aspect — think loooow. Selerity’s systems are colacated at Equinix’s NY1 and NY2 servers, which means we’re talking mere milliseconds.

The reasoning behind incorporating sovereign downgrades is rather interesting, with Selerity CEO Ryan Terpstra specifically citing S&P’s change of the US ratings outlook on April 18 as an instance of rating agencies roiling markets. Market makers or hedge funds can use machine-readable ratings signals as defensive stop-loss generators, or as triggers for new offensive trades, he says.

Most intriguingly, growing investor interest in things like sovereign rating events seems to be translating into the rise of not just “machine readable news” but “machine readable forecasting.”

Selerity calls it “pre-market intelligence,” and when it comes to ratings downgrades it essentially involves the firm tracking events and indicators that might suggest a downgrade is on its way. The model, Terpstra says, is constructed by looking at what happened before previous downgrades.

If you’re wondering what machine readable “pre-market” intelligence looks like, exactly… Well, we are too. Terpstra told us it’s a “visual representation in our reference data platform.”

So maybe this?

Related links:
The redundant trader – Risk, 2008
Rise of the news-reading machines – FT Alphaville
Meet the Google for high frequency trading machines: Selerity – Forbes