Bloomberg News - Fact or Fiction
London Coffee House
News, where would we be without it? It's largely about passing on information and possibly making decisions on the back of it. It's always been around in one form or another.. proclamations from the State, announcements at the Coliseum, talk at the London coffee shops (no, not Starbucks!). During the middle ages, the age of "runners", the event-to-receipt time could be as short as two months. Through history, two challenges lay ahead, latency and distribution/audience.
Newspapers started back in the 1600's and it wasn't until the early 1700's that they were published daily. By the mid 1800's the electrical telegraph ("News Wire") came into being, speeding up the transmission of news across the world. In 1851, Paul Julius Reuter set up the Reuters news agency in London. Then came the Press Association, Associated Press, Telegraph Agency of the Soviet Union (TASS), Xinhua, etc.
In the early 1900's came radio (the BBC started up in 1922), with television to follow on a mass scale in the 1950's.
Matt Winkler
In 1990, Mike Bloomberg and Matt Winkler created Bloomberg News starting with 6 people and eventually several thousand journalists working in News. Later in the 1990's the Internet came into being, and fast forward to today, we're in a an age where you hear and see news, good and bad, in real-time.
One of the key challenges though, is... what is news? Is it "Gizmo the dog went missing in Las Vegas in 2015. He’s been found alive after 9 years", or perhaps "Chicken wings advertised as ‘boneless’ can have bones", or, "Banking crisis: Lehman Brothers files for bankruptcy protection".
For those of us in financial services, the latter is probably a good example of something that might impact us, more so than the others. Whether that's concern about our pension, an investment, or, a signal to close out our positions, when we see news we are constantly in the "What do I do about this?" mode. Now that news is instantaneous, reactions to news is also near real-time.
When I worked with the #quant group in Bloomberg the Journalist's rule book precisely set out the quality and rules that Mike Bloomberg demanded of his team. You can get a taste of that quality by reading the "The Bloomberg Guide to Business Journalism", which not only talks about how different asset classes should be covered but also gives insight into how journalists leverage the Bloomberg Terminal as a part of their function and what happens behind the scenes. In today's world, whilst journalists still produce great quality articles, automation is undoubtedly a key facet of generating news promptly and reliably.
For us consumers, news is available through the Bloomberg Terminal as, TOP<GO> (today's top news), READ<GO> (most popular news), MMN<GO> (Market Moving News). We can search for news using the NI<GO> command (see Ted Merz's article here), so, if you're looking for a quick piece of humour you can easily search for "Trump".
All news articles are constructed from a number of components, components that are being embellished from the very second the news was created. You might find the Headline appears first, and then the Story later. Additionally, the news story is embellished with metadata such as PERSON, TOPIC or COMPANY. If Trump were talking about is admiration of Elon then it's likely the meta data would include TRUMP, ELON MUSK and TSLA N Equity. News is point in time, so as time marches forwards, the story gets embellished. It could be near-instant, it could be seconds, it could minutes, hours.., and what matters to you is... just enough to make a decision.
But whilst the Bloomberg Professional Services Terminal is great, we can't all be sat around, wearing out the big green GO button, refreshing the news waiting for something to happen. Enter technology. Remember, I've spoken about Bloomberg's Real-Time products in the past?
Bloomberg's Real-Time offering is a suite of offerings, available through a single API, the Bloomberg API (BLP API). Whether you're using the Terminal (Desktop API/DAPI), Server API (SAPI), or B-PIPE, they all support the same API. This API provides access to a rich universe of micro-services; mktdata for streaming, refdata for static/reference data, mktdepthdata for Level 2/3 data, srcinfo for Tick Size Tables, Condition Codes, mktlist for identifying pricing sources on a corporate bond, etc, etc.
Well, news is also available as yet another micro-service. You can subscribe to a news wire as though it was a ticker, and receive events of headlines, story updates, metadata updates.
One of the challenges with news is that the market will invariably react, irrespective of whether the news was fact or fiction, until a fact check proves otherwise (the "news" may have come from Twitter/X or an external newswire). But, if you're someone looking for some alpha, you could probably be on the receiving end of the electronic news feed, look for the market drop, then buy stock and wait for it to recover to previous levels.. sell, job done. Perhaps you're a market maker quoting two way pricing, and a piece of news comes in that moves the market (The TOPIC code will be set to Market Making News - MMN, the company/ticker code will also be set), which means you can automatically stop quoting in an instant, with logic to see what happens shortly afterwards. Market Makers would also probably want to understand what happens on an economic (ECO<GO>) event, as well as the timing.
Some of the above is great if you can react to metadata tags, but if you're wanting to understand the text in headlines and stories, then you're in for the long haul in terms of natural language processing (NLP), and even then you'll need data to train your models to understand what you're looking for.
But Bloomberg have an easier path. They have all the data, they have the models, they have an 'analytics' micro-service. It will produce sentiment based on you fictiously being long on a stock and should you feel positive, neutral or negative about the news. Take a look at TREND<GO> on your terminal for sentiment.
This was just a taste of the news feed, but there's way way more. If you want real-time; corporate actions, macroeconomics, textual news, analytics, taxonomy files, history, company filings, events calendar (EVTS<GO>), speak with your friendly Bloomberg representative.... there's a micro-service waiting for you.
