This article was written by Noel Peatfield, Marketing Manager at FINTECH Circle.
I had barely opened my eyes early in the morning of June 24th when my first foggy thought of the day crystallised…’vote results’.
Grabbing my phone off my bedside table without looking, it took minimal effort to find them. The internet was in meltdown with countless voices expressing their opinions on the outcome.
The run up to this moment had provoked one of the largest public debates the UK had ever known. This time many of those involved were doing it on social media suggesting that in amongst this torrent of information the result had been out there all along.
Hedge funds and investment banks spent record sums on polls, yet many of our leading bookmakers and pollsters got it completely wrong.
I was scheduled to have a Skype call later that morning but was told it had to be cancelled as the internet was going down in parts of London due to a huge increase in demand, with news consumption and social media activity happening simultaneously on millions of devices.
What was making this process increasingly interesting was that a new kind of data that could do things polls and bookmakers had never done before appeared to be coming of age. And I wasn’t the only one noticing this.
A few days later it was becoming evident that fund managers had profited from using social media data. It was widely reported that Brevan Howard Asset Management had reduced risk by using social data and that its $16 billion macro fund gained 1 percent on the day of the results. Blurrt’s EURef data hub built for Twitter and the Press Association was consistently indicating a leave vote based on volume of tweets.
The innovative use of Twitter data by financial services is nothing new. Times have moved on since the Derwent Twitter hedge fund and a growing awareness of the sheer scale of secretive algo traders using social data. There are clear signals that behavioral finance can be put into practice for a wider range of applications now that there is an increased ability to read the mood of the crowd.
Sentiment can be an accurate response to the state of the economy whether that be during good times or bad. Among the noise on Twitter it is still easy to find lots of predictably rational views on events. Even so it is plainly evident that curbing impulses driven by cognitive biases is not always possible (even by those who strictly follow the principles of the efficient market) which can translate into tweets and market activity.
During the crashes of Black Monday in 1987, the dot-com bubble of 2000 and the financial crisis of 2008 it became evident that investor sentiment started moving asset prices away from their fundamental values. Now we are in 2016 this sentiment can be analyzed more closely from the most popular social media channel in finance, Twitter.
With the referendum vote result becoming live, panic selling ensued and a cycle of news and market activity began. Twitter went into overdrive with posts about ‘pound at 30 year low’ and ‘a disastrous result for the UK’. The herd had been spooked and the stampede had started.
According to behavioral finance, individuals react more strongly to bad news than good news. Traders examine market consensus, the positions of other traders and news reactions to ascertain market sentiment. Traders then respond to this information. The emotions that the information evokes is a key factor. If information is better than expected they feel good; if it’s worse than expected they feel disappointment and in extreme cases fear. These behavioral biases surface and get shared on social platforms such as Twitter.
Stay Prepared With VPS from InstaForexGo to article >>
Algorithms such as the one that powers Facebook’s news feed are designed to give us more of what they think we want – which means that the version of the world we encounter every day in our own personal stream has been invisibly curated to reinforce our pre-existing beliefs. Sometimes referred to as the ‘filter bubble’. Social networks can lead to a bias and the herd mentality gives way to confirmation bias, with people only looking (and receiving) for information that confirms their preconceived beliefs and retweet them as a commitment to their views.
Weeks after the vote apparent confirmation bias is still running high at all levels regardless of long term self-interests, promoting less value with no end in sight. But there has also been a lot of energy from those trying to look towards a broader landscape.
News analytics used to give us what it claimed was a reflection of crowd sentiment, but with Twitter this isn’t a reflection. This is the real thing, unprocessed opinions of consumers, traders and voters. Granular big data that can be analysed by each individual tweet or summarised by natural language processing in large numbers.
In its simplest form there are two ways social media can be useful when trading in equities and currencies. As a way of reading the crowd’s behaviors for a predictive edge and insight into the origination of your own beliefs and biases. Twitter moves market prices as do news stories. The social media amplification of these events and the market’s response fuels a self-sustaining cycle of confirmation bias activity independent of views that don’t conform to its beliefs.
Warren Buffet produced a treasure trove of quotes and one liners that have been held up by the test of time but one remains at the top of the list – “Be fearful when others are greedy and greedy when others are fearful.”
On the basis of this even with the unprecedented levels of uncertainty in the markets at the moment, one thing is for sure. Traders and investors are fearful. Warren’s upheld theory championed by some of the best in finance tells us that we should be betting the farm on the British pound.
The problem is no one knows where the bottom is. In attempts to identify where it might be technical analysts are starting to explain their findings through behavioral finance. A momentum based strategy that relies on understanding the trend of a security’s price or volume depends on analysing movements in price. There is now a growing school of thought that says that the psychological biases of behavioral finance can help to explain these movements also.
Twitter data offers a front row seat where it is possible to sit back and watch leading influencers and entire crowds, rather than being confined to only watching charts move in irrational directions across their timelines. Twitter can provide an extra layer to correlate with other market data – influencer activity, social sentiment, emotions and tweet volume can create a clearer picture of irrational behavior like overreaction.
In the example of panic selling behavioral finance puts forward explanations of ‘why’ which have become widely accepted. One of its leading concepts is Hindsight Bias which can be seen after all of the major market crashes. This bias follows the belief that with hindsight it is abundantly clear what went wrong and that it should have been easy to predict when in fact the opposite is true. Panic selling though in many cases is irrational, is not always easy to define as it happens.
Predictions and panic
Evidence suggests that even the most seasoned market veterans with nerves of steel are not immune to psychological bias which can drive their behaviors. Considering that around half the population in the US own stocks, it is not then surprising how social media can affect the behaviors of millions and price momentum.
For Stephen Harper, Chief Executive at wealth management company Attivo Group, this phenomenon is not an uncommon feature of the financial landscape.
“Social media can be a great source of information but as an adviser and financial planner I sometimes wish that my clients wouldn’t listen to all the noise that is available. Not every opinion is the right opinion.
Even in the short term many of the predictions of what would happen to the markets on Brexit were completely inaccurate and in numerous cases the opposite happened. Adapting to an increasing amount of information we continue to avoid short term, knee jerk reactions even as we face a sometimes over reactive media both in the press and on social.”
The amplification of news through Twitter can be monitored, and the consequent overreaction of market activity measured, which can in turn enhance predictive models. Equally as important you can begin to ask yourself if your current beliefs are influenced in the same way as the crowd by frequently comparing it with Twitter data.
I know a few individuals in my life that I look to for advice. I have learnt that when one particular individual voices even the slightest concern about something, it’s a solid indication that something serious is on its way. Whilst another of my closest advisors usually ignores the negatives, I know I’m onto a good thing when they give me a tip on a win that is yet to happen. While both have a lot to say about a lot of things I’ve learnt how to listen to them. Twitter is another one of these invaluable voices. It has its own character and is well worth getting to know.