WHITE PAPERS
The impact of news on share price
It is well known that news has an impact on share price, but one BIG question remains: how much? A new service from Spectrum,
Fin-buzz™, can now provide the answer. In this article Spectrum’s founder and product Development Director, Mark Westaby, explains how
On 3rd October, 2008, a ‘citizen journalist’ posted a story on iReport, a website run by CNN that allows anyone to put up posts and videos about the news. The story claimed that Apple’s CEO, Steve Jobs, had suffered a heart attack. Apple rapidly corrected the story, which was not true, but only after the company’s stock had plunged by 10 points, representing a huge fall in its value.
It is well-known that news can have a major effect on share price and company value but what the Apple example graphically reveals is that in today’s internet-world virtually anybody can post a story on virtually any source, which can have a dramatic impact with remarkable speed. In this case the effect of the ‘news’ on Apple’s share price was pretty clear, though it is worth noting that the price after the event was lower than before it happened. But what about everyday news stories and rumours that affect a company, many of which can appear “out of the blue”? How can the financial impact that these and other stories have on share price and market value be determined, especially as they nearly always happen at the same time?
Is ‘gut feeling’ good enough?
Many managers will claim to use gut instinct and have a ‘feeling’ about the news and sentiment that affect their company’s share price, but is this good enough? Given that an unknown ‘citizen journalist’ can have the impact he did on a large quoted company such as Apple, the answer has to be a resounding “no”! Spectrum has developed a solution, Fin-buzz™, that enables companies to understand the impact of different news stories on their share price. This is critical in a world where business confidence is vital for success and even small changes can have a dramatic impact on market value.
Apple’s stock plunged by 10 points on the false news that its CEO, Steve Jobs, had suffered a heart attack
Real-time analysis the key
A major feature of Fin-buzz™ is that it works in real-time, enabling news stories to be tracked against share price simultaneously. This allows much greater accuracy, first, by removing any time lag during which the news content can change; and, second, by removing the possibility of ‘contamination’ from feedback that a change in share price, which could itself be newsworthy, might otherwise create. The following example shows how Fin-buzz™ works. It is based on Rio Tinto, which is one of the world’s largest companies.
Chart 1 shows Rio Tinto’s share price on the London Stock Exchange between 3rd and 9th April, 2009, in two-hourly time slots. Fin-buzz™ uses sophisticated search and web crawls, which 'scrape‘ text from leading financial sites on the web at pre-selected times that coincide with the two-hourly share price slots mentioned above. Automated analysis is then carried out using text mining, which like the highly effective spam filters now being used takes advantage of a special branch of statistics called ‘conditional probability’ to continuously improve accuracy.
Chart 1 showing Rio Tinto share price (RH axis) against volume of coverage for stories shown (LH axis)
Benefits of automated analysis
Automated analysis of news content has come of age with the advent of powerful computers to the extent where it now offers significant advantages over human analysis. No two humans ever analyse content consistently and humans will always introduce errors when large volumes of data have to be coded. In addition, people simply cannot process data fast enough to enable real-time human analysis to be of any practical use. Automated text mining, on the other hand, achieves total consistency and coding accuracy across any volume of content. Add to this the fact that automated text mining allows literally millions of characters to be analysed in a very short time period, typically seconds, and the advantages become obvious.
As well as share price, chart 1 also shows the results of the Fin-buzz™ analysis, which revealed coverage on the following stories during the period in question:
-
The Chinese state-owned metals company, Chinalco, taking a stake in Rio Tinto and buying up to 50% of its key mining asset
-
Aberdeen Asset Management, sees a proposed $19.5 billion investment by Chinese state-owned metals firm Chinalco as good for the miner
-
Rio Tinto pours cold water on BHP Billiton merger
-
Rio Tinto has drawn up plans to ask shareholders for as much as £8bn in a rights issue
From the Fin-buzz™ analysis on chart 1 it appears that the company’s share price was significantly higher during the ‘Aberdeen’ story. It also appears that the share price fell rapidly once the Aberdeen story no longer appeared, which coincided with coverage of the ‘Rights issue’, ‘Chinalco’ and ‘Merger’ stories on the 7th and 8th April. It can also be seen that the share price rallied a little at 13:30 on 8th April – when none of the above stories appeared – before falling again at 15:30, which coincided with further coverage of the merger story.
The positive and negative impact of different stories
As the chart shows, analysing online coverage, or ‘sentiment’ as it is usually called, in real-time clearly reveals that the Aberdeen story had a significantly positive impact on Rio Tinto’s share price. It also appears that the other stories had a negative impact on share price. Given their repetitive nature this is critically important information, which would enable the company to plan and manage its business strategies accordingly. But how can we be confident that the conclusions drawn from the chart are accurate, especially given the confusing overlap between a number of the stories on 7th and 8th April?
To answer this Spectrum uses a well-proven statistical method, which for the technically minded is known as ‘multiple regression’. The non-technically minded need have no fears, however, because the results are extremely easy to understand. Carrying out a statistical analysis using this method, which is commonly used in marketing and economic analysis, reveals that the news stories together explain 81% of the change in share price over the period, with the individual contribution of each story shown in chart 2. From this it can be seen that the ‘Aberdeen’ story is the only one having a positive influence on share price over the period, with each of the others having a negative influence.
Chart 2 shows the contribution of each news story to changes in the Rio Tinto share price between 3rd and 9th April
An obvious question that arises is how confident can we be about the accuracy of the analysis that generates these values? In fact the confidence level of the statistical analysis is remarkably high at 99.99%. Put another way this means we can be 99.99% confident that the results from the analysis cannot be down to chance, which even given the relatively small sample involved is an extremely robust figure. (Again for the technically minded, it is important to note that the statistical model is used only to gain a greater understanding of the underlying trends in terms of which stories are driving share price movements and by how much, which it does very successfully. The model would not be used for forecasting, for which a much bigger data set would indeed be required).
The financial impact
Something we can now do is to determine the financial impact of each of the news stories involved. Let’s first look at this in terms of Rio Tinto’s market value. Over the period the share price fell by 339p, or ~13%, which given Rio Tinto’s market value at the time of ~£23.6bn represents ~£3.1bn of which ~£2.5bn can be explained by the negative stories above. In other words each negative story was responsible for removing ~£800m from the value of Rio Tinto between 3rd and 9th April, 2009!
Such numbers are frightening but should not be a surprise given recent events on the world’s stock exchanges when news stories were wiping literally billions of pounds of value from even the strongest companies. Another way of looking at the financial impact of the Rio Tinto news stories is from the viewpoint of the investor. Doing this reveals that the ‘Aberdeen’ story effectively added 220p to the value of a Rio Tinto share whereas the ‘Chinalco’, ‘Merger’ and ‘Rights issue’ stories effectively removed 43p, 68p and 64p from the value of each share, respectively. Again, these are significant values that would make a large difference to the value of an investor’s portfolio.
In summary, the fact that news has an impact on share price is well known. What is not known in anything but the most obvious cases, however, is just what that impact is. This is rapidly becoming critical as the advent of economic and market volatility is being generated largely by news-driven sentiment. For companies to build confidence in their stability they must first gain in-depth insight and understanding into just how news is driving their market value.
Companies that gain this insight and understanding are much more likely to succeed and prosper while those that fail to do so are themselves more likely to fail.
The critical relationship between online sentiment and share price
The internet has produced an explosion of universally-accessible and instantly-available information. The impact of this information on a company’s share price can be sudden, dramatic and unpredictable. Only by keeping a finger on the pulse of online sentiment can companies understand its potential influence on stakeholders and, ultimately, on market value.
Market sentiment: the traditional definition
In the world of listed companies, “market sentiment” drives share price. Traditionally, the definition of market sentiment has been the views of financial analysts, stockbrokers and those who participate in the activities of the world’s stock markets, with share values reflecting the current accepted wisdom about how likely an organisation is to perform according to expectations. To understand how the company’s share price may behave, executives and investors stay abreast of the predictions of market analysts, bearing in mind what positive and negative news is already factored in to market valuation.
Market sentiment: a definition for the Internet age
Now, there is another dimension to market sentiment – that reflected in the online world, where anybody, anywhere can share an opinion about an organisation and that view is immediately visible globally to anyone searching for information about the company.
Online sentiment can reveal much about a company’s health. The web has become the world’s largest focus group, providing a window on the views of current and potential customers, industry commentators and other stakeholders. The viral nature of the Internet means it takes seconds for feedback posted online to travel across audiences and geographies. In today’s market, a comment on a blog, message board or online review can quickly influence perceptions and start to impact on organisational reputation.
Ignorance is bliss, but not for long
It is astonishing how many publicly quoted companies fail to monitor online commentary relating to their activities, let along evaluate that information or react to it. Company executives may not be aware of the pervasive nature of, for example, a damning product review, or negative feedback about a company’s customer service. They may not realise that a credible blogger can quickly be elevated to a prominent position in an online search, so that the first entry seen by the customer, shareholder, employee or journalist searching for company information is of an unfavourable nature.
As an indication of the power of online opinion, research in the UK shows that “word-of-mouth” opinions published online are now more widely consulted than any other source of information by adults seeking to make purchasing decisions (source: Spectrum 2008). In this environment, companies ignore online commentary at their peril. Stakeholders have the opportunity to be better informed than ever and reputations can be made or destroyed online before the public relations or investor relations teams have even realised there is an issue.
Impact on share price
Online sentiment, whether positive or negative, often reflects very closely trends in a company’s share price. Monitoring and understanding online sentiment not only serves as an early-warning system to detect any potential issues or crises, it also allows organisations to consider their likely effect on market valuation and to manage their response effectively. The key is to know what is being said about the organisation in real-time – at the same time that other stakeholders have the opportunity to read, evaluate and act on it.
Watching the competition
One of the ways analysts track companies is to monitor their performance versus similar organisations in their market sector. Sentiment is an important part of this equation. Currently, analysts rely on charting the rise and fall of share prices in graphs and using technical analysis to determine where the likely highs and lows of the stock trading range may be for the time period they are interested in. Having an earlier indication of prevailing sentiment may forewarn of probable market movements or at least enable them to validate their hypotheses and predictions.
For a management team, knowing how the company’s market sentiment compares to that of other FTSE companies gives valuable insights into the comparisons analysts and shareholders are able to make on a daily or hourly basis. This understanding allows them to be prepared to communicate more effectively through investor relations programmes and to be ready to defend themselves proactively when the tide of opinion seems to be turning against them.
Likewise, for industry watchers such as analysts or management consultancy firms, the ability to gain rapid insights into sentiment from Internet coverage is an added dimension to the complex evaluations they perform when making recommendations about the health and strategy of companies.
In an economic downturn, where listed companies may suffer more than most from a sudden and dramatic decline in their market value; where every announcement about profitability, output, employee relations and market share can have devastating impact, the value of having access to real-time comprehensive information about perceptions in the public domain should not be underestimated.
Market and sentiment drivers
For a listed company there will be a number of market drivers that influence performance on an ongoing basis. These will typically include the state of the economy and whether it is growing or in recession, merger and acquisition activity, the price of oil, commercial property, exchange rates and so on. Tracking online commentary that compares the company’s performance to these key drivers and analyses the sentiment associated with that commentary makes it possible to gain a snapshot of market sentiment at any given time. Similarly, monitoring of sentiment trends over time allows organisations to observe the impact their outbound communications have on “online buzz” and to see correlations between online commentary and stock market response.
Fin-buzz™: online reputation monitoring for FTSE-listed companies
For the first time, Fin-buzz™ offers companies an invaluable source of data relating market sentiment to the factors that make share prices move, providing critical insights into the consequences of business decisions, news and market conditions as they happen. Providing instant feedback from the markets, Fin-buzz captures the mood of the investment community as it changes.
The technology revolution that gave us the Internet has thankfully also given us the necessary tools to understand and manage it. Simple to use, Fin-buzz is available 24 x 7 through any web browser. Advanced text mining systems developed by Spectrum form the platform on which Fin-buzz has been built. They incorporate sentiment ‘dictionaries’ which, like the highly effective Spam filters now coming into use, take advantage of conditional probability based on Bayesian statistics and neural networks to continuously improve their analysis. They do this by building an increasingly detailed understanding, based on ‘learning’ from an accumulation of data over time. This allows Spectrum to rapidly assimilate the vast quantities of data in the online world and translate them into meaningful business information.
What’s more, Spectrum is able to track market sentiment in real-time, helping organisations keep their finger on the pulse and tailor their investor relations programmes more responsively. Now, using automated analysis, organisations can gain in-depth information on competitors too – fast enough for the information to be relevant and at a price that is affordable, even in hard times.
Here are a few examples of how share price can be related to specific business episodes using Fin-buzz:
· Monitoring the immediate and ongoing impact of a company’s financial results
· Evaluating the impact of company news, competitor, government and other announcements
· Monitoring the impact of commercial decisions such as price increases or product launches
· Demonstrating the tangible impact of issues in the public domain, eg environmental concerns
· Showing the repercussions of legislation, regulation and intervention
· Tracking key drivers of online sentiment to reveal their impact on share price at any given time
· Providing early warning of impending crisis and real-time monitoring of crises as they unfold
Flexible Sentiment Monitoring
Fin-buzz provides a subscription-based service via the Internet, allowing listed companies to see, in real-time, how they are performing versus the rest of the market in terms of overall sentiment. Low-cost but incredibly compelling, it charts fluctuations in market sentiment across the entire FTSE100.
Disclaimer: Spectrum is not authorised by the Financial Services Authority to give financial advice. Users of Fin-buzz are reminded that it in no way constitutes any recommendation or otherwise to deal in a company’s shares and that the service should be used as part of an investor relations strategy or in conjunction with qualified professional opinion.
Fin-buzz™ is owned, developed and operated by:
Spectrum
Floor 3, Russell Chambers
The Piazza
Covent Garden
London
WC2E 8AA
Tel: 020 7836 0132
Fax: 020 7240 4849
Email: markw@spectrum-consulting.net
Web: www.spectrum-analysis.com
