Financial Data Mining Literature
Competitors’ Effect Literature
Critical Analysis and Gap in research
Controlling for market
Competitors’ Influence - Regression
Influence on Competitors – Regression
Discussion of the Results
Ratio analysis results
Regression analysis results
Possibility for practical use
Limitations of the research
Recommendations for future study
New information moves the stock prices. Good news makes investors be optimistic about a company and therefore they buy the stock. Bad news, on the other hand, leads investors to sell the shares, which decreases the price. The news forms trends and patterns, and portfolio managers are willing to exploit them.
The goal of active portfolio managers is to earn money. They want to earn money in excess of the market return, but at the same time, maintain low level of risk. To do so, they engage in fundamental and technical analyses, trying to identify mispricings in the market. They look at trading patterns, historical data, financial markets and recent news. The problem with the analyses above is the time required to perform them. In recent years computers became much faster and smarter and programmers taught them to analyse news.
Currently with the release of a news article the stock prices adjust almost instantaneously due to the use of data mining software. These kinds of programs analyse text patterns and financial data from articles and financial statements in a matter of milliseconds and then perform trades accordingly. Obviously, people cannot challenge them in speed and therefore all the direct arbitrage opportunities on recent news releases disappear.
One of the most common types of competition within the technological industry is oligopoly. Oligopoly is a form of competition where few firms have the majority of the market. They produce similar goods and can make economic profit in the long run due to barriers to entry. Due to the size of companies’ market share actions of one company affects all the others in the industry.
Firms in an oligopoly share the overall market demand. Due to that fact, if company A releases an arguably better product than its competitors, the demand would shift from them to company A. Situation X:
Graph 1 – Situation X
DA – company A’s demand
DC – competitors’ demand
The release of the product is good news for company A, but bad news for its competitors. Therefore, the share price of company A should rise, while the share prices of its competitors should fall (all other things remaining constant). Situation X can also be supply-driven. If the costs of production for company A decrease, Marginal Cost will drop, which would increase quantity and decrease price, as well as increase economic profit. The decrease in price for firm A’s goods will result in lower demand for competitors’ products, so it is a bad news for them.
Another situation is when the overall demand for the industry’s goods rises. An example may be a shift in consumer preferences towards portable flash-drive based music players from CD-players. Situation Y:
Graph 2 – Situation Y
DA – company A’s demand
DC – competitors’ demand
In this case the news is good for both company A and its competitors. The demand for all the goods in the industry rises, which increases revenues and profits. Therefore, share prices of all the companies within the industry rise (all other things remaining constant). Situation Y can also be supply driven. Reduction in costs for all firms will increase their profits, which is good news. Therefore share prices of all firms should increase.
Since we know that competitors’ news affect a company in an oligopolistic market we may want to extend the data mining techniques. To do so a program has to identify what type of news arrived, whether it is leading to Situation X or Situation Y. Situation Y arises due to industry-wide news so this news applies to all the firms. Therefore, data mining software has already processed the information and arbitrage opportunities have already disappeared. Situation X, on the other hand, is more subjective. Judging such news technically is more difficult and therefore data mining software is not effective. But if companies in oligopolistic markets happen to have situation X more commonly than situation Y we can try to exploit this arbitrage opportunities.
If situation X occurs more often than situation Y, portfolio managers can get trading programs to do corresponding trades. If a company A releases good news, the trading program will open a short position in competitors’ stocks (assuming situation X occurs). Because of situation X news arrival investors will then sell the shares of the competitors and drive the price down. After the prices adjust the program will close the short position and earn money, therefore earning money from the arbitrage opportunity by reacting faster than human investors.
Two obvious questions arise: “why do not data mining programs trade competitors’ stock in reaction to the company A’s news?” and “why will the suggested program design earn money? What if situation Y happens instead?” As I said earlier – it is hard for data mining programs to judge whether the news is good for one company, but bad for its competitor. This kind of judgement is different in every case. For example, a reduction in the electricity cost in the whole country is good for both companies, but a reduction in the electricity cost in a state, where company A operates results in cost reduction for firm A, making it more competitive. Therefore it is good news for company A, but bad its competitors and such news are controversial for the programs to make a decision due to similar wording. However, if Situation X happens more often than Situation Y, on average a good news for company A will mean bad news for its competitors (and vice versa). If that happens, say, 55% of all the times, 55% of all trades will be profitable. On average the program will be able to earn arbitrage profit.
In part two, Literature Review, the dissertation will cover oligopolistic competition and financial data mining literature. It will also go over empirical studies of competitors’ effect on share prices. The methods of data analysis and sample selection are going to be discussed in part three. The analysis itself will be covered in part four. Then the results of the analysis are discussed and concluded.
Does Situation X occur more often than Situation Y? If yes – for which companies and how do stock markets react? Can arbitrage opportunities arise due to that fact?
Pindyck and Rubinfeld (2013) state that oligopoly is the most common type of competition, which includes electrical and computer goods.
Firms can earn economic profits due to the existence of barriers to entry. The barriers to entry can be either natural or artificial. Natural barriers include patents and brand recognition. Artificial barriers can arise if firms that are already in the business prevent new companies to come. An example can be a drastic price reduction, which would not allow the new firm to survive.
Oligopolies are interdependent and any decision made by a manager of a firm A has to be thought from the point of both company A and its competitors. An example of such interdependence is the game theory situation, where firms’ collusion or competing results in different profit distributions. One of the type of such games is Research and Development expenditure, which is especially important for technological companies. The money spent on R&D can earn them competitive advantage, larger profits and prevent new firms from entry, as patents prevent competitors and newcomers to use the same technology. If that is the case, the share price of the company rises.
+ , +
+++ , -
- , +++
+ , +
Graph 3 – R&D Game
Modified (Pindyk and Rubinfeld 2013, pp. 516)
Firm A and B start at “no new patent” point, where both of them earn some economic profit. If a company A has an R&D development then it is in a competitive advantage, which increases its profits. Meanwhile, company B suffers. These kind of events should be reflected in the stock market by the appreciation of company A’s shares and depreciation of firm B’s stock.