The Quality Of Investing In The Egyptian Stock Market

Published Date: 02 Nov 2017

Chapter 1

Technological advancements and increased customer preferences for convenience have caused an overall restructuring of the financial services sector. Today, financial service providers are customer oriented and focus on improving the satisfaction of their customers by improving the quality of their services (Loonam and O'Loughlin, 2008). The quality of the financial services provided is the main comparative advantage to survive and compete in the financial marketplace (Gonzalez et. al., 2004). This research focus on the stock market since it has become the main investment outlet in most countries. The quality of investment in the stock market is highly subjective, but since stock investment decision are mainly based on predictions of stock prices it is well agreed that improved price predictions would surely improve the quality of investing in the stock market (Perkins Investment Management, 2011). More efficient price discovery contributes to better service quality within the stock market (Hendershott and Moulton, 2011).

Various market analysis techniques have been applied to predict future stock prices. However, these techniques require a certain degree of expertise in finance and economics and require extensive data collection which is too much effort for individual investors. Eventually, many individual investors exit the stock market after losing their invested money in a vicious cycle. A good stock price prediction decision support model for stock investors is indispensable (Chang, Liu, Lin, Fan, and Ng, 2009).

Prediction of stock market prices has been an area of great interest to those who wish to profit by trading stocks and for researchers attempting to uncover the information hidden in the stock market data. The problem has attracted the interest of many researches emerging from field of statistics, economics, and computer science. However, stock prices prediction is still regarded as one of the most challenging problem and all attempts since the 1900s have failed to build an accurate and efficient stock price prediction system (Chang, Liu, Lin, Fan, and Ng, 2009). This is because stock data is data intensive, multidimensional, non-stationary and chaotic in its nature. The autocorrelation for day to day changes is very low, and the noise level and volatility in the stock time series changes as the stock markets move in and out of periods of turbulence (Chang, Fan & Lin, 2011). Stock market prices are influenced by numerous political and economical factors in addition to the influences of the market itself on these factors.

It is well argued that history repeat itself and careful analysis of the past would surely provides good insight about the future. Professional technical analysis stock traders analyze past prices and volumes of stocks, to predict future trends in the market. The artificial intelligence methodology of case-based reasoning (CBR) takes on this line of thinking by attempting to use knowledge of previously experienced situations in the form of cases to solve currently experienced situations with no apparent solutions (Lenz, 1999). Previous experiences are stored in the form of cases and new situations are formatted as a new query case. CBR imitate human problem solving using analogy rather than using probabilistic models. With recent exponential improvements in hardware and software, the CBR methodology could now provide new frontiers for the problem of stock predictions.

The Egyptian Exchange has been among the top emerging stock markets since 2004. It experienced impressive growth in recent years and its annual value of trades grew from $7.8 billion in 2004 to $67 billion in 2007 and $103 billion in 2008 (Oxford Business Group). The monthly average value of trades reached $9 billion. Market capitalization increased from $113 billion in 2007 to $116 billion in 2008 which represents 11% of the Egyptian Gross Domestic Product (GDP) (Egypt Exchange, 2008). However, the Egyptian stock market being one of the emerging stock markets, exhibits high volatility levels (Mecagni and Sourial, 1999).

Research Problem

The quality of investing in the Egyptian stock market is highly dependable on price discovery. However, predicting Stock Market Prices is a complex process governed by large amounts of data and numerous market factors. Moreover, prices of stocks do not follow these factors in an orderly manner, but follow market participants’ sense. The obvious complexity causes great problems for traditional algorithms. Due to complexity of the research problem for predicting Egyptian stock market prices, it is found necessary to explore the feasibility of building a decision support model to aid stock investors predict prices of Egyptian stocks.

. Aim of Study

The aim of this research is to analyze historical Egyptian stock records using lengthily daily data sets from 1998 to 2012. Statistical analysis would be used to investigate the significance and interdependence of different stock variables to give insight towards building a quality controlled price prediction model. This research would then explore the feasibility of integrating the powerful tools of statistical process control with case based reasoning to build a quality controlled decision support price prediction model. The model aims to improve the quality of investing in the Egyptian Exchange by formulating quality controlled next day stock price predictions based on matching historical stock records. Control charts would be formulated to ensure the quality of stock predictions. This research is limited to implementation within the Egyptian stock market being one of the emerging stock markets with high levels of volatility.

Research objectives

This research aims at achieving the following research objectives:

To conduct an in-depth investigation and analysis to the Egyptian Exchange historical stock records (1998 – 2012).

Explore the feasibility of using different control charts to monitor and control variable and attribute features.

Examine the ability of predicting stock prices using case-based reasoning.

Explore the feasibility of integrating the control chart as a powerful statistical process control tool to monitor and control the accuracy of next day stock predictions.

Research Questions

Based on the research objectives, the extensive literature review, brainstorming questions with the author’s supervisors, and informal talks with domain experts, the following research questions were developed:

Question 1:

Based on the statistical analysis for the historical stock records of the Egyptian Stock Exchange, what are the descriptive and correlative values of different stock features to best construct a stock case?

Question 2:

Would it be possible to predict next day stock prices based on matching historical stock records using Case-Based Reasoning?

Question 3:

What are the optimal feature weights and optimal number of nearest neighbors that cause the suggested model to provide the least prediction error?

Question 4:

How to efficiently integrate statistical process control with case-based reasoning to improve the quality of the stock prediction process?

Thesis Structure

This thesis consists of five chapters described briefly as follows: The first chapter describes the research problem, the objectives of study and research questions. The second chapter presents the work related to this study including the Egyptian stock Exchange and description of different CBR approaches used to predict stocks from the respective literature. The third chapter overviews the methodology of statistical process control (SPC) and the use and statistical design of different control charts to monitor and control the natural and assignable causes of variation of any process. The fourth chapter describes the collected data, research methodology, the design choices and implementation of the suggested quality controlled stock prediction model and concludes with model testing and verification. Finally, chapter 5 provides conclusions and recommendations.