Public investors normally make their investing determinations harmonizing to some or all information that is available to them, such beginnings including newspaper, Television intelligence, stock agents, fiscal web sites or even rumor. Some experts encouraged investors to pay more attending to understand their investing portfolio by analyzing relevant information ( e.g. industrial intelligence and fiscal studies ) . They believed that by analyzing past information, investors could better the procedure of determination devising or even predict the hereafter tendencies of stocks. In the same clip, a classical fiscal theory besides told us that in a market with semi-strong or strong signifier efficiency, the monetary values have already reflected all publically available information. That is to state, investors might be unable to do a net income by analyzing all past information they could entree. However, some instances ever could be found to dispute this theory. This state of affairs confuses many investors when they found the imperfectness of fiscal markets. This proposed research is seeking to set up a theoretical account to entree whether a relationship exists between publically available information and stock information for any given fiscal markets and give investors a clear consequence by utilizing a proved quantitative analysis.
The inquiry of whether fiscal markets can be predicted has been studied by many research workers. Some believed that the behavior of fiscal markets are random and unpredictable while some believed that by utilizing information, it is possible to crush the markets and do inordinate net incomes. One celebrated theory, efficient market hypothesis, is normally mentioned when people discussed this subject. Efficient market hypothesis asserts that while investors purchase stocks, they will utilize any available information expeditiously. All known factors which affect the monetary value of stock have been reflected in the monetary value of the stock ( Dimson and Mussavian, 1998 ) . Harmonizing to EMH theory, investors can non surpass the market systematically by merely utilizing the past, public information.
In 1970, Fama ( 1970 ) proposed three types of efficient market signifier in his research: weak signifier efficiency, semi-strong signifier efficiency and strong signifier efficiency. These three signifiers are different in the extent of how the market reflects the information. In a market with weak signifier efficiency, the current stock monetary value has reflected all information provided by past stock monetary value. So, investors can non foretell the hereafter stock monetary value by analyzing the past stock monetary value. Based on the random work hypothesis ( Malkiel Burton, 1973 ) , it has implied that the future information ( future monetary value ) will come to us randomly so that investors have no opportunity to derive inordinate net incomes through past monetary value information ( Fama, 1965 ) . In a market with semi-strong weak signifier efficiency, the current stock monetary value has reflected all publically available information. Investors trusting on the fiscal studies, economic state of affairss, political fortunes, etc. can non foretell the hereafter stock monetary value or derive inordinate net incomes ( Chance, 1985 ) . In a market with strong signifier efficiency, the current stock monetary value has reflected all publically available information and all non-publicly information ( insider cognition ) . Although non-publicly information does n’t let go of to public, some investors can still obtain these information by their ain channels or connexions. In this state of affairs, investors besides can non do inordinate net incomes through inside trading ( Rozeff and Zaman, 1988 ; Finnerty, 1976 ) . However, there has been a difference over the efficient market hypothesis between research workers, because many groundss reflect the imperfectness in fiscal markets and their inefficiency. Some “ behavioral economic experts ” asserted that cognitive prejudices factors such as assurance, reaction, and information prejudice consequence in the inefficient fiscal markets. This besides implied that investors are non ever rational ( Stout, 2002 ; Fama, 1998 ; Rosenberg et al. , 1985 ; Givoly and Lakonishok, 1979 ; Jensen et al. , 1978 ) .
Although research workers realized the fact that there are many factors impacting the efficiency of fiscal markets, it was still less helpful for investors to do determinations. Investors would wish to cognize whether the markets in which they are interesting are efficient and harmonizing to the different features of markets, they can follow different investing schemes. Some research workers tended to pay attending on fiscal indices such as dividend and fiscal ratios and utilize them to measure the mark market. An early survey of the efficiency of the UK stock market by Marsh ( 1979 ) found that the UK stock market has a semi-strong signifier efficiency by detecting the relationship between the timing of rights issue proclamation and portion monetary values ( Marsh, 1979 ) . For measuring the efficiency of 15 emerging capital markets, Karemera et Al. ( 1999 ) evaluated the stochastic belongingss of local currency and US dollar-based returns. The consequence showed that most of these emerging markets seemed to be weak signifier efficient ( Karemera et al. , 1999 ) . Busse and Clifton Green ( 2002 ) described an alone grounds about NYSE ‘s market efficiency. They noticed that when CNBC Television broadcasted an positive intelligence about a single stock, the stock monetary value reflected the intelligence within seconds and bargainers who took action to buy the stock within 15 seconds have obvious opportunity to do important net incomes ( Busse and Clifton Green, 2002 ) . Their grounds showed that to some extent the NYSE fits the semi-strong signifier efficiency.
Research Question and Aims
This research inquiry of this thesis will be to supply a systematic attack to research the relationship between public available information and stocks ‘ properties that particularly mention to monetary value and volume. As mentioned antecedently, many literatures proposed their theories and observation to explicate the construct of market efficiency. However, most of the grounds which extremely depended on few specific fiscal indexs such as past monetary values and inactive fiscal figures seems to merely back up that the most of markets are at least weak signifier efficient. It seems likely that this inquiry can non be easy answered by merely utilizing a individual theory and few peculiar instances to judge to what extent a fiscal market is semi-strong efficient. Because if one would wish to measure the semi-strong market efficiency, the job of managing the legion publically information would be the biggest challenge the research workers have to cover with.
Through utilizing information retrieval engineering, the proposed research is taking to set up a quantitative theoretical account that could measure the semi-strong market efficiency of any given fiscal markets, but due to the clip constrain we will merely take Taiwan ‘s stock exchange as the case. First, the theoretical account should assist us to judge whether a correlativity exists between the publically available information and the stock monetary value or volume. Second, if a correlativity exists, it should give us a more elaborate image about the forms between them. The most of import difference between this proposed research and other surveies is that it would non concentrate on peculiar figures as others did but analyse all related information on varied media.
It is evident that the purpose of the proposed research is to measure the hypothesis of the relationship. Lancaster ( 2005 ) stated that “ Deductive research develops theories or hypothesis and so prove out these theories or hypothesis through empirical observation. ” Using a deductive attack seems sensible. A research methodological analysis which is more nonsubjective and based on facts was the penchant for research workers to carry on this proposed research. Owing to the proposed research ‘s ontological and epistemic orientation ( Garner et al. , 2009 ) , a quantitative methodological analysis ( scheme ) with statistical and mathematical techniques should be our pick.
Harmonizing to the features of the proposed research, the longitudinal design should be the model by which research workers conducted it. Using the definition of Bryman and Bell, a longitudinal design is a research:
“ on a sample on more than one juncture, aˆ¦aˆ¦or may affect content analysis of paperss associating to different clip periodsaˆ¦aˆ¦ ” ( Bryman and Bell, 2007:71 )
The proposed research would wish to happen out whether a relationship exists between publically available information and stock information, so it needs to roll up quantitative informations from secondary informations through newspapers or other official beginnings. This longitudinal design is suited for analyzing the fluctuation and the relationship between several different properties. Under the context of this proposed research, it is appropriate for accomplishing our purposes. One of the advantages of the longitudinal design is that it collects multi-attribute informations in a big sum, so it is more likely to be accepted to transport out a research which makes a general decision in wide contexts. Combined with quantitative techniques, the advantage would be more obvious.
These undermentioned subdivisions are really critical for carry oning this research with a quantitative scheme ( Singh, 2007 ) .
Data Collecting and Handling
The informations information required by this proposed research consists of the public available information and the dealing informations of stock markets. The public available information may come from many beginnings. First, we have to put up the standards in order to filtrate out unwanted beginnings. O’Reilly III ( 1982 ) noted that a direct relationship exists between the quality and handiness of information and the determination devising public presentation. For guaranting the public presentation of analysis consequence, we have to be careful what should be included in the standards. The undermentioned standards may be utile: for quality – authorization, coverage, construction, dependence, consistence and relationship ; for handiness – handiness and stableness.
One of the major information beginnings is day-to-day intelligence. For covering the bulk of investors, we would study what intelligence beginnings are investor ‘s front-runners and take several top newspapers harmonizing to the per centum of their population coverage. Another consideration is whether contents of intelligence are stored digitally and could be accessed through Internet. Paper-based information would increase the trouble of covering with it.
Other cardinal beginnings include corporations ‘ proclamation and their fiscal information ( studies ) . To follow with the demand of authorities ‘s ordinance, any public company in Taiwan has to let go of it fiscal studies on a regular basis and denote important information every bit shortly as possible after of import events go on.
Transaction informations of Taiwan stocks could be gathered from the Taiwan Stock Exchange Corp. ‘s official site ( TWSE, 2010 ) for free. However, the TWSE does non supply real-time connexion to dealing informations. If we would wish to analyze the information on the fly, some securities companies, like Polaris ( Polaris, 2010 ) , could supply the real-time connexion service. Technically, in early phase, the stock dealing information besides could be transmitted into spreadsheet package for initial analysis.
Expect for stock dealing information, other information normally exists in an unstructured or semi-structured signifier. Paring and encoding these unstructured information are critical challenges for this phase ( Ferrucci and Lally, 2004 ) . We will work out this job by using information engineering.
All above information would be downloaded and stored into the dedicated database for farther usage.
In order to research the relationship between the publically available information and stock dealing information, at least two classs of quantitative analyses would be used ( Wisniewski, 2006 ; Bierman et al. , 1991 ) . The first class is statistics. The utile statistical techniques for the proposed research may include correlativity coefficient for correlativity, clip series for tendencies, logistic arrested development for theoretical account, and factor analysis for concentrating on critical properties ( Anderson et al. , 2008 ) . The elaborate parametric quantities and computation demand to be decided after. The 2nd class is data excavation. Sing the demand of the proposed research, classification/decision tree excavation is suited to develop a theoretical account and to foretell the possible hereafter stock result ( Apte and Weiss, 1997 ) .
The ethical issues in the proposed research seem comparatively minor. The major ground is that all needed information and information would be gathered from those beginnings that are available to public. So, it is non necessary for research workers to co-operate with any gatekeepers for entree to information beginnings ( Bruckman, 2002 ) . However, if this research advancement successfully, the findings of this research may dispute some investor ‘s perceptual experience of fiscal markets. Whether it would bring on psychological emphasis or anxiousness is still unsure at the minute ( Smock, 1955 ) .
Another ethical issue we could see is the criterion of hiting information. Although the manner how to judge the information seems non to be harmful to any person, it would still go through judgement on the public presentation of corporate entities, which are comprised of people. Some corporations may non be willing to be judged publically even though it is wholly legal. The lone thing we could make is to guarantee to hit the information in an impartial mode every bit much as possible ( Carse, 1998 ) .
So far the proposed research focuses on the appraisal of semi-strong signifier efficiency so that the merely needed information are freely obtained from public beginnings. However, in the long term, if we consider measuring the strong signifier efficiency of markets, the insider cognition would be one of the biggest ethical issues. The escapes of insider cognition are purely prohibited among most of states around the universe ( Beams et al. , 2003 ) . The trouble of garnering insider information may besides explicate the ground why few surveies have been conducted in this country.
Restrictions and Contingency Plans
First of wholly, there are a big figure of types of information beginnings available to investors, but it is improbable to roll up and analyse all these information in the period of carry oning the research. Under the state of affairs of limited resources, we would be probably to see merely major mainstream media holding big coverage in Taiwan. It may lose rightness to some extent, but maintain its feasibleness.
The 2nd restriction would be the interval between printing information and citing monetary values. If the interval of the most likely available informations is one twenty-four hours, the phenomena which could be observed may be limited to a day-to-day, larger graduated table. The information of forms of relationship within a little graduated table, one twenty-four hours, might lose. For illustration, it will be more hard to analyse the behavior of twenty-four hours bargainers, who trade their stocks in a comparatively very short period. However, this restriction should non do any obvious difference on the theoretical account. Once the real-time beginnings are available, this restriction will non be any longer.
The 3rd restriction is that because most of available information exist or are provided in an unstructured or semi-structured from, it would be a challenge to transform the information into the structured informations, which could be expeditiously dealt with through quantitative methods.
Possibly, the most of import restriction is that the consequence of analysis would be extremely affected by the single difference of perceptual experience of the information. The criterion of hiting the information depends on the subjective judgement, and in some instances, it may non be consistent for each company if we score the information manually. Sing the balance between the clip of fixing the information and the quality of the informations, we might diminish the figure of the companies chosen to be analyzed. This eventuality program would non act upon the cogency and the dependability of the consequence. The critical mission of this research should be to set up a dependable theoretical account instead than to cover all companies every bit many as possible.
Further and Future
As I mentioned in the restriction subdivision, constrained by limited resource, the research is really likely to follow a manual attack to hit the information. However, it would be helpful for practical use if some automatic marking mechanisms can be established. The farther betterment should trust on the advancement of information theory in cross-areas alternatively of pure fiscal country.
The result of this research would non merely uncover the efficiency of specific markets ( e.g. Taiwan stock market ) or specific stocks but besides develop a system that could be applied to other markets from which information would be imported. In the design stage, the architecture and model would be designed carefully for its expandability and scalability in order to be possible for other research workers to construct their ain “ information hiting algorithm ” or “ information indices ” and stop up into the theoretical account the proposed research will construct. Furthermore, after a rigorous confirmation, the theoretical account could besides be integrated into other statistics systems and fiscal programming dealing systems as indexs. It may raise the possibility of economic value in the hereafter.