## Abstraction

The intent of this research is to widen those surveies done by Roberts ( 1959 ) , Keane ( 1983 ) , Fama ( 1970, 1976, 1990 ) , Danthine ( 1977 ) , and Duffy-Deno ( 1996 ) and to look into two facets of a monetary value accommodation to new information, viz. the velocity and quality ( or way and magnitude ) of the accommodation. In this paper, a security pricing theoretical account with the consideration of monetary value alteration dissymmetry is derived foremost, based on Fama ‘s weak signifier efficiency hypothesis and martingales theoretical accounts. To accomplish research intent, hebdomadally adjusted stock monetary value informations of publically traded houses in Taiwan ‘s high engineering industries were collected from TEJ informations bank. Sample period is from 1999/10/22 to 2004/10/22. In entire, 135 stocks were analyzed. Changes in stock monetary value were used as dependent variable in multiple arrested developments, while one to five periods of slowdowns of alterations in stock monetary values were the independent variables. Each stock has all gone through more than 30 different theoretical account specifications. Empirical consequences suggest that efficient market hypothesis can non be rejected for those stocks are to a great extent traded or are major constituents of Taiwan Market Index. However, efficient market hypothesis is rejected for those stocks which are issued by secondary companies.

Keywords: Efficient market hypothesis, weak-form efficiency, monetary value dissymmetry

## Weak Form Market Efficiency Hypothesis Testing –

## A Price Asymmetry Viewpoint

## Introduction

There is a big literature that defines and theoretical accounts monetary values set in efficient markets ( Fama, 1970, 1976, 1990 ; Keane, 1983 ; Lo, 1995i?›Timmermann & A ; Granger, 2004 ) . Ultimately, the effect of the efficiency of a market is that monetary values ever to the full reflect all available information. Fama ( 1970 ) was the first bookman who defined three types of efficient markets. The first 1 is weak signifier market efficiency in which the information subset of involvement is past monetary value histories. Second, semi-strong market efficiency concerns the velocity of monetary value accommodation of other evidently publically available information. Finally, strong signifier market efficiency concerns whether any investor has monopolistic entree to any information relevant for the formation of monetary values have appeared. This survey will concentrate on the weak signifier market efficiency hypothesis testing.

Roberts ‘ ( 1959 ) paper is one of the earliest documents on weak signifier market efficiency. He found that hebdomadal alterations of the Dow Jones Index behaved really much as if they had been generated by a simple opportunity theoretical account. The most of import part of Roberts ‘ paper is to demo that the accretion of random dazes creates evident forms, similar to the behaviour of existent stock monetary values. He makes the point that changes in stock monetary values caused by the reaching of new information should be random, so the degree of stock monetary values follows a random walk. Finally, Roberts ( 1959 ) is an illustration of the random walk theoretical account for stock monetary values.

Many old surveies on weak signifier market efficiency focused merely on the way of degree of monetary value motion, nevertheless. Roberts ( 1959 ) one time criticized, fiscal theories maintain that merely the forms of the past stock monetary values need be studied, because consecutive degrees of stock monetary values can uncover an visual aspect of form or tendency. Few surveies tried to associate tallies of monetary value alterations ( Roberts 1959 ) and magnitude of the accommodation ( Keane 1983 ) into empirical hypotheses scenes. Hence, Roberts ( 1959 ) invited bookmans to analyse monetary value alterations every bit good as monetary value degrees. In add-on, he argued that multiple arrested developments would be sounder if based on independent alterations instead than dependent degrees.

The intent of this survey is to widen those surveies done by Roberts ( 1959 ) , Keane ( 1983 ) , Fama ( 1970, 1976, 1990 ) , Danthine ( 1977 ) , and Duffy-Deno ( 1996 ) and to look into two facets of a monetary value accommodation to new information, viz. the velocity and quality ( or way and magnitude ) of the accommodation. Harmonizing to Fama ‘s weak signifier efficiency hypothesis and martingales theoretical accounts, a security monetary value alteration theoretical account with the consideration of monetary value alteration dissymmetry is derived. Empirical multiple arrested developments are besides established, in which the asymmetric motions of past stock monetary values alterations in the information set are used as independent variables. To accomplish research intent, hebdomadally adjusted stock monetary value informations of publically traded houses in Taiwan ‘s hi-tech industries were collected from TEJ informations bank. Sample period is from 1999/10/22 to 2004/10/22. Empirical findings of this survey would better our apprehension of efficient market literature. Hence, this survey has both theoretical and empirical parts to this subject.

## Related Literature

The history of security monetary values has likely been widely and intensively studied. The followerss will, nevertheless, concentrate on research related to this survey merely. For illustration, Cowles & A ; Jones ( 1937 ) used the amount of waies of consecutive alterations of returns to prove whether stock monetary values move indiscriminately. They defined an index, I, to reflect stock ‘s up or down motions. That is,

It = { 1, if rt = pt – pt-1 & gt ; 0,

{ 0, if rt = pt – pt-1a‰¦0,

where r1, r2, aˆ¦ , rn+1 are known. And specify

Yt= It It+1 + ( 1 – It ) ( 1 – It+1 ) .

Then, amount of Yt peers 1 is the entire figure of monetary value moves up during two clip periods. Similarly, amount of Yt peers 0 represents the entire figure of monetary value moves different waies during two periods. Cowles & A ; Jones ( 1937 ) showed, if a stock moves indiscriminately, so ( pt – pt-1 ) will be a random variable with an independent and indistinguishable distribution. Research workers could utilize the ratio of the entire Numberss of up motions and down motions to prove whether stock monetary value moves indiscriminately. Cowles & A ; Jones ( 1937 ) showed that the above ratio would follow a normal distribution with average 1 and discrepancy 12/n.

Mood ‘ ( 1940 ) attack somewhat differs from that of Cowles & A ; Jones ( 1937 ) . Temper used the size of tallies to bespeak the entropy of a stock monetary value. When the size of tallies ( kn ) is either excessively little or excessively large, a stock monetary value would non hold entropy. Temper has showed that if a monetary value moved either up or down with the same chance ( 0.5 ) , so one could acquire a statistic tn= ( 2kn+1-n ) / ( a?sn ) which follows a standard normal distribution. Therefore, one could utilize Mood ‘s Tennessee statistic to prove the entropy of stock monetary values.

Roberts ‘ ( 1959 ) paper is one of the earliest documents on weak signifier market efficiency. He cited a research done by Kendall ( 1953 ) to do his point, “ hebdomadal alterations of the Dow Jones Index behaved really much as if they had been generated by a opportunity theoretical account ( say a distribution with average 0 and variance 5 ) . He besides criticized fiscal theories focused merely on the forms of the past stock monetary values, because consecutive degrees of stock monetary values can uncover an visual aspect of form or tendency. He makes the point that changes in stock monetary values caused by reaching of new information should be random, so the degree of stock monetary values follows a random walk. In other words, the most of import part of Roberts ‘ paper is to demo that the accretion of random dazes creates evident forms, similar to the behaviour of existent stock monetary values. Roberts besides invited bookmans to analyse monetary value alteration every bit good as monetary value degree motions. In add-on, he argued that multiple arrested developments would be sounder if based on independent alterations instead than dependent degrees. This survey will response to Roberts ‘s call and reference a inquiry which he one time posed: How good do monetary value alteration patterns use to single securities?

The construct of market efficiency is closely related to the procedure of monetary value formation. In order to do the construct testable, Fama ( 1970, 1976 ) used expected return theory to exemplify what he meant about market efficiency. What expected return theory says is that

( 1 ) Tocopherol [ Pt+1a??I©t ] = [ 1 + E ( R t+1a??I©t ) ] * Pt,

where Tocopherol is the outlook operator ; Pt is the monetary value of security at clip T ; Pt+1 ( a random variable ) is its monetary value at t+1 ; I©t is a general symbol for whatever set of information is assumed to be to the full reflected in the monetary value at clip T ; R ( a random variable every bit good ) is the proper price reduction rate. Both random walk and martingale theoretical accounts are two particular instances of expected return theory.

Fama ( 1970 ) was the first bookman who defined three types of efficient markets. The major difference among three types of efficient markets is the contents in the information set. The first 1 is weak signifier market efficiency in which the information subset of involvement is past monetary value histories. To prove whether weak signifier market efficiency hypothesis holds, research workers can carry on consecutive correlativity trials, runs trial, or distribution trials for an single security. One could besides follow CAPM statements to prove multiple security expected return theoretical account ( Fama, 1970 ) . Or, with full usage of past monetary value informations every bit good, one can replace random walk theoretical accounts with the more appealing martingale theoretical accounts ( Sharpe, 1970, Danthine, 1977 ) .

Martingales are one of the cardinal tools in the modern theory of finance ( Neftci, 2000, ch. 6 ) . A procedure { St, tIµ [ 0, a?z ] } is a martingale with regard to the household of information set I©t, and with regard to the probabilityI? , if for all T & gt ; 0,

a. St is known, given I©t.

b.Unconditional prognosis are finite: Ea??Sta?? & lt ; a?z .

c. And if E [ STa??I©t ] = St, for all T & lt ; T, with chance 1. That is, the best prognosis of unobservable hereafter values is the last observation on St.

Harmonizing to the definition, E [ St+u- Sta??I©t ] = 0, u & gt ; 0. This means that martingales are random variables whose future fluctuations are wholly unpredictable given the current information set. The best prognosis of the alterations in random variable over an arbitrary interval is zero. In other words, the waies of the hereafter motions in martingales are impossible to calculate. These belongingss of martingales are really of import and will be used to prove weak signifier market efficiency hypothesis.

Danthine ‘s ( 1977 ) theoretical account is an application of ( sub- ) martingale. He wanted to cognize ( 1 ) Under which conditions is it executable to depict equilibrium in efficient topographic point trade good markets in footings of expected returns? ( 2 ) Is the expected return premise necessary to prove market efficiency and what option can be proposed? Following Fama ( 1970 ) , Danthine conducted a trial of efficiency look intoing the impossibleness of assorted trading system. Based upon the averment that the nothing expected net net income regulation translates itself into a relationship between any monetary value and the outlook of its following period realisation that can be described as E [ e-rt Pta??I©t-1 ] =Pt-1, a conditional outlook on the information available in t-1, the discounted present value of Pt is the last observation on Pt. Making expressed the conditional nature of the outlook and the information set, E [ Pta??I©t-1, Pt-1, Pt-2, aˆ¦ ] a‰§Pt-1. The utility of this preparation is to transform monetary value motions into a sub-martingale and a testable arrested development format

E [ Pta??I©t-1, Pt-1, Pt-2, aˆ¦ ] =I±+ Pt-1,

whereI±is a changeless and the covariance ( Pt+I„- Pt+I„+1, Pt – Pt-1 ) =0, for anyI„a‰ 0. Danthine showed that if a stock monetary value exists a pure martingale belongings, so E [ Pt+1- Pta??I©t ] =0, and cov ( Pt+1 – Platinum, Pt – Pt-1 ) =0.

It would be an interesting research by uniting Cowles & A ; Jones ‘ ( 1937 ) , Mood ‘s ( 1940 ) and Roberts ‘ ( 1956 ) statements about the different waies of stock monetary value motions, and Danthine ‘s ( 1977 ) martingale theoretical account to prove weak signifier market efficiency hypothesis for single securities. Before demoing how to organize this trial, one more piece of literature demands to be brought up. Duffy-Deno ( 1996 ) was interested in how the velocity and magnitude of oil monetary values transmit from sweeping degree to retail degree. He foremost separated sweeping oil monetary value motions into two different forms. One form shows the sweeping oil monetary value goes up if a-?Pupt-1= Pt-Pt-1 if Pt-Pt-1 & gt ; 0. The other form shows sweeping oil monetary value goes down if a-?Pdownt-1=Pt-Pt-1, if Pt-Pt-1 & lt ; 0. Then, regress alterations of retail monetary value on slowdown of sweeping up monetary values and down monetary values. In this attack, Duffy-Deno ( 1996 ) could look into how retail oil monetary value responded otherwise to two different sweeping monetary value alterations and whether there are symmetric effects between up and down motions. Similar attack will be applied to look into how single security monetary value would alter otherwise.

## Model Puting

Harmonizing to Fama ( 1970 ) , efficient market would ensue from an ideal universe in which ( a ) there are no dealing costs, ( B ) all relevant information is costlessly available to all market participants, and ( degree Celsius ) all agree on the deductions of current information for the current monetary value and the distributions of future monetary values. As a consequence of ( a ) and ( B ) , it is claimed that all information potentially taking to expected net incomes is exploited up to the point where these net incomes do non transcend the ‘normal ‘ returns level available in other topographic points ( Danthine, 1977 ) . But though minutess costs, information that is non freely available to all investors, and dissension among investors about the deductions of given information are non necessary beginnings of market inefficiency, they are possible beginnings. Measuring these effects on the procedure of monetary value formation is the major end of empirical survey on market efficiency ( Fama, 1970 ) .

Testing efficient market hypothesis is an empirical issue. As Fama ( 1970 ) has pointed out, the martingale theoretical account is based on two hypotheses: ( a ) the efficient use of information, and ( B ) the possibility of showing market equilibrium in footings of expected returns. Following Fama ‘s efficient market hypothesis ( 1970 ; 1976 ) , a security monetary value at clip T can be expressed as combining weight. ( 1 ) if information set at clip T can to the full reflect past information,

( 1 ) Tocopherol [ Pt+1a??I©t ] = [ 1 + E ( R t+1a??I©t ) ] * Pt,

where Tocopherol is the outlook operator, I©t is the information set at clip T, and R is the proper price reduction rate.

As discussed in Danthine ( 1977 ) , LeRoy ( 1976, 1989 ) , Huang ( 1985 ) and Neftci ( 2000 ) , if a security market can be equilibrium and for certain be a just game, so equation [ 2 ] holds,

( 2 ) Tocopherol [ Pt+1a??I©t ] = Pt.

The best prognosis of unseen hereafter values is the last observation on Pt. In other words, the best prognosis of the alterations in random variable over an arbitrary interval is zero,

( 3 ) Tocopherol [ Pt+1- Pta??I©t ] =0.

The waies of the hereafter motions in martingales are impossible to calculate.

If eq. ( 2 ) right describes the equilibrium status for Pt, it is possible to obtain a testable deduction of efficiency in footings of the ascertained security past monetary value series. That is, when Pt is a pure martingale, combining weight. ( 2 ) or ( 3 ) can be rewritten as ( Danthine 1977 ) ,

( 4 ) Tocopherol [ Pt+1 – Pta??Pt – Pt-1 ] =0.

and

( 5 ) cov ( Pt+1 – Platinum, Pt – Pt-1 ) =a?« ( Pt – Pt-1 ) E ( Pt+1 – Pta??Pt – Pt-1 ) I? ( Pt – Pt-1 ) vitamin D ( Pt – Pt-1 ) =0,

provided a stationary distributionI?exists. Therefore, it is possible that one can look at a simple accommodation of the monetary value series for which pure martingale status holds and prove the zero consecutive correlativity for the sequence of these monetary value alterations. One possible look of past monetary value difference is as combining weight. ( 6 ) ,

( 6 ) Tocopherol [ Pt+1 – Pta??I©t ] = E [ I”Pt+1a?? ( Pt -1- Pt-2 ) , ( Pt-2 – Pt-3 ) , aˆ¦ , ( Pt-r+1 – Pt-r ) ]

= E [ I”Pt+1a??I”Pt-1, I”Pt-2, I”Pt-3, aˆ¦ , I”Pt-r+1 ] .

This survey extends the tallies trial suggested by Hagerman & A ; Richmond ( 1973 ) and Duffy-Deno ‘s ( 1996 ) attack to divide single security monetary value motions into two different forms. That is, we define two dummy variables. One is a-?Pupt-1= Pt-Pt-1 if Pt-Pt-1 & gt ; 0, otherwise a-?Pupt-1=0. The other silent person variable is a-?Pdownt-1=Pt-Pt-1, if Pt-Pt-1 & lt ; 0, otherwise a-?Pdnt-1=0. Therefore, combining weight. ( 6 ) now becomes combining weight. ( 7 ) in a multiple arrested development format

( 7 )

where I±i are arrested development coefficients to be estimated, R and s are the slowdowns of up and down periods, severally. AndIµt is the arrested development mistake term and its distribution is assumed to follow ( 5 ) . It is non necessary to requireIµt to follow an independent and indistinguishable distribution. Nor it is required to be a normal distribution.

It is now possible to setup hypothesis proving for weak signifier market efficiency. Under the premises of ( a ) there are no dealing costs, and ( B ) all relevant information is costlessly available to all market participants, all past monetary value motions information potentially taking to expected net incomes is exploited up to the point where these net incomes do non transcend the normal returns degree available in other topographic points ( Danthine, 1977 ) . It is expected that, under market efficiency hypothesis, allI±i are to be zeros. If non, the underlying stock monetary value is non informationally efficient. For those single stocks are non informationally efficient, farther hypotheses are possible to carry on. Following Duffy-Deno ( 1996 ) , there are three types of monetary value dissymmetry can be tested. First, one can prove whether the figure of up motion R equals the figure of down motion s. If R is non peers to s, so there exists a timing dissymmetry. Second, one can prove magnitude dissymmetry. That is, one tests whether I±1, i=I±2, J or non. Third type of monetary value dissymmetry trial is called pattern dissymmetry. One can prove whether I±1, i=I±2, J when R peers s. The hypothesis of weak signifier market efficiency or pure martingale will be tested with hebdomadal monetary value motion informations.

## Empirical Results and Discussion

This survey uses single securities of publically traded houses in Taiwan ‘s high engineering industries. There are several grounds to make so. First, day-to-day trading volume of these hi-tech stocks histories for more than 80 % of the whole Taiwan capital market. The frequence of turnover is high. Information of past monetary values and volumes are really freely available to all investors. Second, these hi-tech stocks cover several sub-industries, such as semi-conductor fabrication, printed circuit board ( PCB ) fabrication, inactive devices fabrication, computing machine and information fabrication, optical industry and web industry. Firms of up watercourse and down watercourse make a whole and complete representation of a specific industry. Besides in each industry, there are primary prima houses and secondary houses. It is possible to compare the monetary value motions of these houses and state how they change in different formats. Third, the Numberss of hi-tech companies are turning. In any specific industry, there are mature houses and immature houses. So, it is possible to compare the monetary value motions of these immature and ‘old ‘ companies. Finally, merchandising histories of most houses are long plenty to carry on clip series survey.

Samples used in this survey were collected from TEJ informations bank. Sample period is from 1999/10/22 to 2004/10/22. They are hebdomadal adjusted monetary values. In this 5-year period, each stock has more than 250 informations points. In entire, 135 stocks with complete trading history were chosen and analyzed.

The first undertaking demands to be done is to make up one’s mind the Numberss of slowdowns in up and down variables in equation ( 7 ) . Initially, randomly long slowdown for both up and down variables were included ( Duffy-Deno, 1996 ) . Following Conrad & A ; Kaul ( 1988 ) , upper limit of 5 slowdowns for both up and down were chosen. The consequences of this process were verified by ab initio stipulating a one hebdomad slowdown and so adding slowdowns if singly or jointly important. In entire, there are more than 30 different theoretical account specifications waited to be tested. Then working backwards, t- and F-tests were used to corroborate if any of the slowdowns separately or as a group could be eliminated as statistically undistinguished. A 5 % degree of significance was employed in all trials.

A sum-up of multiple arrested development consequences of equation ( 7 ) is listed in Table 1. There are 71 stocks do non reject weak signifier market efficiency hypothesis, while the staying 64 stocks reject the void hypothesis. Since single stocks cover all six major industries, it is difficult to pull a clear line to reason which industry is more likely to back up market efficiency hypothesis. However, if we compare and match houses in each industry, it is possible to see that primary prima houses are more likely to back up the weak signifier market efficiency hypothesis than secondary houses. For illustration, the monetary value alterations of two worldwide recognized foundry houses, UMC ( 2303 ) and TSMC ( 2330 ) , follow what weak signifier market efficiency hypothesis is expected ( see Appendix ) . This should be closely related to their information which is freely available to all investors and understanding among investors about the deductions of given information. Another illustration comes from motherboard sub-industry. Elitegroup ( 2331 ) , Gigabyte ( 2376 ) and MSI ( 2377 ) are the three major participants in this sub-industry and their stock monetary value motions back up weak signifier market efficiency hypothesis. While USI ( 2350 ) , BIOSTAR ( 2399 ) , Shuttle ( 2405 ) , and Chaintech ( 2425 ) are the secondary houses in the same sub-industry. Their stock monetary value motions reject weak signifier market efficiency hypothesis. In computing machine industry, Compal ( 2324 ) , INVENTEC ( 2356 ) , ASUSTEK ( 2357 ) , and QCT ( 2382 ) are the primary houses and their stock monetary value motions support market efficiency hypothesis. Stock monetary value alterations of secondary houses such as Twinhead ( 2364 ) , Arima ( 2381 ) , and Innovace ( 2387 ) do non back up market efficiency hypothesis. More illustrations can be pulled out from other sub-industries such as CRT ( Cathode Ray Tub ) , PCB fabrication and web industry ( see Appendix ) . Since primary houses are normally the ‘old ‘ houses, hence, stock monetary value alterations of ‘old ‘ houses are more likely to back up weak signifier market efficiency hypothesis than immature houses in the same sub-industry.

Table 1 Summary of market efficiency hypothesis testing- figure of houses

Not reject Ho

Reject Ho

Symmetrical houses

High-tech industry

Ob.

Ob.

Code and form ( up, down )

Semi-conductor fabrication

15

8

2388 ( 5,5 ) , 3041 ( 5,5 ) , 2351 ( 4,4 ) , 2434 ( 5,5 ) , 2470 ( 5,5 )

PCB fabrication

8

5

2467 ( 5,5 )

Passive devices fabrication

2

6

2492 ( 5,5 )

Computer and information

27

26

2487 ( 5,5 ) , 2399 ( 4,4 ) , 2378 ( 3,3 ) , 2414 ( 5,5 ) , 2438 ( 5,5 ) , 2476 ( 5,5 )

Optical industry

15

13

Network industry

4

6

Entire Obs=

71

64

13

Although it is by and large true that primary and/or ‘old ‘ houses are more likely to demo weak signifier market efficiency than secondary houses do in the same sub-industry. All seven package companies such as Systex ( 2343 ) , Eten ( 2432 ) , Ares ( 2471 ) , and Ulead ( 2487 ) reject the market efficiency hypothesis. Potential beginnings of market inefficiency may be due to minutess costs, information that is non freely available to all investors and dissension among investors about the deductions of given information.

Table 2 lists a sum-up of consequences of symmetricalness trials for 13 stocks which exhibit possible symmetric forms. In other words, these 13 stocks show timing symmetricalness, while other 51 stocks do non. The hypothesis of clocking symmetricalness is rejected for most instances.

Table 2 Summary of symmetricalness Trials

## Code

Magnitude symmetricalness trials

Pattern symmetricalness trials

I?I±up-i=0

I?I±up-i=1

I?I±dn-i=0

I?I±dn-i=1

I?I±up-i=I?I±dn-i

I±up-1=I±dn-1

I±up-2=I±dn-2

I±up-3=I±dn-3

I±up-4=I±dn-4

I±up-5=I±dn-5

2388

## –

## ***

## –

## ***

## –

## **

## –

## –

## –

## ***

3041

## –

## ***

## **

## ***

## –

## –

## –

## *

## –

## **

2351

## *

## ***

## –

## ***

## *

## –

## –

## –

## ***

2434

## –

## ***

## –

## ***

## –

## *

## –

## *

## **

## ***

2470

## –

## ***

## –

## ***

## –

## –

## ***

## –

## ***

## **

2467

## **

## ***

## –

## ***

## *

## –

## *

## –

## ***

## **

2492

## –

## ***

## **

## ***

## –

## ***

## –

## –

## –

## ***

2487

## –

## ***

## **

## ***

## –

## –

## **

## –

## –

## ***

2399

## –

## ***

## –

## ***

## –

## –

## –

## *

## ***

2378

## –

## ***

## –

## ***

## –

## **

## –

## **

2414

## –

## ***

## *

## ***

## –

## –

## –

## –

## ***

## **

2438

## –

## ***

## –

## ***

## *

## –

## –

## –

## –

## **

2476

## –

## ***

## –

## ***

## –

## –

## **

## –

## –

## ***

Note: * , ** and *** indicate hypotheses were rejected at 10 % , 5 % and 1 % degree, severally.

Following Duffy-Deno ( 1996 ) , two more asymmetric trials were conducted for those 13 stocks with timing symmetricalness. As seen in Table 2, most stocks can non reject the hypothesis of zero magnitude symmetricalness. But all 13 stocks reject the hypothesis of to the full adjustment when monetary values up ( I?I±up-i=1 ) . And the hypothesis of to the full adjustment when monetary values down ( I?I±dn-i=1 ) is rejected every bit good. However, the hypothesis of equal magnitude consequence of up and down can non be rejected in most 13 instances. Based on the consequences of asymmetric trials, it is possible that these 13 stocks with timing symmetricalness could stop up with zero effects on monetary value alterations.

Pattern symmetricalness hypothesis is rejected in about half of the 13 stocks. Particularly when the lagged periods are longer, say slowdown 4 and 5.

## Decisions

There is a big literature that defines and theoretical accounts monetary values set in efficient markets. In this paper, merely weak signifier market efficiency hypothesis was tested. The intent of this survey is to widen research done by Roberts ( 1959 ) , Keane ( 1983 ) , Fama ( 1970, 1976, 1990 ) , Danthine ( 1977 ) , and Duffy-Deno ( 1996 ) and to look into the velocity and quality ( or way and magnitude ) accommodations of stock monetary value to new information.

Harmonizing to Fama ‘s weak signifier efficiency hypothesis and martingales theoretical accounts, a security monetary value theoretical account with the consideration of monetary value alteration dissymmetry is derived foremost. Based on Danthine ‘s ( 1977 ) statement about pure martingale and just game construct of Fama ( 1970 ) , an empirical arrested development scene is proposed and so tested. Datas used in arrested development analysis were hebdomadal informations taken from TEJ informations bank. Sample period is from 1999/10/22 to 2004/10/22. Empirical consequences from publically traded companies in Taiwan ‘s hi-tech industries during sample period suggest that weak signifier efficient market hypothesis can non be rejected by more than half of 135 sample houses, but rejected by the other half sample houses.

A major empirical determination of this survey is that if we compare and match houses in each sub-industry, we can see primary prima houses be more likely to back up the weak signifier market efficiency hypothesis than secondary houses. Since primary prima houses are normally the ‘old ‘ houses, hence, stock monetary value alterations of those ‘old ‘ houses are more likely to back up weak signifier market efficiency hypothesis than immature houses in the same sub-industry.