Printer Friendly

BEHAVIOR OF ISLAMIC AND CONVENTIONAL HOLIDAY-EFFECT AND ADAPTIVE MARKET HYPOTHESIS: A FIRM LEVEL EVIDENCE FROM EMERGING MARKET OF ASIA.

Byline: Muhammad Naeem Shahid, Dr. Abdul Sattar, Dr. Faisal Aftab and Sumaira Aslam

Keywords: Adaptive Market Hypothesis; Efficient Market Hypothesis; Holiday Effect; Islamic Holiday Effect; Pakistan Stock Exchange.

INTRODUCTION

Through investigating the varying degree of well-known holiday effect, the study attempts to add literature on Adaptive Market Hypothesis (AMH) proposed by Lo (2004). Market conditions and the way the market participants incorporate novel information into the prices of equity impact investors' psychology in the market which in turn may change the behavior of holiday effect over time. Thus, the study is intended to investigate the time-varying behavior of holiday effect in the context of Adaptive Market Hypothesis (AMH). The holiday effect can be defined as the effect due to which the average returns become higher and statistically significant on the trading day immediately prior to the off-trading-days (holidays). These holidays are apart from Saturday and Sunday, means other public holidays on which the stock exchanges remain closed.

Thus, the study considers the holidays in the Islamic years which predominantly include: 9th and 10th of Muharram (Ashura); 12th Rabi-ul-Awwal (Eid-Milad-un-Nabi), 1st Shawwal and 10th Zilhaj (Eid-ul-Fitr and Eid-ul-Adha respectively); while holidays in the Gregorian calendar include 5th of February (Kashmir-day), 23rd March (Pakistan-day), 1st May (Labor-day), 14th August (Independence-day) and 25th December (Quaid-e-Azam-day). Considering the importance of market efficiency is imperative to understand the working of the stock market. Poshakwale (1996), asserted that the efficiency of the emerging markets assume greater importance as the trend of investments is accelerating in these markets, as a result of regulatory reforms and removal of other barriers for the international equity investments. The notion of EMH explores that if the market is weak in efficiency then stock prices must be independent of each other and returns will be unpredictable (Fama, 1970).

Additionally, Fama (1970), also classified the market efficiency into three different categories, each category is characterized in terms of different forms of information as; (i) weak form efficiency, which defines that equity prices fully reflect all available information about the historical trading; (ii) semi- strong form efficiency, which delineates that the publicly available information is fully reflected by the equity prices; and (iii) the strong form efficiency which proclaims that the equity prices fully reflect possible relevant information along with inside information of the company. Accordingly, all these three types portray that all possible available information is reflected by the equity prices, thus, any forecasting about future price changes is not possible.

Therefore, both the fundamental analysis1 (predicting equity/security prices on the basis of economic variables), and the technical analysis2 (predicting equity/security prices on the basis of historical trading and performance of eq uity/securities) are useless and would not be beneficial for the market participants to gain abnormal returns3. In the literature, all three sorts of the efficient market hypothesis (EMH) have captured great attention but the weak form of EMH is widely studied and also is the primary focus of this study. Against the proposition of EMH, if the prices of stocks are predictable and not independent, the investors can gain abnormal returns by using the historical information of the past trading trend.

Recent literature contradicts EMH preposition, as many studies (Shahid and Mehmood, 2015; Hashmi, 2014; Halari, 2013), expound the stock returns to have a dependent nature and substantiate that there exist some profitable investment opportunities in the markets, thus, market anomalies do exist in Pakistan. Grossman and Stiglitz (1980), expressed that it is impossible for a capital market to be perfectly efficient as investors otherwise would have no benefit to acquire costly information if the markets were not inefficient and the profit-making opportunities were available. Keeping in view the argument of Grossman and Stiglitz (1980), of "impossibility of perfectly efficient market", Campbell, Lo, and MacKinlay(1997), offered the notion of "relative efficiency" rather than the "perfect efficiency" which leads a swing from measuring efficiency of market from an all-or-nothing condition to test it over the period of time (Shahid and Sattar, 2017).

Recent studies (Rehman and Rizwan, 2014; Haque, Liu, and Nisa, 2011), provide the evidence that the stock markets in Pakistan are inefficient while some studies show that Pakistani equity markets are effectual as well. Nisar and Hanif (2012), found that the monthly return data identifies PSX as weak form efficient, similarly, Rabbani, Kamal, and Salim (2013), suggest that PSX was weak form efficient in sub-period 1999-2001 and 2005-2007, while Riaz, Hassan, and Nadim (2012), identified that the efficiency of market changes with the application of different tests which means that market efficiency may change from time to time. Thus, a contradiction exists about efficiency and inefficiency of the markets. Therefore, it is essential to explore the stock market efficiency through AMH (Adaptive Market Hypothesis) which states that efficiency (return predictability) changes over time.

To incorporate the varying degree of return predictability, Lo (2004), proposed a new model "Adaptive Market Hypothesis (AMH)" that facilitates market anomalies to co-exist with market efficiency and enables market efficiency to evolve over time. Moreover, the AMH proclaims that the market efficiency is not a guaranteed outcome as to gain abnormal profit, the arbitraging opportunities also arise from time to time. Hence, Lo (2004), characterized the six attributes of AMH as; i) investors perform in favor of their self-benefits to protect their own self-interest; ii) investors make wrong judgments and make mistakes; iii) investors pick up learning from their mistakes and adapt them to their behavior which is not explored by EMH; iv) rivalry energizes adaptation as well as innovation; v) market ecology is shaped by natural selection; vi) evolution determine the dynamics of the market.

Susequently, Shahid and Sattar (2017); and Urquhart (2013), argue that the earlier studies apparently clarify the efficiency and inefficiency of the market over a pre-determined time, while market conditions may change from time to time causing changes in efficiency, which is consistent with AMH. Currently, AMH is receiving great attention, thus, this study aims to explore if AMH is the better elucidation of behavior of holiday effect than traditional EMH at firm level in Pakistan. The findings of this study will be useful for individual investors and security organizations for accurate forecasting and a better understanding of the market. To conduct the study, individual firms listed in Pakistan Stock Exchange have been selected which were list during the time period of January 1996 to December 2015, using subsamples of five years of fixed length, to inspect the behavior of the holiday effect.

Investigation of the varying behavior of holiday effect is conceded by sub-sample analysis across the time period of the study. However, the choice/selection of sub-samples as well as the range of their size is of subjective nature (Shahid and Sattar, 2017; Urquhart and Hudson, 2013). Thus, the data set is split into four sub-samples of 5-years, of equal length to investigate how the holiday effect has behaved over time. Sub-samples consist of enough observations to produce reliable results which enable comprehensive analysis of the varying degree of the holiday effect. Along these lines, an attempt to enhance the literature on AMH is undertaken by fulfilling the missing link of varying degrees of holiday effect through AMH in multifarious ways. Firstly, this study is the first attempt to investigate the Islamic and conventional holidays' effect anomaly with AMH, which alters the behavior of returns during the holidays, over time.

Secondly, this is the first study which investigates the performance of holiday effect at the firm level under the umbrella of AMH. Finally, the paper examines the behavior of the holiday effect with the application of a GARCH (1,1) regression model which facilitates the time-varying nature of volatility in equity returns. On the other hand, to handle the non-normal nature of stock returns data, the Kruskal-Wallis test statistic is used. The rest of the paper is organized as follows; the subsequent segments offer the relevant review of holiday effect literature; the data and methodology used to conduct the study; empirical results and summaries; findings and conclusions respectively.

LITERATURE REVIEW

The Holiday effect anomaly have been rigorously tested in previous studies. Fields (1934)4 first documented the holiday effect and found that "the stock returns on trading days before the religious and secular closed- market holidays, are significantly higher than returns on other trading days". Seminal studies of Lakonishok and Smidt (1988); and Ariel (1990), report significantly higher returns on pre-holidays as compared to post- holiday returns. Furthermore, they found abnormal returns not only on weekend closing but for other gaps in trading. Ariel (1990), found an eight-time greater return on pre-holidays than post-holiday returns. He further proved that the eight holidays per year account for 38% of the total annual rate of returns. Also, Lakonishok and Smidt (1988) reported that the pre-holiday returns occupied 30% to 50% of the total return of US equity markets before the year 1987. Agrawal and Tandon (1994), found the pre-holiday effect in seventeen markets.

Similarly, Kim and Park (1994); Brockman and Michayluk (1998), investigated AMEX and NASDAQ over the period of 1963-1987 and 1987-1993 respectively and found holiday effect's impact on the market. Boyle et al. (2002), analyzed the New Zealand stock market. They selected five economically different events which had an impact on the emotions and moods of the investors (as claimed by psychology researchers). They found that pre-holiday returns are statistically different from other days (i.e. non-events). Similarly, Chong et al. (2005), noticed the pre-holiday effect in the UK, US and Hong Kong markets which are considered as the most important markets of the world. They construed that the average expected returns before specific holidays were significantly greater than the average expected returns before other holidays. The same effect of holidays was discovered in the Kuwait Stock Exchange from the period of 1984 to 2000 (Al-Loughani, Al-Saad, and Ali, 2005).

Picou (2006), studied the stock return behavior in stock exchanges of six countries including the All Ordinaries Index from Australia, Index of TSE from Canada, HIS-Hang Seng Index from Hong Kong, Nikkei-225 from Japan, Financial Times Stock Exchange -FTSE from the UK, and SandP-500 from the US. By calculating the daily return for ten years (1989- 1999), he found ex-post-holiday anomaly in all the exchanges, this was because the investors sell more before the holiday to avoid the risk after the holiday. Wong, Agarwal, and Wong (2006), examined the Singapore Stock Exchange to investigate the holiday effect. They divided the sample into two periods; pre-crisis period and post-crisis period and found that the preholiday return was higher than the other trading days in the pre- crisis period, but the trend was inverse in post-crisis periods. Marrett and Worthington (2007), examined the holiday effect in Australian Stock Exchange for the period of 1996 to 2006.

They selected eight annual holidays that were ANZAC day, Australia day, Boxing day, Easter Friday and Monday, new-year days, Queen's birthday, and Christmas day, and confirmed the pre-holiday effect. Cao et al. (2009), estimated the holiday effect in the stock market of New Zealand. To test the variance, pre- holiday returns were considered along with the non-preholiday returns. For the purpose, they took data for the period of 1967 to 2006 of NZSE40 and NZSE50 indices. The results of this study illustrated significant positive returns before holidays in New Zealand. Zafar et al. (2012), examined the half-month effect as well as holiday- effect at Pakistan Stock Exchange (PSX) over the period of 1991-2007. They calculated the daily logarithmic returns from the KSE-100 index to test these calendar effects. They concluded the Pakistan Stock Exchange as an inefficient market by elaborating that the pre-holiday has significant positive returns than post-holidays.

They further argued that the investors in the market react very certainly and take more part in trading activities before holidays, thus gains in the time prior to holidays is significantly greater than gains after holidays. By using ARMA (1,1) model as well as GARCH (1,1) model over the period of 1999-2012 Yuan and Gupta (2014), presented a robust evidence of positive CLNY-pre-holiday effect in almost all major indices of Asia5 except for Malaysia, where the post- CLNY effect was greater, significant, and positive than the pre-CLNY6 effect. Huang (2017), investigated the Chinese stock market to examine the holiday effect returns over the period of 2006 to 2017. With the application of GARCH (1,1) and GARCH (1,1)-M models, the study found evidence of holiday effect in Chinese stock market. Moreover, Shahid and Sattar (2017), investigated the Pakistan Stock Exchange over the period of 1992 to 2015 and found that the holiday effect fluctuates over time and is consistent with AMH.

Hassan and Sarker (2018), investigated the Dhaka Stock Exchange to examine pre-and post-holiday returns over the period of 2013 to 2017. With the application of Wilcoxon-signed rank test, they found significantly higher returns in pre-holidays than post-holidays. The literature suggests the prevalence of holiday effect in different stock markets, but a limited number of studies have investigated the varying degree of holiday effect through AMH. Thus, adding more literature on the subject will help to have a comprehensive view of the behavior of the holiday effect in different markets.

DATA COLLECTION AND RESEARCH METHODOLOGY

To observe the presence of holiday effect and how this effect has influenced over time, we investigated the daily-returns of companies listed in the Pakistan Stock Exchange. There were 560 companies listed on PSX in December 2015. Out of the 560 companies only 540 had data available on the data stream database. Thus, the daily share price data was downloaded for all 540 firms. In order to explain the adaptive nature of the behavior of the holiday effect, a large substantial time frame is required for the study to investigate the individual companies. Thus, a sample of 20 years' data from January 1996 to December 2015 was selected. Furthermore, a sample of 107 companies7 was selected out of 540 companies which had the data available from January 1996 to December 2015. To investigate the varying degree of the behavior of the holiday effect, data of individual firms are more appropriate than using national indices.

Thus, the analysis provides a more accurate sign of whether equity returns are foreseeable for investors on holidays and whether this effect has cyclic nature of efficiency. The following regression equation was estimated:

Rt = c+ [beta]Dt+ Iut, t = 1,...,T

Where Rt represents the stock index return, Dt represents an indicator of holiday effect as adopted by (Urquhart and McGroarty, 2014; Shahid and Sattar, 2017), while Iut is the error term. Instead of using OLS regression, we use GARCH (p, q) model proposed by, to investigate the existence of the holiday effect in Pakistan stock exchange. Across our analysis, we employ GARCH (1, 1) regression model because GARCH (1, 1) model is the most robust and simplest model of the family of volatility models as well as it is most widely used in the literature. Whereas the GARCH (1, 1) model allow researchers to model variance as conditional on the past variance and error, rather than fixed through the series (Urquhart and McGroarty, 2014). Therefore, to capture the time-varying behavior of return of individual firms, we run the following GARCH (1, 1) regression:

ht = [alpha]0 + [alpha]1 Iu2t-1 + I,ht-1

Where, for equity returns at time t, ht is the conditional variance, ht-1 represents the conditional variance of equity returns at time t-1 while [alpha]0, [alpha]1 and I, are the coefficients of the GARCH model. The GARCH model is an appropriate model and possesses the potential ability to capture the desirable features of equity market returns but it is not appropriate to use to capture the non-normality feature of returns series. Therefore, we also employ a non-parametric Kruskal-Wallis (K.W) test to examine predominant sensitivity of the population to the difference in mean and whether the population has identical distributions from which the samples are drawn. Thus, we investigate the mean differences in the stock returns on holidays and non- holidays:

(Equation)

Where represents the total number of observations, denotes the number of groups, and indicate the total number of observations and the average rank of observations in the group respectively. Therefore, to investigate how exactly holiday effect has behaved/performed through time we employ the Kruskal-Wallis test and GARCH regression model to the full-sample as well as to subsamples of fixed length. We split our data into sub-samples of 5 years, thus generate 4 subsamples of identical lengths. A sub-sample of 5-years holds a sufficient set of observations to offer reliable and sufficient results for investigating the behavior of holiday effect and observe how this anomalous effect has behaved/performed through time. We employ the empirical tests discussed above on the returns of 107 companies listed at Pakistan Stock Exchange (PSX). We calculate daily returns for 20 years (from January 1996 to December 2015) using the following formula;

Rt = [ln(Pt) - ln(Pt-1)] x 100

Where at time t, the natural logarithm of the price of individual companies is represented by ln(Pt), while at time t - 1 natural logarithm of price is represented by ln(Pt-1), series of returns for each of 107 companies comprising 5219 observations. Kurtosis, skewness and Jarque-Bera-statistics are used to detect the normality of data which show that 107 companies deviate from the normal distribution which indicates that the distributions of companies' return series are not normal (A normal distribution should have a zero- skewness statistic and a kurtosis statistic of three). In order to investigate the series further, three most common types of unit root tests are also conducted (ADF, PP and KPSS) for all 107 companies. Both the ADF test (Augmented Dickey-Fuller) and the PP test (Phillips Perron) have non- stationarity as their null hypothesis while the alternative hypothesis is being stationarity.

KPSS test (Kwiatkowski-Phillips-Schmidt-Shin) is also conducted in order to avoid the over the rejection of the null hypothesis. KPSS has stationarity as the null hypothesis while the alternative hypothesis is being non-stationarity. ADF test and PP-test reveal that price level for around 83% firms is non-stationary, as the first difference is taken (returns), the series of return of all the companies become stationary at 1% significance in each case of Pure Random-Walk, Random-Walk with drift and Random-Walk with drift and deterministic trends. The results of KPSS test reveal that price levels of all 107 firms reject the null hypothesis of stationarity at 1% significance in full- sample; indicating price levels are nonstationary with both Random-walk with drift and Random-walk with drift and deterministic trends.

Similarly, the results reveal that when the first difference (return) of the series is taken 99% firms accept the null hypothesis of stationarity at 1% significance in full-sample indicating return series are stationary with Pure Random-Walk, Random-Walk with drift and Random-Walk with. The results of Kurtosis, skewness, Jarque-Bera-statistics and unit root tests are calculated for full and all sub-samples and are kept with the author may be provided on demand.

Table 1. Descriptive Statistics of Holiday effect in all firms during full sample period while *** shows the significance level at 1%.

###Mean###Std. Deviation###t-statistic###W-statistic

Holiday###0.1681###0.3008###4.884***###36.655***

Non-Holiday###0.0249###0.0385

Table 2. Mean Returns on Holidays and non-Holidays of individual firms over the period 1996-2015.

Holiday Effect Firms###Mean###Firms###Mean###Firms###Mean###Firms###Mean

Holiday###PK:ABB 0.037###PK:DEG 0.246###PK:JIN 0.262###PK:TLM###0.071

Non-Holiday###0.052###0.031###0.06###-0.016

Holiday###PK:ADI 0.086###PK:ETU 0.319###PK:KIE 0.63###PK:PTC###0.103

Non-Holiday###0.044###0.055###-0.053###0.063

Holiday###PK:AGR 0.509###PK:ERO 0.135###PK:KRM -0.075###PK:PSM###0.207

Non-Holiday###0.051###0.045###0.01###-0.022

Holiday###PK:AGT 0.13###PK:FSM -0.196###PK:KWG 0.489###PK:LAK###0.108

Non-Holiday###0.065###0.055###0.012###0.101

Holiday###PK:ACB 0.196###PK:FAU 0.252###PK:KNR 0.317###PK:PCT###0.391

Non-Holiday###0.031###0.037###0.026###0.029

Holiday###PK:ATH 0.095###PK:FZM 1.296###PK:LDP 0.042###PK:POC###0.271

Non-Holiday###0.101###-0.031###-0.003###-0.013

Holiday###PK:ATR 0.142###PK:FEC 0.086###PK:MLC 0.639###PK:RMP###0.252

Non-Holiday###0.044###0.02###0.001###0.07

Holiday###PK:BKP 0.306###PK:NAK -0.103###PK:MBK 0.135###PK:RUP -0.057

Non-Holiday###0.015###0.026###0.067###-0.023

Holiday###PK:BAP 0.124###PK:GAI -0.157###PK:MIR -0.127###PK:STM###0.134

Non-Holiday###0.088###0.038###0.041###-0.009

Holiday###PK:BHA 0.091###PK:GTR 0.401###PK:MRB 0.047###PK:CCB###0.548

Non-Holiday###0.036###0.045###0.091###-0.066

Holiday###PK:BOC 0.029###PK:GWC 0.131###PK:NAR 0.055###PK:SAN###0.098

Non-Holiday###0.05###-0.004###0.034###0.008

Holiday###PK:CAL 0.326###PK:GLT -0.214###PK:NPK 0.265###PK:HPN###0.08

Non-Holiday###-0.003###0.051###0.084###0.032

Holiday###PK:CPB 0.055###PK:GRY 0.388###PK:NAT 0.464###PK:SPP###0.103

Non-Holiday###0.025###0.009###-0.04###0.065

Holiday###PK:CTC 0.33###PK:GUL -0.005###PK:NHT 0.25###PK:SAP###0.178

Non-Holiday###0.019###0.042###0.05###0.041

Holiday###PK:CSA -0.016###PK:GSM -0.07###PK:NON -0.247###PK:SEA###0.067

Non-Holiday###0.044###-0.019###0.048###0.087

Holiday###PK:CTX -0.23###PK:HAB 0.12###PK:ORI -0.047###PK:SER###0.184

Non-Holiday###0.018###0.036###0.021###0.045

Holiday###PK:CYA 0.13###PK:MET 0.188###PK:PAC 0.073###PK:SHA###0.197

Non-Holiday###0.058###0.055###0.043###0.023

Holiday###PK:DAC 0.02###PK:HSM 0.151###PK:PET 0.205###PK:SCM -0.252

Non-Holiday###-0.028###0.041###0.045###0.017

Holiday###PK:DAE 0.106###PK:HAE 0.189###PK:PSM -0.031###PK:SHJ -0.022

Non-Holiday###-0.012###0.008###0.073###0.025

Holiday###PK:DAN 0.365###PK:HPM 0.067###PK:PNC -0.212###PK:SHK###0.361

Non-Holiday###-0.024###0.062###0.072###-0.002

Holiday###PK:DDH 0.206###PK:HUB 0.03###PK:PEN 2.164###PK:PBS -0.002

Non-Holiday###0.058###0.029###-0.032###0.033

Holiday###PK:DAW 0.31###PK:HUF -0.222###PK:PAL 0.437###PK:SIT -0.097

Non-Holiday###0.056###0.066###-0.021###0.046

Holiday###PK:DKT 0.397###PK:ICI 0.016###PK:PNS 0.291###PK:SON###0.384

Non-Holiday###-0.059###0.021###0.047###0.025

Holiday###PK:DMT -0.041###PK:IMO 0.293###PK:POF 0.035###PK:SNG###0.285

Non-Holiday###-0.034###0.069###0.064###-0.001

Holiday###PK:DES 0.326###PK:INI -0.068###PK:PRE 0.119###PK:SUI###0.167

Non-Holiday###-0.07###0.059###0.023###0.019

Holiday###PK:DSM 0.521###PK:ASB 0.173###PK:PSO 0.146###PK:TRP -0.39

Non-Holiday###-0.062###-0.039###0.027###-0.006

Holiday###PK:DEW -0.165###PK:JAV 0.082###PK:PSC 0.376

Non-Holiday###-0.06###0.018###-0.025

EMPIRICAL RESULTS

Table 1 presents the analysis of Holiday effect covering the whole sample period from 1996 to 2015 on all 107 companies. A non-parametric test Kruskal-Wallis statistic along with a standard t-statistic for differences in mean are calculated. Pre-holidays mean returns are higher than mean returns on non-holidays. Further, both test statistics support robust evidence of holiday effect by indicating significant mean differences between holiday and non-holidays returns. Therefore, we find holiday effect over the full sample period which is statistically significant. Table 2 shows the mean return of holidays and non-holidays for individual companies over the period of full-sample. Holidays mean returns are higher than mean returns on non-holidays in 71.1% firms. Therefore, in the majority of the firms we find holiday effect in the whole sample period on the basis of mean returns.

Tables 3, 4, 5 and 6 presents the behavior of holiday effect in full-sample as well as in sub-samples through GARCH (1,1) model and KW test. The results of full-sample reveal that holiday effect is significantly positive in 12 firms8 over the period of 20 years, comprising 1996 to 2015. This behavior shows that the returns of these firms are significantly higher and positive prior to the holidays time. Similarly, 66 firms9 show that the pre-holiday return is positive but insignificant over the full sample. However, 78 (12+66) firms show that the returns are positive before the holiday in the full sample period. On the other hand, firms like PK:DAW, PK:GAI, PK:GLT and PK:INI generate significant but negative coefficient prior to holidays, while 25 firms10 reflect insignificant and negative returns before the holidays.

As far as the sub-sample analysis are concerned, Table 3 reveals that the coefficients of the holiday effect are insignificant (independent) in first sub-sample (1996-2000) for the companies PK:BOC, PK:FSM, PK:FZM, PK:MIR, PK:NPK, PK:PAL, PK:PNS PK:PRE, PK:PCT, PK:SPP, PK:SER and PK:SNG. The behavior of the holiday effect then turns to dependency (inefficiency) during the period of 2001-2005 for these companies as the coefficients are significant. While the effect then reverses and turns to independence and market becomes efficient for the companies in next two sub-samples (from 2006-2010 to 2011-2015), thus supporting AMH which states that market efficiency varies over the time and encounters the periods of efficiency and inefficiency. Table 4 reveals that firms PK:CTX, PK:DAN, PK:DDH, PK:HSM, PK:INI, PK:KIE, PK:NAR, PK:PNC, PK:LAK, PK:SEA and PK:SHJ show independence of holiday effect in first two sub-samples (1996-2000 and 2001-2005).

The behavior of holiday effect reverses in third sub-sample (2006-2010) and becomes dependent which completely reverses and show independent behavior in the last sub-sample, consistent with AMH. Holiday effect remains insignificant (independent) in first three subsamples (from years 1996-2010) for the firms PK:ADI, PK:AGT, PK:BAP, PK:DAE, PK:DSM, PK:ETU, PK:ERO, PK:GAI, PK:POF, PK:RMP and PK:SHK and then reverts, predictable and moving towards dependency (market inefficiency) in last sub-sample (2011-2015) thus supporting AMH (Table 5). Similarly, Holiday effect for PK:GTR, PK:IMO, PK:JIN, PK:NON, PK:PEN, and PK:PSC (Table 6) also illustrates the behavior consistent with AMH. Therefore, 40 firms show the behavior of holiday effect consistent to AMH, means holiday effect fluctuates over time.

While the holiday effect in 67 firms1 1 remains independent and does not evolve over time as all the sub-samples produce insignificant coefficient (the results of firms generating insignificant holiday effect are not reported in the study but may be provided on demand).

Table 3. Results of the Holiday-Effect with the application GARCH (1,1) regression model and Kruskal-Wallis (K.W) test in full-sample as well as in subsample periods for companies listed at PSX (PK:BOC, PK:FSM, PK:FZM, PK:MIR, PK:NPK, PK:PAL, PK:PNS PK:PRE, PK:PCT, PK:SPP, PK:SER and PK:SNG). Where ***, ** and * represent significance at levels of 1%, 5% and 10%, while "ss" represents Holiday effect and "c" represents returns in non-holiday and number of observations are represented by "N".

###N###Firms###Period###c###[beta]###[alpha]1###[alpha]2###I,###K.w

5219###PK:BOC###Full-Sample 0.0235###-0.0656###4.028***###0.09***###0.435***###0.0169

1305###1996-2000###-0.0201###0.4228###9.65***###0.075***###0.4169***###1.9203

1305###2001-2005###0.015###0.5901**###0.1982***###0.0639***###0.9056***###0.1428

1304###2006-2010###-0.060***###-0.2262###-0.0128***###-0.0038***###1.0082***###0.272

1305###2011-2015###-0.0006###0.0368###0.4595***###0.167***###0.7403***###0.0703

5219###PK:FSM###Full-Sample 0.0183###-0.2883###2.6657***###0.0586***###0.7165***###0.0001

1305###1996-2000###-0.0911###-0.4604###26.2518***###0.0923***###-0.1092***###0.2456

1305###2001-2005###0.1396*###-1.075***###0.0372***###0.0296***###0.9702***###1.1267

1304###2006-2010###-0.0567***###-0.1378###0.0067***###-0.0071***###1.0052***###1.016

1305###2011-2015###0.0394###0.1587###0.6515***###0.0992***###0.7891***###0.9165

5219###PK:FZM###Full-Sample 0.3666***###-0.4074###9.2275***###2.7492***###0.014***###0.0822

1305###1996-2000###1.121***###0.6892###98.1191***###3.2804***###-0.0003###0.2567

1305###2001-2005###-0.0031###-0.6561**###0.0227***###0.0158***###0.9815***###0.6669

1304###2006-2010###0.0926*###-0.3383*###0.0211***###-0.0079***###1.0068***###0.369

1305###2011-2015###-0.0142###-0.0516###0.0544***###0.0507***###0.945***###0.2276

5219###PK:MIR###Full-Sample -0.0223###-0.1915###1.7477***###0.0839***###0.7242***###1.415

1305###1996-2000###-0.0701###-0.2557###10.9595***###0.0395***###-0.1844***###0.1657

1305###2001-2005###-0.045###-1.334***###2.4695***###0.1126***###0.7467***###1.9986

1304###2006-2010###0.0481###0.2968###4.7766###-0.0106***###0.5641*###0.106

1305###2011-2015###0.0107###-0.4618*###0.9088***###0.1579***###0.6889***###1.5395

5219###PK:NPK###Full-Sample 0.0498**###0.0751###0.0482***###0.0433***###0.9508***###0.2868

1305###1996-2000###0.0336###0.0517###0.0747***###0.0755***###0.9214***###0.3217

1305###2001-2005###0.0729###0.6624**###4.212***###0.1413***###0.306***###0.7665

1304###2006-2010###-0.0275###0.2621###0.1699***###0.0926***###0.8691***###0.143

1305###2011-2015###0.0156###0.2151###0.5153***###0.1562***###0.72***###0.5815

5219###PK:PAL###Full-Sample -0.0439###0.3319###0.8653***###0.0959***###0.8444***###4.2956**

1305###1996-2000###-0.0487###-0.3454###1.1573***###0.1039***###0.8374***###0.0828

1305###2001-2005###-0.0321###0.9311***###0.6235***###0.0824***###0.8703***###2.8266*

1304###2006-2010###-0.1608*###0.1593###1.8078***###0.17***###0.6736***###0.205

1305###2011-2015###0.0406###0.4519###0.6967***###0.0868***###0.8729***###2.5274

5219###PK:PNS###Full-Sample 0.0269###-0.1374###0.0134***###0.0228***###0.9775***###0.1973

1305###1996-2000###-0.1423###-0.2746###1.5114***###0.0604***###0.8904***###0.0268

1305###2001-2005###0.2211**###1.0824**###1.1035***###0.1049***###0.8616***###3.0079*

1304###2006-2010###-0.0623###-0.4079###0.5801***###0.198***###0.728***###2.202

1305###2011-2015###-0.0439###-0.0933###1.499***###0.1992***###0.5602***###0.0011

5219###PK:PRE###Full-Sample -0.0331###0.1881###0.2109***###0.0486***###0.9261***###1.7336

1305###1996-2000###-0.2121**###-0.0332###0.3252***###0.0333***###0.9369***###0.0011

1305###2001-2005###0.0684###0.5228**###0.0947***###0.0499***###0.9407***###0.291

1304###2006-2010###-0.0013###0.0638###0.2874***###0.1901***###0.7737***###0.508

1305###2011-2015###-0.056###0.3093###0.5432***###0.1925***###0.7038***###1.7496

5219###PK:PCT###Full-Sample 0.0586###0.4193**###0.0472***###0.049***###0.9512***###2.7984*

1305###1996-2000###-0.298**###0.4972###1.3841***###0.0583***###0.8932***###1.7603

1305###2001-2005###0.1056###1.2545***###0.695***###0.0701***###0.8973***###0.9387

1304###2006-2010###-0.0187###0.428###0.273***###0.1802***###0.8009***###0.588

1305###2011-2015###0.1347**###0.1702###0.0912***###0.0631***###0.9277***###0.8858

5219###PK:SPP###Full-Sample -0.0058###-0.128###0.1066***###0.0769***###0.9426***###0.0069

1305###1996-2000###-0.029###0.4543###7.3799***###0.1016***###0.1734***###0.4377

1305###2001-2005###-0.1052###2.2698***###0.114***###0.3702***###0.8783***###0.0795

1304###2006-2010###-0.0342###-0.3984###1.0722***###0.1266***###0.7033***###1.143

1305###2011-2015###0.0381###-0.0142###0.0338***###0.0581***###0.9387***###0.4796

5219###PK:SER###Full-Sample -0.043###-0.1122###0.0725***###0.0506***###0.9656***###0.8704

1305###1996-2000###-0.0542###0.0769###0.2632***###0.0209***###0.8707***###0.4521

1305###2001-2005###-0.3517***###-1.490***###0.0771###0.9287***###0.8797***###0.1427

1304###2006-2010###0.0406###-0.1832###1.599***###0.0811***###0.7614***###0.833

1305###2011-2015###-0.0144###0.2199###0.5919***###0.221***###0.6672***###4.0494**

5219###PK:SNG###Full-Sample -0.0026###0.2151###0.5385***###0.1333***###0.788***###2.7583*

1305###1996-2000###-0.0261###-0.0463###0.5194***###0.1519***###0.8096***###0.6661

1305###2001-2005###0.0385###0.6616**###0.4442***###0.0684***###0.8687***###1.114

1304###2006-2010###-0.0279###0.1355###0.7991***###0.2071***###0.6545***###0.247

1305###2011-2015###-0.0039###0.138###0.6567***###0.1252***###0.7221***###0.8842

Table 4. Results of the Holiday-Effect with the application GARCH (1,1) regression model and Kruskal-Wallis (K.W) test in full-sample as well as in subsample periods for companies listed at PSX (PK:CTX, PK:DAN, PK:DDH, PK:HSM, PK:INI, PK:KIE, PK:NAR, PK:PNC, PK:LAK, PK:SEA and PK:SHJ). Where ***, ** and * represent significance at levels of 1%, 5% and 10%, while "ss" represents Holiday effect and "c" represents returns in non- holiday and number of observations are represented by "N".

N###Firms###Period###c###[beta]###[alpha]1###[alpha]2###I,###K.w

5219###PK:CTX###Full-Sample 0.0053###-0.2381###0.066***###0.0296***###0.966***###0.1938

1305###1996-2000###0.1144###-0.3979###26.1685***###0.1369***###-0.1747***###0.0983

1305###2001-2005###0.0342###0.3194###7.8322***###0.1101***###0.1144###1.3419

1304###2006-2010###-0.0611###-0.8696***###0.0143***###0.1245***###0.8883***###0.436

1305###2011-2015###-0.0082###-0.2484###0.03*###0.0209***###0.9755***###0.5267

5219###PK:DAN###Full-Sample -0.0419###0.0437###0.0527***###0.015***###0.9843***###1.0317

1305###1996-2000###-0.246###0.3194###-0.2184***###-0.0018***###1.0083***###1.1596

1305###2001-2005###-0.0928###0.2646###9.2593***###0.1784***###0.552***###0.2024

1304###2006-2010###-0.0283###0.9146***###0.0115***###0.1372***###0.8868***###0.022

1305###2011-2015###0.0544###0.4571###0.0103###0.0144***###0.9846***###0.6778

5219###PK:DDH###Full-Sample 0.0711**###0.0719###1.5143***###0.2109***###0.5708***###1.1575

1305###1996-2000###0.1276*###-0.1988###2.4057***###0.2901***###0.5138***###0.2299

1305###2001-2005###0.075###0.302###1.794***###0.2803***###0.4656***###1.4557

1304###2006-2010###0.0239###-0.472**###0.1313***###0.153***###0.8344***###0.497

1305###2011-2015###0.0073###0.2613###0.8094***###0.167***###0.6762***###2.9479*

5219###PK:HSM###Full-Sample 0.0232###0.0696###1.251***###0.0765***###0.7135***###2.2075

1305###1996-2000###-0.0111###-0.1711###3.8677***###-0.0078***###0.4479**###0.2124

1305###2001-2005###-0.0092###0.7972###15.5545***###0.0248***###-0.2737***###2.1292

1304###2006-2010###0.0731###-0.4887**###0.0808***###0.1153***###0.8771***###1.08

1305###2011-2015###0.0917**###0.2564###0.6699***###0.3134***###0.5648***###3.1251*

5219###PK:INI###Full-Sample -0.0076###-0.1953*###0.7622***###0.2313***###0.639***###0.0246

1305###1996-2000###-0.0591###-0.1545###0.6214***###0.1946***###0.6941***###0.48

1305###2001-2005###0.0645###-0.2594###1.028***###0.3343***###0.6157***###0.0247

1304###2006-2010###0.0671###-0.42*###0.4495***###0.296***###0.6361***###2.372

1305###2011-2015###-0.0445###0.1634###0.7156***###0.1318***###0.677***###4.1979**

5219###PK:KIE###Full-Sample -0.0571###0.5072**###1.0302***###0.1335***###0.8018***###5.7382**

1305###1996-2000###-0.1388###-0.0317###0.8913***###0.1539***###0.8059***###1.0869

1305###2001-2005###-0.0311###0.3437###0.7067***###0.1365***###0.8203***###0.1372

1304###2006-2010###-0.1748*###0.6773*###2.0821***###0.1692***###0.704***###2.641*

1305###2011-2015###0.0164###0.699*###0.5783***###0.1166***###0.8456***###2.3849

5219###PK:NAR###Full-Sample 0.0107###-0.0151###1.5882***###0.1474***###0.6213***###2.4781

1305###1996-2000###-0.16**###0.2904###0.3414***###0.0952***###0.8768***###0.0026

1305###2001-2005###0.1326*###0.6146*###0.5532***###0.123***###0.8031***###1.3714

1304###2006-2010###-0.0636###-3.2254***###2.1235***###0.7291***###0.2533***###0.092

1305###2011-2015###0.0081###0.2742###0.9319***###0.1417***###0.5525***###4.0226**

5219###PK:PNC###Full-Sample 0.0603###-0.2674###4.1068###-0.0024***###0.5931*###4.2185**

1305###1996-2000###-0.124###0.1029###4.6679###-0.0017***###0.396###0.3722

1305###2001-2005###0.2242*###-0.5656###0.8702***###-0.0045***###0.9229***###2.4281

1304###2006-2010###0.0509###-0.4765***###0.4199***###0.2344***###0.6842***###5.113**

1305###2011-2015###0.0059###-0.1111###0.8553***###0.1492***###0.6936***###0.7227

5219###PK:LAK###Full-Sample 0.0569**###0.14###0.1614***###0.0672***###0.9132***###0.543

1305###1996-2000###0.0731###-0.0427###2.8121***###0.1059***###0.5747***###0.0336

1305###2001-2005###0.1241*###-0.1907###0.1102***###0.0224***###0.9616***###0.0662

1304###2006-2010###0.0627###0.5436**###0.0516***###0.1097***###0.9204***###0.148

1305###2011-2015###0.0117###0.0961###0.0562***###0.0859***###0.9044***###2.0765

5219###PK:SEA###Full-Sample 0.0495*###-0.1622###0.1742***###0.111***###0.867***###0.061

1305###1996-2000###-0.1382*###-0.0612###2.0167***###0.2245***###0.5225***###0.0056

1305###2001-2005###0.1356**###0.0008###0.4736***###0.1433***###0.8008***###0.4626

1304###2006-2010###-0.0358###-0.4854***###0.0951***###0.1217***###0.8702***###0.658

1305###2011-2015###0.1075**###0.0506###0.0648***###0.079***###0.9069***###0.2499

5219###PK:SHJ###Full-Sample -0.026###-0.0196###1.3599***###0.2328***###0.4937***###0.1822

1305###1996-2000###-0.0029###-0.023###2.0408***###0.1341*###0.4828***###0.0027

1305###2001-2005###0.0489###0.0566###0.8301***###0.2099***###0.5719***###0.4724

1304###2006-2010###-0.0208###-0.4873*###2.0948***###0.1289***###0.4853***###2.406

1305###2011-2015###-0.0711###0.2031###1.7514***###0.2779***###0.465***###0.0742

Table 5. Results of the Holiday-Effect with the application GARCH (1,1) regression model and Kruskal-Wallis (K.W) test in full-sample as well as in subsample periods for companies listed at PSX (PK:ADI, PK:AGT, PK:BAP, PK:DAE, PK:DSM, PK:ETU, PK:ERO, PK:GAI, PK:POF, PK:RMP and PK:SHK). Where ***, ** and * represent significance at levels of 1%, 5% and 10%, while "ss" represents Holiday effect and "c" represents returns in non- holiday and number of observations are represented by "N".

N###Firms###Period###c###[beta]###[alpha]1###[alpha]2###I,###K.w

5219###PK:ADI###Full-Sample 0.0708**###0.1547###0.5557***###0.1526***###0.7872***###0.9545

1305###1996-2000###-0.0288###-0.1163###0.7853***###0.1299***###0.8026***###0.0373

1305###2001-2005###0.1435*###0.0541###1.3866***###0.2477***###0.6549***###0.7828

1304###2006-2010###0.1798**###0.045###1.1105***###0.1891***###0.6877***###0.334

1305###2011-2015###0.0126###0.4502**###0.2525***###0.1051***###0.8309***###1.5345

5219###PK:AGT###Full-Sample 0.0027###0.1699###0.9197***###0.1373***###0.6897***###0.6773

1305###1996-2000###-0.0775###0.0815###2.1518***###0.1571***###0.5849***###0.4967

1305###2001-2005###0.0606###0.0614###0.4091***###0.1169***###0.8294***###0.0252

1304###2006-2010###0.0222###0.1705###0.2264***###0.2277***###0.7343***###0.004

1305###2011-2015###-0.0171###0.4851**###0.3944***###0.1291***###0.7632***###1.1823

5219###PK:BAP###Full-Sample 0.0076###-0.0468###0.1336***###0.0568***###0.921***###0.326

1305###1996-2000###-0.0588###-0.1249###0.183***###0.0585***###0.8997***###0.0374

1305###2001-2005###0.0394###0.0195###1.3721***###0.0725***###0.686***###0.8996

1304###2006-2010###0.0659###-0.0822###1.9666***###0.1699***###0.5041***###0.208

1305###2011-2015###0###-0.0001**###0###0.114***###0.8971***###0.0992

5219###PK:DAE###Full-Sample -0.0598*###0.1028###1.236***###0.0995***###0.7204***###0.4999

1305###1996-2000###-0.1148###0.042###5.1726***###0.0841***###0.5094***###2.5337

1305###2001-2005###0.0273###-0.1488###2.8268***###0.0769***###0.2938***###0.6936

1304###2006-2010###-0.0467###0.1519###3.789###-0.0049***###0.3647###0.162

1305###2011-2015###-0.105###0.3514###1.1167***###0.1306***###0.7586***###1.5924

5219###PK:DSM###Full-Sample -0.0749###0.3726###0.0869***###0.0222***###0.9729***###6.0797**

1305###1996-2000###-0.1237###0.2842###2.1638*** -0.0046###0.5947***###1.9813

1305###2001-2005###0.0057###0.78###5.0981*** -0.0135***###0.5156***###1.4337

1304###2006-2010###-0.1494###-0.438###0.7913***###0.0993***###0.8672***###0.118

1305###2011-2015###-0.063###1.6184***###1.4046***###0.0774***###0.8903***###7.002***

5219###PK:ETU###Full-Sample 0.0242###0.473***###0.9696***###0.1259***###0.7068***###3.3177*

1305###1996-2000###-0.1055*###0.2301###3.8088***###0.028**###0.2346###0.5755

1305###2001-2005###0.1276*###0.2244###1.0173***###0.0514***###0.7923***###0.0241

1304###2006-2010###0.0988*###0.1845###0.5947***###0.2461***###0.6777***###0.115

1305###2011-2015###-0.0428###0.6624***###0.8996***###0.2036***###0.5922***###12.7528***

5219###PK:ERO###Full-Sample 0.0716***###0.1099###0.4451***###0.174***###0.7508***###1.7814

1305###1996-2000###0.0251###0.1871###0.3706***###0.1582***###0.8053***###2.1831

1305###2001-2005###0.0626###-0.2158###0.6358***###0.3281***###0.6053***###1.118

1304###2006-2010###0.1256**###-0.1481###0.7251***###0.2418***###0.6309***###2.321

1305###2011-2015###0.0344###0.5921**###0.1998***###0.0865***###0.8668***###4.5855**

5219###PK:GAI###Full-Sample -0.0375###-0.7034***###0.0822***###0.0635***###0.9543***###0.0301

1305###1996-2000###-0.1551###-0.6366###11.2847*** 0.0938***###0.6127***###0.2085

1305###2001-2005###0.0733###0.3412###0.9577***###0.1091***###0.8043***###0.0276

1304###2006-2010###-0.2319***###0.0298###0.33***###0.0599***###0.893***###0.121

1305###2011-2015###0.0231###-0.2561**###0.0356***###0.0285***###0.954***###0.459

5219###PK:POF###Full-Sample 0.0539**###0.2573*###0.0991***###0.0691***###0.9155***###2.077

1305###1996-2000###-0.0507###0.1047###0.0208***###0.0609***###0.9523***###0.129

1305###2001-2005###0.0795###0.1346###0.5209***###0.2147***###0.7643***###0.5436

1304###2006-2010###0.1258**###0.0285###0.6328***###0.1786***###0.7085***###0.559

1305###2011-2015###0.0101###0.3565**###0.0876***###0.0585***###0.8959***###8.7933***

5219###PK:RMP###Full-Sample 0.0289###0.0164###0.0451***###0.0334***###0.9542***###2.9248*

1305###1996-2000###-0.1241***###0.1129###0.0011*** -0.005***###1.008***###0.0676

1305###2001-2005###0.1054**###-0.1104###0.058***###0.0282***###0.9414***###0.7962

1304###2006-2010###0.0464###0.2902*###0.0854***###0.0538***###0.918***###0.024

1305###2011-2015###0.0167###0.6353**###1.7707***###0.2321***###0.425***###2.0456

5219###PK:SHK###Full-Sample -0.017###0.1243###0.3017***###0.0294***###0.9504***###0.7855

1305###1996-2000###-0.065###0.3257###6.1137***###0.0556***###0.3899***###0.2361

1305###2001-2005###0.105###-0.4875###0.2469***###0.0199***###0.9649***###1.6014

1304###2006-2010###-0.1313###-0.3088###1.4565***###0.161***###0.7187***###0.188

1305###2011-2015###-0.064###0.8206**###0.507***###0.1092***###0.8575***###1.361

Table 6. Results of the Holiday-Effect with the application GARCH (1,1) regression model and Kruskal-Wallis (K.W) test in full-sample as well as in subsample periods for companies listed at PSX (PK:ADI, PK:AGT, PK:BAP, PK:DAE, PK:DSM, PK:ETU, PK:ERO, PK:GAI, PK:POF, PK:RMP and PK:SHK). Where ***, ** and * represent significance at levels of 1%, 5% and 10%, while "ss" represents Holiday effect and "c" represents returns in non-holiday and number of observations are represented by "N".

N###Firms###Period###c###[beta]###[alpha]1###[alpha]2###I,###K.w

5219###PK:JIN###Full-Sample -0.1589***###0.3718*###0.3903***###0.142***###0.8646***###3.2195*

1305###1996-2000###-0.018###0.2464###2.4122***###0.108***###0.5969***###0.6566

1305###2001-2005###-0.4433*** -0.4891**###0.0938**###0.3527***###0.8715***###0.0333

1304###2006-2010###-0.0164###0.1921###0.766***###0.2337***###0.5983***###1.129

1305###2011-2015###0.0398###0.369**###0.3468***###0.0956***###0.8131***###0.1432

5219###PK:NON###Full-Sample -0.0072###-0.1441###0.5886***###0.0822***###0.8587***###0.0023

1305###1996-2000###-0.0002###-0.5603*###0.1087***###0.0207***###0.9619***###1.0223

1305###2001-2005###0.0923###0.107###0.6236***###0.0425***###0.8969***###0.1206

1304###2006-2010###-0.0473###-0.3939*###0.1766***###0.1749***###0.824***###0.662

1305###2011-2015###0.0657###0.2006###6.3678###-0.007***###0.5746###1.0309

5219###PK:PSC###Full-Sample -0.0325###0.4385**###0.2962***###0.0553***###0.9229***###0.7705

1305###1996-2000###-0.1447###0.2617###0.6447***###0.0591***###0.901***###0.1721

1305###2001-2005###-0.0129###0.7016**###5.0654***###0.1181***###0.3082***###1.8082

1304###2006-2010###-0.2104*###1.0972**###0.735***###0.0684***###0.9007***###0.162

1305###2011-2015###0.0049###-0.2646###0.3661***###0.1675***###0.7935***###0.0208

5219###PK:GTR###Full-Sample -0.0201###0.1757###0.7986***###0.15***###0.7605***###2.4734

1305###1996-2000###-0.1937***###0.4306###0.9768***###0.1399***###0.8042***###1.7285

1305###2001-2005###0.0199###-0.2455###1.7227***###0.1849***###0.5977***###0.0002

1304###2006-2010###-0.0431###-0.3898**###0.1801***###0.175***###0.8099***###0.133

1305###2011-2015###0.0491###0.721**###1.6121***###0.1776***###0.5503***###6.2667**

5219###PK:IMO###Full-Sample###0.0606**###0.2933**###0.1939***###0.1009***###0.8713***###3.0898*

1305###1996-2000###-0.137###1.0107***###3.1405***###0.1984***###0.5304***###2.5407

1305###2001-2005###0.1467**###-0.2781###0.8212***###0.1225***###0.7474***###0.1483

1304###2006-2010###0.0256###0.3494###0.1991***###0.1303***###0.8385***###0.015

1305###2011-2015###0.0688###0.2908###0.9386***###0.1855***###0.5065***###4.7209**

5219###PK:PEN###Full-Sample -0.1392###3.6821***###68.2632###-0.0008***###0.5396###1.6291

1305###1996-2000###-0.4707###8.731***###196.3981###-0.0027***###0.5974*###1.8545

1305###2001-2005###-0.0008###0.1188###1.6502***###0.0909***###0.9122***###0.4753

1304###2006-2010###0.0718###-0.514**###0.4029***###0.1612***###0.7913***###0.92

1305###2011-2015###-0.1055**###0.3573###0.7586***###0.1323***###0.6976***###1.9692

CONCLUSION

Although the recent studies support the fact that calendar anomalies have reversed or even diminished over time, the voluminous literature is evident of the fact that calendar anomalies are accepted in almost all equity markets of the world. This paper examined the holiday-effect across time to explore whether this anomaly can be used to exploit the excess returns. The study finds around 72% of firms exhibit positive returns before holidays thus, supporting the presence of the holiday effect through average returns and GARCH (1,1) model in the whole-sample period of 1996-2015. Thus, this anomaly can be used to earn abnormal returns. Finally, it is clear from sub-sample analysis that holiday-effect in 40 firms has shifted from periods of predictability/market inefficiency to the periods of no-predictability/market efficiency or vice versa, while 67 firms exhibit no swing in holiday effect during sub-samples.

As the predictability of holiday effect swings under periods of dependency/inefficiency and independency/efficiency, we conclude that AMH provides a better description of behavior of holiday effect in Pakistan than the classical/traditional EMH. In summary, we conclude that the holiday effect in firms' exhibits time- varying behavior across time through sub-samples. The sign of varying behavior of holiday effect is consistent and supporting AMH while opposing the traditional EMH. We believe a sub-sample analysis of long time period may be a more appropriate method to elucidate the idea of Adaptive Market Hypothesis (AMH) in future research and suggest the current method could be adapted and would be helpful to examine other calendar and market anomalies in different equity markets in the world.

Appendix. Names of sample companies and their codes

ABBOTT LABS. (PAK.)###PK:ABB###JUBILLE INSURANCE###PK:JIN

ADAMJEE INSURANCE###PK:ADI###KARACHI ELECTRIC SUPP.###PK:KIE

AGRIAUTO INDUSTRIES###PK:AGR###KARAM CERAMICS###PK:KRM

AL-GHAZI TRACTORS###PK:AGT###KOHINOOR MILLS###PK:KWG

ASKARI BANK###PK:ACB###KOHINOOR TEX.MILLS###PK:KNR

ATLAS HONDA###PK:ATH###LINDE PAKISTAN###PK:LDP

ATTOCK REFINERY###PK:ATR###MAPLE LEAF CMT.FACTORY###PK:MLC

BANK OF PUNJAB###PK:BKP###MCB BANK###PK:MBK

BATA PAKISTAN (~PR) (#T)###PK:BAP###MIRPURKHAS SUGAR###PK:MIR

BHANERO TEXTILE MILLS###PK:BHA###MURREE BREWERY COMPANY###PK:MRB

BOLAN CASTINGS###PK:BOC###NATIONAL REFINERY###PK:NAR

CAPITAL ASSETS LSG.###PK:CAL###NESTLE PAKISTAN###PK:NPK

CENTURY PAPER###PK:CPB###NIB BANK###PK:NAT

CHEARAT CEMENT COMPANY###PK:CTC###NISHAT (CHUNIAN)###PK:NHT

CRESCENT STEEL###PK:CSA###NOON SUGAR MILLS###PK:NON

CRESCENT TEXTILE MILLS###PK:CTX###ORIX LEASING PAK.###PK:ORI

CYAN LIMITED###PK:CYA###PACKAGES###PK:PAC

DADABHOY CEMENT###PK:DAC###PAK ELEKTRON###PK:PET

DADEX ETERNIT###PK:DAE###PAK SUZUKI MOTOR###PK:PSM

DANDOT CEMENT###PK:DAN###PAKISTAN CABLES###PK:PNC

DAWOOD HRC.CHEMS.CORP.###PK:DDH###PAKISTAN ENGINEERING###PK:PEN

DAWOOD LAWRENCEPUR###PK:DAW###PAKISTAN INTL.AIRLINES###PK:PAL

DEWAN KHALID TEX.###PK:DKT###PAKISTAN NAT.SHIP.###PK:PNS

DEWAN MUSHTAQ TEX.###PK:DMT###PAKISTAN OILFIELDS###PK:POF

DEWAN SALMAN FIBRE###PK:DES###PAKISTAN REFINERY###PK:PRE

DEWAN SUGAR###PK:DSM###PAKISTAN STATE OIL###PK:PSO

DEWAN TEXTILE MILLS###PK:DEW###PAKISTAN SYNTHETICS###PK:PSC

DG KHAN CEMENT COMPANY###PK:DEG###PAKISTAN TELECM.###PK:TLM

EFU GENERAL INSURANCE###PK:ETU###PAKISTAN TOBACCO###PK:PTC

ENGRO###PK:ERO###PARAMOUNT SPNG.MLS.###PK:PSM

FAISAL SPINNING MILLS###PK:FSM###PHILIP MORRIS PAKISTAN###PK:LAK

FAUJI FERTILIZER###PK:FAU###PIONEER CEMENT###PK:PCT

FAZAL TEXTILE MILLS###PK:FZM###POWER CEMENT###PK:POC

FECTO CEMENT###PK:FEC###RAFHAN MAIZE PRDS.###PK:RMP

FEROZE1888 MILLS###PK:NAK###RUPALI POLYESTER###PK:RUP

GATRON INDUSTRIES###PK:GAI###SAIF TEXTILE MILLS###PK:STM

GENERAL TYRE and RUBBER###PK:GTR###SAMBA BANK###PK:CCB

GHARIBWAL CEMENT###PK:GWC###SANA INDUSTRIES###PK:SAN

GLAXOSMITHKLINE PAK.###PK:GLT###SANOFI AVENTIS PAKISTAN###PK:HPN

GRAYS OF CAMBRIDGE###PK:GRY###SAPPHIRE FIBRES###PK:SPP

GUL AHMED TEXTILE MILLS###PK:GUL###SAPPHIRE TEX.MLS.###PK:SAP

GULISTAN SPNG.MILLS (~PR) (#T)###PK:GSM###SEARLE###PK:SEA

HABIB ADM LIMITED###PK:HAB###SERVICE INDUSTRIES###PK:SER

HABIB METROPOLITAN BANK###PK:MET###SHABIR TILES###PK:SHA

HABIB SUGAR###PK:HSM###SHADMAN COTTON MILLS###PK:SCM

HALA ENTERPRISES###PK:HAE###SHAHTAJ SUGAR MILLS###PK:SHJ

HINOPAK MOTORS###PK:HPM###SHAKARGANJ MILLS###PK:SHK

HUB POWER COMPANY###PK:HUB###SHELL PAKISTAN###PK:PBS

HUFFAZ SEAMLESS PIPE###PK:HUF###SITARA CHEMICAL###PK:SIT

ICI PAKISTAN###PK:ICI###SONERI BANK###PK:SON

INDUS MOTOR COMPANY###PK:IMO###SUI NORTHERN GAS###PK:SNG

INTERNATIONAL INDS.###PK:INI###SUI SOUTHERN GAS###PK:SUI

INVEST CAPITAL INV.BANK###PK:ASB###TRI-STAR POLYESTER###PK:TRP

JAVEDAN###PK:JAV

REFERENCES

Abdul Karim, B., Abdul Karim, Z., and Tang, A. N. (2012). Holiday effects in Malaysia: An empirical note. International Journal of Research in Economics and Business Management, 1(1), 23-26.

Abidin, S., Banchit, A., Sun, S., and Tian, Z. (2012). Chinese New Year effects on stock returns: Evidence from Asia-Pacific stock markets. In the Asian Finance Association 2012 International Conference, Taiwan.

Agrawal, A., and Tandon, K. (1994). Anomalies or Illusion? Evidence from Stock Markets in Eighteen Countries. Journal of International Money and Finance, 13, 83-106.

Al-Loughani, N. E., Al-Saad, K. M., and Ali, M. M. (2005). The holiday effect and stock return in the Kuwait Stock Exchange. Global Competitiveness, 13(1), 81-91.

Ariel, R. A. (1990). High Stock Returns before Holidays: Existence and Evidence on Possible Causes. The Journal of Finance, 45(5), 1611-1626.

Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307-327. doi:10.1.1.468.2892

Borowski, K. (2015). Analysis of Selected Seasonality Effects in Market of Barley, Canola, Rough Rice, Soybean oil and Soybean Meal Future Contracts. Journal of Economics and Management, 21(3), 73-89.

Boyle, G., Hagan, A., O'connor, R. S., and Whitwell, N. (2004). Emotion, fear and superstition in the New Zealand stock market. New Zealand Economic Papers, 38(1), 65-85.

Brockman, P., and Michayluk, D. (1998). The Persistent Holiday Effect: Additional Evidence. Applied Economics Letter, 5(4), 205-209.

Brown, D. P., and Jennings, R. H. (1989). On Technical Analysis. Review of Financial Studies, 2(4), 527-551.

Campbell, J. Y., Lo, A. W., and MacKinlay, A. C. (1997). The econometrics of financial markets (Vol. 2, pp. 149-180). Princeton, NJ: Princeton University Press.

Cao, X. L., Premachandra, I. M., Bhabra, G. S., and Tang, Y. P. (2009). Firm size and the pre-holiday effect in New Zealand. International Research Journal of Finance and Economics, 32, 171-187.

Chong, R., Hudson, R., Keasey, K., and Littler, K. (2005). Pre-Holiday Effects: International Evidence on the Decline and Reversal of a Stock Market Anomaly. Journal of International Money and Finance, 24(8), 1226-1236.

Engle, R. (2001). GARCH 101: The use of ARCH/GARCH Models in Applied Econometrics. Journal of Economic Perspectives, 15(4), 157-168.

Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The Journal of Finance, 25(2), 383-417.

Fields, M. J. (1934). Security Prices and Stock Exchange Holidays in Relation to Short Selling. The Journal of Business of the University of Chicago, 7(4), 328-338.

Grossman, S. J., and Stiglitz, J. E. (1980). On The Impossibility of Informationally Efficient Markets. The American Economic Review,70(3), 393-408.

Halari, A. (2013). An Analysis of Monthly Calendar Anomalies in the Pakistani Stock Market: A Study of the Gregorian and Islamic calendars (Doctoral dissertation, University of Dundee).

Haque, A., Liu, H. C., and Nisa, F. U. (2011). Testing the Week Form Efficiency of Pakistani Stock Market (2000-2010). International Journal of Economics and Financial Issues, 1(4), 153-162.

Hashmi, M. A. (2014). January Effect in Pakistan: A Time Series Analysis. Market Forces, 9(1).

Huang, C. C. (2017). New Evidence on the Holiday Effect in the Chinese Stock Market. Journal of Applied Business and Economics, 19(9), 83-98.

Hassan, M., and Sarker, A. (2018). Holiday Effect on Stock Market Return: Evidence from Dhaka Stock Exchange (DSE). Anik, Holiday Effect on Stock Market Return: Evidence from Dhaka Stock Exchange (DSE)(September 10, 2018).

Kim, C. W., and Park, J. (1994). Holiday effects and stock returns: Further evidence. Journal of Financial and Quantitative Analysis, 29(1), 145-157.

Kothari, S. P. (2001). Capital markets research in accounting. Journal of accounting and economics, 31(1-3), 105-231.

Lakonishok, J., and Smidt, S. (1988). Are Seasonal Anomalies Real? A Ninety Year Perspective. The Review of Financial Studies, 1(4), 403-425.

Lo, A. W. (2004). The adaptive markets hypothesis: Market efficiency from an evolutionary perspective. Journal of Portfolio Management, 30, 15-29.

Lucey, B. M. (2005). Are Local or International Influences Responsible for the Pre-Holiday Behavior of Irish Equities? Applied Financial Economics, 15, 381-389.

Marrett, G. J., and Worthington, A. C. (2009). An empirical note on the holiday effect in the Australian stock market, 1996- 2006. Applied Economics Letters, 16(17), 1769-1772.

McGuinness, P. B., and Harris, R. D. (2011). Comparison of the 'turn-of-the- month'and lunar new year return effects in three Chinese markets: Hong Kong, Shanghai and Shenzhen. Applied financial economics, 21(13), 917-929. doi:http://dx.doi.org/10.1080/09603107.2010.548782

Nisar, S., and Hanif, M. (2012). Testing Weak form of Market Hypothesis: Empirical Evidence from South Asia. World Applied Sciences Journal, 17(4), 414-427.

Picou, A. (2006). Stock Returns Behavior During Holiday Periods: Evidence from Six Countries. Managerial Finance, 32(5), 433-445.

Poshakwale, S. (1996). Evidence on weak form efficiency and day of the week effect in the Indian stock market. Finance India, 10(3), 605-616.

Rabbani, S., Kamal, N., and Salim, M. (2013). Testing the Weak Form Efficiency of the Stock Market: Pakistan as an Emerging Economy. Journal of Basic and Applied Scientific Research, 3(4), 136-142.

Rehman, S., and Rizwan, Q. M. (2014). Testing weak form efficiency of capital markets: A case of Pakistan. International Journal of Research Studies in Management, 3(1), 65-73.

Riaz, T., Hassan, A., and Nadim, M. (2012). Market Efficiency in its Weak- Form; Evidence from Karachi Stock Exchange of Pakistan. The Journal of Commerce, 4(4), 9-17.

Shahid, M. N., and Sattar, A. (2017). Behavior of Calendar Anomalies, Market Conditions and Adaptive Market Hypothesis: Evidence from Pakistan Stock Exchange. Pakistan Journal of Commerce and Social Sciences, 11(2), 471-504.

Shahid, N. M., and Mehmood, Z. (2015). Calendar Anomalies in Stock Market: A Case of KSE 100 Index. International Journal of African and Asian Studies, 7, 16-23.

Urquhart, A. (2013). An Empirical Analysis of the Adaptive Market Hypothesis and Investor Sentiment in Extreme Circumstances.

Urquhart, A., and Hudson, R. (2013). Efficient or Adaptive Markets? Evidence from Major Stock Markets using very Long-Run Historic Data. International Review of Financial Analysis, 28, 130-142.

Urquhart, A., and McGroarty, F. (2014). Calendar Effects, Market Conditions and the Adaptive Market Hypothesis: Evidence from Long-Run Data. International Review of Financial Analysis, 35, 154-166.

Wong, W. K., Agarwal, A., and Wong, N. T. (2006). The disappearing calendar anomalies in the Singapore Stock Market. Lahore Journal of Economics, 11(2), 123-139.

Yuan, T., and Gupta, R. (2014). Chinese Lunar New Year effect in Asian stock markets, 1999-2012. The Quarterly Review of Economics and Finance, 54(4), 529-537.

Zafar, N., Urooj, S. F., Chughtai, S., and Amjad, S. (2012). Calendar Anomalies: Case of Karachi Stock Exchange. African Journal of Business Administration, 6(24), 7261-7271.
COPYRIGHT 2018 Knowledge Bylanes
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Journal of Business Strategies (Karachi)
Article Type:Report
Geographic Code:9PAKI
Date:Dec 31, 2018
Words:10539
Previous Article:EXAMINING DETERMINANTS OF ENTREPRENEURIAL SUCCESS: THE CASE OF SMALL and MEDIUM ENTREPRENEURS IN KP, PAKISTAN.
Next Article:SUPPLIER SELECTION CRITERIA IN THE CONTEXT OF CITY GOVERNMENT: EVIDENCE FROM RENOWNED ORGANIZATIONS OF KARACHI CITY.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters