Calendar Anomalies: Definition and Background Information - Research Paper Example

?2. Literature Review 2 Calendar Anomalies: Definition and Background Information Calendar anomalies, or seasonal effects, as applied to stock markets can be defined as “the tendency of financial assets returns to display systematic patterns at certain times of the day, week, month or year” (Brooks 2008, p. 454). Although it was only recently that substantial studies were devoted to look seriously and closely into this phenomenon, calendar anomalies have always been regarded as part of the market folklore. As early as the 1930s, a few studies had already broached the idea of seasonal effects (Jacobs & Levy, p. 135). Recently, a spate of studies looked particularly into this phenomenon despite the admitted dominance of Professor Farma’s (1970) efficient market hypothesis (EMH hereafter), which is said to largely underpin stock market and securities returns rationale. The EMH is “A market in which prices always ‘fully reflect’ available information” (cited Scheurle, p. 19). Thus, while the information-reliant EMH makes stock markets largely unpredictable, calendar anomalies results in the opposite. Moreover, they breach the EMH principle that stock returns should be random and unpredictable because they allow market participants to anticipate potential rise and fall of the market and use that knowledge to make profits (Gao and Kling 2005, pp. 75-76) In the United States, for example, the day-of-the week and January effects have been widely studied, as a consequence of which a large amount of evidence has been found that supports their validity (Lean et al, 2007, p. 2). The studies conducted by Rozeff and Kinney (1976), Gultekin and Gultekin (1983), Keim and Stambaugh (1984) and Kato and Shallheim (1985), among others, validated the observation that stock returns are higher on the month of January as compared to other months. On the other hand, the day-of-the-week effect suggests that stock returns between the close of Friday and that of Monday are comparatively lower than that of the other days of the week. These were the results of the studies conducted by the likes of Gibbons and Hess (1981), Mills and Coutts (1995) and Al-Loughani and Chapell (2001) (Alagidede and Panagiotidis 2006, p. 2). Most of these studies were conducted in the stock markets of the US, Australia, Canada, Tokyo Italy and London. However, Alagidede and Panagioditis (2006) belied the existence of both January and day-of-the week effects in their study of the Ghana stock Exchange, where they instead observed that the month of April shows a comparatively higher yield of returns of all months in a year - an observation that can be readily explained by the usual submission of company reports in late March during the year (pp. 75-88). Some of the so-called calendar anomalies, aside from the day-of-the-week and January effects, are: the turn-of-the-month effect; the holiday effect; the time-of-day effect (Jacobs & Levy 1988, pp. 28-36), the Ramadan effect (Seyyed et al, 2005), the school-is-out effect (Coakley et al 2007), among others. The day-of-the-week effect refers to the anomalies observe where average returns show a pattern of being higher on some days of the week compared to other days of the same week (Brooks & Persand 2001, p. 155). The turn-of-the-month effect as documented by Ariel (1984) in his study of 19 years stock index returns shows that the entire market cumulative advance for those years was contributed on the first half of the month as compared to the zero contribution in the next half of it. The Ramadan effect is another seasonality effect that is based on the Ramadan, the holy month of the Muslims where trading activities are observed to be comparatively low than in other months (Seyyed et al 2005). Various authors have a variety of explanations to explain away calendar anomalies ranging from psychological to social and to traditions. Thaler (1987) offers three suggestions he thinks are worth looking into. First, it could be custom-related. An example would be the custom of distributing pension and mutual funds at certain periods that affects portfolio sizes and movements or the propensity of small firms to buy and sell towards the end of the year. A second possibility is ‘window dressing’ or the cleaning up of portfolios before certain dates of the year to present a more attractive one as is reflective of the practice of investment managers before reporting dates. Finally, calendar anomalies may have a link to news arrival, whether good or bad, which is particularly applicable to the weekend effect (pp. 169-177). Hansen and Lunde (2003) noted that calendar effects are most significant in return series and that the largest calendar anomalies are those that take effect at the end of the year. However, the two also noted that as far as standardised returns, the calendar effect is less and less becoming significant (pp. 1-16). Heinnen (2010) also reached a similar conclusion and attributed the waning effect of calendar anomalies in the transition economies of Europe to “EU accession, the growing awareness of the importance of the standards of corporate governance, gradual integration with developed market capitals markets,” among others (p. 21). Notwithstanding the existence of a depth of literature and research on the subject of calendar effects there are those who maintained that stock markets are unpredictable in accordance with the efficient market hypothesis. Calendar anomalies defy the principle of standard economic theory that stock prices should abide by a martingale process and returns should not indicate any pattern at all (Sullivan et al, 1998, p. 2). Zweig (2000), for example, points out a study in human behavioral psychology that indicates that humans have a propensity to look for patterns in things and objects when, in fact, none exist and that research on calendar anomalies are based on erroneous statistical methods or are the result of random patterns (Kunkel et al 2003, pp. 1-2). Rubenstein (2000) believes that anomalies, in general, are “empirical illusions” that came about as a result of “data mining, survivorship bias, selection bias, short-shot bias, trading costs and the high variance of sample means” (p. 17). Finance theory, in an effort to counter this escalating amount of evidence on the subject of calendar anomalies have come up with various logical rationales for them. However, none of them have succeeded in fully explaining away the existence of these anomalies especially those that are being experienced outside of the US (Cao et al 2007, pp. 1-2). 2.2 Types of Calendar Anomalies 2.2.1 Short Term Anomalies: Weekend Effect (or Day-of-the-Week Effect) The weekend effect was first brought to the attention of the public by Fields (1931) at a time when many stock markets still traded on Saturdays. The day-of-the week effect or weekend effect shows Mondays as having consistently comparative lower returns (Mitchell 8). It was however, in 1973 when Frank Cross published an article that pointed out the variation of returns during the week that spurred a rush of researches into this particular phenomenon. Using 844 sets of Fridays and their subsequent Mondays from 1953 to 1970, Cross drew attention to the fact that the stock prices advanced 523 times, or 62% as compared to Mondays 333, or 39.5% (pp. 67-69). Likewise, Harris (1986) observed that Mondays have the lowest returns in stock markets compared to the other days of the week. These developments have encouraged many researchers to conduct studies on weekend effect, also known as day-of-the-week effect. The weekend or day-of-the-week effect is largely documented in the US with the studies conducted on the subject by French (1980), Gibbons and Hess (1981), Keim and Stambaugh (1984) and Linn and Lockwood (1988), which all validated the existence of this phenomenon. Outside of the US, Jaffe and Westerfield (1985) found a substantial negative Monday returns in the stock markets of Australia, Canada, Japan and the UK. Jaffe and Westerfield (1985), in particular, wanted to find out of the US experience has any bearing on the stock markets of other countries considering the differences in time zones. After establishing the presence of weekend effect in these countries, although in the case of Australia and Japan the lowest mean returns fall on Tuesdays, the team further established that such effects are independent of the US weekend influence. (pp. 433-454). The independence of the weekend effect in other countries from the US was validated by the study conducted by Pena (1995) on the Spanish Stock Exchange (pp. 419-423). According to Savva et al (2006) the day-of-the-week effect can be studied in three ways: the first is to assume that it is present only in returns and negating calendar anomalies in volatility equations; the second is to consider its effects only in equality equations through GARCH model variants, and; the last is to employ a combination of both. The first is documented by the studies of Cross (1973), French (1980), Gibbons and Pees (19890), Keim and Stambaug (1984), and Rogalski (1984). Agrawal and Tandon (1994) conducted a study on 15 stock markets outside of the USA to find out if calendar anomalies are also present outside it. These include the stock markets of the following countries: Australia; Belgium; Brazil; Canada; Denmark, France, Germany; Hong Kong; Italy; Japan; Luxembourg; Mexico; Netherlands; New Zealand; Singapore; Sweden; Switzerland; and; the UK. As far as the weekend effect is concerned, the study had a mixed result. Only nine out of the 18 countries had lowest Monday returns similar to the US, but in the rest of the countries the lowest returns fell on Tuesdays. Fridays remained to have the highest returns of the week, however, as is the case in the US, except for Luxembourg. The study also validated other studies that a particularly low-returns Monday is usually preceded by a previous week decline and no negative Monday occurs when the preceding week rose except in the UK (pp. 83-106). In the study conducted by Savva et al (2006) to inquire into the validity of the day-of-the-week effect (also known as weekend effect), among the fifteen stock markets studied were DAX-30 of Germany, FTSE-100 of UK, IBEX-35 of Spain and the total indices in the stock markets of Italy, Portugal, Luxembourg, Greece, Finland, Belgium, Austria, the Netherlands, Switzerland, Denmark and Norway. On the other hand, Connoly (1991) attempted to contradict the aforesaid finding by employing the Bayesian model rather than the classical approach used by most studies in analysing data. He concluded that inferences on a weekend or day-of-the-week effects made by previous studies validating its existence can be attributed to the shortcomings of the classical approach than an indicative pattern of returns. This is because, according to him, the classical approach has a tendency to excessively reject null hypothesis in cases involving large amount of data in what is known as the Lindley Paradox (pp. 51-104). The logic behind the weekend effect has been attributed to various factors. These reasons include settlement effect, timing of earnings announcement, measurement error or specialist-related bias (Jacobs and Levy 1988, p. 6), statistical errors, micro market effects, information flow effects and order flow effects (cited Lean et al 2009, p. 2). Some explained the day-of-the-week effect as the tendency of human nature to announce good news immediately while delaying bad news. Thus, the observed low returns on the close of Fridays is attributed to bad news that may have been announced only at the last trading day of the week that naturally affects the opening of the next trading day, which is Monday (Jacobs and Levy 1988, pp. 6-7). On the other hand, Chen and Singal (2003) partially attributed the weekend reports to speculative short sellers who are averse to long periods of non-trading and therefore, buy on Fridays and sell on Mondays (pp. 685-705). Some researchers also proposed the individual investor pattern as the driving spirit behind the weekend effect. This rationale is underpinned by the following: the tendency of institutional investors not to trade on Mondays and; a bias towards a buy recommendation by brokerages to individuals during weekdays and the absence of recommendations after Friday’s close and into the weekends, which thus leave individual investors to initiate a high possibility to sell on early Mondays (Brooks, 1997, p. 725-726; Sias and Starks, 1995, p. 58). Using intraday data of 276 firms in the year 1989, Brooks (1997) observed that the total dollar volume on Mondays is considerably less than that of the other trading days of the week, especially for medium to large-size trades. Small-size trades, on the other hand, have a higher trading activity on Mondays than any other day of the week with more than 50% of it selling than buying. The largest-size trades reflected the lowest volume of activities, with selling constituting 44.4% of such activities. Brooks (1997) concluded that both institutional and individual investors contribute to the weekend effects if small-size trades are made to represent individual investors and large-size trades to institutional investors. Specifically, Brooks (1997) sees their respective contributions as follows: the individual investors directly through their selling, and; the institutional investors indirectly through their absence, which lessens liquidity (pp. 725-735). Minimising the role of individual investors, Sias and Starks (1995) particularly zeroed in on institutional investors as the primary reason behind the so-called weekend effect. Institutional investors, as opposed to individual investors, are large corporations that have large cash reserves participating in stock and securities trading. Sias and Starks (1995) pinned their conclusion on the following assumptions: institutional investors use Monday to plan strategically for the week rather than engage in actual trading, thus, decreasing market depth; institutional investors received information before the public through the week and hence, have more information than other traders on Mondays; institutional investors, just like individual investors, also received asymmetric recommendations from brokers since a majority of them also use the help of investment professionals, and; some research suggested a close connection between weekend effects and autocorrelation in portfolio returns and on the other hand, researches suggesting serial correlation of institutional investors. Using NYSE data from 1977 to 1991 with similar-sized holdings differentiating between high and low institutional holdings, Sias and Starks (1995) tested whether stocks with higher institutional investors display comparatively higher weekend effects than other kinds and whether they exhibit greater return seasonality. The team concluded in the affirmative in both instances (pp. 58-59). The weekend effect, however, failed to be detected in Australia in the study conducted by Marrett (2008). He studied data from the year 1996 up to 2006 and found only one instance of the weekend effect at the industry level and none at the market level. This led him to conclude that the Australian market is weak-form efficient, which may be underpinned by such factors as growth in derivative markets, liberalisation/internalization of the domestic capital market and the presence of more institutional investors over individual investors engaged in trading activities (pp. 3-8). Some studies suggest that the weekend effect is either non-existent in some markets or is gradually disappearing. Chang (1993), who conducted a study of the subject using data from 23 countries, observed that only 13 of these countries reflected a weekend effect. Davidson and Faff (1999), on the other hand, concluded after a similar study on the Australian market that this anomaly is gradually vanishing (cited Lean et al, 2009, p. 3). Fortune (1999), who conducted a study on the weekend effect using a jump diffusion analysis similarly concluded that the difference between weekday and weekend performance is gradually diminishing over time and will likely correct itself in the long run (pp. 1-19). To validate the latter, Lean et al (2009) conducted a study using the non-parametric stochastic dominance approach (SD hereafter), which does not omit any information but directly takes into account all returns data, on the indices of the stock markets of Indonesia, Malaysia, Hong Kong, Singapore, Taiwan and Thailand. The team contradicted the findings of other researches that the weekday effect is disappearing after obtaining results showing weaker Monday returns and strong Friday returns dominating the aforesaid markets (pp. 1-27). 2.2.2 Long Term Anomalies: Gone Fishin’ Effect and Related Calendar Anomalies The gone fishin’ effect is another calendar anomaly and is associated with summer vacation. According to this phenomenon, there are comparatively lower trading returns during summer vacation than the non-summer months because for some reason, or another, the investors are preoccupied with other activities that are closely linked to this period. Although the term was just recently created by a group of researchers, it has its roots in some other calendar anomalies such as the Halloween effect, the holiday effect, the ‘sell in May and go away’ effect and a host of similar anomalies built around the same concept. In 2000, Kamstra et al conducted a study on the so-called seasonal affective disorder (SAD hereafter) that seems to emotionally influence traders and investors and their activities. According to that study, the shorter daylights during fall and winter make Americans depressed, the kind of depression that pushes them towards more risk-taking, thereby affecting market equilibrium. On the other end of the scales, SAD is said to also encourage risk aversion especially around winter time. Applying this psychology principle to stock markets, Kamstra et al (2000) set out to prove a pattern manifesting this behavior that would reveal unusually low returns before winter solstice and unusually high returns after winter solstice. An associated phenomenon is the sunshine effect, which purportedly creates an opposite mood in people, which is manifested when there are longer hours of sunshine than usual. Using the indices of seven countries such as the US, Canada, Britain, Germany, Sweden, Australia and New Zealand, Kamstra et al (2000) concluded at the end of their study that there is basis to support a finding that SAD affects indices of the world’s stock markets. Thus, in countries in the Northern Hemisphere, the tax-loss effect is seen in the month of January when SAD is specifically harsh whilst in Southern Hemisphere countries it is manifested in the month of July. In Australia, which is the closest to the equator of the countries studied, there was little evidence of the SAD effect (pp. 1-30). Bouman and Jacobsen (2002) also examined the “sell in May and go away” anomaly where May is believed to be the beginning of a bearish market because of the practice of traders to sell the stocks on that month until September. Thus, it is believed, according to this calendar anomaly, that stock returns are lowest in these summer months. Bouman and Jacobsen (2002) were able to establish that in 36 of the 37 countries they studied using MSCI data, the “sell in May” is present even as far back as 1694, that is there is substantial differences in return from the periods covered by the months of May to October and the rest of the year. Thus, whilst returns from October to April constitute the larger returns the May-October returns are no “different from zero and are often even negative” (p. 1630). Moreover, the team also noted that this effect persists not only in developed markets, but also in emerging ones, which cannot be said of other calendar anomalies such as the more popular January effect (pp. 1618-1635). Another “sell in May” or summer vacation effect study was earlier conducted by Marquering (2002) on the monthly indices of the UK, Germany, the Netherlands, Belgium and the US and found that the returns corresponding to the summer months of May until October are smaller compared to the returns in winter months or the months of November up to April. Moreover, he observed that while the January effect seems to be waning, the summer effect shows no indication of the likelihood of tapering off (pp. 557-576). Maberly and Pierce (2004) also did a study on the “sell in May and go away” effect on S&P futures covering the periods from 1982 to 2003 in the US and discovered that such an effect vanished after adjustments from the impact of several outliers such as stock market crash of 1987 and 1998 (pp. 29-42). The ‘sell in May and go away” is also known as the Halloween effect and was named by Bouman and Jacobsen (2002) and is said to be within the period from November through April characterised by lower volatility than the other months or periods of the year. Haggard and Witte (2010) conducted a study to determine whether the Halloween effect is what others accused it of: a result of outliers or the January effect in disguise. The team concluded that the Halloween effect was present only between 1954 and 2008, but not previous to that, which should not be the case for genuine anomalies. Moreover, the presence of a January effect renders it insignificant and so do the presence of outliers (pp. 379-387). The aforesaid studies preceded and some became the basis for the study conducted by Hong and Yu (2009) on the gone fishin’ effect. The gone fishin’ effect is actually very similar to the “sell in May” phenomenon because it focuses on the same summer vacation effect, except that it focuses on the months of July, August and September for the Northern Hemisphere and the months of January, February and March for the Southern Hemisphere countries. The primary study on this phenomenon was conducted by Hong and Yu (2009) who used seasonality in stock trading activity in connection with summer vacation to link volume and expected returns. The research team employed data from 51 countries between the periods of 1993-1999 to validate suspicions about a gone fishin’ effect. For Northern Hemisphere countries, summer is equated to the months of July, August and September and for countries in the Southern Hemisphere, summer means January, February and March. These countries come from the regions of Asia, Africa, Middle East, Europe, North America, Oceania and South America. Hong and Yu (2009) concluded that stock returns for the summer months of both Northern and Southern Hemisphere countries are significantly lower than the rest of the year, although countries near the tropics do not exhibit much of that phenomenon (pp. 672-702). Another study that goes by another name but has the same effect and feature as the gone fishin’ anomaly is the ‘school-is-out’ effect, which is the focus of research by Coakley et al (2007). There are differences, however, the first being that SO is shorter than gone fishin’ although both are built around the same concept of summer vacation. Coakley et al (2007) observed that study of Hong and Yu (2006) was not able to find a strong indicator of gone fishin’ in Asian countries, but the former explained this as stemming from the fact that tropical Asian countries do not necessarily observe the same period of summer vacation. Using data from some Asian countries with descending order latitude, viz. Korea, Japan, China, Taiwan, Hong Kong, the Philippines, Thailand, Singapore, Malaysia and Indonesia, the group concluded that share turnover, return volatility and stock returns were comparatively down during summer vacation from schools than when classes are being held and a contradicting pattern of high returns immediately after the period. At the same time, the team also tested for the gone fishin’ effect, which yielded the same results as that of Hong and Yu (2006). The team explained this as being evidence that in tropical Asian countries go on family vacations during school breaks that do not necessarily coincide with summer vacation (pp. 1-39). Closely allied with the gone fishin’ and summer effects is the earlier work of Lamb et al (1997), which catalogued stock market returns in the US before, during and after Congressional sessions. After conducting a study of 97 years of Dow Jones Industrial Averages (DJIA), the team concluded that there is evidence to prove that there is a connection between Congressional sessions and market behavior so that during the period when Congress is closed the market rises and the opposite trend can be evidenced when it is open. Lamb et al (1997) surmised that this could be due to the uncertainty that pending bills bring during sessions and the certainty of laws that have been resolved and passed just before Congress closes its sessions (pp. 19-25). The effect of a US holiday on European markets is the subject of study by Casado et al (2010). The group recently validated the reverse effect of the holiday anomaly in the European markets by studying DATASTREAM data of opening and closing prices between the periods of 1991 and 2008. The study shows that the returns in European markets are higher when the NYSE is closed because of a holiday as much as 15.53 times higher than when there is no such a holiday in the US describing the situation with a wry remark “when the US cat is away, the European mice will play” (p. 5). References: Agrawal, A. (1994) Anomalies or illusions: evidence from stock markets in eighteen countries. Journal of International Money and Finance. 13, pp. 83-106. Alagidede, P. and T Panagiotidis. Calendar anomalies in emerging market: evidence from the Ghana stock exchange. Ariel, R. (1984) A monthly effect in stock returns. Alfred P. Sloan of Management, MIT. Bouman, S. and Jacobsen, B. (2002) The Halloween indicator, “sell in May and go away:” another indicator. The American Economic Review. 92(5), pp. 1618-1635. Brooks, C. (2008) Introductory Econometrics for Finance. 2nd ed. Cambridge University Press, p. 454. Brooks, R. (1997) The individual investor and the weekend effect: a reexamination with intraday data. The Quarterly Review of Economics and Finance. 37(3), pp. 725-737. Brooks, C. and Persand, G. (2001) Seasonality in Southeast Asian stock markets: some new evidence on day-of-the week effects. Applied Economics Letters. Francis & Taylor, Cao, Z., Harris, R. and Wang, A. (2007) Seasonality in the returns, volatility and turnover of the Chinese stock markets. Finance Letters. 5(6), pp. 1-11. Casado, J., Muga, L. and Santamaria, R. (2010) The effect of US holidays on the European markets: when the cat’s away… Chen, H. and Singal, V. (2003) Role of speculative short sales in price formation: the case of the weekend effect. The Journal of Finance. 58(2), pp. 685-705. Coakley, J., Kuo, J. and Wood, A. (2007). The school’s out effect in East-Asian stock markets. Department of Accounting, Finance and Management and Essex Finance Centre, University of Essex, pp. 1-39. Cross, F. (1973) The behavior of stock prices on Fridays and Mondays. Financial Analysts Journal. 29(6), pp. 67-69. Fortune, P. (1999). Are stock returns different over weekends? A jump diffusion analysis of the ‘weekend effect.’ New England Economic Review, pp. 1-19. French, K. (1980) Stock returns and the weekend effect. Journal of Financial Economics. 8, pp. 55-69. Gao, L. (2005) Calendar effects in Chinese stock market. Annals of Economics and Finance. 7, pp. 73-88. Gibbons, M. and Hess, P. (1981) Day of the week effects and asset returns. Journal of Business. 54, 579-506. Haggard, K. and Witte, H. (2010) The Halloween effect: Trick or treat? International Review of Financial Analysis. 19(5), pp. 379-387. Hansen, P and Lunde, A. (2003) Testing the significance of calendar effects. Brown University, Department of Economics, Working Paper No. 2003-03, pp. 1-37. Heininen, P. and Puttonen, V. (2008). Stock market economies in the transition economies through the lens of calendar anomalies. EACES 10th Conference “Patterns of Transition and New Agenda for Comparative Economics”, Higher School of Economics, Moscow, Russia, pp. 1-47. Hong, H. and Yu, J. (2009) Gone fishin’: seasonality in trading activity and asset prices. Journal of Financial Markets. 12, pp. 672-702. Jacobs, B. and Levy, K. (1988) Calendar anomalies: abnormal returns at calendar turning points. Financial Analysts Journal. pp Jaffe, J. and Westerfield, R. (1985) The week-end effect in common stock returns: the international evidence. Journal of Finance. 40 (2), pp. 433-54. Kamstra, M., Kramer, L. and Levi, M. (2003) Winter blues: seasonal affective disorder (SAD) and stock market returns. Federal Reserve of Atlanta Working Paper No. 2002-13a, Sauder School of Business Working Paper, pp. 1-19. http://ssrn.com/abstract=208622 or doi:10.2139/ssrn.208622. Keim, D. and Stambaugh, R. (1984). A further investigation of weekend effect in stock returns. The Journal of Finance. 39 (3), pp. 819-835. Kunkel, R., Compton, W. and Bayer, S. (2003) The turn-of-the-month effect still lives: the international evidence. International Review of Financial Analysis. 137, pp. 1-15. Lamb, R., Ma, K.C., Pace, R. and Kennedy, K. (1997) The Congressional calendar and stock market performance. Financial Services Review. 6(1), pp. 19-25. Lean, H-H., Wong, W. and Smyth, R. (2007) Revisiting calendar anomalies in Asian stock markets using a stochastic dominance approach. Berkeley-NUS Risk Management Institute. Leontitsis, A. and Siriopoulos, C. (2006) Calendar corrected chaotic forecasts of financial time series. International Journal of Business. 11(4), pp. 368-374. Mabelry, E. and Pierce, R. (2004) Stock market efficiency withstands another challenge: solving the “sell in May/buy in Halloween” puzzle. Econ Journal Watch. 1(1) 29-46. Marquering, W. (2002). Seasonal predictability of market returns. Review of Business and Economics. 67(4), pp. 557-576. Marrett, G. and Worthington, A. (2008) The day-of-the-week effect and the Australian stock market: an empirical note on the market industry and small cap effects. Internal Journal of Business and Management. 3(1), pp. 3-8. Pena, J. (1995) Daily seasonalities and stock market reforms in Spain. Applied Financial Economics. 5, pp. 419-423. Rubinstein, M. (2000) Rational markets: yes or no? The affirmative case. UCBerkeley Research Program in Finance Working Papers RPF-294, pp. 1-24. Savva, C., Osborn, D. and Grill, L. (2006) The day of the week in fifteen European stock markets. University of Manchester, pp. 1-24. www.uni-konstanz.de/micfinma/conference/Files/.../Savva_Osborn_Gill.pdf - Scheurle, P. ( ). Predictability of the Swiss Market with respect to style, pp.18-19. Seyed, F., Abraham, F. and Al-Hajji, M. (2005) Seasonality in stock returns and volatility: the Ramadan effect. Research in International Business and Finance. 19, pp. 374-383. Sias, R. and Starks, L. (1995) The day-of-the-week anomaly: the role of institutional investors. Financial Analysts Journal, pp. 58-66. Sullivan, R. Timmerman, A. and White, H. (1998). Dangers of data-driven inference: The case of calendar effects in stock returns. UCLA Department of Economics. pp.1-47. Thaler, R. (1987) Anomalies: seasonal movements in security prices II: weekend, holiday, turn of the moth and intraday effects. The Journal of Economic Perspectives. 1(2), pp.169-177. Read More