Is The Small Firm January Effect An Anomaly
The main target of this paper is to find evidence of January effect especially in small size firms and to look for proper explanation for this phenomenon from existing literature. The structure of the paper divided into three parts, first introduce the January effect and review the first literature based on this issue by Donald Keim. The second part is some empirical evidence of January effect including different explanations for January, data base on this topic and some other study in international capital markets. The last part is based on the different aspects to the January effect to make a comparison between the viewpoints from literatures, and raised some further proposals that have not been given an explanation yet.
Small firms tend to generate excess returns in January especially in the first several trading day, since this phenomenon was first found by Keim in early 1980s, it become a compelling topic in financial markets. The phenomenon has got a consensus viewpoints and has been evidenced by many researchers, however, the reason for January effect is so controversial that even no one could give a certainly explanation.
The existing literature shows that no conclusive explanations were given, but some possible reasons for January effect were raised by different researchers. Some notable explanations include tax-loss selling hypothesis, liquid profits hypothesis, and behavior-based hypothesis. It seems that tax-loss selling hypothesis the most convinced explanation for January effect before 1986, however, after 1986 this explanation was weaken by tax reform carried out. Since then more studies believed the January effect is caused by behavior-based hypothesis.
To compare with the January effect in financial markets in developed countries, January effect is not found in some emerging market such as China, Malaysia. Some studies gave the explanation for this phenomenon is due to the different financial system and the markets were not mature enough.
The January effect is a big challenge to the efficient market hypothesis, and some paradoxes based on January also seemed so unexplained, so the January calls for further studies.
The January effect is a calendar-related anomaly in the financial market that the increase of security prices during the month of January especially in small size firms. The January Effect was first raised by Donald Keim.
Donald Keim (1983) study the relationship between market value and abnormal returns based on the data from NYSE and AMEX common stocks during the period of 1963 to 1979, and pointed the phenomenon that the means of daily abnormal return distribution in January were larger than the other eleven months, especially the first five trading days, and the firm size and abnormal returns always have a negative relation. And in his study he also listed several possible explanations for January effect based on existent literatures.
Tax-loss selling hypothesis.
The first one was based on Wachtel (1942) and Branch (1977). This two previews study indicated that the large returns in January were attributed to year-end tax loss selling of stock that have declined in value during the past year. Rozeff and Kinney (1976) also provide some evidence to the loss-selling hypothesis. However, Keim pointed this explanation have poor evidence support unless other countries have the similar tax codes system to the US's.
The second one was based on Rozeff and Kinney (1976). They pointed one possible explanation is that January is a special month because it is a signal of significant financial and information to release. And it is marks as an increased uncertainty and anticipation of new information.
3. Other explanation
Keim (1983) also gave some explanation by himself. He pointed that there may not have an economic cause to lead to the January. It is just because some spurious reasons such as the wrong database or outliers.
A further study by Tinic and West (1984) analyzed the risk and return of January and the other month base on Rozeff and Kinney (1976) and Fama and MacBeth (1973), using the data from 1935 to 1982, and they found a significant relationship between return and risk in January, however to the rest 11 months, there is not a significant risk premiums, this phenomenon reveals as January effect but Tinic and West didn't gave an explanation for it but remained it to whom would considered why only January shows up a consistently risk-return tradeoff.
According to the existing literature, January effect has become a compelling phenomenon but it is hard to give a determinate explanation.
2. Literature Review
2.1 Theoretical Evidence and Existing Explanation of “January Effect”
Rozeff and Kinney (1976) found evidence from NYSE that the average returns are higher than that of the other 11 months, what more, the further founding was that the January effect is generally a small firm phenomenon, however, in this paper, they didn't test the explanation for their discovery but raised some possible ramifications. One was tax-selling hypothesis. This hypothesis was supported by Reinganum (1982), his research pointed that the small firms which have excess returns tend to have a decline of price the previous year, because the firm's stock price declined in the previous year have a absence of selling pressure and the price would bounce up and lead to an abnormal increasing but he just thought that tax-loss cannot explain the January completely, just partially. Roll (1983) also provide some evidence that stocks which have higher returns in January tend to got a negative returns in the previous year. Branch (1977)'s original study shows that a company's yearend tax-loss selling activities and later on the repurchase activities caused the abnormal return on January. Shleifer (1986) study the relationship between demand curves and stock price, and pointed the demand curves for stock slope down hypothesis strongly corroborated to tax-loss selling hypothesis to explain the January effect. Ritter (1988) studied the buying and selling behavior of investors, and provided evidence that small firms which did tax-loss selling in the year end outperform other small firms, but it cannot explain why they outperform large firms. So he indicated the tax-loss selling can partially explain the January effect, not completely. However, Constantinides (1983) argued that yearend tax-loss selling hypothesis cannot be an explanation of January effect, Chan (1986), Pettengill (1986) and Jone et al. (1987) also provided inconsistent empirical evidence with tax-loss selling hypothesis. According to Kato and Schallheim (1985), tax-loss selling did not exist in Japan, however, Japan have January effect, the same research also based on Canadian capital market, Bergers et al. (1984) pointed that Canada just had capital gains tax after 1972, but Canada did have January effect before 1972.
Another group of works indicated that the explanation for January is Liquid profits hypothesis. Brock and Evans (1989) examined the small firm economics and they found that a excess returns of small firms in January, and gave a different explanation to this phenomenon, they demonstrate the January effect was due to the small size firms have a high liquidity in January after the Christmas sales period. And this viewpoint was similar to Brock (1988), who reported small firms have a fast liquidity dries up speed than large firms. Ritter (1988) in his study also considered that most small firm stocks were hold by individual investors while large firm stocks were hold by institutional investors, and small firm stocks earn more liquid profits at the yearend than large firm stocks. So he claimed that liquid profits hypothesis also a partial explanation for January Effect. A further study by Ogden (1990) also provides evidence to the liquid profits hypothesis. In his study he pointed that small firms tend to be more liquidity and obtain an overall aggregate liquid profits and this phenomenon is related to monetary policy.
In Rozeff and Kinney (1976)'s study, they raised another un-tested explanation, accounting information hypothesis. According to Chambers and Penman (1984), stock price behavior is related with reporting behavior. The stock returns variability related to the release of kinds of reporting, both expected and unexpected. And there exist an inverse relationship return and firm size; usually earning reports tend to have large reaction to small firms and small reaction to large size firms. Similarly, Atiase (1985)'s research also indicated that there is a inversely relationship between the capitalized value of the firm and stock price, and his explanation for the phenomenon is that due to the different capitalized value, the firms tend to got a different levels of effect of pre-disclosure information.
Some behavior-based hypothesis also mentioned to explain January effect. Bondt and Thaler (1987) using the US data from 1926 to 1982 to examined the winner firms and the loser firms, and they conclude the both kinds of firms showed the abnormal returns in January related with overreaction, they called this an overreaction hypothesis, but they admit that this hypothesis cannot explain the January effect satisfied. Haugen and Lakonishok (1988) indicate that the portfolio managers' behavior in the end of December and the start of January is the primary reason of January effect, it is a gamesmanship hypothesis. And their study was supported by Porter and Weaver (1998), which agreed that the abnormal returns in January is a result of hedging performance. Ritter (1988) also pointed that the January can partially explained by some behavior-based activities.
A more recent research by Haug and Hirschey (2006) reviewed the explanation to January and did some further study on this issue. They reviewed the January effect before 1986 and after 1986 which the year of the passage of the tax reform. They claimed that January effect before 1986 seemed can be explained by tax-loss selling activities, however the tax-loss selling hypothesis not stand for January effect after 1986 because no tax-motivated selling activities were allowed in the yearend. After 1986, the behavior-based explanations were more relevant to this issue.
2.2 Data set of the empirical study
Rozeff and Kinney (1976) used the data From New York Stock Exchange from 1904 to 1974 to examine the seasonal patterns in an equal-weighted index, and they found that the average monthly return in January was 3.79%, compared with January, the other months was only 0.5%. What more, they used two-parameter capital asset pricing model and discovered a large risk premium in January which were not significant in other months.
Reinganum (1982) focused on the relationship between tax effects and the excess returns of small firms in January. He used the daily data from University of Chicago's Center for Research in Security Prices (CRSP) from 1962 to 1979. He divided the small firm stocks to two parts, one is ‘winners' in the previous year, and the other is ‘losers' in the previous year. And this divide is related to tax-loss. His study showed that the losers' price outperform the winners in the first several trading days, especially the first day, which losers got an average 13.9%. He thought tax-loss can be one reason to lead the January effect but not the whole reason.
Schultz (1985) also tested the tax-loss hypothesis in explain the January effect. He examined the small firms' January effect and used data from 1900 to 1980, by measuring the small firm stock change and the Dow Jones Industrials change during 1900 to 1929, he found that there is no evidence of January effect before 1917, which the year of tax act, and in the year 1900 to 1918, the average Dow Jones Industries outperformed small firm stock in 10 of the 18 January during that period. However, after the year 1931, it appears excess returns in small size firms. But he claimed that although excess returns found after the tax act, but it cannot link the personal income taxes to the abnormal return in January, so Schultz (1985) just found the January but did not gave an explanation for that.
Bhardwaj and Brooks (1992) used data from NYSE and AMEX stocks which cross-classified by firm size during 1967 to 1988. The result showed a strong relationship between firm size and January returns, so they assert the January effect is a firm size effect, however, in their study when took the transaction costs and some other elements into consideration, no abnormal excess returns exists. And they argued that some prior studies were affected by incomplete consideration.
2.3 Empirical evidence of January effect in other regions
2.3.1Empirical evidence in UK stock market
Gultekin et al. (1983) studied the performance of stock return in 16 countries and found that a January effect in UK stock market. And Clare, Psaradakis and Thomas (1995)'s study, they examine the UK stock market and using monthly data of FTSE-A All Share from 1955 to 1990. And they found the January effect exists in UK and Clare and Thomas (1995) gave their explanation for the UK January is based on Bondt and Thaler (1987), due to the overreaction. Later study by Levis (2002) provides evidence of January effect in UK. Levis studied the anomalies in UK stock market and considered January effect is one of the anomalies, and also he indicated that the phenomenon is related to firm size.
2.3.2Empirical evidence in Chinese stock market
Li (2003) studied the January effect in Chinese stock market using the daily data from Shanghai Stock Exchange from January 1st 1993 to December 31th 2002 and daily data from Shenzhen Stock Exchange from January 1st 1994 to December 31th 2002. However, he found that the January effect was not exist in both Shanghai and Shenzhen stock market, and he gave an explanation that China does not have January is due to the different accounting system and a different economic circles. A later study by Liu (2004) provides the same conclusion, by he gave a different explanation is that there is no arbitrage opportunity in the yearend, and the Chinese capital market is not mature enough.
2.3.3Empirical evidence in other Asian regions
Campbell and Hamao (1992) tested the long-term predictability in Japan and compare to the US capital using the data from Tokyo Stock Exchange from 1971 to 1990 and data from Morgan Stanley Capital International World Index during the same period, and they found evidence that the stock returns in Japan highly related to stock returns in US and found January effect in Japanese stock market. Chui and Wei (1998) examined five emerging markets in Asian, and their discovery is interesting, they found that small size firms have larger January returns in Korea, however large size firm have more January returns in Hong Kong, and the firm size premium in January is significantly positive in Hong Kong and negative in Korea, and stock prices in Malaysia, Taiwan and Thailand were not related to the January effect. Later on another study by Lean, Smyth and Wong (2006) examined the January effect in Hong Kong, Indonesia, Japan, Malaysia, Singapore, Taiwan and Thailand using daily data from 1988 to 2002 based on SD test (Davidson and Duclos). However, they found only Japan and Thailand have the highest stock returns in January, other regions are not the highest in the 12 months, and Hong Kong even got a very low returns in January, that is -0.097%. According to their study, only Japan and Thailand got January effect.
3. Conclusion and Further Proposals
The January effect has become a notable phenomenon in not only stock market but also other financial instruments. Pinegar and Chang (1986) observed January effect in bond market, and later Wilson and Jones (1990) studied the corporate bond and commercial paper return by specifically analyzing for January using the past 131 years data and also found that January effect existing in both corporate bond and commercial paper. And they thought the explanations for January effect in corporate and commercial are so complex, but they also indicated two possible reasons, one is the income taxes and the other is gold standard in US. Some further study got the similar result. Maxwell (1998) found January effect especially exist in non-investment grade bonds.
So the January effect is a common phenomenon in many capital markets in developed countries (see Gultekin, 1983), however January effect is not exist in many emerging markets like China, Malaysia. The existing literature showed many possible explanations for January effect but no one is sure about a certain explanation. From the previous literature, the existing explanation is diversified in different period. The tax-loss selling hypothesis seems to overwhelming the other explanation like accounting information hypothesis, liquid profits hypothesis and behavior-based hypothesis before 1986, when the new tax law issued. However, the tax-loss selling hypothesis also cannot explain January effect perfectly; it cannot explain why Japan is a country without tax-loss selling but has the January (see Kato and Schallheim, 1985). And the similar took place in Canada (see Bergers et al., 1984). And after the year 1986, more research holds the opinion that January effect is caused by behavior-based activities and liquid profits.
Admittedly, although the tax-loss selling hypothesis has some defects, but it still considered as one of the most acceptable explanation for January effect, the causes of January effect is complex and uncertain, further study on this topic is necessary and meaningful, because it brought some new prospects to the efficient market hypothesis, it is a challenge to the EMH (see Malkiel and Burton, 2003). The January effect is still a ‘mystery' worthy further study.
An interesting question is the January effect offers investors a profitable trading strategy, according to (Malkiel and Burton, 2003), if investors buy stock in the end of December especially small size firm stock and sell them in the first two weeks of January, they could obtain an excess return, however, theoretically speaking, the small firms got a small trading volume and offer limited opportunities for individual investors, so the individual investors who want to earn a profit based on January could buy stock earlier in December and sell stock earlier in January, and if all the investors do so, the January would disappear. But the January effect now exists!
And according to the EMH, stock price should fully reflect the public available information, so theoretically, the January effect would not happen due to it is well known by all the investors.
However, no one gave these two questions a convinced explanation, it is puzzled, just as the same as January effect, still a mystery after 30 years it has been discovered.