Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Standard view
Full view
of .
Look up keyword
Like this
0 of .
Results for:
No results containing your search query
P. 1
The Calendar Structure of The Japanese Stock Market

The Calendar Structure of The Japanese Stock Market

Ratings: (0)|Views: 46|Likes:
Published by calibrateconfidence
The Calendar Structure of The Japanese Stock Market
The Calendar Structure of The Japanese Stock Market

More info:

Published by: calibrateconfidence on Jul 04, 2013
Copyright:Attribution Non-commercial


Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less





The Calendar Structure of the Japanese Stock Market: The ‘Sell in May Effect’ versus the
, T
School of Commerce, Kwansei Gakuin University, Nishinomiya, Japan
Graduate School of Business Administration, Kobe University, Kobe, Japan and
Institute of Business and Accounting, Kwansei Gakuin University,Nishinomiya, Japan
We report on a seasonal pattern that has persisted in the Japanese stockmarket for more than half a century: Mean stock returns are significantlypositive for months during the first half of the calendar year and significantlynegative for months during the second half. Dubbed the
, this seasonality is independent of other known calendar anomalies,such as the so-called January effect. The
effect should bedistinguished from the ‘sell in May effect,’ because Japanese stocks performwell in June and poorly in November and December. The
effect varies in magnitude among firms and is particularly significant amongsmall firms with high book-to-market ratios. Nonetheless, the effect exists,regardless of a company’s size or book-to-market ratio.
 Japanese monthly stock returns are significantly higher during the January to June versus the July to December periods. We call this seasonality the
effect, after a traditional Japanese ballad,
which advocates that peoplework only the first half of the year and spend the second half in leisure. Out of our 59-year time span study of the popular index of Nikkei 225, the index
The authors thank the following people for their helpful comments: Kenya Fujiwara, NobuyukiIsagawa, Yasuo Kakuta, Hideaki Kato, Hideo Kozumi, Yusaku Sakaguchi, Kengo Shiroshita, HitoshiTakehara, Roger Van Noorden, Katsunari Yamaguchi, and an editor-in-chief, anonymous associateeditor, referee, and seminar participants at the Asian Finance Association International Conferencein Hong Kong, along with colleagues at Kwansei Gakuin University and Kobe University. TheGrants-in-Aid for Scientific Research by Japan Society for the Promotion of Science provided finan-cial support. Any errors are our own.
is a well-known folk song traditionally sung by people since the Edo era(1603–1868) in the Sasayama district, in western Japan.
 International Review of Finance,
13:2, 2013: pp. 161–185DOI: 10.1111/irfi.12003
© 2013 International Review of Finance Ltd. 2013
cumulatively advanced during the first half of the year while retreated duringthe second half in 39 years. The impact of this effect on stock returns wasconsiderable. From 1950 to 2008, the price-weighted Nikkei 225 showed acumulative gain of 3887.4% for a buy-and-hold strategy during just the first half of each year, versus a cumulative gain of 102.2% for just the second half of eachyear. Using the value-weighted Tokyo Stock Price Index (TOPIX), the disparitybetween these two strategies is even more dramatic: a gain of 3900.6% versus again of just 69.7%.
Studies of financial markets in several countries have documented empiricalregularities that appear to be inconsistent with the efficient market hypothesis.In the United States, Reinganum (1983) finds that small stocks outperform largestocks in January, while Tinic and West (1984) find that high-beta stocksoutperform low-beta stocks in January. French (1980) finds that returns onMondays are lower and are higher on Fridays, in what is termed the weekendeffect. Ariel (1990) finds that returns on the days before holidays are higher, theso-called holiday effect. Ariel (1987) also reports a monthly effect on stockreturns: Stocks are higher in the first half of the month and flat during thesecond half.As for empirical evidence regarding the Japanese stock market,Kato and Schallheim (1985) report that the January effect is indeed at workthere and Sakakibara (1994) confirms the presence of the weekend effect in theindex call options market. Bouman and Jacobsen (2002) report a ‘sell in May’effect in 36 of the 37 countries including Japan. Maberly and Pierce (2005)examine the Japanese popular price-weighted average index (Nikkei 225) andreport that the sell in May effect disappeared since the introduction of Nikkei225 index futures in September 1986.The findings reported in this article add to this list of regularities that areindependent of previously reported seasonal patterns. In particular, our findingsare independent of the well-known January effect because our results remainrobust even when January is excluded from the sample months. Although the
pattern is similar to that of the calendar anomalies reported byBouman and Jacobsen (2002), the Japanese stock market investor would bebetter off selling in July rather than ‘sell in May and go away.’
On the basis of our observation of 25 reference portfolios of similar size and book-to-marketratio, most Japanese stocks perform well until June, but lag from July to the endof the year. This also follows the ‘Twain effect’ (Twain 1894), in which Octoberis a dangerous month in terms of stock prices, but by no means the only one.The
pattern persists in various indices and reference portfolios,even since the internationalization of the Japanese market. In fact, in contrast
2 The return over each half-year is defined as the sum of one and the monthly return over thatperiod.3 According to the saying, the month signals the start of a bear market, so investors are betteroff selling their stocks in May and holding cash. The adage ends thus: ‘But buy back on St.Leger Day.’ St. Leger Day refers to a horse race run at Doncaster in England every September.
 International Review of Finance
© 2013 International Review of Finance Ltd. 2013
with the findings of Maberly and Pierce (2005), the
effectbecame slightly more conspicuous after the index future was introduced.
The study of such patterns often follows a path in which the popular pressmentions a supposedly profitable trading rule, which in turn prompts a schol-arly inquiry. The regularity reported in this article, however, was first docu-mented in our working paper and then reported in the popular Japanese press,suggesting that the pattern in question is not well known among investors inthe Japanese stock market.
This article was organized as follows. Section 2 reports the results of severaltests that indicate a half-year pattern in Japanese stock market returns. Section3 discusses these results and considers possible biases that could be responsiblefor the observed effect. Section 4 concludes the article.
A. Monthly returns of various indexe
To represent the returns accruing to stocks, the following tests employ theNikkei Economic Electronic Database Systems (NEEDS) Financial Quest toobtain the returns of the value-weighted TOPIX and the price-weighted Nikkei225 index, the two most commonly quoted Japanese stock indexes. In addition,we obtain the Tokyo Stock Exchange First Section Arithmetic Stock PriceAverage and the Nikkei All Stock Average. The data span the years 1950–2008(708 months) for both the TOPIX and Nikkei 225.The data portrayed in Table 1 and Figure 1 indicate the superiority of thetrading environment in Japan in the first half of the year compared with thesecond half. When each trading year is divided into two halves, the meanmonthly return for the first half is significantly greater than the mean monthlyreturn for the second half. Indeed, the first-half monthly mean is alwayspositive, while the second-half monthly mean is negative in some indexes. The
-statistics for the difference in the mean monthly returns for the two popula-tions are 1.910 for the price-weighted index and 2.168 for the value-weightedindex.Figure 1 is a graphic representation of Table 1, which shows the statistics forthe entire period under study. We calculate the mean monthly returns for boththe TOPIX and the Nikkei 225 since January 1950, which includes Japan’spostwar high-growth period. Other indexes cover the maximum period, as longas the data are available. The Tokyo Stock Exchange First Section Arithmetic
4 The Nikkei 225 index performs 1.0% per month better, on average, in the first half of the yearthan the second half for the period 1970–1986. During 1987–2008, the average monthlydifference between the first and second halves of the year was 1.1%.5 The half-year seasonality in the Japanese stock market was first reported by our working paperon September 29, 2003; subsequently, the first article mentioning this Japanese stock marketseasonality appeared in the popular
Nihon Keizai Shimbun
on January 15, 2009.
Calendar Structure of the Japanese Stock Market 
© 2013 International Review of Finance Ltd. 2013

You're Reading a Free Preview

/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->