Quantitative finance collector
C++ Matlab VBA/Excel Java Mathematica R/Splus Net Code Site Other
Oct 27

Sell in May and Go Away: Evidence from China

Posted by abiao at 01:04 | Paper Review | Comments(0) | Reads(9543)
I have co-authored a short paper with a friend in Zhejiang University, forthcoming in the Finance Research Letters, titled "Sell in May and Go Away: Evidence from China".

Using the Chinese stock market data from 1997 to 2013, this paper examines the “Sell in May and Go Away” puzzle first identified by Bouman and Jacobsen (2002). We find strong existence of the Sell in May effect, robust to different regression assumptions, industries, and after controlling for the January or February effect. However, part of the puzzle is subsumed by the seasonal affective disorder effect. We then construct a trading strategy based on this puzzle, and find that it outperforms the buy-and-hold strategy and could resist the market downside risk during large recession periods.


As the abstract suggests, basically we aim to examine whether the sell-in-may phenomenon existed in developed country also happens in China, and if Yes, if there is any special reason to explain it, which has implications for those international investors as MSCI plans to add Chinese A shares to its emerging index from May 2015, and as the recent China's stock market plan that permits Hong Kong investors to trade designated stocks in Shanghai Exchange market directly. People would expect investing in China provides a diversified strategy.

“Sell in May and Go Away” puzzle means that stocks have higher returns in the November-April period than the May-October period, in this paper we first run a dummy regression that assign dummy=0 when the date t is in the May-October period, and dummy=1 when otherwise. We find the dummy variable is highly significant, not driven by a specific industry, and cannot be explained by well-known January or February effect, nor by time-varying risk, nevertheless, time-varying risk aversion approximated by the SAD (seasonal affective disorder) effect by Kamstra, et al. (2003) subsumes part of the Sell in May effect.

Then we test whether such a phenomenon could generate any economic benefit, We construct a trading strategy that buys the Chinese stock market at the beginning of November and sells it at the end of April of the next year. We save the capital in a bank earning a risk-free floating deposit rate from the beginning of May to the end of October . Our benchmark is a buy-and-hold strategy. This simple trading strategy is shown to outperform the buy-and-hold strategy and can protect investors from dramatic losses during large recession, as shown in belowing Table and Figure.

  Sell in May strategy  Buy-and-hold strategy
Return  13.03%  7.50%
Sharpe ratio  0.6002  0.2199
Maximum drawdown  27.00%  69.30%
Downside deviation  2.98%  5.34%
Historical VaR (95%)  6.86%  11.20%
Leland’s alpha  8.69%  

The short paper is at http://www.sciencedirect.com/science/article/pii/S1544612314000579


Add a comment
Emots
Enable HTML
Enable UBB
Enable Emots
Hidden
Remember
Nickname   Password   Optional
Site URI   Email   [Register]