ABSTRACT
This study tests the market efficiency of the South Korean stock market by examining returns on stocks of the constituents of the KOSPI 50 from 2000 to 2014 following large 1-month price decreases and increases. An exponential GARCH (EGARCH) event study framework is used to analyse the stock returns. The results show that large price shocks, positive and negative, are likely to be followed by positive market returns. Moreover, the results show an increase in the beta of stocks in the years following a large price shock. The overall results therefore support the Uncertain Information Hypothesis. However, beginning in 2008, return patterns more closely reflect those hypothesised by the Efficient Market Hypothesis, possibly due to increased participation by international investors. The observed returns following large price increases and decreases can be partially explained by changes in the Korean won to US dollar exchange rate and the trading behaviour of foreign investors.
Acknowledgements
The authors would like to thank the participants of the Financial Markets session at the Korea and the World Economy XIII Conference on 21-22 June 2014 in Seoul for their helpful comments and suggestions. They would also like to thank Byung-Joo Lee for his advice on an earlier draft of this paper and Seong Kook Kim and Jungsoon Shin helping with the data.
Notes
1 If an index member was excluded during a particular month, the firm was treated as if it were excluded during the entire month. Inclusions are treated in a similar fashion: if a firm was included during a particular month it was treated as if it had been an index member during the entire month.
2 In addition, stock return patterns after price increases and decreases in excess of 30% were investigated. No discernable patterns can be observed after increases. Price decreases, on the other hand, are followed by strong reversal patterns, confined mainly to short event windows.
3 In addition, by firm clustered regression analyses were conducted. The results are qualitatively the same to the ones presented, which are controlled for heteroscedasticity using the White statistic.
4 Since the variable IDEC is negative by definition, a negative sign for the coefficient will result in an overall positive effect on CARs.