Examining Event Study Methodologies in Foreign Exchange Markets
- PDF / 8,537,295 Bytes
- 36 Pages / 426 x 706.77 pts Page_size
- 66 Downloads / 214 Views
This content downloaded from 158.37.87.164 on Mon, 1 Sep 2014 05:51:43 AM All use subject to JSTOR Terms and Conditions
190 JOURNAL OF INTERNATIONALBUSINESS STUDIES, SECOND QUARTER 1990
differencesin results are found, caution will be needed in adopting a specific methodology based on tests of either common stock returnsor currencymarketreturnsif anythingotherthancommonstocksor currencies are being subjectedto event study analysis.The potential for either false acceptancesor rejectionsof tests for abnormalreturnsmay then not be determinableuntil a Brownand Warner-typesimulationis performedfor the specific type of security(or non-security)returnseries. Marketefficiencytestsof weakand semi-strongformshavebeen conducted in the foreign exchangearea. Among the semi-strongform tests, studies such as Frenkel[1981],Ito and Roley [1987]and Hardouvelis[1988]use the multipleregressionapproachemployingchangesin exchangeratesas the dependentvariablein testing the effects of public announcementson foreignexchangemarkets.On the other hand, Cosset and Doutriauxde la Rianderie[1985]and Sheffrinand Russell[1984]use the eventstudymethodology as an alternativeapproachin examiningthese effects. The results of the presentstudy will be useful to foreignexchangemarketefficiency tests which use the event study approach. The discussionof the paper is dividedinto six major sections. After this introductorysection, the second section reviewsthe literatureon event studies while the third section describes the experimentaldesign. The fourth section examinesthe propertiesof excess returnson event dates. Varioussimulationsare then conductedand reportedin the fifth.section. The findings are then summarizedin the last section. LITERATURE REVIEW
Throughsimulationprocedures,Brownand Warner[1980]examinevarious event study methodologies.Their objectiveis to determinethe power of variouseventstudytechniquesto correctlyclassifyboth the occurrenceand nonoccurrenceof abnormal events with common stock returns under differentenvironmentalconditions.Their 1980studyexamineseventmethodologiesusing monthlysecurityreturns.In Brownand Warner[1985]the authors extend their evaluation of event study methodologies to daily common stock returns. They examine three return estimation models returnmodel, includingthe mean-adjustedreturnmodel, a market-adjusted and a marketmodelthat employsan OLSprocedure.The resultsshowthat simplermodels areratherrobustin correctlyclassifyingthe occurrence,or nonoccurrence,of abnormalreturns.Brownand Warner[1980,1985]also examinethe sensitivityof the threemodelsto a host of otherenvironmental conditions. For example, they examine the impact of the size of the abnormalreturn,the certaintyof the exact date of the abnormalreturn, change in variance of returns where there is an abnormal event, and changesin samplesize. The currentpaperalso examinesmanyof the same conditionsexaminedby Brownand Warnerin additionto specific conditions that are found in the foreign exchangemarket.
This content downloaded from 158.37.87.164 on Mon, 1 Sep 2014
Data Loading...
