![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
Economic theory in the context of floating exchange rates has focussed on underlying medium and long term directions of exchange rate movements. Daily volatility is less well understood. One theory that offers an explanation for short-term exchange rate movements is that of the efficient market hypothesis or EMH. Its application to the forex market allows exchange rate movements to be understood as the reaction of traders to relevant news. In an efficient market traders react to news and specifically to surprise news events which necessitate a re-evaluation of the currency value. We test for the validity of this hypothesis in the context of the daily rand/dollar forex market over a three-year period, adding an emerging market case to the literature. We test the significance of macroeconomic news surprises -measured by the difference between actual and forecast data - in driving daily exchange rates. We find that surprises in both real and nominal variables cause a statistically significant reaction in the exchange rate. The results support an asymmetry between news of different origin as only surprises that originate in the U.S. prove significant. Good news also seems to receive greater attention from traders than bad news in our sample. Finally, we find that the statistical significance of variables is time-varying.
Notes
1 Famously, CitationJohansen and Juselius (1992) demonstrated the operationalization of the VECM estimation strategy which distinguishes clearly between long-run equilibrium relationships and short-run dynamics on the estimation of a model that combined PPP and UIP theory for the UK.
2 The link to CitationFama (1970, Citation1991) efficient markets is immediate.
3 See CitationAlmeida et al. (1998), CitationAndersen and Bollerslev (1998), CitationEdison (1996), CitationFaust et al. (2003), Galati and Ho (2003), CitationHarris and Zabka (1995), CitationPearce and Solakoglu (2003), CitationPrast and de Vor (2001).
4 See CitationAlmeida et al. (1998), CitationAndersen and Bollerslev (1998), CitationEdison (1996), CitationGalati and Ho (2001), and Ito and Roley (1996).
5 See CitationGalati and Ho (2001).
6 See CitationAlmeida et al. (1998), CitationAndersen and Bollerslev (1998), CitationFaust et al. (2003), CitationGalati and Ho (2001).
8 See CitationAndersen and Bollerslev (1998), CitationEdison (1996), Engel and Frankel (1985).
9 For instance Harris and Zabka (1985) point to the importance of US employment data, and CitationAndersen and Bollerslev (1998) to the importance of real variables in the US, and nominal variables in Germany.
10 See CitationAlmeida et al. (1998), CitationAndersen and Bollerslev (1998), CitationPearce and Solakoglu (2003).
11 We have not been able to locate a study which does so.
12 Part of the reason for the volatility was contagion from the Asian crises of the late 1990's. See the discussion in Fedderke (2004). But this is only part of the story, and much remains to be done in trying to understand short term movement in the South African currency.
13 A further candidate variable was the dollar/euro exchange rate. However, since the regressors included in the study already included news events from the US relevant to the $/€ rate, and since a range of studies have shown US to dominate European news events in exchange rate movements, we chose not to include the $/€ rate.
14 The authors are grateful to Donovan Byrne and Factiva.com for access to historic Reuters and Dow Jones articles as well as exchange rate data free of charge. Actual data announcements were taken from the same sources via the website of Factiva.com.
15 Given the weak empirical link between M3 and inflation, in this case the interpretation of the surprise event requires corresponding modulation.
16 Possibly due to the high interest rate differentials with the U.S. and resultant capital inflows into South Africa.
17 Which loads on the appreciation.
18 Recall that data are daily. We tested for numerous alternative lag structures, without obtaining substantively different results. Full results are available from the authors on request.
19 See CitationAlmeida et al. (1998), CitationAndersen and Bollerslev (1998), CitationPearce and Solakoglu (2003).
20 Full results are available from the authors on request.
21 Recall that the definition of “good” and “bad” here was event-specific, and relied on a careful consideration of prevailing market perceptions at the time of the event. Running the good/bad news regression over the shorter frequency exchange rate data we constructed, did add good news about the SA money supply and good news about non-farm payrolls to the list of significant variables though it renders good news about U.S. industrial production insignificant. This was one instance where results were marginally affected by the use of slightly higher frequency data.
22 These thus cover the 1/6/2001-30/12/2001, 1/1/2002-30/6/2002, 1/7/2002-30/12/2002, 1/1/2003-30/6/2003, 1/7/2003-30/12/2003, 1/1/2004-30/6/2004 time periods.