Introduction

Share prices are difficult to predict since prices always reflect relevant information according to the efficient market hypothesis (Fama, 1965, 1991, 1998). However, this viewpoint is challenged by the overreaction hypothesis (Debondt & Thaler, 1985, 1987), which supposes that (i) investors may overreact to the available and private information released owing to excessive self-confidence (Chuang & Lee, 2006; Daniel, Hirshleifer, & Subrahmanyam, 1998), (ii) herd behaviors result in chasing rising or falling prices (Chalmers, Kaul, & Phillips, 2013; Mendel & Shleifer, 2012), and (iii) investors’ sentiments and behavioral pscyhology have an impact on price movements (Da, Engelberg, & Gao, 2015; Dergiades, 2012; Huang et al., 2014; Huang & Ni, 2017; Kumar, Page, & Spalt, 2013).

In this study, in divergence from the commonly explored causal relationship between share prices and exchange rates (Ajayi & Mougouė, 1996; Basher, Haug, & Sadorsky, 2012; Dornbusch, 1987; Grammig, Melvin, & Schlag, 2005; Pan et al., 2007a, b), we mainly investigated whether market participants are able to beat the market. Instead of employing time-series models, we used the standard event study approach to examine whether investors are able to gain abnormal returns (ARs) and cumulative abnormal returns (CARs). Specifically, we explored whether market participants outperform the markets right after a sharp rise or fall in exchange rates represented by USDX, GBP/USD, and USD/CNY.Footnote 1

We included the constituent stocks of the Dow Jones Industrial Average 30 index (DJ 30), Financial Times Stock Exchange 100 index (FTSE 100), and the Shanghai Stock Exchange 50 index (SSE 50) as our samples, since investors are able to buy the constituent stocks of these stock indexes instead of index spots. We then explored whether our results are different among these representative stock markets.

We define sharp rises or falls in exchange rates (i.e., the rises or falls in exchange rates over 0.25%, 0.5%, 0.75%, and 1%) as our events. The events explored in this study are different from those explored in other studies, such as the declaration of dividend yield (Charest, 1978; Firth, 1996), merger or acquisition (Barney, 1988; Hackbarth & Morellec, 2008), and stock repurchase (Lie, 2005; Netter & Mitchell, 1989), among others. That is, we tested whether investors are able to beat the market following sharp movements in the exchange rate instead of these other events tested in previous studies. Our study expands the applications of the event study approach.

The contributions of this study are as follows: first, we explored whether a sharp rise (fall) in exchange rates provides valuable and unrevealed information to market participants for trading our selected sample stocks, which is rarely examined in the literature. Second, we compared the profitability of trading the constituent stocks of the DJ 30, FTSE 100, and SSE 50, as these three stock markets are representative markets. Third, this study links the occurrence of sharp rises (falls) in exchange rates with market efficiency and inefficiency, overreaction, investors’ sentiments, and herd behaviors, which is not adequately studied in the existing literature.

We found that share prices rose when the USD and GBP depreciate or appreciate sharply, and positive CARs are observed without exception. Share prices were more volatile following sharp movements in the CNY, especially with sharp depreciation. Our results are beneficial for investors who trade these constituent stocks around occurrences of sharp currency movements. We argue that these stock markets might be inefficient, as sharp movements in exchange rates result in stock price overreaction possibly because of investors’ heightened sentiments and/or herd behaviors.

The rest of the study is organized as follows: we survey the relevant literature in the second section. The third section describes the study design and data. The fourth section presents the empirical results. The concluding remarks and recommendations are provided in the final section.

Literature review

We reviewed the relevant literature related to market efficiency, market inefficiency, herd behaviors, and positive feedback trading. Given that whether investors are able to beat the market around sharp movements in exchange rates is related to investing strategies, we also reviewed the relevant literature related to investing strategies and the relationship between exchange rates and stock markets.

With regard to market efficiency and inefficiency, we found no consistent conclusion documented in the literature (Debondt & Thaler, 1985, 1987; Fama, 1965, 1991; Kahneman & Tversky, 1979; Malkiel & Fama, 1970; Nguyen, 2016). Technical trading rules that are widely applied in the real world seem to challenge the market efficiency hypothesis (Bessembinder & Chan, 1995; Brock, Lakonishok, & Lebaron, 1992). Schwert (2003) argued that there exist anomalies that seem to be inconsistent with the market efficiency hypothesis, indicating that market inefficiency likely exists in the stock markets.

Herd behaviors, such as simultaneous buying or selling of the same stocks by many investors, have been reported (Balcilar, Demirer, & Hammoudeh, 2013; Economou, Kostakis, & Philippas, 2011; Frey, Herbst, & Walter, 2014; Grinblatt, Titman, & Wermers, 1995; Lakonishok, Shleifer, & Vishny, 1992; Ni, Liao, & Huang, 2015; Nofsinger & Sias, 1999; Wu, Huang, & Ni, 2017a), which indicates that momentum strategies may be useful in trading. Herd behaviors are often observed in trading by institutions, such as mutual and pension funds (Brown, Wei, & Wermers, 2013; Dennis & Strickland, 2002; Kremer & Nautz, 2013; Raddatz & Schmukler, 2013; Sias, 2004; Zheng et al., 2015). Griffin, Harris, and Topaloglu (2003) found that strong contemporaneous daily patterns can be largely explained by net institutional trading following the past intraday excess stock returns.

Positive feedback trading refers to buying past winners and selling past losers, in which investors are likely to chase rising (falling) prices, which, in turn, will result in share prices rising (falling) continuously. Regarding positive feedback trading, Nofsinger and Sias (1999) showed that institutional investors engage in more positive feedback trades than individual investors. Koutmos and Saidi (2001) also reported evidence of positive feedback trading and selling/buying during market declines/advances in developed stock markets. Kurov (2008) showed that traders in the index futures markets are positive feedback traders. Dennis and Strickland (2002) also found positive feedback trading by some institutions, particularly mutual and pension funds. Wang (2002) found that large hedgers behave like positive feedback traders, but small traders are liquidity traders. Antoniou, Koutmos, and Pericli (2005) supported the view that futures markets help stabilize the underlying spot markets by reducing the effect of feedback traders, thereby attracting more rational investors who make the markets more informationally efficient and providing investors with superior ways of managing risk.

Regarding momentum strategies, Menkhoff et al. (2012) reported a significant cross-sectional spread in excess returns between past winner and loser currencies in the foreign exchange market. Ni, Liao, and Huang (2015) found that investors might adopt momentum strategies in trading stocks in the Chinese stock exchanges because of over-optimism that likely exists because individual investors account for 80% of the trading volume. With respect to contrarian strategies, De Haan and Kakes (2011) pointed out that several institutional investors in the Netherlands, such as pension funds and life and non-life insurers, tended to be contrarian traders. Malin and Bornholt (2013) found that the reversal of long-term returns in international equity markets may be stronger and more persistent than is generally understood, indicating that contrarian strategies might be appropriate in long-run investments.

Concerning the relationship between exchange rates and stock markets, Aggarwal (1981) found that US stock prices and the dollar were positively correlated. Donnelly and Sheehy (1996) documented a significant contemporaneous relation between exchange rates and the market values of large UK exporters. Bahmani-Oskooee and Sohrabian (1992) showed that there is bidirectional causality between the S&P 500 stock index and the effective exchange rates of the dollar. Chiang and Yang (2003) found that stock returns and currency values were positively related in nine Asian markets.

However, Ajayi et al. (1999) showed unidirectional causality from the stock markets to the currency markets for advanced economies but not for emerging economies. Ramasamy and Yeung (2002) found that the results are inconsistent for bivariate causality between stock prices and exchange rates. Bartov and Bodnar (1994) did not find a significant contemporaneous relation between USD movements and stock returns for US firms. Griffin and Stulz (2001) found that weak exchange rate shocks had negligible impacts on the performance of industries in six industrialized countries.

The relationship between exchange rates and stock markets has been explored in numerous studies (Aggarwal, 1981; Ajayi et al., 1999; Bahmani-Oskooee & Sohrabian, 1992; Bahmani-Oskooee & Saha, 2017; Bartov & Bodnar, 1994; Chiang & Yang, 2003: Donnelly & Sheehy, 1996; Hamao et al., 1990; Jorion, 1990; Griffin & Stulz, 2001; Karolyi, 1995; King & Wadhwani, 1990; Lin, 2011; Lin et al., 1994; Ramasamy & Yeung, 2002; Schwert, 1990; Theodosiou & Lee, 1993). The market efficiency hypothesis has been challenged by researchers (Chuang & Lee, 2006; Daniel et al., 1998) who claim that overreactions, herd behaviors, positive feedback, investors’ sentiments, and other factors oppose the hypothesis (Antoniou, Koutmos, & Pericli, 2005; Baker & Wurgler, 2006; Chiang & Zheng, 2010; Debondt & Thaler, 1985, 1987; Kurov, 2008; Liao, Huang, & Wu, 2011; Nofsinger & Sias, 1999; Scharfstein & Stein, 1990; Schmeling, 2009).

We argue that currency appreciation (depreciation) results from capital inflows (outflows) that result in rising (falling) share prices in accordance with the viewpoints about the behaviors and impacts of the capital account (Calvo, Leiderman, & Reinhart, 1993, 1996; Wu, Huang, & Ni, 2017b). However, currency depreciation (appreciation) might also increase exports (imports) by enhancing (reducing) the competitiveness of domestic products, resulting in rising (falling) share prices based on the viewpoints about the current account (Chinn, 2006; Corsetti et al. 2000; Fang, Lai, & Miller, 2006; Gruber, & Kamin, 2007; Zia & Mahmood, 2013).Footnote 2

Aside from the above mechanism, the short-term stock price performance (CARs) might be affected by the interpretation of individual investors and institutional market participants. In addition, short-term share price performance is likely affected by the heightening of investors’ sentiments due to sharp rises (falls) in exchange rates (Heiden, Klein, & Zwergel, 2013; Rehman, 2013), which might result in herd behaviors as per the theories of behavioral finance (Bekiros et al. 2017; Ramiah, Xu, & Moosa, 2015). Accordingly, we argue that investors’ sentiments would be another mechanism. We regard that share prices are likely affected by not only economic factors but also psychological factors. In sum, we examined the subsequent performance of trading stocks immediately after sharp movements in the USDX, GBP/USD, and USD/CNY. This relationship is insufficiently explored in the existing literature.

Study design and data

Design of this study

We argue that a sharp rise (fall) in exchange rates conveys valuable information. We first classified sharp rises (falls) in exchange rates according to the thresholds of 0.25%, 0.5%, 0.75%, and 1% as our events. We then measured the one-, two-, three-, four-, and five-day cumulative abnormal returns around the occurrence of these events. We used the standard event study approach to examine stock price movements following these events.

We defined day 0 as the day when the above-mentioned events occurred. Abnormal return was defined as the difference between the actual return and the expected return generated by the market model. We used three indices, namely DJ 30, FTSE 100, and SSE 50, as the proxies for market returns and then estimated the parameters of the market model by using data over the period from − 155 to − 6 days (i.e., 150 days) before day 0.

Cumulative abnormal returns at t days (CAR(t)) are the cumulative abnormal returns from AR (1) to AR (t). Given that investors might hold stocks for a short period, we measured one-, two-, three-, four-, and five-day CARs considering them as short windows in line with the literature (Ahern, 2009; Chopra, Lakonishok, & Ritter, 1992; Marshall, Young, & Rose, 2006). Finally, we tested whether these CARs are statistically different from zero based on their t statistics.

We present the empirical results for the constituent stocks of the DJ 30, FTSE 100, and SSE 50 in Columns 1–3 of Tables 3 and 4 in Section 4. The CAR results associated with the sharp rises and falls in exchange rates are shown in Panels A–D of Tables 3 and 4, respectively.

Data

We used the daily data for USDX, GBP/USD, and USD/CNY and the constituent stocks of the DJ 30, FTSE 100, and SSE 50 from Datastream (a financial database) from 2006 to 2015. Figure 1 shows an upward trend in the US stock market after the 2008 stock market crisis, a stable trend in the UK stock market, and two peaks, in 2007 and 2015, for the China stock market. Figure 1 indicates that the China stock market is more volatile than the US and UK stock markets; however, the volatility of USD/CNY seems less than that of USDX and GBP/USD, as shown in Fig. 2.

Fig. 1
figure 1

Trends of DJ 30, FTSE 100, and SSE 50 indices from 2006 to 2015

Full size image
Fig. 2
figure 2

Trends of three currencies from 2006 to 2015

Full size image

Table 1 lists the means, medians, standard deviations, minima, and maxima for the currencies and representative indices for data over the period of 2006–2015.

Table 1 Summary statistics. This table reports the means, standard deviations, maxima, and minima for USDX, GEP/USD, and USD/CNY as well as DJ30, FTSE100, and SSE50 indices over the data period 2006–2015
Full size table

Panel A of Table 1 shows that the ranges of these stock indices are rather wide as revealed by their minima and maxima, especially for the China stock index. The minima may be a result of the global financial crisis in 2008. USD/CNY is less volatile than USDX and GBP/USD, as shown in Panel C of Table 1. These statistics suggest that the China stock market is more volatile than the other two markets, whereas the currency volatility of the CNY is less than the USD and GBP. The results indicate that the stock market in China seems to be sensitive to sharp currency movements even though exchange rate movements are managed by the government. We argue that these statistics might contain valuable and unreleased information for stock trading.

We investigated the association between currency movements and stock market performance and examined whether investors are able to beat the markets in the trading of the constituent stocks of the DJ 30, FTSE 100, and SSE 50 right after a sharp rise (fall) in the USDX, GBP/USD, and USD/CNY. We defined our events as the rises (falls) in exchange rates of more than 0.25%, 0.5%, 0.75%, and 1%.Footnote 3 We have listed the numbers of the sharp rise (fall) in exchange rates that occurred for the constituent stocks of these indices over the data period in Table 2.

Table 2 The samples of the sharp rise (fall) in exchange rates. Panel A of Table 2 lists the numbers of the sharp rise in exchange rates (USDX, GBP/USD, and USD/CNY) over 0.25%, 0.5%, 0.75%, and 1% for the constituent stocks of DJ 30, FTSE 100, and SSE 50. Panel B of Table 2 lists the numbers of the sharp fall in exchange rates over 0.25%, 0.5%, 0.75%, and 1% for the constituent stocks of DJ 30, FTSE 100, and SSE 50
Full size table

The samples of the sharp rise in exchange rates for the four thresholds (i.e., over 0.25%, 0.5%, 0.75%, or 1% rise) are similar to those of the sharp falls, indicating that the currency appreciation and depreciation shocks occurred similarly over the data period from 2006 to 2015. An over 1% fall did not occur in the data period for China, which might be a result of the central management of the CNY. Sharp movements in the GBP occurred often in the United Kingdom, which might be related to the share prices and constituent companies of the FTSE 100.

Empirical results and analysis

We measured the one-, two-, three-, four-, and five-day CARs around sharp exchange rate movements. We have presented the CARs and their t values for the constituent stocks of the DJ 30, FTSE 100, and SSE 50 immediately after sharp rises and falls in exchange rates in Tables 3 and 4, respectively.Footnote 4

Table 3 CARs of diverse sharp rises in exchange rates. We investigate whether these CARs including 1-, 2-, 3-, 4-, and 5-day CARs would be different from 0 if investors take the long positions in the constituent stocks of DJ 30, FTSE 100, and SSE50 as the sharp rise in exchange rates over 0.25%, 0.5%, 0.75%, and 1% in a day. We also present the statistics of t-tests for these CARs. In addition, *, **, and *** represent 10%, 5%, and 1% significance levels, respectively
Full size table
Table 4 CARs of diverse sharp falls in exchange rates. We investigate whether these CARs including 1-, 2-, 3-, 4-, and 5-day CARs would be different from 0 if investors take the long positions in the constituent stocks of DJ 30, FTSE 100, and SSE50 as the sharp rises in exchange rates over 0.25%, 0.5%, 0.75%, and 1%. We also present the statistics of t-tests for these CARs. In addition, *, **, and *** represent 10%, 5%, and 1% significance levels, respectively
Full size table

Results of sharp rises in exchange rates

Table 3 shows that most of the CARs are positive for the constituent stocks of the DJ 30 and FTSE 100 for over 0.25%, 0.5%, 0.75%, and 1% rises in the exchange rates in a day, especially for CAR (4) and CAR (5). The results indicate that taking long positions might be appropriate for these stocks for four to 5 days after a sharp appreciation of the USD and GBP.

As for the result for China, most of the CARs are negative for the constituent stocks of the SSE 50 for over 0.25%, 0.5%, 0.75%, and 1% rises in the exchange rate in a day. Especially, CAR (5) was − 2.04% and − 5.10% for over 0.75% and 1% rises in the exchange rate in a day, respectively, which are sizeable movements. The results indicate that taking a short position would be appropriate for trading the constituent stocks of the SSE 50 after a sharp depreciation of the CNY.

In summary, the CARs rise after the appreciation of the USD and GBP, whereas the CARs decline after the depreciation of the CNY.

Results of sharp falls in exchange rates

Table 4 shows that almost all of the CARs are positive for trading the constituent stocks of the DJ 30 and FTSE 100 for over 0.25%, 0.5%, 0.75%, and 1% falls in the exchange rate in a day, implying that a contrarian strategy might be appropriate for trading the constituent stocks of the DJ 30 and FTSE 100 following a sharp appreciation of the USD and GBP. However, the highest CAR was less than 1%, indicating that the profits might be limited for this trading strategy.

As for the result for China, we found positive and significant CARs following sharp falls in the exchange rate in China, indicating that the momentum strategy may be productive for trading the constituent stocks of the SSE 50 following more than 0.25%, 0.5%, and 0.75% falls in the exchange rate in a day.

A fall of more than 1% in the exchange rate did not occur over the data period from 2006 to 2015 in China; as such, there are no results to be shown in Panel D of Table 4. We infer that the low volatility of the CNY stems from the actions of the monetary authority of China.

Similar to the results for the United States and United Kingdom, the CARs are positive following falls of more than 0.5% and 0.75% in the exchange rate in China. Furthermore, the CARs in China are slightly higher than those in the United States and United Kingdom, especially for CAR (5). For example, CAR (5) was 1.35% and 0.77% following 0.5% and 0.75% falls in the exchange rate for China, which are higher than those shown for the United States and United Kingdom.

As all of the CARs are positive without exception for the United States and United Kingdom, the results indicate that taking a long position would be appropriate for trading the constituent stocks of the United States and United Kingdom following both sharp currency appreciation and depreciation for these two countries. However, the CARs are all less than 1%.

For China, significantly negative CARs were observed after a sharp depreciation of the CNY. CAR (5) was − 5.01% for the constituent stocks of the SSE 50 after a 1% drop in the CNY in a day, whereas it was 1.35% after a 0.5% appreciation in the CNY. These CAR results are more volatile than those for the United States and United Kingdom. In summary, for China, share prices increased after a sharp appreciation in the CNY, and they weakened after a sharp depreciation in the CNY. We infer that the SSE 50 in not as mature as the DJ 30 and FTSE 100, resulting in market inefficiency.

Conclusions

We hypothesized that heightened sentiments of investors’ following sharp exchange rate movements in a day would impact the movements in the respective stock markets. We examined whether investors would realize profits in trading the constituent stocks of the DJ 30, FTSE 100, and SSE 50 right after sharp appreciations or depreciations of over 0.25%, 0.5%, 0.75%, and 1% rises in the respective exchange rates in the countries. We regard that such currency movements have informational value for market participants.

Our hypotheses are likely connected to market efficiency or inefficiency. Sharp changes in exchange rates might cause herd behaviors and overreactions among market participants. The theories of behavioral finance are applicable to this situation. Our results can help guide investing and trading strategies.

By studying the constituent stocks of the DJ 30, FTSE 100, and SSE 50, we found that the prices of the stocks in those indexes rose following sharp appreciation and depreciation in the USDX and GBP/USD. In the case of China, although the CNY is less volatile than the other two currencies, the share prices are more volatile than the other two markets. CAR (5) was − 5.01% for trading the constituent stocks of the SSE 50 after a 1% sharp depreciation in the CNY, whereas it was 1.35% after a 0.5% sharp appreciation in the CNY.

Our study adds to the existing literature in several ways. First, to our knowledge, whether market participants can beat the market following sharp exchange rate movements is seldom explored in previous studies. Second, our results can help devise investing strategies triggered by sharp currency movements. We used the constituent stocks of the DJ 30, FTSE100, and SSE 50, since investors are able to trade them instead of the index spots.

There are two implications of our study: First, investors might be able to outperform the market if they are able to track sharp rises and falls in these exchange rates. Second, our empirical results provide valuable information for investors for trading in such situations.

In addition to the event study approach, several other approaches are worthy of further investigation. For example, Kou et al. (2012) proposed an approach based on Spearman’s rank correlation to resolve disagreements among multiple criteria decision-making (MCDM) methods. Kou et al. (2014a, b) further demonstrated the effectiveness of MCDM methods in evaluating clustering algorithms for financial risk analysis. By proposing a cosine maximization method for the priority vector derivation in the Analytic Hierarchy Process (AHP), Kou and Lin (2014) demonstrated that cosine maximization is more flexible and efficient than other prioritization methods. They also proposed a Hadamard model to mitigate judgment contradiction and showed that the model enhances data consistency in the decision matrix.

Recently, because AHP has been widely adopted for group decision-making (GDM), Lin et al. (2020) proposed aggregation of the nearest consistent matrices with the acceptable consensus in AHP-GDM. Zhang et al. (2019) also proposed consensus models involving both the consensus degree and cost in GDM and showed the usability of the models in real-world contexts. Employing these approaches would be beneficial for further research in this area.