See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/229400207

Contrarian strategy and overreaction in foreign exchange

markets

Article  in  Research in International Business and Finance · September 2008

DOI: 10.1016/j.ribaf.2007.09.003

CITATIONS READS

12 79

2 authors, including:

Theodore Syriopoulos
University of the Aegean
64 PUBLICATIONS   1,020 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Applying the Corporate Social Responsibility to the Shipping Industry View project

Hedge Ratio estimation method for Commodity Markets View project

All content following this page was uploaded by Theodore Syriopoulos on 12 May 2019.

The user has requested enhancement of the downloaded file.

 Available online at www.sciencedirect.com

Research in International Business and Finance 22 (2008) 319–324

Contrarian strategy and overreaction in foreign

exchange markets
George S. Parikakis a,∗ , Theodore Syriopoulos b
a Department of Business, University of the Aegean and Piraeus Bank, Corporate Division,
4 Amerikis Street, 10564 Athens, Greece
b Department of Shipping, Trade and Transport, University of the Aegean, 2A, Korai Street,
Chios 82100, Greece
Received 14 June 2007; accepted 25 September 2007
Available online 2 October 2007

Abstract
This paper investigates patterns to assist investors to forecast future exchange rate movements. We test for
overreaction and underreaction examining exchange rate changes following excess 1-day fluctuations for
currencies in two emerging (Turkey, Brazil) and two developed (US, UK) countries. Using euro as the base
currency, we identify that the Turkish lira, the Brazilian real and the US dollar overreact, while the British
pound underreacts. In the case of British pound, asymmetric responses and lack of volatility are two crucial
factors to reject overreaction. Also, we find that contrarian strategy can be used in all currency markets for
profitable investments.
© 2007 Elsevier B.V. All rights reserved.

JEL classification: F31; G14

Keywords: Overreaction; Underreaction; Exchange rate; Contrarian investment

1. Introduction

Overreaction and underreaction phenomena have been widely studied during the last decade.
However, most of the studies focus on stock markets. Hence, it is of great interest to examine
whether these phenomena can be applied in currency markets where large fluctuations occur
mainly when national economic cycles are out of synchronisation. Increasing synchronisation
and convergence during the last years, made these fluctuations to appear rarely. Hence, in this

∗ Corresponding author.
E-mail addresses: ParikakisG@Piraeusbank.gr, gpgp@otenet.gr (G.S. Parikakis).

0275-5319/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ribaf.2007.09.003
320 G.S. Parikakis, T. Syriopoulos / Research in International Business and Finance 22 (2008) 319–324

paper we study not only extreme movements for developed markets (EU, US, UK), but also for two
emerging markets which during the last decade experienced at least one blow-out. Particularly,
we study the Turkish lira, i.e. TR (collapse in 2001) and the Brazilian real, i.e. BRL (collapse in
1999). In this study, the exchange rates are defined as the number of foreign currency units per
euros.
In an overreaction event the price level is way beyond where it should be. A key feature of
the overreaction is that this phenomenon conveys a contrarian or a momentum strategy (see Nam
et al., 2001). The response in the price of a currency to these phenomena is of great interest
for fund managers who during the last years undertake diverse approaches to managing assets.
Indeed, hedge funds are growing in popularity and attract a large portion of investor’s portfo-
lios, and as a consequence the contrarian style of managing funds is becoming more and more
prevalent.
The objective of this study is to determine whether: (i) exchange rates respond properly to
information, giving rise to overreaction hypothesis; and (ii) a contrarian strategy can be used in
currency markets.
Studying extreme 1-day fluctuations, our findings provide sufficient evidence that all currencies
except for the British pound, overreact. Moreover, we find that contrarian investment strategy holds
for all exchange rates.
This paper contributes to the existing literature in two ways: (i) sheds more light to exchange
rate fluctuations, studying overreaction hypothesis which so far has not been used widely in the
academic society; and (ii) investigates that contrarian strategy can be used very profitable from
investors to foreign exchange markets.
The structure of the paper is organized as follows: Section 2 briefly presents the literature
review. Section 3 presents the data. Section 4 analyzes methodological issues. The empirical
results are reported in Section 5. Section 6 contains the concluding remarks.

2. The literature review

DeBondt and Thaler (1985) are among the first who find that stock prices behaved as if indi-
vidual investors overreacted to given information. They find that in the stock market a contrarian
strategy which buys losers and sells winners based on their returns over a 3- to 5-year horizon
performs well in subsequent holding periods of 3–5 years. Their finding provides assistance on
security selection for practitioners.
The phenomenon of overreaction and underreaction has also been studied from Clare et al.
(1995), Barberis et al. (1998) and Nam et al. (2001), among others, providing supportive results
for the overreaction hypothesis for the UK and the US stock markets. Akhigbe et al. (1998) and
Peterson (1995) examined post-event abnormal returns with extreme 1-day stock prices changes
for US stocks. Their findings support the overreaction hypothesis. Also, Larson and Madura (2003)
used extreme 1-day returns for a sample of losers and winners by selecting daily stock price returns
in excess of 10%. They found that for winners, there is overreaction in response to uninformed
events but no overreaction on average in response to informed events. This finding according to
Larson and Madura (2003) suggests that the degree of overreaction to new information depends
on whether the cause of the extreme stock price change is publicly released.
Moreover, Fung et al. (2000) and Grant et al. (2005) used futures market as “lead” contract to
assess the question of intraday price reversals following large price changes at the market open.
They find highly significant intraday price reversals as well as significant intraday reversals in
yearly and day of the week investigations.
 G.S. Parikakis, T. Syriopoulos / Research in International Business and Finance 22 (2008) 319–324 321

Larson and Madura (2001) identify that overreaction and underreaction also holds in the for-
eign exchange market. They used five emerging currency markets (Hong Kong, Israel, Malaysia,
Singapore, South Korea) and 10 industrial currency markets (Belgium, Britain, Canada, France,
Germany, Italy, Japan, Spain, Sweden, Switzerland), expressed in US dollars. They provide evi-
dence that following 1-day extreme fluctuations, there is evidence that at least one emerging
currency overreacts, while at least one industrial currency underreacts.

3. Data

We use daily closing prices for the euro/US dollar (EUR/USD), the euro/British pound
(EUR/GBP), the euro/Brazilian real (EUR/BRL) and the euro/Turkish lira (EUR/TR) exchange
rates. Our data is from Bloomberg database from January 1999 till February 2007. The day’s
closing price reflects the last transaction price rather than the settlement price.

4. Methodology

The stock market overreaction hypothesis asserts that investors tend to overreact to new infor-
mation which results in exaggerated movements in share prices; as a result prices deviate from
the actual values implied by the new information. Once investors have considered the news
in more detail, the overreaction wanes causing share prices to move back to their equilibrium
levels.
We test two competing hypotheses. Hypothesis 1 states that currency market participants
respond too strongly to new information and subsequently revise their estimates of currency
value (overreaction). On the other hand, hypothesis 2 assumes that currency market participants
do not respond strongly enough to new information and subsequently revise their estimates of
currency value. These two hypotheses are tested by assessing exchange rate changes following
extreme 1-day fluctuations in currency value.
If overreaction hypothesis holds, then extreme 1-day fluctuations will be followed by significant
changes in the opposite direction. On the other hand, if underreaction hypothesis holds, extreme
1-day fluctuations will be followed by significant changes in the same direction.
We use a filter size which appears when the closing price is equal to or widen than ±0.6%.
Also, we use cumulative return for all currencies so that CARt at time t is defined as:
CARt = αi + βi−1 log(Rt−1 ) + βi,0 log(Rt ) + βi+1 log(Rt+3 ) + εt , (1)
where Rt−1 is the return of the exchange rate in the previous day; R1 is the return on the event
day; and Rt+1 is the return on the next day. The average cumulative return 3 days after the event
day is calculated as:
1 
ACAR3 = ∗ CARt , (2)
3
to N

where N is the total number of event days satisfying the given filter size. A price reversal exists
if the sign of the 3-day average cumulative return is different from the sign of the event day t.
Also, we use a standard t-test for the entire sample period. The null hypothesis states that the
average cumulative return for 3 given days after the event day (where return is higher or lower
than 0.6%) is not significantly different from zero. This implies that, the t-statistic on the average
cumulative return is statistically insignificant in a contrarian investment strategy. The formula for
322 G.S. Parikakis, T. Syriopoulos / Research in International Business and Finance 22 (2008) 319–324

Table 1
Descriptive statistics
Max Min Mean S.D.

EUR/USD 0.075 0.027 0.061 0.048

EUR/GBP 0.064 0.011 0.029 0.027
EUR/BRL 0.086 0.038 0.048 0.066
EUR/TL 0.093 0.042 0.053 0.081

the t-calculation is the following:

Sample mean − 0
tcalc = √ . (3)
Sample standard deviation/ N

5. Empirical results

Table 1 reports the descriptive statistics. In Table 2, we present the number of event days for
the ±0.6% filter. From Table 2, it is clear that the number of event days varies widely between
currencies. The higher number of extreme positive (for the euro currency) returns is observed
for the EUR/TR exchange rate. On the other hand, higher number of days with extreme negative
returns for the euro is observed for the EUR/BRL exchange rate. For the EUR/GBP exchange
rate we find insignificant price reversals with a p-value rejecting the overreaction hypothesis
and supporting underreaction. It is clear that the small number of extreme days is due to lack
of volatility or declining volatility. This is mainly caused from two reasons: (i) the associated
gamma and vega hedging that the Bank of England needs to undertake, pushes down volatility;
(ii) the GBP is on an uptrend channel against the euro for many years. For trends in foreign
exchange, we need macroeconomic dislocations across countries. In the case of EUR/GBP seems
that macroeconomic conditions favour “trend following” investments.
On the other hand, the number of days with extreme return is more than double for all other
exchange rates. This may be due to lack of directionality. At the simplest level, currency moves
happen because of interest rate or inflation differentials. As a result, we either have not much event
risk due to global convergence (like in the cases with EUR/GBP and EUR/USD), or geopolitical
event risk has not been factored in (like in the cases with EUR/BRL and EUR/TR).
In Table 3 we report the average cumulative returns and the results of t-tests.
When there is positive extreme return for the euro, the p-value drifts below the 0.05 test level
for the first 3 days after the event occurred. This indicates that ACARt values are statistically
significant. The same finding occurs for negative filter size.

Table 2
Number of event days
Filter sizes Days of extreme return
+0.6% −0.6%

Whole period
EUR/USD 274 109 383
EUR/GBP 62 126 188
EUR/BRL 285 193 478
EUR/TR 321 104 425
 G.S. Parikakis, T. Syriopoulos / Research in International Business and Finance 22 (2008) 319–324 323

Table 3
Tests of significance for average cumulative returns (ACAR%)
Day Filter ≥+0.6% Filter ≤0.6%
ACAR% p-Value ACAR% p-Value

EUR/USD
1 −0.10 (0.028)* 0.12 (0.031)*
2 −0.26 (0.034)* 0.19 (0.035)*
3 −0.17 (0.042)* 0.20 (0.042)*
4 −0.03 (0.065) −0.05 (0.58)
EUR/GBP
1 −0.13 (0.052) 0.04 (0.057)
2 −0.09 (0.057) 0.02 (0.050)
3 0.04 (0.058) −0.05 (0.051)
4 0.08 (0.052) −0.10 (0.59)
EUR/BRL
1 −0.08 (0.033)* 0.14 (0.018)*
2 −0.09 (0.036)* 0.21 (0.027)*
3 −0.02 (0.039)* 0.16 (0.039)*
4 0.04 (0.061) 0.02 (0.55)
EUR/TR
1 −0.06 (0.024)* 0.16 (0.011)*
2 −0.01 (0.030)* 0.25 (0.018)*
3 0.11 (0.032)* 0.19 (0.027)*
4 0.12 (0.057) 0.12 (0.50)
* Indicates significance at the 10% level.

Moreover, from Table 3 we observe that contrarian strategy holds for all exchange rates.
Indeed, the sign of the ACARt is different for at least 2 days after the day of the extreme return
has occurred. As a result, a fund manager who sells past winners and buys the worst performing
currencies will experience superior profits.

6. Conclusion

This paper presents new insights in forecasting currency movements. Our tests are based on
extreme 1-day fluctuations in exchange rate changes. Using the euro as the base currency, we
observe overreaction for US dollar, Brazilian real and Turkish lira. On the other hand, we identify
that British pound underreacts. This may be due to lack of volatility, asymmetric responses and/or
strong directionality. Indeed, in the last years the GBP seems to be in an upward trend against the
euro. This directionality favours underreaction instead of overreaction.
Also, using a filter size we identify that contrarian strategy is well suited for all exchange rates.
For at least 2 days after the event day, currencies follow a different movement (the return has
different size). Hence, investors can take advantage from this strategy, to sell past winners and
buys worst performing currencies.

References

Akhigbe, A., Gosnell, T., Harikumar, T., 1998. Winners and losers on NYSE: a re-examination using daily closing bid—ask
spreads. J. Financ. Res. 21, 53–64.
324 G.S. Parikakis, T. Syriopoulos / Research in International Business and Finance 22 (2008) 319–324

Barberis, N., Shleifer, A., Vishny, R., 1998. A model of investor sentiment. J. Financ. Econ. 49, 307–343.
Clare, A., Maras, M., Thomas, S.H., 1995. Overreaction hypothesis in the U.K. stock market. J. Bus. Financ. Account.
22, 961–973.
DeBondt, W.F.M., Thaler, R.H., 1985. Does the stock market overreact? J. Financ. 40, 793–805.
Fung, A.K.-W., Mok, D.M.Y., Lam, K., 2000. Intraday price reversals for futures in the US and Hong Kong. J. Bank.
Financ. 24, 1179–1201.
Grant, J.L., Wolf, A., Yu, S., 2005. Intraday price reversals in the US stock index futures market: a 15-year study. J. Bank.
Financ. 29, 1311–1327.
Larson, S.J., Madura, J., 2001. Overreaction and underreaction in the foreign exchange market. Global Financ. J. 12,
153–177.
Larson, S.J., Madura, J., 2003. What drives stock price behaviour following extreme one-day returns. J. Financ. Res. 26,
1–127.
Nam, K., Pyun, C.S., Avard, S.L., 2001. Asymmetric reverting behaviour of short-horizon stock returns: an evidence of
stock market overreaction. J. Bank. Financ. 25, 807–824.
Peterson, D.R., 1995. The influence of organized options trading on stock price behaviour following large one-day stock
price declines. J. Financ. Res. 18, 33–44.

View publication stats