Keywords: Uncovered Interest Parity (UIP), interest rate differentials, Generalized Method of Moments (GMM)

Transcription

1 Exchange Rate and Interest Rate Linkage: the Validity of Uncovered Interest Parity (UIP) in Sri Lanka. Champika Dharmadasa (Faculty of Science & Engineering, Saga University, Japan) Abstract The objective of this paper is to examine the link between interest rates and exchange rates in Sri Lanka under the framework of uncovered interest rate parity (UIP) hypothesis between the periods from January 1990 to December To explore the validity of UIP hypothesis in the Sri Lankan setting, the paper used three foreign financial markets to utilize three bilateral exchange rates namely US Dollar, Japanese Yen and Indian Rupee which are defined by the domestic price of foreign currency and three types of short term (with three month maturity) interest rates that are denoted by interbank call money rates, treasury bills rates and discount rates for India, US and Japan respectively. Generalized Method of Moment (GMM) estimation is used as the main estimation technique of the paper and prior to that several other tests were carried out including co-integration test and a test on excess returns. Empirical results revealed that the coefficient of interest rate differentials is significant with a negative sign in all cases which in turn implies the UIP condition does not hold in Sri Lanka within the given time frame. The implication drawn from the study therefore is that interest rate differential is a poor predictor of exchange rate yields and the consideration of other possible policy variables under long time span along with interest rate differentials would be advantageous to obtain more accurate results. JEL classification: E43, E44, F30 Keywords: Uncovered Interest Parity (UIP), interest rate differentials, Generalized Method of Moments (GMM) 1. Introduction The degree of financial market integration is significantly increasing across the globe at present creating cross global investment opportunities and risk diversification. The resulting dynamics have made the transmission of interest rates crucial for the macroeconomic performances and the other related counterparts in the development process. According to some experts, the transmission of

2 volatile global interest rates gradually creates volatilities in developing countries financial markets (Edwards et al. 1985). However, for some others this might be the case that financial market integration reduces the divergence between domestic and foreign interest rates and makes similar market movements over time (Throop 1994). In this setting, the concept of interest rate parity appears as an important as well as an essential policy related concept for both international financial and investments decisions as it provides the information regarding the link between interest rates and foreign exchange market prices. Theoretically, there are two forms of interest rate parity known as covered interest rate parity (CIP) and uncovered interest rate parity (UIP) which relies on investor s desire of holding assets. Generally, there are two basic alternatives of holding assets i.e. assets denominated in their local currency and assets denominated in foreign currencies. The earnings of both assets depends on the interest rate that offered by both local and foreign authorities. To benefit from the foreign interest rate, the investor must convert the earnings in the domestic currency in foreign currency unit using the spot exchange rate at that time. If both domestic and foreign assets differ only in terms of the currencies of denomination and if the investors always have opportunity to cover from the exchange rate risks by converting their earnings in foreign currencies at a future time, at forward exchange rate, then the market equilibrium situation is called CIP. It can simply be explained as the equilibrium relationship between two countries interest rates where there is no arbitrage condition hold in any country. The UIP clarifies the condition that nominal interest rate differential between two countries must be equal to the expected change in the exchange rate. This happens when investors hold their positions at the particular time and wait for a future date in order to convert their earnings at a spot rate prevailing in the decided future date. When UIP holds, it implies there is free capital movement among countries and investors are willing to allocate their international investments without exchange rate risks. The fundamental assumption behind UIP is the efficient Market Hypothesis (EMH) which postulates that information should be fully available to the market participants and therefore no excess returns is possible in the market. As described by Taylor (1995) EMH is the joint hypothesis which reflects rational expectations and risk neutrality of market participants. Relying on the above discussion, this paper is attempted to test the validity of UIP hypothesis in the Sri Lankan setting over the period January 1990 to December 2011 while assessing the interest rates and exchange rates linkage in Sri Lanka. The paper used randomly selected exchange rates and interest rate differentials that appear as important for Sri Lanka in both regional and global wise. There is only a few sources can be found relating to UIP interest rate and exchange rate movements in the Sri Lankan setting and yet none of them has directly tested the presence of UIP in Sri Lanka (Weerasinghe et al. 2006; Sivarajasinham et al. 2012).

3 Considering Sri Lanka, the country introduced the policy liberalization process in late 1977 by eliminating trade and exchange rate restrictions in response to the worldwide policy deregulation process that emerge in 1970s. Since then it has been practicing a flexible exchange rate system which turned to free floating in the second quarter of 2001 to boost the depleted foreign reserves (CBSL 2003). Due to the flexibility of exchange rate system in Sri Lanka many foreign investments in the forms of both direct and portfolio has been flowing into the country creating many investment opportunities for local investors within and outside the country (Athukorala et al. 2000). In addition, a massive infrastructure projects that appeared in the country after the liberalization, created considerable development of the financial markets i.e. money and the capital market (Athukorala et al. 1994). However in the early 1990s, high interest rate and exchange rate appreciation become major impediments for this development process (jayasooriya 2004). The process of globalization, the Asian currency crisis and other related international shocks are partly responsible for this condition. Under this setting, it is worthwhile to investigate the impact of foreign interest rates and exchange rates movements on domestic financial variables as they are considered as important assets prices. The organization of the paper is as follows. The next section provides a discussion about the existing literature on UIP which is used to derive theory and methodology for the present paper. Section 3 describes the major theoretical standpoint of UIP and possible model for the estimation. Section 4 explains the data and the sources of data that used for the estimation. Section 5 presents results and discussion and finally section 6 gives summary and conclusion. 2. Review of the Literature UIP is known as one of the most empirically tested theoretical relations in international finance literature. It has been tested using various time frames, for many different currencies and interest rates relating to both developed and developing countries. The majority however has shown the failure of UIP condition ( =0, =1) 1 in practice (Taylor 1995; Meredith and Chinn 1998; Flood and Rose 2002). It is partly because it cannot test directly, and therefore it should be tested using the Rational Expectation Hypothesis (REH) as the unbiasedness conjecture. A survey has reported that in a few industrialized countries the sign of coefficient of interest rate differentials consistent with unbiasedness hypothesis, but none of them shows the theoretical value of unity in the focused coefficient (Froot et al Reviewing the literature, one of the most common features found, is that almost all who reject the UIP condition empirically tested the hypothesis in short-time horizons. The existence of a time-varying foreign exchange risk premium, market inefficiencies, sampling problems and autocorrelation 1 Detailed explanation on UIP hypothesis and its parameters can be found in the next of this paper.

4 problem in the interest rate differentials are some reasons in this regard (Huisman et al. 1998; Frankel et al. 1989; De Grauwe et al. 1993; Lewis 1995; Benassy-Quere et al. 1999; MacDonald 2000; Francis et al. 2002; Chinn et al. 2004; Lily et al. 2011; Ray 2012). In addition, forward premium puzzle is another remarkable finding of most empirical studies which explains that interest rate differentials show the change in the expected spot exchange rate in the wrong direction (Taylor 1995). Several explanations have been made that UIP works differently for countries and also vary with time as well as the level of financial market development (Fama 1984; Flood and Rose 1996; MacDonald and Taylor 1992; Isard 1995). Country specific studies which focus especially on Turkey, Malaysia and India in this setting provide good examples (Erdemlioglu et al. 2007; Lily et al. 2011; Ray 2012). According to them, violation of joint hypothesis of rational expectation, the forward premium condition and existence of arbitrage opportunities due to inefficient market conditions might be the possible facts for the UIP failure. Studies with reference to long time-horizon also provided mixed results. Exclusion of policy variables, common currency system in some regions and central bank reactions on interest rate volatility has caused the failure of UIP condition (McCallum 1994; Christensen 2000; Weber 2010). Contrary to that, some others have pointed that there is a possibility to that UIP condition hold when the time span is larger and some currency pair. The large interest rate differentials have stronger forecasting power for currency movements. Studies which used five to ten year time horizons obtained the focused coefficient ( ) near to the unity (Meredith and Chinn 1998; Lothian and Simaan 1998; Chinn and Meredith 2001; Lothian and Wu 2003; Bekaert et al. 2005; Chinn 2006; Sarantis 2006). On the other hand, some have emphasized that UIP works better in fixed exchange regimes rather than flexible exchange rate regimes (Flood and Rose 1996). However, these findings restricted to industrial countries like UK, U.S., Japan, Germany and Canada, and studies on UIP condition in developing countries is less frequent and inconclusive with the time horizon. There can be seen limited amount of literature on UIP relating to emerging markets as well (Ito and Chinn 2007; Backe and Schardex 2009; Di Giovanni and Shambaugh 2008; Backe and Schardex 2009). The findings confirmed that the diversity in the focused coefficient ( ) is attributed to differences in inflation volatility, financial development, capital account openness, legal condition, the prevailing exchange rate regime and the presence of forward premium puzzle. As mentioned in the earlier section, when it comes to the Sri Lankan context however, studies on UIP condition are very limited in number. Thus, this paper fulfills the above gap by testing the UIP hypothesis along with the assessment of the link between interest rates and exchange rates in the Sri Lankan setting.

5 3. Theoretical Framework, Estimation Methods and Data Derivation of Uncovered Interest Parity Basically, UIP can be derived using the covered interest rate parity (CIP), unbiasedness hypothesis, rational expectations and risk neutrality assumptions (Chinn 2007). Following that, the CIP can be expressed as follows: f n t s i i (1) t d t * t Where n ft is the logarithm forward exchange rate for maturity n period ahead, s t is the logarithm of spot exchange rate at time t, and i d * t, i t are domestic and foreign interest rate respectively. It should be noted here that CIP holds regardless of the investors preference. Nevertheless, this condition is not always practical. If the market participants are risk averse, the forward rate will be differ from the future spot exchange rate with risk premium which comes from the risk of holding domestic verses foreign assets. This situation can be explained using the following formula; E s s i i (2) d * t t 1 t t t t 1 Where, s t 1 is the logarithm of future spot exchange rate, Et expectations of future exchange rate at time t and t 1 is the future risk premium. As mentioned before in this paper, UIP is based on the joint hypothesis which relies on rational expectations and risk neutrality. Thus these assumptions can be used to form UIP. In this setting, risk neutrality can be expressed as t 1 =0 and rational expectation hypothesis is given that: s E s (3) t 1 t t 1 t 1 Substitution the equation (3) to (2), the UIP can be derived as: s s i i (4) d * t 1 t t t t 1 The equation (4) is explained the UIP relationship. It is said that UIP hold only if there is open capital market. In this case, this paper is assumed that it can be applied to countries with minimal capital

6 restrictions. Although it is not fully liberalized, capital transactions are free between countries in the South Asian region and most of the time it imposes minimum restrictions to other countries. Therefore utilizing UIP theory would not be inappropriate. Based on the aforementioned theoretical concerns, following empirically testable model which derives from the equation (4) is used for the purpose of checking the existence of UIP in the Sri Lankan setting. s d * t 1 ( it it ) t 1 (5) s t 1 denotes change in exchange rate changes (logarithms) i d is domestic interest rate and i * denotes foreign interest rates. represents the intercept term and represents the sensitivity of local interest rates to foreign interest rates. The acceptance or the rejection of UIP holds depends on the hypothesis testing and it can be said that the null hypothesis: H 0 : =0, =1. However, in the related literature usually focus only on (the slope coefficient) and according to them should be significant with a positive sign (Flood et al. 2002). The above model is tested using the Generalized Method of Moment (GMM) estimation technique of Hansen (1982) as it is famous for consistent parameter selecting. This makes the sample correlation between instruments and the error term of the model is close to zero (Sarantis 2006). J-statistic proposed by Hansen (1982) is employed to test the over identifying restriction and NW-test (Newey et al. 1987) is used to measure the robustness of the estimation Generalized Method of Moment (GMM) as an efficient estimator The theory of GMM estimation postulates the usage of two sets of population moment conditions in a way that minimize the asymptotic variance among the method of moment estimators of population mean. It is well-known that many common econometric estimators including ordinary lease square (OLS), two-stage least square (TLS), maximum likelihood (ML) and instrumental variables estimators are mostly derived based on the method of moment. However, it has been argued that GMM estimator is relatively more advantageous and more efficient over all other aforementioned estimators when it is dealing with the time series data (Hansen 2001). Several reasons can be found from the related literature in this regards. First is related to the problem of serial correlation in the error term of linear time series data. According to the literature, GMM allow accounting for serial correlation of unknown form and heteroskedasticity (Hansen 1982; Newey et al. 1987). However, some studies have pointed out that the problem of serial correlation can be solved using Least Square estimations under strict

7 exogeneity assumptions such that corvariates in some time periods other than the current period that consider in the estimation (t) are uncorrelated with the error term (Wooldridge 2000). Nonetheless, these assumptions do not have a base under time series context. Second, GMM is efficient estimator than OLS, TLS and maximum likelihood (ML) as it can be tested for issue of over-identification. This can be easily done with the time series data as it allows to add more moment conditions by assuming that past value (lagged value) of both dependent and independent variables are not correlated with the error term even though they are do not include in the model (Hansen 2001; Wooldridge 2001). On the other hand, it is said that GMM estimator is more suitable for models that are based on rational expectations. Since the model of this study is based on the rational expectation hypothesis, the usage of GMM would help to draw reliable inferences on interest rate-exchange rate linkage in Sri Lanka. It is empirically proven that the use of GMM for time series data is significantly advantageous over other related estimators as the GMM allow the estimations to be done in accordance with the restrictions that implied by the theory itself. It should be noted here that the paper conducted several unit root tests to avoid the issues of non-stationarity of data and tested the presence of co-integration utilizing Johansen Co-integration test before moving to the main estimation. It is also carried out the informal test for UIP before the GMM estimation The Data and the Sources of Data This paper utilized monthly data from January 1990 to January 2011 to investigate the presence of UIP in Sri Lanka. Since 1990s is well-known for remarkable global financial market integration, investment boom as well as currency crises, this time frame would be perfect for the particular investigation. Data on interest rates and exchange rates are collected from three foreign financial markets to obtain the interest rate differentials and these financial markets are assumed to have a close connection with the financial market in Sri Lanka.US treasury bills rate, discount rate of Japan, interbank loan rates for India and Sri Lanka were used as interest rates. All interest rates are with the same maturity level of three months. Three bilateral exchange rates i.e. US Dollar, Japanese Yen, and Indian Rupee were used and they were defined as the domestic price of foreign currency. The selection of exchange rates was based on the pattern of recent trade flow and foreign direct investments of both developed countries and

8 countries in the South Asian region. Data are collected from various publications including annual reports and monthly bulletins of the Central Bank of Sri Lanka, as well as from central bank websites of all relevant countries in the sample. In addition, internet sources like and Bloomberg were used. 4. Results and Discussion To test the presence of stationarity of time series data, the paper has conducted Augmented Dickey Fuller (ADF) test and Phillips-Perron (PP) 2 test for series of domestic interest rate, foreign interest rates (US, Japan and India respectively), interest rate differentials, exchange rates (Both domestic and foreign) and exchange rates yields. Decisions were made following the null hypothesis (H 0 ) of presence of the unit root and the alternative hypothesis (H 1 ) of stationarity. It is expected that three variables i.e. domestic interest rates, foreign interest rates and exchange rates are to be integrated order I(1) and interest rates differentials and exchange rate yields should be at order I(0). Table 01- Augmented Dickey-Fuller (ADF) test for interest rates and interest rates differentials Level Test First Difference Intercept Intercept & Trend Intercept Intercept & Trend Lag 0 Lag 1 Lag 0 Lag 1 Lag 0 Lag 2 Lag 0 Lag 2 Interest Rates SLR USR JPR IR Interest Rate Differentials DUSR DJPR DIR Source: Author s own estimation H 0 : series has unit root, H 1 : series are trend stationary Note 1: 1%, 5% and 10% critical values for ADF level test are -3.46, and respectively. For the first difference they are -3.99, and respectively. Note 2: SLR- Sri Lankan interest rate, USR- US interest rate, JPR- Japan interest rate, IR- Indian 2 Test results did not report here as they did not alter much from ADF test results.

9 interest rate and interest rate differentials are shown by DUSR, DJPR and DIR for US, Japan and India respectively. Test results are shown that Sri Lankan interest rate and Indian interest rate reject the null hypothesis while Japanese and US interest rate series have proven the presence of unit root. Domestic interest rate differentials too fail to reject the null hypothesis for Japan and US while it shows interest rate differential between Sri Lankan and India has no unit root at the level test. Table 02- Augmented Dickey-Fuller (ADF) test for exchange rates and exchange rates yields Level Test First Difference Intercept Intercept & Trend Intercept Intercept & Trend Lag 0 Lag 1 Lag 0 Lag 1 Lag 0 Lag 1 Lag 0 Lag 1 Exchange Rates USD JPY INR Exchange Rate Yields YUSD YJPY YINR Source: Author s own estimation H 0 : series has unit root, H 1 : series are trend stationary Note 2: 1%, 5% and 10% critical values for ADF level test are -3.46, and respectively. For the first difference they are -3.99, and respectively. Note 2: USD- US dollar, JPY- Japanese yen, INR- Indian rupee and exchange rate yields are shown by YUSD, YJPY and YINR for US, Japan and India respectively. Here the test results are in line with the standard theoretical position emphasizing that exchange rates have unit root while exchange rate yields are trend stationary. However, it should be noted here that it is worthwhile to move to a co-integration test as some series of the sample shows the presence of unit root and it may cause to obtain spurious results which leads to incorrect judgments. Hence, Johannes co-integration test was adapted to test whether series i.e. interest rate differentials and exchange rates are interdependence in the long run. Lag interval is selected using both Akaike Information Criteria (AIC) and Schwarz Criteria (SC) and trace statistic

10 ( trace ) is considered for the inference. Table 03- Johansson Co-integration Test Sample: 1990M M12 Included observations:259 Series: DUSR DJPR DIR USD JPY INR Lag interval: 1 to 2 Hypothesized No. of CEs Eigen Value Trace Statistic 0.01 Critical value 0.05 Critical value 0.1 Critical Value None *,**,** * At most *,**,**** At most * At most Source: Author s own estimation Note 01: Test assumption- linear deterministic trend in the data series; H 0 : Series are not co-integrated; H 1 : Series are co-integrated; *, **, *** shows the level of significance 10%, 5% and 1% respectively. Source: Author s own estimation As shown in the co-integration test results, it is evident that at least one co-integrating relationship exists between the variables. Hence, utilization of GMM to test the UIP in the Sri Lankan setting would not be harmful and it would yield super consistent estimators. Thus, equation (5) was tested using GMM to check the validity of the UIP in the Sri Lankan setting. The choice of instrument variables for the GMM test was done using Akaike Information Criteria (AIC). As indicated by the AIC, three lags of interest rate differentials and exchange rates were utilized as instrument variables. Before carrying out the main estimation, the paper did an informal way of testing the presence of UIP d following the literature (Ray 2012). In this case excess return { i i ) s } is taken into ( t, k t, k t, t k account. It is said that if the excess return is stable overtime, it implies the UIP condition holds in practice. In line with the above method, this paper examines the behavior of excess returns that come from foreign sources is examined. It is shown in the following Figure 1.

11 US IND JPY Fig.1. Excess Returns (Sri Lanka vs. exchange rate and interest rates of US, India and Japan) Source: Author s own estimation However, the movement of the excess returns overtime does not show stable condition relating to any currency or exchange rates in the Sri Lankan setting. As shown in the Fig. 1 there are heavy fluctuations can be seen in the excess returns throughout the time frame which the study is utilized for the test. The above situation can also be explained using the changes of exchange rates with the corresponding interest rate differentials separately. The following three graphs are used for the above purpose. The right axis of each shows the interest rate differentials while the left axis shows the exchange rate movements. Fig. 2 (a, b) shows the changes interest rate and exchange rate in Sri Lanka with reference to the US, and Indian markets. Here the high fluctuation of exchange rate can be observed between while relatively high interest rate changes can be observed during and after This might be due to the effect of the Asian currency crisis and 2008 global financial turmoil. Although both situations did not directly affect the Sri Lankan financial market, it seems the fluctuations of these macroeconomic variables in US and India had an effect on the domestic interest rate and exchange rate. Fig. 2 (c) reflects the Japanese currency and interest rate influence on domestic variable changes. There can be seen a high volatility of exchange rates but in response to that interest rates do not show heavy fluctuations. To the end, monthly exchange rate fluctuations display large movements compared to the changes of interest rates.

12 US/SLR USD IND/SLR INR (a) (b) JP/SLR JPY (c) Figure 2 (a) LKR/US Dollar- Exchange rate changes (USD) and interest rate differentials US/SLR, (b) LKR/Indian Rupee- Exchange rate changes (INR) and interest rate differentials IND/SLR, (c) LKR/Japanese Yen- Exchange rate changes (JPY) and interest rate differentials JP/SLR. Source: Author s own estimation The results of the GMM estimation too seems support for the above findings and as well as previous literature of failure of UIP showing not only the focused coefficient ( ) shows negative signs in all cases, but also showing their values are too small even though they are significant in two cases except the Indian currency. The results of the GMM estimation are shown in the following Table 4. According to Table 4, the values in the parenthesis indicate t-statistics of the estimated coefficients. R 2 represents the coefficient of determination adjusted for the degrees of freedom; SER indicates the standard error of the regression; J(4) indicates the Hansen statistic for the test of over identification restrictions. This has a 2 distribution with the degrees of freedom equal to the number of instruments minus the number of estimated parameters. Conclusions were drawn based on 5 percent and 1 percent critical values (9.488 and respectively) of Hansen statistic for J(4). The hypothesis: =0 and =1 is tested with Wald coefficient restriction test with the 5 percent critical value with one degree of freedom Results confirm the failure of UIP with significant negative sign of coefficient and the insignificant low value of R 2. The assumption of zero constant is rejected in all four cases according to the values of the t-statistic.

13 Table 4 GMM Test Results LKR/USD LKR/JPY LKR/INR (4.966) (2.674) (1.7498) (-3.372) ( ) ( ) 2 R SER DW J(4) = = Source: Author s own estimation On the other hand, value of J-statistic rejects the over-identifying restriction at 5percent significance level. This provides a proof for the validity of the chosen instruments of the estimated model. The significant low values of the standard error of the regression (SER) interpret the model is best fit with all available data. 5. Summary and Conclusions This paper addressed the link between exchange rate changes and interest rate differentials in the uncovered interest parity framework in Sri Lanka during January 1990 to December It is utilized three major foreign currencies i.e. US dollar, Japanese yen and Indian rupee and three interest rates namely US treasury bills rate, Indian call money rate and Japanese discount rate. Sri Lankan call money rate is used as domestic interest rates to obtain the interest rate differentials of foreign and domestic interest rates. All interest rates are in same three months maturity. The testing of UIP is done in two steps. First, the movements of excess returns of three exchange rates and interest rates were tested as an entry point to the investigation. Second, the paper adopted the GMM estimation technique to test the presence of UIP accurately. The overall test results show that the rejection of UIP hypothesis within the given time frame in Sri Lanka confirming the previous findings relating to practical situation of UIP condition during 1990s. GMM estimation results show that the focused coefficient ( ) is too small with a negative sign; this implies that UIP condition between interest rate and exchange rate is weak. Several implications can be drawn from this study. First, the UIP condition has a lack of support in the Sri Lankan setting during the given time period, which may implies that there is a possibility for arbitrage opportunity between Sri Lanka financial market and the foreign financial markets

14 considered in the study. Since Sri Lanka financial market is not well developed, information asymmetry may also cause for the failure of the UIP. On the other hand, since the study used short term-interest rate differentials, they may not sufficient to predict the exact results. Consideration of long term variables and exchange rate risk premium and a longer time span would be useful to obtain consistent results for the UIP condition. Third, the study is assumed that the prevailing capital market condition would not be a problem for utilizing the standard UIP model. However, findings suggested that it would be useful to include capital restrictions into the standard UIP model since Sri Lankan capital market is not fully liberalized. Therefore, further investigations have to be conducted with the new version of UIP model to understand the interest rate-exchange rate linkage in Sri Lanka in general and the deviation from interest rate parity in particular. Bibliography Athukorala, P. & Jayasooriya, S. (1994). Macroeconomic Policies, Crises, and Growth in Sri Lanka, World Bank, Washington D. C. Athukorala, p. & Rajapathirana, S. (2000). Liberalization and Industrial Transformation: Sri Lanka in International Perspective, Oxford University Press: Oxford and New Delhi. Backe, P. & Schardax, F. (2009). European and Non-European Emerging Market Currencies: Forward Premium Puzzle and Fundamentals. Focus on European Integration, (2), Bakaert, G. & Hodrick, R. (1993). On Biases in the Measurement of Foreign Exchange Risk Premiums, The Journal of International Money and Finance, 12, Bakaert, G. & Hodrick, R. (2001). Expectations Hypotheses Tests, Journal of Finance, 56(4), Bekaert, G., Wei, M. & Xing Y. (2005). Uncovered Interest Rate Parity and the Term Structure, Journal of Money and Finance, (26), Benassy-Quere, A., Larribeau, S. & MacDonald, R. (1999). Model of Exchange Rate Expectations. University Strathclyde Discussion Paper, (10). Central Bank of Sri Lanka (CBSL) ( ). Annual Report. Central Bank of Sri Lanka. Chinn, M. (2006). The Partial Rehabilitation of Interest Rate Parity in the Floating Rate Era: Longer Horizons, Alternative Expectations and Emerging Markets, Journal of International Money and

15 Finance, (25), Chinn, M. (2007). Forward Premium Puzzle, World Economy Forward Premium Puzzle (entry written for the Princeton Encyclopedia of the World Economy), 1-8. Chinn, M. & Meredith, G. (2001). Testing Uncovered Interest Parity at Short and Long Horizons, HWWA Discussion Paper (102). Chinn, M. & Meredith, G. (2004). Monetary Policy and Long Horizon Uncovered Interest Parity, IMF Staff Papers, 51(3), Christensen, M. (2000). Uncovered Interest Parity and Policy Behavior: New evidence, Economics Letters, 69(1), De Grauwe, P., Dewachter, H. & Embrechts, M. (1993). Exchange Rate Theory: Chaotic Models of Foreign Exchange Markets. Blackwell, Oxford. Di Giovanni, J. & Shambaugh, J. C. (2008). The Impact of Foreign Interest Rates on the Economy: The Role of the Exchange Rate Regime. Journal of International Economics, 74(2), Domowitz, I. & Hakkio, C. S. (1985). Conditional Variance and the Risk Premium in the Foreign Exchange Market. Journal of International Economics, (19), Dreger, C. (2010). Does Nominal Exchange Rate Regime affect Real Interest Rate Parity Condition?. North American Journal of Economic and Finance, (21), Dubey, Muchkund (2007). SAARC and South Asian Economic Integration. Economic and Political Weekly, 42(14), Edwards, S. & Khan, M. S. (1985). Interest Rate Determinations in Developing Countries: A Conceptual Framework. IMF Staff Papers, 32, Erdemlioglu, D. M. & Alper, C. M. (2007). A New Test of Uncovered Interest Rate Parity: Evidence from Turkey, MPRA Paper, Fama, E. F. (1984). Forward and Spot Exchange Rates. Journal of Monetary Economics, 14(3),

16 Flood, R. P. & Rose, A. K. (2002). Uncovered Interest Parity in Crisis. IMF Staff Papers, 49, Flood, R. P. & Rose, A. K. (1996). Fixes: Of the Forward Discount Puzzle. The Review of Economics and Statistics, 78, Foy, T. M. (2005). An Empirical Analysis of Uncovered Interest Rate Parity and the Forward Discount Anomaly. Department of Economics, Queen s University, Ontario,Canada, K7L 3N6, 1-8. Francis, B. B., Hasan, I. & Hunter, D. M. (2002). Emerging Market Liberalization and the Impact on Uncovered Interest Rate Parity. Journal of International Money and Finance, (21), Frankel, J. A. (1992). Measuring International Capital Mobility: a review. American Economic Association Papers and Proceedings, 82, Frankel, J. A. & Froot, K. A. (1987). Using Survey Data to Test Standard Propositions Regarding Exchange Rate Expectations. American Economic Review, 77(1), Frankel, J. A. & Froot, K. A. (1989). Forward Discount Bias: Is it an Exchange Rate Premium?. Quarterly Journal of Economics, 53, Frankel, J. A. & Froot, K. A. (1990). Exchange Rate Forecasting Techniques, Survey data, and Implications for the Foreign Exchange Market, NBER Working Paper no Froot, K. A. & Thaler, R. H. (1990). Anomalies: Foreign Exchange. Journal of Economic Perspectives, 4(3) Hansen, L. P. (1982). The Larger Sample Properties of Generalized Method of Moments Estimator. Econometrica, (32), Huiman, R., Koedijk, C. K. & Nissen, F. (1998). Extreme Support for Uncovered Interest Parity. Journal of International Money and Finance, (17), Isard, P. (1995). Exchange Rate Economics. Cambridge: Cambridge University Press. Isard, P. (2006). Uncovered Interest Rate Parity, IMF Working Paper, WP/06/96 (April). Ito, H. Chinn, M. (2007). Price-based Measurement of Financial Globalization: A Cross-country Study of Interest Rate Parity. Pacific Economic Review, 12(4),

17 Lewis, K. (1995). Puzzles in International Financial Markets. In Grossman, G. and Rogoff, K (Eds.), The handbook of International Economics. North Holland. Lily, Jaratin, Mori, Kogid, Karim, M. R. A., Rozilee, A. & Dullah Mulok. (2011). Empirical Testing on Uncovered Interest Rate Parity in Malaysia. Journal of applied Finance and Banking, 1(2), Lothian, J. R. & Simaan, Y International Financial Relations under the Current Float: Evidence from Panel Data. Open Economies Review, 9(4), MacDonald, R. (2000). Expectation Formation and Risk in Three Financial Markets: Surveying what the Surveys Says. Journal of Economic Survey, 14, MacDonald, R. & Taylor, M. P. (1992). Exchange Rate Economics: A survey. IMF Staff Papers, 39(1), McCallum, B. T. (1994). A Reconsideration of Uncovered Interest Parity Relationship. Journal of Monetary Economics, 33, Meredith, P. R. & Chinn, M. D. (1998). Long Horizon Uncovered Interest Parity. NBER Working Paper (6797). Newey, W. & West, K. (1987). A Simple Positive Definite, Heteroscedasticity and Autocorrelation Consistent Matrix. Econometrica, 55, Ray, Sarbapriya. (2012). Testing the Validity of Uncovered Interest Rate Parity in India. Advances in Applied Economics and Finance (AAEF), 1(4), Sarantis, Nicholas. (2006). Testing the Uncovered Interest Parity using Traded Volatility, a Time-varying Risk Premium and Heterogeneous Expectations. Journal of International Money and Finance, 25, Sivarajasingham, S. & Balamurali, N. (2012). The Dynamic Behavior of Sri Lankan Exchange Rates: Evidence from Five Currencies of Main Trading Partners. Annual Academic Sessions 2012, Taylor, M. P. (1995). The Economics of Exchange Rates. Journal of Economic Literature, 33,

18 Throop, A. W. (1994). International Financial Market Integration and Linkages of National Interest rates, Federal Reserve Bank, USA. Weber, E. (2010). What Happened to Transatlantic Capital Market Model? Econometric Model, doi: /j,econmod, Weerasinghe, E. A. & Ratnayake, R. M. A. K. (2006). The Time Series Analysis: the Behavior of the Exchange Rate in Sri Lanka. Proceedings of International Conference of Business Management, 3. Wu, L Modeling Financial Security Returns Using L evy Processes. in Handbook of Financial Engineering, J. Birge, and V. Linetsky (Eds.), Elsevier, New York.