Can BRICS Form a Currency Union? An Analysis under Markov Regime-Switching Framework

Abstract

The article applies Markov Regime-Switching Model (MRSM) to explore the prospects of forming currency union among BRICS countries. Our data span the period before and after the formation of the group, and the study compares the regime-switching behaviour of their real exchange rate markets accordingly. The analysis found divergent real exchange market behaviour of the member countries before the formation of the group. However, after the integration of economies, the convergences in central bank’s direct intervention behaviours are evident especially among India, China and South Africa. The study concludes that the inclusion of the stronger policy interaction in the region now and future, especially in monetary management, unveils the chance of a strong currency union among BRICS members.

Keywords

Markov Regime Switching real exchange rate currency union

Introduction

Exchange rate regimes gain significant traction in modern economy and most often appear to be the essence of discussion, debate and analysis in various fields including academics, politics, finance, investments and geostrategic. Currency market movements or exchange rate regime changes are transitory in nature and often sensitive to the economic damage and financial disorder resulting from the regional or global crises. Similar regime changes sometimes are recurring and some other times can be permanent because of structural breaks. Regime-switching models can explain the abrupt changes in currency prices and is useful in describing the non-linearity features of exchange rates (Ismail & Isa, 2006).

The choice of exchange rate regime for emerging markets is important not only in economic literature, but also in policy circles. Lessons from the recent crises in emerging markets led some economists, including Obstfeld and Rogoff (1995) and Eichengreen (1998), to suggest bipolar view of exchange rate regimes in a world of increasing international capital mobility. Newly industrializing emerging economies should settle at one of the two extremes; either opting for a freely floating currency or moving to a common currency like euro, since they are in the process of integrating into global capital markets. A middle of the road approach, that is pegged exchange rate regime, is inherently crisis prone for emerging market economies and these countries should be encouraged, in their own interest and for the broader interests of the international community to adopt floating rate regimes (Eichengreen, Masson, Savastano & Sharma, 1999). This would be a step subsequent to the adoption of floating regimes and would enhance trade and financial integration in the region. Moreover, this political decision would help economies to stabilize their local currencies (Pereira & Holland, 2009). However, the determinants of exchange rate regimes could explain better by the macroeconomic and structural variables on the exchange rate policy (Klein & Shambaugh, 2010; Rose, 2011).

BRICS constitute the world’s leading emerging economies. In particular, in the last decade, they have exhibited rapid economic growth and industrialization. Degree of economic integration, which is explained by the depth of trade in goods and services, and the level of factor mobility decide the magnitude of benefits from this country group. Symmetrical currency regimes and exchange management presuppose the success of economic integration. However, the wide relative fluctuations of the prevailing exchange rates appear to be a major barrier to increased trade within the region. Exchange rate stability in the region can make light of the price disturbance disrupting economic activities there. Moreover, asymmetric disturbances place less importance to exchange rate as a price adjustment tool and provide more thrust to monetary unification. Strong economic integration expects to make a common currency which will best serve the economic interests of the countries participated.

The rest of the article proceeds as follows. The second section reviews the literature relating to the regime-switching behaviour of real exchange markets and the usefulness of Markov Switching Model (MSM) in capturing such behaviour in advanced/emerging market context. While the third section presents the objectives and rationale of the study, fourth section explains the data, sample and empirical model used in this research. In fifth section, we document the data descriptive and the empirical results. Finally, the sixth section concludes the article with managerial implications and limitations.

Review of Literature

Purchasing power parity (PPP) explanation has a substantial explanatory power for the behaviour of exchange rates like US Dollar and Japanese Yen (Kulkarni & Ishizaki, 2002). The fundamental notion of PPP theory is that the law of one price should always apply in the long run so that the cost of representative portfolio will be same for economies after bringing into a common currency (Froot, Kim & Rogoff, 1995). In economic literature, we should say that the real exchange rate should be stationary. However, a significant number of studies (e.g., Frankel, 1986, 1990; Meese & Rogoff, 1988; Mark, 1990; Rossi, 2006) failed to reject the unit root null hypothesis relating to exchange rate and thus challenging the PPP theory. Even the ADF test, the most versatile among the group of unit roots has the low power in the presence of structural break, since it is difficult to separate the stationary process subject to structural break and a unit root process. To investigate this possibility, one should estimate a MSM (Bergman & Hansson, 2005).

Goldfeld and Quandt (1973) first proposed MSM, but the model did not gain popularity until the publication of Hamilton (1989). Now the model found immense application in finance, especially in price forecasting and volatility dynamics. This approach is much useful when we think of a series to undergo shifts from one state of behaviour to another and back again, but where the forcing variable, which causes that shift, is unobservable (Brooks, 2008). MSM is more specific while explaining the dynamic switching structure of variables with two or more states and helps researchers to make out how they have developed in the past and how they may adjust in the future (Goutte & Zou, 2011).

Engel and Hamilton (1990) are the first among the researchers who have found that MSM is a good approximation to the currency data series. Engel (1994) extends this work by describing the behaviour of 18 exchange rates. His research found that the model fits well for many exchange rates despite it failed to produce superior forecasts to a random walk or the forward rate. Later, Engel and Hakkio (1996), using MSM, were able to perfectly match the changes in exchange rate with the periodic extreme volatility. Marsh (2000) could make a well estimate of exchange rate on daily basis by MSM, though the out-sample forecasting is very poor due to parameter instability.

Kirikos (2000) makes a vivid discussion on the fitness of MSMs in exchange rate process and their soundness in generating superior forecasts relative to the random walk models. Bollen, Gray and Whaley (2000), Dewatcher (2001) and Fiess and Shankar (2005) made accurate portraits of the ability of regime-switching model to capture the dynamics of foreign exchange rates. Ichiue and Koyama (2007) found the influence of regime switch on exchange rate returns in the relationship between returns and interest rate differentials. Lee and Chen (2006) observe that Markov kind of time series process is consistent with the most popular system of dirty floating exchange rate regime. The recent research of Naszodi (2011) could identify well-defined switching behaviour between high and low intervention regimes.

In literature, we find some studies using MSMs for the prediction of exchange rate behaviour in different country contexts. MSM classifies the currency regimes and delivers information about the change of currency prices in some Asia-Pacific economies (Wu, 2015). Bergman and Hansson (2005) prove that the MSM is successful in forecasting the exchange rates of industrialized nations against the US dollar. Cheung and Erlandsson (2005) present a systematic and extensive empirical study on the presence of Markov switching dynamics in three dollar-based exchange rates—DEM, Deutsche mark, GBP, British pound and FFR, French franc. Monthly data used in the study offer explicit evidence of the presence of Markov switching dynamics in the observed currency series. Their study also suggests that data frequency, in addition to sample size, is crucial for determining the number of regimes under Markov-switching dynamics.

Studies exhibit evidence in favour of converging currency behaviour in emerging markets too. The currency depreciation in Thailand at the time of East Asian crisis quickly developed into an all-out financial crisis and spread over to major economies in the region (Rao, 2000). MSM as the best-fitted model in identifying the structural breaks in exchange markets of three ASEAN countries—Malaysia, Singapore and Thailand (Ismail & Isa, 2006). Real effective exchange rates and money supply ratios are most important in understanding the exchange rate turbulence in South East Asia (Brunetti, Mariano, Scotti & Tan, 2008). Parikakis and Merika (2009), using Markov-switching Monte Carlo approach and rejecting random walk hypothesis, provide evidence in favour of MSMs for precisely predicting the exchange rate movements of Euro against US dollar and British pound. However, the model loses its forecasting power against Brazilian real and Mexican peso. The empirical results of Kumah (2007) find the periods of exchange market pressure in the Kyrgyz Republic and substantiate the statistical superiority of the non-linear regime-switching model over linear econometric forecasting models in making out foreign exchange market volatility there.

Objectives

This article attempts to identify the volatility and regime shifts in real exchange markets in the BRICS region. More specifically, the study addresses the potential of more comprehensive economic partnership agreement among BRICS members in the form of a currency union that eventually leading to the formation of a common currency which can possibly play a proactive role in global trade and capital mobility in the future.

Rationale of the Study

The research investigating the dynamic switching structure of currency systems are not much extensive in literature. Moreover, most of the academic exercises on this issue have taken place in developed market context. This research is an improvement over previous research in many respects. First, this study adds the construct validity of Markov-switching approach in currency markets predictions that are less investigated in emerging markets. Second, most empirical research does not formally consider the number of regimes in the data that we propose to do in this article. Finally, the study indirectly explores central banks’ intervention behaviour among BRICS region, the largest emerging market group constituted by the fastest growing economies in the world. We do not find any other research focusing on this issue in a detailed manner.

Methodology: Data, Sample and Empirical Model

Data and Sample

The data include monthly observations on the period average real effective exchange rates for five countries (Brazil, Russia, India, China and South Africa) taken from the BIS Financial Statistics database. The sample spans from the first month of 1996 to the fourth month of 2015. The effective sample of observations is 232. We normalize the real exchange rate to unity and employ 100 times the natural logarithm of the real exchange rate for the analysis. The research estimates switching models separately for two distinct phases. At first, MSM analyses 180 observations, 1996:1 to 2010:12; this helps us to make out the real exchange market conditions of BRICS countries before their integration. We keep the remaining 52 observations of the sample (2011:1 to 2015:4) for evaluating the effect of integration in bringing the unification of currency regimes into the region. We use Gretl (version 1.9.4) to derive the needed empirical results.

Empirical Model

The study uses MSM for analysing the regime-switching behaviour of real exchange markets of BRICS countries. Under the Markov-switching approach, the universe of possible occurrence is split into ‘m’ states of the world, denoted as s_i, I = 1… m corresponding to ‘m’ regimes. In other words, it is assumed that y_t switches regime according to some unobserved variable, s_t that takes on integer values. If s_t = 1, the process is in regime 1 at time t, and if s_t = 2, the process is in regime 2 at time t. Movements of the variable between regimes are governed by a Markov process.

Markov-switching process suggests performing under either dynamic switching (MS-DR) structure or auto regressive (MS-AR) structure (Bergman & Hansson, 2005). The MS-DR model that allows a quick adjustment after a state change often used to model monthly and higher frequency data (Sanchez, 2016). MS-AR model that allows a gradual adjustment after the process state change is often used to model quarterly and low-frequency data. Since our research used monthly series of exchange rates in BRICS region, we follow dynamic structure of Markov-switching process.

MSM fits dynamic regression models that exhibit different dynamics across unobserved states using state-dependent parameters to accommodate structural break or other multiple state phenomena. Since the transition between the unobserved states follows a Markov chain, we call it MSM. Now the model found immense application in social science research especially in price forecasting and volatility dynamics.

MSM assumes that the observed change in a variable between period t and t + 1 is a random draw from one of two or more distributions. Hence, we conveniently consider it as a switching regression (see Hamilton, 1994, Chapter 19).

\begin{array}{l} η r_{t} = β_{1} z_{1 t} + ε_{t}, i f s_{1} = 1 \\ β_{2} z_{2 t} + ε_{t}, i f s_{2} = 2 \end{array}

(1)

where ηr_t is the series to be explained, Z_t is a vector of exogenous repressors’, β_t is the vector of real numbers, whose value depends on the non-observable state variables s_t and ε_t is the Gaussian white noise.

The regime-generating process assumes an Ergodic Markov chain and is characterized by the matrix η consisting of the transition probabilities p_ij from state i to state j.

η = Pr (s_t = 1/s_t-1 = 1) = p₁₁, Pr(s_t = 1/s_t-1 = 2) = p₂₁, Pr(s_t = 2/s_t-1 = 1) = p₁₂, Pr(s_t = 2/s_t-1 = 2) = p₂₂

Thus, P_ij = Pr (s_t = j/s_t_-1 = i).

Regime-switching models are most often estimated by maximum likelihood given the relatively simple form of the distribution of the data (Bollen et al., 2000). The log likelihood function to be maximized is

\sum_{i = 1}^{t} I n [f (y_{t} / {\overset{─}{y}}_{t - i})]

(2)

where:

{\overset{─}{y}}_{t - i} = {y_{t - 1}, y_{t - 2} \dots \dots}

(3)

which is equivalent to

\sum_{i = 1}^{t} I n [\sum_{i = 1}^{t} f (y_{t} / {\overset{─}{y}}_{t - i}; s_{t} = i) \Pr (s_{t} = i / {\overset{─}{y}}_{t - i})]

(4)

In brief, regime-switching model allows for shift in mean and variance to occur independently, that is, for periods of stable and unstable appreciation and for periods of stable and unstable depreciation.

In this article, we apply MS-DR model to log changes in real exchange rates in BRICS region using equation (5),

y_{t} = I n (s_{t} / s_{t - 1})

(5)

Equation (6) expresses our switching model in more specified form,

Y_{t} = μ s_{t} + α X_{t} + Z_{t} β s_{t} + ε t, ε_{t} \to N [0, σ_{s}^{}_{}^{2}]

(6)

where, S_t represents the real effective exchange rate in USD, Y_t is dependent variable, X_t is vector of exogenous variables with state invariant coefficients α, Z_t is vector of exogenous variables with state-dependent coefficients β_s. We model Y_t as being conditionally normal when the mean and variance depend on the regime that is operative.

Analysis

Real Exchange Rate Movement in BRICS Region: Trend and Pattern

Figure 1 depicts the logarithm of the five real exchange rates. All real exchange rates, except Ruble, showed a great deal of volatility over time, despite the time points of their changes are uneven. Brazilian real exhibited a steep declining trend until 2003 and then a sharp rising trend, which was broken in 2008. Long swings are evident for the Indian rupee despite the overall trend is upward until 2007 before it fell sharply over the next few months. Ruble and rand registered quite opposing behaviour during the period. While the real exchange rates of ruble were continuously appreciating except some major but short-lived corrections in 1998/1999, the exchange values of South African rand steeply declined throughout the period barring its trend corrections between the periods 2002 and 2005. The behaviour of Chinese yuan real exchange rates in general is quite different from the other currencies. In fact, increase was steep until 1998, thereafter declined moderately through the next three years before reverting to its previous height in 2003. Next three years the real exchange values of yuan were depreciating and finally it regained its previous momentum in the year 2006 onwards. Then, these dissimilarities seemed to be a sign of the need of BRICS currency cooperation once the member economies go for economic integration in future.

Figure 2 compares the real exchange rate movement in BRICS countries after the formation of the group. All real exchange rates, except yuan, were continuously depreciating at least until 2013. Yuan real exchange rates, this period also, has come out with odd behaviour against others. Yuan registered steep rising trend throughout the period even with some moderate corrections in 2014. Ruble found almost stable until 2013 and then came to sharp reductions for the next few months before making some quick corrections. As in the previous phase, long swings are evident for the Indian rupee despite the overall trend is downward until 2013 and modest recoveries afterwards. Rand registered somewhat similar behaviour of rupee during the period. Rand was in down trend during most periods with the exception of its corrections during the last two years. The convergences in exchange rate behaviour appear to be the result of economic cooperation among the BRICS group with regard to the intervention policies of central banks there.

Estimates of the Markov-Switching Model for the Real Exchange Rates

Three Regimes

How many regimes we should allow in the modelling? We should answer this question at first. Previous researchers have argued for the use of two regimes. In our study, we would like to follow a normal modelling approach given the question of whether the real exchange rates in the BRICS region are having inimitable regime or not. Statistically, it is difficult to test for the number of regimes (Dahlquist & Gray, 2000). Standard asymptotic theory does not apply (Hamilton, 1994) and non-standard tests are much exhaustive for complex modelling. Even with these limitations, we extend two-regime model into its three-regime version under the impression that the third regime can capture the outlier associated with market extremities. Here we do not have our say on the ‘right’ number of regimes. Instead, we just want to spot out the analogy of the different regime states of exchange rates in the BRICS region especially after the integration. We therefore estimate a three-regime state model for each of the sub period analysis.

Smoothed Probabilities

Figures 3–7 cover plots of the smoothed probabilities measuring the probability that next month’s real exchange rate innovation will be drawn from the high-volatility regime, conditional on the entire sample information. The vertical lines in the figures mark special events. Solid lines denote a shift involving the country under observation, and dashed lines denote repositioning elsewhere in the system. The figures confirm highly persistent trends in BRICS exchange rates intercepted by unexpected changes that regime-switching models capture well.

Estimates of the MSM for the Real Exchange Rates: Pre-BRICS Formation Phase

Table 1 presents the standardized likelihood ratio (LR) statistics for this model. The test statistics and their associated p-values are reported for all the five real exchange rate series of BRICS region. The asymptotic p-values are calculated using Monte-Carlo samples. The standardized LR statistics are relatively high for yuan, ruble and rand. Moreover, the DR(3) model is rejected at 1 per cent level for all series and thus the standardized LR statistics find evidence in favour of the three-state MSM during the preceding years of BRICS formation.

The model is able to split the real exchange rate into three distinct regimes for each series, with the intercept in all regimes—μ₁, μ₂ and μ₃ are being positive for all countries, corresponding to depreciation of the domestic currency against the dollar. Interestingly, one can find that the currencies of Brazil and China are more variable at times when they were in regime 1, evidenced by their higher standard error. The standard deviations of these two currencies were almost four times that of Indian rupee and Russian ruble.

The parameters 5 and 7 in Table 1 give the values of P₁₁ and P₂₂, respectively, that is the chance of falling in state 1 given that the real exchange rate was in state 1 in the immediate preceding month and the probability of staying in state 2 given that the real exchange rate was in state 2 previously. The high values of these parameters signal that the regimes are highly stable with less than a 5 per cent chance of moving from a lower real exchange rate and vice versa for all the series of observations.

Table 1.

Estimates of the Markov Switching Model for the Real Exchange Rates: Pre-BRICS Formation Phase

Parameter	Brazilian Real	Russian Ruble	Indian Rupee	Chinese Yuan	South African Rand
µ ₁	3.999[0.016]	4.458[0.003]	4.478[0.003]	3.891[0.017]	4.324[0.011]
µ ₂	4.263[0.013]	4.527[0.004]	4.529[0.002]	4.209[0.001]	4.530[0.009]
µ ₃	4.561[0.009]	4.597[0.003]	4.612[0.003]	4.507[0.008]	4.656[0.006]
σ	0.078[0.004]	0.024[.001]	0.018[0.001]	0.0768[0.004]	0.055[0.003]
P₁₁	0.949	0.969	0.919	0.9585	0.929
P₁₂	0.034	0.019	0.044	0.021	0.035
P₂₂	0.934	0.924	0.934	0.981	0.928
P₂₃	0.027	0.067	0.039	0.010	0.022
LR test statistic	207.31	243.94	179.68	274.36	228.11
p-value	[0.000]	[0.000]	[0.000]	[0.000]	[0.000]

Source: Author’s calculations.

Estimates of the Markov Switching Model for the Real Exchange Rates: Post-BRICS Formation Phase

Statistical inference regarding the empirical validity of the three-state regime-switching process in second phase analysis also is carried out using LR tests. Here also, the standardized LR statistics are relatively high for rand, ruble and yuan. The test provides a rejection of the null of linearity for all countries at usual significance levels. Thus, LR statistics reported in Table 2 support the regime shift during the period of post-formation of BRICS.

At this phase also, the model is able to separate the exchange rates into three distinct regimes for each of the currency series except Indian rupee. The currencies are continuously depreciating against dollar, which is evident from the positive μ values. Still, Brazilian real and Chinese yuan found more volatile despite their rate of fluctuations almost halved. Indian rupee and South African rand moved at relatively stable rate.

Table 2.

Estimates of the Markov Switching Model for the Real Exchange Rates: Post-BRICS Formation Phase

Parameter	Brazilian Real	Russian Ruble	Indian Rupee	Chinese Yuan	South African Rand
µ ₁	4.373[0.019]	4.288[0.016]	4.480[0.005]	4.636[0.033]	4.363[0.011]
µ ₂	4.499[0.008]	4.646[0.005]	4.554[0.006]	4.670[0.010]	4.495[0.009]
µ ₃	4.632[0.014]	4.482[0.015]	–	4.798[0.009]	4.591[0.006]
σ	0.047[0.005]	0.035[0.005]	0.029[0.002]	0.043[0.005]	0.027[0.003]
P₁₁	0.849	0.747	0.999	0.954	0.999
P₁₂	0.032	0.219	0.001	0.047	0.001
P₂₂	0.924	0.976	–	0.953	0.879
P₂₃	0.001	0.023	–	0.000	0.063
LR test statistic	50.628	101.890	40.684	97.499	102.331
p-value	[0.000]	[0.000]	[0.000]	[0.000]	[0.000]

Source: Author’s calculations.

There exists significant difference between parameter values of P₁₁ and P₂₂ for most of the series after the formation of BRICS group. The higher values of P₁₁ for rupee, yuan and rand convince that there is only lower chance of moving from a lower real exchange rate. Higher P₂₂ values of real and ruble imply more stability in higher mean regime state and lower probability to moving to a lower mean regime state. Among the series, Yuan exhibited more stability at both ends of high-and low-mean regime states.

During first few months of the study period, the real exchange market of India and China were in the low-mean regime state (Table 3). Meanwhile the rest of the group—Russia, Brazil and South Africa were in high-mean regime state and they remained in the same state for the next many months. The number of observations for which the probability that the real exchange rate market is in the high-mean regime state exceeds 0.5 is 101 for Russia (56.10 per cent of the total) and 94 for South Africa (52.22 per cent of the total). Contrary to this, exchange market of India stayed only for a small fraction of the period (15 per cent of the total) and in more than 50 per cent of the period of observation, it continued in medium regime state. In fact, Indian rupees reached the high-mean regime state only by the ending months of observation. The number of observations has almost been evenly distributed over the three regimes for China. The figures corresponding to Brazil reveal that the proportions of the period during which its exchange market stayed in different regimes progressing from low-mean state to high-mean state. Thus, overall, the exchange rate market is more likely to be in the high-mean regime for Russia and South Africa, while it is likely to be medium in India.

During the initial months of the post-formation phase, the real exchange market of all members was in moderate regime state (Table 4). Such convergence in exchange rate movement conveys a strong message on the willingness of BRICS members to deepen and consolidate their partnership in the economic-financial area. When the rand has stayed in the moderate state for most part of the period (78.85 per cent), Brazil took almost one year before its exchange market adjusted to that state and then remained in the state for almost two-third of the post-formation phase. An interesting point to be noted here is that the countries such as China and Russia reached the high-mean regime state only by the ending months of observation and none of the group stayed in that state for a longer period. Indian rupee has never entered into that regime after the formation of the group. When the number of observations for which the probability that the real exchange rate market was in the low-mean regime state exceeds 0.5 for India and South Africa, it is more than 0.40 for that of China. A close analysis of these facts reveals stronger convergence among the real exchange markets of India, China and South Africa, the same we could consider as the outcome of special focus to financial sector by BRICS members as a new front of cooperation among them.

Table 3.

Distribution of Regimes: Pre-BRICS Formation Phase

Country	Regime 1			Regime 2			Regime 3
Regime	Months	No.	%	Months	No.	%	Months	No.	%
Real	60–68 74–107	9 34	23.89	33–59 69–73 108–133 150–156	27 5 26 7	36.11	1–32 134–149 157–180	32 16 24	40
Yuan	1–9 85–140	9 56	36.11	11–19 30–56 73–84 142–147	9 27 12 6	30	10–10 20–29 57–72 141–141 148–180	1 10 16 1 33	33.89
Rupee	1–11 25–46 73–86 92–94 152–161	11 22 14 3 10	33.33	12–24 47–72 87–91 95–131 145–151 162–166	13 26 5 37 7 5	51.67	132–144 167–180	13 14	15
Ruble	29–52	24	13.33	53–107	55	30.56	1–28 108–180	28 73	56.11
Rand	66–79 142–156	14 15	16.11	49–65 80–87 122–141 157–168	17 8 20 12	31.67	1–48 88–121 169–180	48 34 12	52.22

Source: Author’s calculations.

Table 4.

Distribution of Regimes: Post-BRICS Formation Phase

Country	Regime 1			Regime 2			Regime 3
Regime	Months	No.	%	Months	No.	%	Months	No.	%
Real	193–193 226–232	1 7	15.38	194–225	32	61.54	181–192	12	23.08
Ruble	224–227	4	7.69	181–221	41	78.85	222–223 228–232	25	13.46
Rupee	206–232	27	51.92	181–205	25	48.08
Yuan	201–222	22	42.31	181–200	20	38.46	223–232	10	19.23
Rand	206–232	27	51.92	186–189193–205	413	32.69	181–185191–193	53	15.38

Source: Author’s calculations.

Conclusion

This article used Markov-switching approach to model the real exchange rate of BRICS countries. The ultimate aim of the research is to investigate the potential of setting up a currency union in the region, for which the study employed switching models on the real exchange rate data of member countries of the group before and after its formation separately.

Even though, most central banks in the region do not reveal their intervention data, the estimated smooth probability of the regime-switching model provides good inference regarding their interventions in specific periods. Initially, before the actual formation of the group, the study finds that larger divergences exist in the real exchange market behaviour of the BRICS members. However, after the formation of the group, our regime-switching model predicts converging real exchange price trend in the region that acknowledges the integrative behaviour of the central banks in their monetary and market intervention policies. Interestingly, more amount of unification is visible in market intervention behaviour among the central banks of India, China and South Africa.

Managerial Implications

The inclusion of the stronger policy interaction in BRICS region during recent times, especially in monetary management, brings convergence in their real exchange market behaviour. Deviations from the Law of One Price are considerably low within a currency union (Cavallo, Neiman & Rigobon, 2015) and such currency union reduces its trade cost at least at intra-regional level (Macedoni & Davis, 2017). Hence, forging closer economic integration among BRICS members in the form of a currency union gives their monetary cooperation a transcontinental dimension, making its role more proactive and significant in promotion of global trade and investment. Such a strategic economic partnership, complementing and strengthening the bilateral and multilateral relations between member states, gives their cooperation a better title and image, and will contribute to the sustainable and faster economic growth and competitiveness in the global arena (BRICS Summit, 2015). The economies may converge for policy aiming to strengthen geographically distant, but economically powerful regions. The region can be open enough to take advantage of their capabilities.

Limitations

The findings and implications of this research are limited to the real exchange market behaviour of BRICS countries. Many other parameters like monetary policy, money supplies, etc. are also vital to refer before concluding with the possibilities of formation of currency union in the region. Only five years have passed after the formation of the group and we should wait for longer period to enhance the validity and significance of the study in this regard.

Footnotes

Acknowledgements

The authors are grateful to the anonymous referees of the journal for their extremely useful suggestions to improve the quality of the article. Usual disclaimers apply.

References

Bergman

U. M.

& Hansson

(2005). Real exchange rates and switching regimes. Journal of International Money and Finance, 24(1), 121–138.

Bollen

Gray

& Whaley

(2000). Regime switching in foreign exchange rates: Evidence from currency option prices. Journal of Econometrics, 94(1–2), 239–276.

BRICS Summit. (2015, July 9). 7th Summit in Ufa, Russia.

Brooks

(2008). Introductory econometrics for finance. New York, NY: Cambridge University Press.

Brunetti

Mariano

R. S.

Scotti

& Tan

A. H.

(2008). Markov switching GARCH models of currency turmoil in Southeast Asia. Emerging Markets Review, 9(4), 104–128.

Cavallo

Neiman

& Rigobon

(2015). The price impact of joining a currency union: Evidence from Latvia. IMF Economic Review, 63(2), 281–297.

Cheung

& Erlandsson

(2005). Exchange rate and Markov switching dynamics. Journal of Business and Economic Statistics, 23(3), 314–320.

Dahlquist

& Gray

S. F.

(2000). Regime-switching and interest rates in the European monetary system. Journal of International Economics, 50(2), 399–419.

Dewatcher

(2001). Can Markov switching models replicate chartist profits in the foreign exchange market? Journal of International Money and Finance, 20(1), 25–41.

10.

Eichengreen

(1998, December). The only game in town. The World Today. Retrieved 21 August 2016 , from https://eml.berkeley.edu/~eichengr/policy/worldtod.pdf

11.

Eichengreen

Masson

Savastano

& Sharma

(1999). Transition strategies and nominal anchors on the road to greater exchange-rate flexibility (Essays in International Finance No. 213). Princeton, NJ: Princeton University.

12.

Engel

C. M.

(1994). Can the Markov switching model forecast exchange rates? Journal of International Economics, 36(1–2), 151–165.

13.

Engel

C. M.

& Hamilton

J. D.

(1990). Long swings in the dollar: Are they in the data and do markets know it? American Economic Review, 80(4), 689–713.

14.

Engel

& Hakkio

(1996). The distribution of exchange rates in the EMS. International Journal of Finance and Economics, 1(1), 55–67.

15.

Fiess

& Shankar

(2005). Regime switching in exchange rate policy and balance sheet effects (Policy Research Working Paper Series No. 3653). Washington, DC: The World Bank.

16.

Frankel

J. A.

(1986). International capital mobility and crowding-out in the U.S. economy: Imperfect integration of financial markets or of goods markets? In Hafer

R. W.

(Ed.), How open is the US economy (pp. 33–67). Lexington, KY: Lexington Books.

17.

Frankel

J. A.

(1990). Zen and the art of modern macroeconomics: A commentary. In Haraf

W. S.

& Willett

T. D.

(Eds), Monetary policy for a volatile global economy (pp. 117–123). Washington, DC: American Enterprise Institute for Public Policy Research.

18.

Froot

K .A.

Kim

& Rogoff

(1995). The law of one price over 700 years (Working Paper No. 5132). Cambridge, MA: National Bureau of Economic Research.

19.

Goldfeld

S. M.

& Quandt

R. E.

(1973). A Markov model for switching regressions. Journal of Econometrics, 1(1), 3–16.

20.

Goutte

& Zou

(2011). Foreign exchange rates under Markov regime switching model (CREA Discussion Paper Series). Luxembourg City: University of Luxemburg.

21.

Hamilton

J. D.

(1989). A new approach to the economic analysis of non-stationary time series and the business cycle. Econometrica, 57(2), 357–384.

22.

Hamilton

J. D.

(1994). Time series analysis. Princeton, NJ: Princeton University Press.

23.

Ichiue

& Koyama

(2007). Regime switches in exchange rate volatility and uncovered interest rate parity (Working Paper 07-E-22). Tokyo: Bank of Japan.

24.

Ismail

M. T.

& Isa

(2006). Modeling exchange rates using regime switching models. Sains Malaysiana, 35(2), 55–62.

25.

Kirikos

D. G.

(2000). Forecasting exchange rates out of sample: Random walk vs Markov switching regimes. Applied Economics Letters, 7(2), 133–136.

26.

Klein

M. W.

& Shambaugh

J. C.

(2010). Exchange rate regimes in the Modern Era. Cambridge, MA: MIT Press.

27.

Kulkarni

K. G.

& Ishizaki

(2002). Purchasing power parity theory revisited: Recent evidence from US Dollar and Japanese Yen. Global Business Review, 3(1), 53–61.

28.

Kumah

F. Y.

(2007). A Markov switching approach to measuring exchange market pressure (Working Paper No. WP/07/242). Washington, DC: International Monetary Fund.

29.

Lee

H.-Y.

& Chen

(2006). Why use Markov-switching models in exchange rate prediction? Economic Modelling, 23(4), 662–668.

30.

Macedoni

& Davis

U. C.

(2017). Has the Euro shrunk the band? Trade costs in a currency union. Retrieved 13 July 2017 , from www.lucamacedoni.com/uploads/7/2/1/9/72195123/euro_tar.pdf

31.

Mark

N. C.

(1990). Real and nominal exchange rates in the long run: An empirical investigation. Journal of International Economics, 28(1–2), 115–136.

32.

Marsh

I. W.

(2000). High-frequency Markov switching models in the foreign exchange market. Journal of Forecasting, 19(2), 123–134.

33.

Meese

R. A.

& Rogoff

K. S.

(1983). Empirical exchange rate models of the seventies: Do they fit out of sample? Journal of International Economics, 14(1), 3–24.

34.

Naszodi

(2011). Exchange rate dynamics under state-contingent stochastic process switching. Journal of International Money and Finance, 30(5), 896–908.

35.

Obstfeld

& Rogoff

(1995). The mirage of fixed exchange rates. Journal of Economic Perspectives, 9(4), 73–96.

36.

Parikakis

G. S.

& Merika

(2009). Evaluating volatility dynamics and the forecasting ability of Markov switching models. Journal of Forecasting, 28(8), 736–744.

37.

Pereira

L. C. B.

& Holland

(2009). Common currency and economic integration in Mercosur. Journal of Post Keynesian Economics, 32(2), 231–234.

38.

Rao

V. V. B.

(2000). The East Asian crisis of 1997–98: Implications for India. Global Business Review, 1(1), 1–10.

39.

Rose

A. K.

(2011). Exchange rate regimes in the modern era: Fixed, floating, and flaky. Journal of Economic Literature, 49(3), 652–672.

40.

Rossi

(2006). Are exchange rates really random walks? Some evidence robust to parameter instability. Macroeconomic Dynamics, 10(1), 20–38.

41.

Sanchez

(2016). Introduction to Markov-switching regression models using the mswitch command. Retrieved 15 January 2017 , from http://www.stata.com/meeting/mexico16/slides/Mexico16_Sanchez.pdf

42.

J.-T.

(2015). Markov regimes switching with monetary fundamental based exchange rate model. Asia Pacific Management Review, 15(2), 79–89.