FOREIGN EXCHANGE MARKET INTERVENTION AND NEUTRALIZATION The Indian Experience under Controlled Floating of Rupee K.G. SAHADEVAN, Ph.D Associate Professor Indian Institute of Management Prabandh Nagar Off Sitapur Road Lucknow – 226 013 INDI e-mail: [email protected] This paper has benefited greatly from the discussions with Prof. Bandi Kamaiah of Department of Economics at University of Hyderabad. The author wishes to acknowledge him and an anonymous referee for their valuable suggestions on the earlier draft of this paper. However, the usual disclaimer applies. This is the modified version of the paper titled “Balance of payments, exchange rates and neutralisation: The Indian experience under controlled floating of rupee” presented at the Second Conference on Money and Finance in the Indian Economy held in IGIDR, Mumbai during November 30 to December 2, 1999.

SUMMARY This study attempts to analyze the impact of monetary policy on the behavior of rupee exchange rate and international reserves during the controlled floating of rupee. It analyses how the Reserve Bank of India offsets the pressure that monetary shocks exert on exchange rates and reserves. Based on the estimates of Girton -Roper model o exchange market pressure for the period between 1992:4 and 1999:3 the study examines the RBI’s policy of maintaining exchange rate and reserves. The estimates do indicate that RBI offsets though not completely the domestic monetary expansion (contraction either by depreciation (appreciation) of rupee or by running down (accumulating foreign exchange reserves or by some combination of both. The values of offset coefficient ranging between –0.81 and –0.93 signify that the pressure on exchange rate and reserve level is partially being neutralized by some other means. The controls on international trade and capital flows do provide significant insulation from exchange market pressure. When exchange rates and reserve levels are considered to be the indicators of government’s performance and when they are being maintained at ‘politically correct’ levels, the economic reasoning underlying the model becomes irrelevant. The statistically significant intercept term as against the postulation of the model is a manifestation of these institutional realities. Key Words: Controlled float of rupee; Exchange market pressure; Foreign exchange market intervention; Neutralization; Sterilization.

FOREIGN EXCHANGE MARKET INTERVENTION AND NEUTRALIZATION The Indian Experience unde Controlled Floating of Rupee I. INTRODUCTION The Keynesian framework for analyzing balance of payments (BoP) and exchange rates suffered a setback with the collapse of Bretton Woods system consequent up on the massive inflationary pressure unleashed by the US government through its liberal monetary policy especially during the later half of 1960s. The Smithsonian Realignmen of 1971 by which dollar was devalued by 8 per cent against gold and a revaluation of other leading currencies in the world including J apanese yen and deutsche mark proved insufficient to compensate the harm that has already done to the system by monetary indiscipline. This has once again turned the attention to the classical prescriptions. During the early 1970s the monetary approach, which is more of classical in spirit gained prominence and offered a widely acceptable explanation to the then prevailing exchange rate and BoP conditions. This new approach provided a broad framework for analyzing BoP problem through explicit specificatio n of monetary behavior, and its integration wit ‘real’ factors. Any change in domestic component of monetary base, according this approach is ultimately offset by an equal and opposite changes in international reserve component through BoP in a fixed exc hange rate system while under floating rates system, monetary policy changes are transmitted to exchange rates through purchasing power parity channel. In its true sense, however, no currencies in the world are freely floated. Since the beginning of floating exchange rate system in 1973, most of the world’s major central banks have intervened frequently, and at times forcefully, in the foreign exchange markets to influence the path that their respective currencies have taken. Of 155 IMF member nations, only 26 have adopted independent free float and the path all those currencies have taken is often managed though the frequency and volume of each county’s central bank intervention in the market vary. Also, there are instances in which coordinated intervention by two or more central banks has been carried out in order to hold currencies from further adverse moves. The sharp drop in the dollar value following the September 1985 Plaza Agreement is often viewed as a major success story for such concerted coordinated intervention. Similarly, G-7 members intervened on behalf of the dollar, when it was under heavy downward pressure in 1987 and, when it rose sharply in 1989 the central banks of G-7 leaned heavily against the wind to limit the dollar’s rise. In 1991, Federal Reserve and Bundesbank bought dollars to prevent further depreciation of the dollar against deutsche mark. The Federal Reserve’s large-scale purchase of yen in June 1998 during the yen crisis and the Bank of Japan’s (BoJ) purchase of US dollar during August-September 1999 in its bid to prevent the yen from further appreciation against dollar are examples of central banks’ leaning against the wind for the sake of other currencies and for protecting the currency of its own. Moreover, currently talks are on about the possibility of joint intervention by BoJ and Federal Reserve to dampen the yen’s recent sharp rise against dollar. The central bank’s foreign exchange market intervention for making deliberate corrections in exchange rate has various implications. The monetary pressures that would normally influence just the exchange rate in a freely floating exchange rate regime would influence both the exchange rate and the level of foreign exchange reserves of 2

the central bank in a controlled floa ing exchange rate regime. In order to accommodate this new dimension of monetary policy in a controlled floating exchange rate system, Girton and Roper (1977) has formulated a variant of the original monetary model of exchange rates by taking account of the fact that central bank intervention often distorts the market forces to maintain exchange rate at a desired level. The objective of this paper is to examine the practices of Reserve Bank of India (RBI) in absorbing the pressure that its monetary policy exerts on exchange rates of rupee and international reserve position under the liberalized and market-related exchange rate system for the period between April 1992 and March 1999. It investigates empirically into how and to what extent the RBI offsets the pressure of monetary imbalance by a change in exchange rate or by change in reserve position or a combination of both so as to ensure monetary equilibrium in the context of present controlled floating exchange rat system. It estimates the offset coefficient using Girton-Roper (G-R) model of exchange market pressure (EMP). The monthly measures of bilateral and multilateral EMP and indices of absorption are calculated over the period 1992 – 1999 and these values are used to analyze the RBI’s conduct of exchange rate policy during this period. The remainder of this paper is organized as follows. Section 2 presents the rationale of central bank’s intervention in foreign exchange market and discusses the RBI’s practice of offsetting the pressure of monetary expansion or contraction through intermittent and deliberate corrections in exchange rate and/or changes in the level of foreign exchange reserves. An outline of the G -R model of EMP and its econometric implementation is summarized in Section 3. Section 4 carries discussion on the estimates of the model and the implications of findings. A summary of conclusions of the study is presented in section 5. A note on definition of variables, sources of data and methodology of the study is provided in appendix-I. II.

THE EXCHANGE MARKET PRESSURE AND CENTRAL BANK’S INTERVENTION

In the Indian context, in addition to the trade and capital controls imposed by the government, RBI uses its foreign exchange reserves for market intervention so as to align the market rate of rupee with its desired rate consistent with certain macroeconomic parameters. This official exchange rate management has a conventional objective of ensuring the currency not to deviate far away from the long-run equilibrium rate because as Krugman (1989) argued, “financial markets are not to be trusted; they can drive exchange rates far away from a sensible value, doing real harm in the process.” However, other considerations, including maintaining export competitiveness, guarding currency against speculation, etc., often outweigh this objective and necessitates official intervention to lean against the wind. For instance, RBI has allowed rupee to depreciate against the US dollar around 2.5 per cent between April and June 1999 by deliberately abstaining from market. This was in view of the prevailing historically low level of inflation. Since the rupee value is already kept artificially high against the dollar, a small amount of depreciation is justified with an inflation rate of less than 2 per cent. This depreciation, it is believed that, will not harm the price level significantly considering slow industrial recovery. However, as industrial growth picks up it would become necessary to keep the rupee stable without further downward move so as to maintain 3

domestic price situation under control These objectives differ by country and from one period to the other. BoJ has consistently pursued a policy of leaning against the wind. Whenever yen rose against the dollar, the BoJ bought dollars and so ld yen to moderate the yen’s rise; and vice versa. While BoJ intervenes in order to moderate the trends in yen over time, the Bundesbank does intervene primarily for domestic monetary control. However, at times, Bundesbank has compromised this objective in order to get the exchange rate in alignment with its long-run rate.1 In the case of dollar, however, Federal Reserve has never maintained a uniform policy for exchange rate management. While during 1978-80 it carried out major intervention to arrest dollar’s decline, the period between 1981 and 1984 witnessed benign neglect toward dollar’s rise. In line with the Plaza Accord dollar was encouraged to decline during 1985 -86 while during 1987 -92 Federal Reserve promoted greater stability of dollar in li ne with Louvre Accord. From 1993 onward, Federal Reserve has encouraged the dollar to decline. The market intervention has various implications, and it is designed to fulfill certain intentions of the central bank depending on the choice between sterilized intervention and non-sterilized intervention. The sterilized intervention through open market operations offsets the change in net foreign assets by a corresponding change in net domestic assets. This in turn helps the central bank to adhere to monetar targets. The non-sterilized intervention, on the other hand, creates a mismatch between supply of and demand for money eventually leading to change in exchange rate in the medium term. This intervention is effective only if its volume is sizable relative to the outstanding stock of domestic money holding. However, “most studies conclude that the direct effect of intervention on exchange rates is either statistically insignificant or quantitatively unimportant” [Rosenberg (1996)]. 2 The rupee though remained managed to a great extent until 1993 has been vulnerable to changes in monetary conditions from time to time. An alarming growth of broad money (M3) at annual average rate above 18 per cent from mid-1980s has mounted enormous pressure on exchange rate as well as on BoP. In spite of a gradual release of this pressure by RBI through changes in rupee rates or foreign exchange reserve position or a combination of both over a period of time, a major revision of exchange rate has taken place in 1991. Though no serious attempt has taken place so as to reduce money supply significantly thereafter up to 1994 -95, the government has alleviated the pressure on currency primarily by trade and capital controls and by a reasonable amount of sterilized intervention by RBI. This is evident from figure 1, 2 and 3. Figure 1 It is clear from figure 1 that there is no apparent relationship between exchange rates and reserves. Rupee has been consistently depreciating against dollar since September 1995 while reserves have shown an increasing trend. This diverging trend indicates that not much of reserves have been used to prevent the rupee slide. Against the sharp increase of foreign exchange reserves from Rs. 30,745 crore to Rs. 74,812 crore the exchange rate of rupee was held almost constant at around 31.5/dollar during the period between March 1993 and August 1995. During this period RBI has very actively intervened in the market to maintain a stable rupee aiming at building confidence among the internationa community which was necessitated after the introduction of unified and market related exchange rate system in March 1993. 4

Figure 2 As the figure 2 shows, there are instances in which RBI has turned out to be a net buyer of dollar when rupee was sliding. For instance, between October 1998 and March 1999, rupee depreciated from 42.25/dollar to 42.43/dollar while RBI has undertaken a net purchase of dollar to the tune of Rs. 10,879 during this period. All these indicate that the pressure that monetary growth exer ts on currency to depreciate (appreciate) is not always relieved by running down (accumulating) foreign exchange reserves. When government considers the size of foreign exchange reserves as one of the indicators of its success and not too wide is the exch ange rate variations to question the credibility of government policies, the only alternative for neutralizing the monetary growth is trade and capital controls. Figure 3 The figure 3, which plots monetary growth and exchange rate also confirms the fact that exchange rate of rupee does not really reflect the growth rates of money supply. Ideally, if the transmission mechanism works, the M3 growth and currency value should move in opposite directions i.e., a positive growth in money supply should be offset y a depreciation of exchange rates. The period of stable exchange rate between March 1993 and August 1995 has been the period of wild fluctuations in money supply growth ranging between 14 per cent to as high as 22 per cent. In spite of this, rupee remai ned relatively stable at around Rs. 31.50/dollar until August 1995. In September 1995, rupee closed at 34.00 per dollar and subsequently, remained stable at around 35.00 per dollar until July 1997. During 1995-97, a tighter monetary policy has brought down the growt of M3 to around 16 per cent on an average. Further, rupee depreciated substantially to move from 36.50/dollar to 42.50/dollar between August 1997 and June 1998, and thereafter it stabilized. Though M3 has not grown beyond 16 per cent during this period the fall in rupee value was on account of declining capital inflow and weak expor growth. III NEUTRALIZATION OF MONETARY SHOCKS THROUGH CHANGES IN EXCHAGE RATES AND RESERVES : AN EMPIRICAL FRAMEWOR The monetary model of EMP originally develop ed by Girton and Roper (1977) designs a measure that can represent more accurately the extent of monetary pressure bearing down on the foreign exchange market. Girton and Roper define the sum of the rate of change in international reserves, deflated by the level of the monetary base, plus the rate of change in exchange rate itself, as a proxy measure of the true amount of “exchange market pressure” i.e., the pressure on foreign exchange reserves and exchange rates when there is an imbalance between demand for and supply of national monies. The present study has used this model to examine the simultaneous adjustment of exchange rate and international reserves in the Indian context. The EMP model proposes that an increase in the rate of domestic credit, for a given rat of growth of world prices and permanent income, will result in an equ -proportionate loss in reserves with no change in exchange rate, or an equi-proportionate depreciation of domestic currency with no change in reserves or some combination of the two. Connolly and Silveira (1979), Modeste (1981), Kim (1985), Wohar and Lee (1992) and Weymark (1995) have applied the EMP model in the context of Brazil, Argentina, Korea, Japan and Canada respectively. An attractive feature of this model is tha t it draws upon 5

monetary models of BoP and of exchange rate determination. The two basic elements of the EMP model are the demand for and supply of money. The demand for money is a stable function of real income (Y), price level (P) and the Cambridge constant (k) and is specified as : Md = kPY … (a) Money supply (M s) is defined as money multiplier (m) times monetary base (R+D). Thus, Ms = m(R+D) … (b) Where R is net foreign asset and D is domestic credit money supply. The model assumes that in the long run there is equilibrium in money market. That is, Md ≡ Ms … (c) The model assumes that purchasing power parity (PPP) relationship holds continuously. The conditions for this international price parity is given by P = E Pf … (d) Where P is domestic price, E is exchange rate (in units of domestic currency) and f is foreign price. Taking logs on both sides of equations (a) and (b) and using equation (c), we ge ln P + ln k +ln Y = ln m + ln (R+D) … (e) Further, taking logs on both sides of equation (d) and using it in equation (e), we ge … (f) ln Pf + ln E + ln k + ln Y = ln m + ln (R+D) Differentiating3 equation (f) with respect to time t and writing p f, e, y and m * respectively for changes in logarithms of P f, E, Y and m, we ge 4 r – e = pf + y – m* – d … (g) 5 where r = (dR/dt)/(R+D) and d = (dD/d t)/(R+D). In equation (g), r and d are ratios of changes in reserves to monetary base and changes in domestic credit to monetary base respectively, p f is the rate of growth of world price index, y is rate of growth of income, m is the rate of growth of money multiplier and e is growth rate of exchange rate. Since the theory indicates the absence of an intercept term and if exchange rate is measured in terms of foreign currency units6, then equation (g) may be estimated in the following form rt + et = β1pf t + β2yt + β3m*t + β4dt + εt … (h) where εt is the error term and βs are coefficients to be estimated. The equation (h) is als known as EMP equation. The EMP i.e. the sum of r and e measures the pressure on foreign exchange reserves and the exchange rates when there is an excess of domestic money supply over money demand in a managed floating exchange rate system. This provides a measure of the volume of intervention necessary to achieve any desired exchange rate and foreign exchange reserves in a controlled floating exchange rate regime [Sahadevan (1993)]. The model hypothesizes the following restrictions on coefficients of the variables : β1 = β2 = 1 and β3 = β4 = -1 IV. ESTIMATES OF THE MODEL AND DISCUSSIONS The main purpose of this study is to empirically examine how RBI gets rid of the pressure that monetary expansion/contraction puts on exchange rate of rupee and foreign exchange reserve position under the liberali zed and market-related exchange rate system. The period chosen for this study is April 1992 – March 1999. The lower-end of the dat 6

period coincides with the introduction of market -related dual exchange rate system (recall that this system was replaced by a completely marke -related system in March 1993). The two long-run equilibrium conditions i.e., monetary equilibrium and purchasing power parity are imposed on estimates of the model which has utilized monthly data (see appendix-I for definition of var iables and methodology of the study). In spite of the fac that narrow money (M1) is a better representative of the transaction behavior of the public the broad money (M3) on which monetary target is set and has more multiplier impact in the economy, has been used for the purpose of empirical test. The estimates of EMP equation [equation (h)] using M3 are reported in Table 1 (line 1 and 5). For measuring the EMP, both real and nominal effective exchange rates (REER/NEER) with 36 country bilateral trade weights have been used. The upper pane of the table pertains to REER-based EMP and the lower panel to NEER -based EMP. All estimates of the parameters of equation are statistically significant with expected signs. As the theory indicates intercept term is not included in the equation which in turn, makes R2 as an inappropriate measure of the performance of the model. Figure 5 has plotted the actual and predicted EMP consistent with the model. More importantly, the offse coefficient [i.e. β4 in the equa ion (h)] which is of crucial importance in the model, and the coefficient of money multiplier ( β3) are close to minus one (-0.87 and –0.76 respectively) as predicted by the model signifying thereby that the central bank offsets though not completely the d mestic monetary expansion (contraction) either with depreciation (appreciation) of rupee in terms of REER or running down (accumulating foreign exchange reserves or some combination of both. However, there is a caveat in this when central bank undertakes sterilization. The monetary targeting often necessitates central bank to contract (expand) domestic credit proportionate to an autonomous reserve inflow (outflow) so as to prevent reserve inflow or outflow from affecting money supply. In such cases, changes in reserves trigger offsetting changes in domestic credit and makes the offset coefficient biased. A larger coefficient value for world prices (p f) indicates that the exchange rate and reserve levels are highly sensitive to the average price inflation of 23 industrial countries. This is consistent with the fact that more than 90 per cent of trade and capital transactions are with these industrial countries. In a very competitive marke a small change in price can make substantial changes in the market share. The coefficien of domestic income (y) is found to be relatively less significant with much smaller values. The use of a sector index does not proxy the overall growth of the economy. Similar results are obtained when NEER-based EMP equation is estimated. However, EMP being a composite measure, this result does not say whether r+e is sensitive to its composition between exchange rate and reserve changes because it is assumed that the implied trade-off between reserves and exchange rate changes is one-to-one. To test whether EMP is independent of its composition, the original EMP equation is modified as given below by including a measure of absorption (Q) as suggested in Connolly and Silveira (1979). rt + et = β1pf t + β2yt + β3m*t + β4dt + β5Qt + εt … (i) Where Q = (e-1)/(r-1). It indicates whether central bank absorbs EMP in exchange rat or in foreign exchange reserves. As they suggest, “the variable Q is a good measure of the way in which the monetary authorities absorb EMP because the m ore they let pressure be alleviated by depreciation relative to reserve losses, the greater is Q.” 7 A higher (lower) value of Q implies that the central bank alleviates EMP more by 7

depreciation (appreciation) relative to reserve losses (gains). The values of Q are reported in Appendix-II. The estimates of the modified equation [equation (i)] (line 2 and 6 in table 1) indicate tha the inclusion of Q has improved the explanatory power of the equation only marginally. While Q is statistically significant when REER has used, it is insignificant in NEER based EMP equation. This implies that the central bank does make choice between changes in REER and foreign exchange reserves and not between changes in NEER and foreign exchange reserves in response to monetary shocks. This is expected because of the fact that REER is a target variable while NEER serves as an instrument variable. A table containing the absolute values of EMP and its components along with Q is provided in appendix-II. It is clear from the able that the value of Q is more than unity in most cases implying that the central bank preferred exchange rate depreciation to the foreign exchange reserve losses in order to alleviate the EMP. To test whether the central bank has any choice between exc hange rate and foreign exchange reserves in alleviating the monetary pressure, the EMP equation for each component of EMP namely, r and e is separately estimated and the results of which are reported in table 1 (line 3, 4 and 7). When foreign exchange reserve were treated as dependent variable, the overall performance of the model has improved with relatively higher R2 of 0.62 by retaining the statistical significance of all independent variables except domestic income. This is expected when money and foreign exchange markets are getting better integrated each other and more foreign exchange is coming into the country through foreign institutional investments and Euro issues. However, the estimates of exchange rate equation have shown a poor performance o f the monetary model. These results indicate that foreign exchange reserves have taken more pressure than exchange rate (REER). This finding conflicts with the observations made on the basis of the absolute values of each component of EMP, and Q. This l eads us to conclude that the underlying specific economic reasoning of the model has not taken into account certain policy options available with the government through which it can neutralize the pressure on exchange rate and reserve position. Though dom estic monetary conditions influence exchange rates and reserve positions, their movements are controlled by various qualitative and quantitative restrictions on exports, imports and capital flows. In the presence of such controls, the pressure that the m netary imbalances exert on currency and foreign exchange reserves will get neutralized. A liberal monetary policy along with tighter controls on imports and/or an indirect subsidy on exports, for example, can balance the pressure on currency and foreign exchange reserves. The persisting capital controls also provide significant insulation from EMP. Figure 4 Figure 4 that has plotted M3 growth and foreign exchange reserves further confirms that there are some missing links between monetary growth and rese ve position. The figure indicates that up to the beginning of 1995 there has been no apparent negative relationship between these two. After March 1995, M3 growth has come down from a peak of 22 per cent to a range between 14 to 18 per cent until August 1998. This slackening in monetary growth has however shown a consistent increase in foreign exchange reserves. In spite of a further peak of M3 growth in October 1998, the foreign exchange reserves have grown continuously. This partial insensitiveness of exchange rate and foreign exchange reserves to monetary growth is expected when government 8

tries to keep exchange rate at ‘politically correct’ level because a fall in rupee or a reduction in foreign exchange reserve level are considered to be indications of government’s failure. This along with the wariness of inflation and of debt burden effects of the depreciation of currency often leads RBI to undertake of -market deals to keep the pressure on currency out of the market. Alternatively, the EMP model has been estimated using the bilateral nominal exchange rates of rupee against dollar and pound sterling. The results are presented in table 2. The estimates of rupee-dollar-based EMP equation have shown better results than that of the rupee-pound sterling equation. In both cases the variable Q has turned out to be statistically insignificant. This indicates that the EMP is not sensitive to its distribution between change in rupee -dollar or rupee-pound sterling rate and foreign exchange reserves. This finding leads us to conclude as confirmed earlier by the estimates of the multilateral EMP equation that the central bank alleviates the EMP either by changes in foreign exchange reserves or by changes in composite currency index REER and not by changes in rupee-dollar or rupee-pound sterling rates individually To summarize, the conflicting results of the study make one to conclude that the exchange rate and BoP are not essentially driven by monetary factors in India. The inclusion of an intercept term as against the specification of the EMP model has in addition to improving its explanatory power, proved itself to be statistically significant across the equations. For the sake of comparison with figure 5, which has plotted the actual against the mode -consistent predicted EMP, the actual against predicted (with intercept term) EMP are plotted in figure 6. Interestingly, the variable Q also found to be significant in all cases. These supplementary results are presented in appendix - III. These results further confirm that monetary imbalances are only being partially compensated by changes in exchange rates and international reserves. The EMP to either depreciate (appreciate) the currency keeping the foreign exchange reserves in tact or to run down (accumulate) reserves without any change in exchange rate or to choose som combination of both due to a liberal (tight) monetary policy can be absorbed by imposing higher duties or quantitative restrictions (liberalizing trade) on international trade and capital flows. Figure 5 Figure 6 V. CONCLUSIONS The present study has estimated the neutralization coefficient using Girton -Roper model of exchange market pressure. The monthly measures of bilateral and multilateral EMP and indices of absorption are calculated o ver the period 1992 -1999 and these values are utilized for analyzing RBI’s conduct of exchange rate policy during this period. The estimates of the model indicate that monetary policy has important bearing on exchange rate and reserve position in the Indi an context. The offset and money multiplie coefficients are found to be highly significant with expected signs across equations. These estimates signify that central bank offsets though not completely, the domestic monetary expansion (contraction) either by depreciation (appreciation) of rupee in terms of REER or by running down (accumulating) foreign exchange reserves or by som combination of both. In the further analysis it is found that foreign exchange reserves have absorbed more pressure than REER. This finding conflicts with that of the observations made on the basis of the values of measure of absorption Q. The value o Q, which is more than unity in most of the cases, implies that the central bank preferred 9

exchange rate depreciation to reserve losses to alleviate EMP. The inclusion of intercept term in model estimation as against the theory underlying the EMP model has not given results consistent with the predictions of the model. The statistically significant intercept term does caste doub s on the empirical validity of the model in the Indian context. This aberration however may be considered as a manifestation of the institutional realities that limit the application of the model in the Indian context. As long as exchange rate and reserv es are maintained at ‘politically correct’ levels, it is necessary to find alternative channels for absorbing monetary shocks. The restrictions on international trade and strict controls on capital flows are often resorted to for absorbing such pressures which otherwise would not have served political interests by leaving exchange rates and reserve levels untouched. Table 1 Estimates of Multilateral EMP Equations

Coefficient of Line

Dependent variable

1

pf

y

m*

d

Q

R2

r + ereer

6.20 (6.12)*

0.07 (1.94)**

-0.76 (-4.90)*

-0.87 (-5.73)*

-

0.40

18.48* (4,80)

1.97

2

r + ereer

3.84 (2.48)*

0.05 (1.39)

-0.84 (-5.33)*

-0.93 (-6.12)*

0.007 (1.98)**

0.42

16.11* (5,79)

2.02

3

r

5.13 (9.56)*

0.03 (1.47)

-0.79 (-9.58)*

-0.81 (-10.1)*

-

0.62

47.09* (4,80)

1.88

4

ereer

1.07 (1.18)

0.04 (1.30)

0.03 (0.20)

-0.06 (-0.42)

-

0.06

1.30 (4,80)

1.93

6

r + eneer

4.86 (4.80)*

0.09 (2.41)*

-0.80 (-5.12)*

-0.85 (-5.58)*

-

0.41

16.23* (4,80)

2.04

7

r + eneer

3.85 (2.44)*

0.08 (2.11)*

-0.83 (-5.15)*

-0.87 (-5.62)*

0.003 (0.83)

0.42

13.07* (5,79)

2.05

8

eneer

-0.27 (-0.29)

0.06 (1.8)***

-0.008 (-0.06)

-0.04 (-0.25)

-

0.05

1.32 (4,80)

1.93

F-statistics D-W

ereer and eneer is the real and nominal effective exchange rates of rupee respectively D-W is the Durbin-Watson statistic and values in parentheses indicate t -statistic. One, two and three asterisks respectively indicate significance at 1%, 5% and 10% levels The values in parentheses given under F-statistic are its degrees of freedom.

10

Table 2 Estimates of Bilateral EMP Equations having Rupee/dollar (e USD) or Rupee/pound (eGBP) Exchange Rates Component Dependent variable

Coefficients of pf

y

m*

d

Q

R2

F-statistics

D-W

ρ

r - eUSD

3.94 (4.36)*

-0.003 (-0.01)

-1.17 (-7.5)*

-1.10 (-7.3)*

-

0.47

18.76* (4,80)

1.95

-

r - eUSD

2.30 (1.50)

-0.011 (-0.30)

-1.22 (-7.56)*

-1.15 (-7.39)*

0.005 (1.31)

0.48

15.49* (5,79)

1.94

-

r

4.49 (9.28)*

0.010 (0.49)

-0.803 (-9.56)*

-0.797 (-9.82)*

-

0.61

45.04* (4,80)

1.75

-

eUSD

0.56 (0.77)

0.013 (0.44)

0.37 (2.89)*

0.30 (2.48)*

-

0.07

2.44** (4,80)

1.90

-

r - eGBP

1.70 (2.12)**

0.14 (2.36)**

-0.59 (-2.45)*

-0.467 (-2.0)**

-

0.13

3.82* (4,80)

1.68

-

r - eGBP

1.16 (1.25)

0.119 (1.91)**

-0.65 (-2.64)*

-0.53 (-2.24)*

0.005 (1.17)

0.15

3.34* (5,79)

1.68

-

r

-0.06 (-0.19)

0.032 (1.8)***

-0.723 (-8.50)*

-0.716 (-8.43)*

-

0.53

29.77* (4,79)

2.22

0.65 (7.48*

eGBP

-0.04 (-0.04)

-0.068 (-1.23)

-0.170 (-0.73)

-0.216 (-0.93)

-

0.07

0.45 (4,79)

1.90

0.22 (2.0)**

D-W is the Durbin-Watson statistic and value s in parentheses indicate -statistic. One, two and three asterisks respectively indicate significance at 1%, 5% and 10% levels The values in parentheses given under F-statistic are its degrees of freedom. ρ is the estimate of auto-correlation coefficient. Notes 1. After the formal unification of Europe, Bundesbank implements exchange rate policy and conducts foreign exchange operations consistent with the provisions of Article 109 of the Treaty of European Union. 2. Against the consensus view in the early 1980’s that central bank intervention is ineffective, Dominguez and Frankel (1993) in their seminal study argue that when the authorities are prepared to intervene at a particular upper or lower limit they will achieve a higher degree of success in stabilizing t he currency with a smaller amoun of intervention if they publicly announce these limits ahead of time. In the light of these findings, the Committee on Capital Account Convertibility in India headed by Shri. S.S. Tarapore recommended that a REER -monitoring band be declared to 11

enable the participants to anchor expectations on when RBI would intervene and when it would not for making its intervention more effective [Tarapore (1997)]. 3. Since k is constant and R and D are measured as (dR/dt)/(R+D) and (dD/dt) /(R+D) respectively, the equation (f) can be expressed as: d ln Pf d ln E d ln Y d ln m 1 dR 1 dD + + = + + …(f.i) dt dt dt dt R + D dt R + D dt 4. When ∆ replaces the term d/dt, the equation (f.i) becomes ∆R ∆D + p f + e + y = m* + R+D R+D ∆R ∆D − e = p f + y − m* − …(f.ii) or R+D R+D 5. This definition according to Weymark (1998) h as however posed certain ambiguit “because the monetary units in which reserve and domestic credit changes are measured are not commensurate with the units in which exchange rate changes are measured it is difficult to know how to interpret the magnitude of changes in terms of the underlying external imbalance that generated them”. 6. In this case, the term E in equation (d) will be replaced by (1/E) and then the equation (f) will read as : ln Pf – ln E + ln k + ln Y = ln m + ln (R+D …(f.iii) Accordingly, final equation (g) will be modified as …(g.i) r + e = pf + y – m* - d 7. According to Connolly and Silveira (1979), “the simple ratio e/r used in Girton and Roper (1977) does not have this desirable monotonic property since it is discontinuous f r values of r equal to zero”. Appendi - I Definition of Variables and Methodology of the Study a. Data The present study is carried out on monthly data for a period between April 1992 and March 1999. The variables used in the study are defined as follows: NEER Nominal effective exchange rate is the trade (exports plus imports) weighted average of the bilateral nominal exchange rates of rupee in terms of currencies of India’s 36 major trading partners with base year 1985. The exchange rates have been defined in indirect quotes so that the appreciation/depreciation of the rupee is directly reflected by a rise/fall in index value. A rise in index represents an appreciation of rupee relative to these 36 currencies and a fall represents depreciation. REER Real effective exchange rate is a weighted average of NEER adjusted by the ratio of domestic inflation rates to foreign inflation rates. Rupee/dollar and The Foreign Exchange Dealers’ Association’s (FEDAI) indicative rupee/pound rates exchange rates of the rupee vis-à-vis the US dollar and pound sterling respectively. They represent the month-end direct quotes for buy and hence, an increase in the rates indicates depreciation o rupee and decrease indicates appreciation of rupee. 12

B M1 M3 R D

Reserve money. Narrow money composed of demand-deposit money and currency money with the public. Broad money i.e. M1 plus time deposit liabilities of banks. Domestic currency value of net foreign exchange assets. Domestic credit component of monetary base (B) and it is defined as D = B – R. Broad money multiplier i.e., M3/B is used.

Money multiplier (m) Foreign exchange The rupee value of SDR, gold and foreign currency assets with RBI. reserves RBI’s purchase of Rupee equivalent of the net purchase (purchase minus sale) a dollar contract rate. It includes only spot sale/purchase of US dollar by RBI World price level The EMP equation having REER/NEER component uses the average of Consumer Price Indices (CPI) of 23 industrial countri es viz., USA, Canada, Australia, Japan, New Zealand, Euro area countries (11), Denmark, Greece, Iceland, Norway, Sweden Switzerland and UK. These countries put together cover more than 90 per cent of India’s international transactions. The EMP equation h aving rupee/dollar, and rupee/pound rates use the CPI of the respective countries. Domestic income Due to the non-availability of monthly income data, the index o (Y) industrial production (1980-81 = 100) is used as a proxy. The sources of data excep for world price and the CPI of US and UK are RBI Bulletin and RBI Report on Currency and Finance . The data on world price and the CPI of US and UK are collected from International Financial Statistics of the International Monetary Fund. b. Methodology The ordinary least square (OLS) method is used for estimating EMP equation. The possibility of serial correlation problem has been verified by using Durbin -Watson test statistics the values of which are reported against all estimated equations. In those cases where serial correlation is detected, the coefficient estimates are adjusted by using Cochrane-Orcutt method. In such cases the serial correlation coefficient ρ) is reported along with its t-statistics.

13

Appendix - II Values of Multilateral Exchange Market Pressure and its Components Year & month 1992 4 1992 5 1992 6 1992 7 1992 8 1992 9 1992 10 1992 11 1992 12 1993 1 1993 2 1993 3 1993 4 1993 5 1993 6 1993 7 1993 8 1993 9 1993 10 1993 11 1993 12 1994 1 1994 2 1994 3 1994 4 1994 5 1994 6 1994 7 1994 8 1994 9 1994 10 1994 11 1994 12 1995 1 1995 2 1995 3 1995 4 1995 5 1995 6 1995 7 1995 8 1995 9

r (1) -0.004 -0.002 0.020 0.005 -0.007 -0.013 -0.009 -0.014 -0.004 -0.003 -0.012 0.070 0.016 -0.010 -0.006 0.010 0.013 0.011 0.008 0.008 0.028 0.021 0.044 0.084 0.024 0.011 0.021 0.026 0.007 0.027 0.017 0.003 -0.011 0.003 0.000 0.033 -0.005 -0.002 -0.013 0.004 0.000 -0.004

Q ereer EMP (1+2) (e-1/r-1) (2) -0.045 0.019 -0.008 0.021 0.005 0.013 0.029 0.028 -0.016 0.003 -0.041 0.039 -0.020 0.003 0.015 0.023 0.011 0.005 0.024 0.008 0.003 0.008 -0.001 -0.007 0.022 0.004 -0.003 -0.007 0.000 -0.008 -0.006 0.007 0.023 0.001 -0.005 -0.035 -0.003 0.018 -0.001 0.007 0.019 -0.034

-0.049 0.017 0.012 0.026 -0.002 0.000 0.019 0.014 -0.020 0.000 -0.052 0.109 -0.004 -0.007 0.008 0.033 0.024 0.016 0.033 0.017 0.031 0.029 0.043 0.077 0.046 0.015 0.018 0.019 0.007 0.018 0.011 0.010 0.012 0.004 -0.005 -0.002 -0.008 0.015 -0.014 0.011 0.019 -0.038

1.04 0.98 1.03 0.98 0.99 0.97 0.96 0.96 1.01 0.99 1.03 1.03 1.04 0.99 0.98 0.99 1.00 1.01 0.98 1.00 1.03 1.01 1.05 1.10 1.00 1.01 1.03 1.03 1.01 1.04 1.02 1.00 0.97 1.00 1.01 1.07 1.00 0.98 0.99 1.00 0.98 1.03

Year & month

r (1)

1995 10 1995 11 1995 12 1996 1 1996 2 1996 3 1996 4 1996 5 1996 6 1996 7 1996 8 1996 9 1996 10 1996 11 1996 12 1997 1 1997 2 1997 3 1997 4 1997 5 1997 6 1997 7 1997 8 1997 9 1997 10 1997 11 1997 12 1998 1 1998 2 1998 3 1998 4 1998 5 1998 6 1998 7 1998 8 1998 9 1998 10 1998 11 1998 12 1999 1 1999 2 1999 3

0.019 -0.004 0.003 -0.002 0.002 0.019 0.005 -0.003 -0.003 0.014 0.006 0.034 0.010 0.010 0.001 0.003 -0.005 0.049 0.004 0.019 0.018 0.010 0.020 -0.013 0.010 -0.004 -0.006 0.007 -0.001 0.035 0.001 0.001 -0.004 -0.003 0.011 0.029 0.006 0.000 0.004 0.012 0.005 0.028

14

ereer (2)

EMP Q (1+2) (e-1/r-1)

-0.047 0.002 -0.008 -0.016 -0.023 0.066 0.019 -0.015 0.006 -0.001 -0.004 0.010 0.008 -0.006 0.008 0.010 0.024 0.003 0.008 -0.008 0.003 0.015 0.017 -0.005 0.007 -0.023 -0.017 0.029 -0.006 -0.017 0.003 -0.010 -0.020 0.007 0.004 -0.014 -0.019 0.006 -0.017 -0.002 0.018 0.009

-0.028 -0.002 -0.006 -0.018 -0.022 0.085 0.024 -0.017 0.003 0.014 0.003 0.044 0.018 0.003 0.009 0.013 0.019 0.052 0.011 0.010 0.021 0.025 0.036 -0.018 0.016 -0.027 -0.023 0.036 -0.007 0.019 0.004 -0.009 -0.024 0.005 0.015 0.015 -0.013 0.005 -0.012 0.010 0.023 0.037

1.07 0.99 1.01 1.01 1.03 0.95 0.99 1.01 0.99 1.02 1.01 1.02 1.00 1.02 0.99 0.99 0.97 1.05 1.00 1.03 1.02 0.99 1.00 0.99 1.00 1.02 1.01 0.98 1.00 1.05 1.00 1.01 1.02 0.99 1.01 1.04 1.03 0.99 1.02 1.01 0.99 1.02

Appendi - III Estimates of Multilateral and Bilateral EMP Equations with Intercept Term

Dependent variable

Coefficients of y m*

d

Q

R2

D-W

ρ

-0.86 (-5.42)

-0.94 (-6.20)

-

0.43

2.02

-

0.029 (0.94)

-1.39 (-8.97)

-1.42 (-9.74)

-0.54 (-6.3)

0.62

2.14

-

2.20 (3.05)

0.004 (0.248)

-0.90 (-12.12)

-0.894 (-12.6)

-

0.72

2.14

-

-0.001 (-0.346)

1.44 (1.02)

0.046 (1.34)

0.041 (0.284)

-0.046 (-0.33)

-

0.06

1.94

-

r - eUSD

0.006 (1.44)

2.14 (1.40)

-0.012 (-0.323)

-1.23 (-7.60)

-1.16 (-7.43)

-

0.48

1.93

-

r - eUSD

0.563 (4.93)

2.69 (1.99)

-0.010 (-0.315)

-1.57 (-9.90)

-1.52 (-9.75)

-0.55 (-4.88)

0.60

1.97

-

r

0.009 (5.05)

1.54 (2.12)

-0.005 (-0.255)

-0.907 (-11.88)

-0.89 (-12.1)

-

0.71

2.05

-

eUSD

0.004 (1.16)

-0.608 (-0.49)

-0.008 (0.25)

0.324 (2.47)

0.266 (2.12)

-

0.08

1.96

-

r – eGBP

0.006 (1.45)

1.04 (1.12)

0.113 (1.83)

-0.666 (-2.72)

-0.55 (-2.33)

-

0.15

1.67

-

r – eGBP

0.789 (12.32)

1.18 (2.17)

0.039 (1.06)

-1.63 (-9.92)

-1.55 (-9.64)

-0.769 (-12.2)

0.71

2.15

-

r

0.011 (10.16)

0.871 (3.09)

-0.001 (-0.034)

-0.997 (-13.66)

-0.954 (-14.0)

-

0.71

1.99

-0.25 (-2.0)

eGBP

0.005 (1.07)

-0.406 (-0.44)

-0.084 (-1.46)

-0.218 (-0.911)

-0.272 (-1.15)

-

0.08

1.91

0.20 (1.7)

intercept

pf

r + ereer

0.008 (2.17)

3.64 (2.37)

0.051 (1.35)

r + ereer

0.55 (6.38)

4.13 (3.27)

r

0.009 (5.30)

ereer

D-W is the Durbin-Watson statistic and values in parentheses indicate t -statistic. ρ is the estimate of auto-correlation coefficient

15

REFERENCES Connolly, M and Silveira, J.D. (1979) “Exchange Market Pressure in Post-war Brazil Application of the Girton-Roper Model”, American Economic Review, Vol. 69 (3), pp.448-454. Dominguez, Kathryn. M and J.A. Frankel. (1993), “Does Foreign Exchange Intervention Work?”, Institute for International Economics, Washington D.C. Girton, L and Roper, D. (1977), “A Monetary Model of Exchange Market Pressure Applied to the Post-war Canadian Experience”, American Economic Review, Vol. 67 (4), pp.537-548. Johnson, H.G. (1972), “The Monetary Approach to the Balance of Payments Theory”, in Further Essays in Monetary Theory, George Allen and Unwin, London. Kim, I. (1985), “Exchange Market Pressure in Korea : An Application of the GirtonRoper Monetary Model”, Journal of Money, Credit and Banking, Vol. 17(2), pp.258-263. Krugman, Paul A. (1989), Exchange-Rate Instability, Cambridge, MA: MIT Press. Modeste, N.C. (1981), “Exchange Market Pressure during the 1970s in Argentina : An Application of the Girton-Roper Monetary Model”, Journal of Money, Credit and Banking, Vol. 13 (2), pp.234-240. Rosenberg, Michael Roy. (1996), Currency Forecasting: A Guide to Fundamental and Technical Models of Exchange Rate Determination, McGraw-Hill, New York. Sahadevan, K.G, (1993), An Econometric Analysis of India’s Balance of Payments , unpublished Ph.D thesis, Department of Economics, University of Hyderabad, Hyderabad. Tarapore, S. S. (1998), “Exchange Rate Policy Reforms: Recent Initiatives”, Vikalpa, Vol. 23(1), pp.23-26. Weymark, D.N. (1995), “Estimating Exchange Market Pressure and the Degree of Exchange Market Intervention for Canada”, Journal of International Economics , Vol. 39(3,4), pp.273-295. Weymark, D.N. (1998), “A General Approach to Measuring Exchange Market Pressure”, Oxford Economic Papers, Vol. 50 (1), pp.106-121. Wohar, M.E and Lee, B.S. (1972), “An Application of the Girton -Roper Monetary Model of Exchange Market Pressure Model : The Japanese Experience, 1959 -1986”, Indian Journal of Economics, Vol.LXXII(287), pp.379-407.

16

forex reserves (Rs. Crore) 140000

120000

100000

80000

60000

40000

20000 1992 3 1992 7 1992 11 1993 3 1993 7

Figure 1 RBI's Exchange Rate Policy

1993 11 1994 3 1994 7

17

1995 3 1995 7 1995 11 forex reserves

Year and Month

1994 11

1996 3 1996 7 1996 11 1997 3 1997 7

Rs/US$ rates

1997 11 1998 3 1998 7 1998 11 1999 3

28

30

32

34

36

38

40

42

44

46

48

50

Rs/US$ rate (reverse scale)

Rs/US$ rates (reverse scale) 28

30

32

2000

34

38

0

36

40

42

44

46

48

50 1995 6

Rs/US $ rates

1995 8 1995 10 1995 12 1996 2 1996 4

Figure 2 RBI's Leaning Against the Wind

18

Year and Month

Purchase of US $

1996 6 1996 8 1996 10 1996 12 1997 2 1997 4 1997 6 1997 8 1997 10 1997 12 1998 2 1998 4 1998 6 1998 8 1998 10 1998 12 1999 2

10000

8000

6000

4000

-2000

-4000

-6000

-8000

RBI's Purchase of US $ (Rs.Crore)

Rs/US dollar rates 44

42

40

38

36

34

32

30

28

Rs/US$ rates

Figure 3 Exchange Rate and Monetary Growth

Year and Month

19

1992 3 1992 6 1992 9 1992 12 1993 3 1993 6 1993 9 1993 12 1994 3 1994 6 1994 9 1994 12 1995 3 1995 6 1995 9 1995 12 1996 3 1996 6 1996 9 1996 12 1997 3 1997 6 1997 9 1997 12 1998 3 1998 6 1998 9 1998 12 1999 3

M3 growth 24.00

22.00

20.00

18.00

16.00

14.00

12.00

10.00

Growth rates of M3

Forex reserves (Rs.crore) 140000

120000

100000

80000

60000

40000

20000 forex reserves

Figure 4 Monetary growth and Forex Reserves

Year and Month

20

1992 3 1992 6 1992 9 1992 12 1993 3 1993 6 1993 9 1993 12 1994 3 1994 6 1994 9 1994 12 1995 3 1995 6 1995 9 1995 12 1996 3 1996 6 1996 9 1996 12 1997 3 1997 6 1997 9 1997 12 1998 3 1998 6 1998 9 1998 12 1999 3

M3 growth 24.00

22.00

20.00

18.00

16.00

14.00

12.00

10.00

Growth rates of M3

Actual/Predicted EMP (r + e) 0.120

0.100

0.080

0.060

0.040

0.020

0.000

-0.020

-0.040

-0.060 1992 4 1992 7 1992 10 1993 1 1993 4 1993 7 1993 10 1994 1

1994 7 1994 10 1995 1

21

Year and Month

1995 4 1995 7 1995 10 1996 1 1996 4 1996 7 1996 10 1997 1

1997 7

Actual

1997 4

1997 10 1998 1 1998 4

1998 10 1999 1

Predicted

1998 7

Figure 5 Exchange Market Pressure in India

1994 4

Actual/Predicted EMP (r+e) 0.120

0.100

0.080

0.060

0.040

0.020

0.000

-0.020

-0.040

-0.060 Actual

Figure 6 Actual Vs Predicted (with intercept) EMP in India

Year and Month

22

Predicted

1992 4 1992 7 1992 10 1993 1 1993 4 1993 7 1993 10 1994 1 1994 4 1994 7 1994 10 1995 1 1995 4 1995 7 1995 10 1996 1 1996 4 1996 7 1996 10 1997 1 1997 4 1997 7 1997 10 1998 1 1998 4 1998 7 1998 10 1999 1