CHAPTER 8. The Evolution of Monetary and Exchange Rate Policy

James Yao, Gamal El-Masry, Padamja Khandelwal, and Emilio Sacerdoti
Published Date:
March 2005
  • ShareShare
Show Summary Details

As in many parts of the world, including developed countries, the modalities of conducting monetary policy in Mauritius evolved from strong reliance on controls to more market-based mechanisms. As in other countries, the 1970s were turbulent years for the conduct of monetary policy, and external price shocks, together with excessively expansionary domestic polices, led to an acceleration of inflation. Macroeconomic stability was regained in the early 1980s through the pursuit of appropriate adjustment policies and was maintained successfully thereafter. Overall, the monetary authorities exhibited a remarkable capacity to correct rapidly for slippages. They were steadfast in promoting the deepening of the financial system and in creating more sophisticated instruments for the conduct of monetary policy. They abandoned direct controls, moving to rely on interventions through the money markets, which gradually matured and became a more efficient channel for the transmission of monetary policy impulses. In all, the sound monetary and exchange rate policies contributed to foster an environment favorable to sustained growth.

The Early Years—A Controlled System

After being established in 1967, the BOM gave prominence to the objectives of diversifying the economy and fostering economic growth and employment through supportive actions. An expansionary policy was followed by keeping interest rates low and providing concessional credit to “priority sectors.” The “bank rate” was the rate at which the BOM lent to banks to signal the stance of monetary policy. Other interest rates in the system were to be determined on the basis of the bank rate. Concessional credit was provided to industrial enterprises, which were being supported to promote growth, diversification, and employment. EPZ companies and manufacturing companies were the targeted beneficiaries. Concessional credit was provided through two channels: rediscounting of commercial bills and direct lending to commercial banks for financing of exports and imports of raw materials. In addition, commercial banks were urged to direct credit toward priority sectors and away from consumption.

In the early 1970s, high international sugar prices resulted in large balance of payments surpluses that gave rise to a fast rise of liquidity and of the money supply. At this time, the exchange rate of the Mauritian rupee was pegged vis-à-vis the pound sterling. As credit expanded, import demand soared, and inflationary pressures started to build up. A wage-price spiral took hold, as the national wage-bargaining process led to large pay increases to workers to compensate them for increases in the cost of living.

The BOM responded to inflation by introducing in the early 1970s credit ceilings and cash ratio requirements, that were gradually tightened. Credit to the priority sectors was protected and exempted from credit ceilings, and special credit lines were created with the explicit purpose of financing these sectors at below-market rates. Moreover, the level of the official discount rate (the so-called bank rate) was such that real interest rates were negative over most of this period (Figure 8.1), while commercial banks were subject to regulations on the maximum deposit and lending rates.31 As a result of these factors, growth of credit to the private sector remained high through the 1970s (Figure 8.2).

Figure 8.1.Mauritius Monetary Policy—Bank Rate and Inflation, 1967–2003

(In percent)

Source: IMF, International Financial Statistics.

1 LIBOR is the London inter-bank offer rate on six-month dollar deposits.

Figure 8.2.Credit to Private Sector, 1967–2003

(Annual growth, in percent)

Source: IMF, International Financial Statistics.

In 1974, as sugar prices rose to record levels and the world suffered its first oil price shock, inflation in Mauritius reached a high mark of 29 percent. World sugar prices collapsed in 1975, and balance of payments surpluses turned into large deficits (Figure 8.3). Pressure mounted on international reserves as the credit expansion continued.32 In order to stem the losses of international reserves, the BOM raised interest rates in 1977 and tightened credit ceilings. A liquidity ratio requirement was also introduced in 1978. However, the second international oil shock in 1979, combined with the absence of adequate fiscal tightening, led to a severe balance of payments crisis. As the need of comprehensive adjustment policies became apparent, in 1979 Mauritius entered into a Stand-By Arrangement with the IMF in support of an economic and financial stabilization program. Under the program, the rupee was depreciated substantially, interest rates were increased, the fiscal deficit was reduced, and the wage policy was tightened to ensure competitiveness.

Figure 8.3.Balance of Payments, 1967–2003

(In millions of U.S. dollars)

Source: IMF, International Financial Statistics.

Despite an increase in interest rates under the stabilization program, the real bank rate stayed negative until 1983 and remained below the London inter-bank borrowed rate (LIBOR); see Figure 8.1. The velocity of quasi money, which had declined in the early 1970s, reflecting an increased willingness to hold time deposits, rose somewhat in the late 1970s; see Figure 7.1. The income velocity of money, which had been declining sharply since the late 1960s, reflecting growing confidence in the banking system, leveled off. It was not until the mid-1980s that the declining trends in income velocity of money and quasi money resumed. A statistical analysis of the demand for money in Mauritius is presented in Box 8.1.

Box 8.1.Money Demand in Mauritius

We examine the hypothesis that there exists a stable long-run relationship among the logarithm of real broad money (m-p), the logarithm of real gross domestic output (y), the “own” rate of return of money, and the rate of return on alternative investments. The rate of return proxies that we use are the following: domestic inflation (INF), the average interest rate offered on time deposits with maturities of six months to one year (DEPO), the weighted-average yield of treasury bills (TBILL), and the annualized three-month London interbank offered rate (LIBOR).1 The yield on treasury bills (TBILL) is a proxy for the opportunity cost of money after 1983, given that the government made increasing recourse to this instrument to meet its funding needs. Similarly, the London interbank market became an alternative investment for Mauritian investors after exchange controls were lifted in 1994.

Following the literature, the demand for real money balances is specified as follows (with lower-case symbols indicating logarithms):

(m-p)t = a0 + a1yt + a2INFt + a3DEPOt + a4 TBILLt ·D 1 + a5LIBORt · D 2 + εt,

where εt is the error term. We have also included two interactive dummy variables, D1 and D2, to test for major institutional changes: the dummy variable D1 takes a value of 1 after 1983, when Mauritius moved to a managed floating exchange rate system, and 0 otherwise; the dummy variable D2 takes a value of 1 after 1994, when exchange controls were removed, and 0 otherwise.

The empirical analysis is based on quarterly data from 1976: Q1 to 2001: Q2. We use cointegration analysis to examine the long-run demand for money (Johansen and Juselius, 1990; Johansen, 1992). Johansen’s (1988 and 1991) maximum likelihood procedure for cointegrated vectors in a vector autoregression (VAR) is used to set up a model with four lags and a constant term.2 The results of the Johansen cointegration procedure strongly reject the null hypothesis of no cointegrating vector (r = 0); at the same time, they fail to reject the null hypothesis of no more than one cointegrating vector, thereby implying that there is exactly one cointegrating relationship.

The estimated cointegrating relationship may be written as

(m-p)t = 2.2 + 2.11 yt – 0.009INFt + 0.07DEPOt – 0.02TBILLt · D 1 – 0.03LIBORt · D 2 + εt,

where εt is the error term and the constant is calculated so that the errors sum to zero. This equation has the properties of a money demand function in that real money demand is positively related to output and the own rate of return on money, and negatively related to inflation, the treasury bill rate after 1983, and LIBOR after 1994. The coefficients of all the variables entering the cointegrating equation (except inflation) are significantly different from zero at the 99 percent confidence level; the coefficient of inflation is significantly different from zero at the 90 percent confidence level.

The coefficient on y implies an income elasticity of 2.11; this, in turn, suggests that, on average, in Mauritius the increase in real money demand associated with a given increase in real output was twice the size of the increase in real output. This elasticity is an indicator of the rapidity with which financial deepening occurred in Mauritius over the sample period—a result of increasing confidence in the domestic financial system, owing to successful development policies, high growth rates in output, declining inflation, and positive real interest rates maintained over an extended period of time.

The null hypothesis of weak exogeneity can statistically be rejected at the 90 percent confidence level for real money balances, DEPO and LIBOR, while it fails to be rejected for real output, the treasury bill rate, and inflation. This suggests that if a shock were to cause the system to deviate from its long-run equilibrium, real money and time deposit interest rates would adjust over time to restore long-run equilibrium.

1 For a resident investor, the remuneration on a foreign deposit is the foreign nominal interest rate plus the rate of depreciation of domestic currency. However, according to our results the rate of depreciation of domestic currency is stationary, implying that it can be excluded from the long-run cointegration analysis.2 The number of lags was chosen to ensure that the residuals are serially uncorrelated and nonheteroskedastic.

Initial Steps Toward Liberalization

In 1983, as the international economy stabilized and inflation declined, Mauritius moved from a fixed to a managed exchange rate regime. Over the next few years, the BOM exchange rate policy was aimed at protecting competitiveness, and the real effective rate was steadily depreciated until 1987 (Figure 8.4). The real depreciation of the currency contributed to a sharp turnaround in the balance of payments, which swung into surplus in 1986, bringing about an increase in international reserves (Figure 8.5). Successive Stand-By Arrangements with the IMF were in effect until 1986. For the duration of the IMF-supported programs, overall credit ceilings remained the primary instrument of monetary policy and were established annually by the BOM in consultation with the IMF.33 These were fairly effective in restraining the growth in credit and inflation over the program period.

Figure 8.4.Real and Nominal Effective Exchange Rate Indices, January 1979–June 2004


Source: IMF, Information Notice System.

Figure 8.5.Net Foreign Assets of Bank of Mauritius, 1967–2004

(In millions of U.S. dollars)1

Source: IMF, International Financial Statistics.

1 End year data, but for 2004 end-October data.

As the stabilization program came to an end, there was a cautious move toward liberalizing interest rates. In 1987, commercial banks were allowed to determine their own deposit and lending rates. Credit ceilings continued to be the primary instrument of monetary policy, although the bank rate was modified more frequently and followed changes in international interest rates. There was also a reduction in cash and liquidity ratio requirements and some minor liberalization in exchange controls. Real interest rates were brought to a positive level in the mid- and late 1980s, leading to resumption of the declining trend in the income velocity of money. As the net international reserves situation became comfortable, the BOM started targeting a stable real effective exchange rate, interrupting the previous policy of depreciation of the currency in real terms. Credit ceilings were eased in the late 1980s.

From the mid-1980s to mid-1990s, the fiscal deficit (after grants) was small, at less than 1 percent of GDP on average. As a result, the BOM’s credit to the government remained modest and did not hamper the conduct of monetary policy. With the balance of payments in surplus and credit ceilings eased, while the real exchange rate was pegged, Mauritius was subject again to inflationary pressures in 1989 and 1990. In contrast to the early 1970s, the BOM acted promptly and decisively in this instance, by raising interest rates and tightening credit ceilings.


After the inflationary episode in the late 1980s, the BOM hastened its move to an indirect, market-based system of monetary policy implementation. Weekly treasury and BOM bill auctions were introduced in 1991. In these primary auctions, a predetermined quantity of bills was offered for sale, and financial institutions tendered their bids; bids below an arbitrary cutoff yield were not accepted. The bank rate was linked to the weighted average yield in the auctions.

Credit ceilings were abolished in July 1993. Exchange controls were also liberalized, as Mauritius accepted in 1993 the obligations of Article VIII of the Articles of Agreement of the International Monetary Fund, and liberalized the exchange regime for current account transactions. This was followed by a complete liberalization of capital controls in 1994 and by the establishment of an interbank foreign exchange market. In 1997 the foreign exchange surrender requirement for the export proceeds of the Mauritian Sugar Syndicate was abolished. The liberalization of capital movements did not result in any dramatic movements in the exchange rate or in international reserves.

Following the abolition of the credit ceilings in 1993, the central bank introduced a reserve-money-programming and liquidity-forecasting framework in order to control the overall money supply. The main objective was to maintain the monetary base on a path consistent with the central bank’s inflation target and the Central Statistics Office’s economic growth forecast. This resulted in a forecast for the demand for money, and the central bank would set a domestic credit target consistent with the inflation objective. The monetary programming continued to be accompanied by an exchange rate objective, with a view to keeping the rupee broadly stable in real terms on a trade-weighted basis. Thus, the central bank stood ready to buy or sell foreign exchange to maintain the exchange rate at the targeted level, and under its monetary base programming policy would sterilize these operations by modulating the sale of treasury bills on the government’s behalf.

Operationally, in the absence of open market operations, the instruments used were the treasury bills auction process and the establishment of rediscount tranches. Effective July 1994, the bank rate was set as the average treasury bill auction rate over the previous 12 weeks, plus a margin determined by the BOM, and commercial banks’ access to central bank credit was recast and set within rediscount tranches, plus an unlimited tranche at a penalty rate.34 In this system, the stance of monetary policy was set by the amount of treasury bills sold at the weekly auctions, which could exceed the borrowing need of the government for monetary policy purposes. The lack of an active secondary market in treasury bills meant that the BOM could not use secondary market operations to implement monetary policy, and financial institutions faced difficulties in liquidity management. To address the situation, the BOM established a secondary market cell in 1994 to stimulate trading outside the weekly auctions. However, the secondary market cell at the BOM was fairly passive, and the BOM was reluctant, except under extreme circumstances, to conduct outright purchases of treasury bills to meet liquidity shortfalls of banks. As a result, the secondary market operations have been relatively modest, and banks typically hold treasury bills to maturity.

To establish open market operations for monetary policy purposes and to improve liquidity management, the BOM in December 1999 introduced a Lombard facility and a framework for repurchase and reverse repurchase transactions. The Lombard facility is a standing facility to provide overnight collateralized advances to banks at a preannounced interest rate; the rate is used by the BOM as a signaling mechanism for its monetary policy stance. The repurchase and reverse repurchase transactions have become a tool for fine-tuning liquidity, complementing the primary auction of treasury bills; the amount of transactions has been modest, and transactions have halted since mid-2003, as the interbank rate fell significantly below the Lombard rate.

Inflation as a Monetary Policy Objective in Mauritius: Challenges in Recent Conduct of Monetary and Exchange Rate Policy

The liberalization of capital controls in 1994 necessitated an increased emphasis by the BOM on maintaining price stability and ensuring confidence in the currency. Annual inflation had averaged close to 10 percent over the years 1989 to 1993, accompanied by considerable variability. Thus, there was a need for monetary policy to achieve lower inflation and to do so consistently. To this end, the BOM tightened monetary policy by raising in 1994 the level of interest rates at the same time as capital controls were lifted.35 After the monetary-programming framework was implemented in 1993, the BOM took a lead in influencing inflationary expectations, particularly since inflation is the key input in the wage-bargaining process. In fiscal year 1996/97 (July to June), the BOM started announcing an inflation objective in its annual report. As shown in Table 8.1, the BOM has been able to meet the preannounced inflation objective in all years except in 2001/02, when unanticipated increases in the value-added tax and a cyclone resulted in the target being missed.

Table 8.1.Inflation Target and Outcome, 1997/98–2003/04
YearAnnounced Inflation TargetRealized Inflation
Source: Annual Reports of the Bank of Mauritius.
Source: Annual Reports of the Bank of Mauritius.

It should be noted that the monetary regime prevailing in Mauritius in recent years cannot be considered a fully fledged inflation-targeting regime. Such a regime requires a clear mandate to the central bank to achieve price stability, together with an institutional framework in which all other monetary policy objectives are subordinated to inflation targeting and in which the central bank is held accountable for reaching the target and for having to explain the reasons for any deviations between outcomes and targets. This requires that high-frequency inflation reports be prepared, and appropriate analytical models elaborated, for inflation forecasting and for highlighting the channels between policy instruments and inflation. While many countries preannounce inflation targets, many at the same time adhere to other policy objectives such as monetary targets or some form of exchange rate targeting. Fully fledged inflation-targeting countries, on the other hand, commit to subordinating other policy objectives to the inflation target, using continuously updated inflation forecasts as the intermediate guide to monetary policy.36

In the case of Mauritius, while annual inflation targets have been announced by the central bank since 1997, the inflation forecasting is still rudimentary. It is not formulated in the framework of a medium-term trajectory that takes into account the lags between policy changes and inflation (only a forward-looking 12-month target is announced), and an analytical framework specifying the channels between monetary policy instruments and inflation is not yet available. The monetary authorities, however, plan to further refine in the period ahead their inflation-forecasting methodology, together with the analysis of the transmission mechanism between monetary policy and inflation.

In a small open economy like Mauritius, the exchange rate has a large impact on inflation, so that the exchange rate policy of the authorities has been closely interlinked with the behavior of inflation. In the 1980s and the 1990s the authorities let the nominal effective exchange rate depreciate steadily, with a view, up to the early 1990s, to achieving a depreciation of the real effective exchange rate in order to promote competitiveness. From the mid-1990s the objective shifted to maintaining a broadly stable real effective exchange rate (Figure 8.4), with the authorities aiming at a gradual depreciation of the nominal exchange rate to compensate for the positive inflation differential with partner countries. The gradual nominal depreciation, together with the backward-looking characteristics of the wage-bargaining process, in which wage increases were based on past increases in the cost of living, is a critical factor in explaining why annual inflation has declined only slowly through the 1990s and early 2000s, from 10.5 percent during 1993 to 6.7 percent during 2002.

With the balance of payments in significant surplus since mid-2000, owing mainly to an improvement in the current account, the monetary policy of the BOM has been confronted with additional challenges. In this environment, the BOM opted up to mid-2003 for significant interventions in the foreign exchange market to preserve competitiveness, resulting in a sharp increase of its net foreign assets (from US$665 million at end-June 2000 to US$1.34 billion at end-June 2003 and US$1.5 billion at end-December 2003), but was unable to fully sterilize the resulting increase in bank liquidity. Reserve money and money growth thus significantly exceeded the growth in nominal GDP—the annual average growth rates from end-1999 to end-2003 were, respectively, 9.9 and 10 percent—and velocity of money declined. At the same time, beginning in early 2002, the central bank lowered the Lombard rate in five steps, from 11½ percent to 9½ percent at end-January 2004. The rate of consumer price inflation in Mauritius remained broadly stable during this period, from 5.3 percent in the year ending June 2000 to 5.1 percent in the year ending June 2003.37 In part, the stability in domestic inflation in Mauritius may be attributed to the declining and remarkably low inflation worldwide that has been witnessed over the past few years. The stability in inflation rates also in part reflects the confidence of investors in the Mauritian economy.

Since mid-2003 the BOM has shifted to a policy of less intervention in the exchange market, resulting in some nominal appreciation of the exchange rate and a fall in inflation, to 3.9 percent in the 12 months to December 2003. In the 12 months to September 2004, consumer price inflation picked up somewhat, to 4.8 percent, reflecting higher oil prices, leading the BOM to raise the Lombard rate to 9.75 percent in October.

The experience in many countries shows that in the presence of a strong balance of payments surplus an appreciation of the real exchange rate is ultimately unavoidable, since sterilization operations can be very costly and ultimately self-defeating, as the higher interest rates caused by sterilization are likely to attract capital inflows. In fact, since 2003 the Bank of Mauritius has not fully sterilized the increase in net foreign assets and in government deposits, leading to excess liquidity. The commercial banks’ reluctance to expand credit, because of concerns over the viability of the EPZ firms, also contributed to excess liquidity. Thus the treasury bill rates and the interbank rates have fallen sharply since mid-2003 and are now well below the Lombard rate (see Figure 8.6). The treasury bill yield curve has also shifted downward considerably since early 2003 (see Figure 8.7).

Figure 8.6.Principal Interest Rates, January 2000–February 2004

(In percent)

Source: Bank of Mauritius.

1 The bank rate is the weighted average of 1-, 6-, and 12-month treasury bill rates. It is used for central bank advances to the government.

1 Discounted by the annual average consumer price inflation rate.

Figure 8.7.Yield Curve for Treasury Securities

(In percent)

Source: Bank of Mauritius.

An effort to model the Bank of Mauritius monetary policy in recent years has been presented in a paper by Porter and Yao (2004), whose findings are reported in Box 8.2. The analysis shows that the central bank, as can be expected, modified its objective for the short-term interest rates on the basis of the behavior of inflation, the real exchange rate, and the output gap. As the inflation target was reduced only very gradually, it is not surprising that changes in the short-term interest rate targets have also been small over time.

Box 8.2.Econometric Analysis of Bank of Mauritius Monetary Policy Rule

An effort to estimate econometrically the Bank of Mauritius monetary policy rule is contained in Porter and Yao (2004). It is assumed that the monetary rule of the central bank is similar to that recommended by Ball (1999) for small open economies: nominal interest rate i is set as a function of the output gap y, inflationary expectations are captured by L, the deviation between actual and expected inflation is πL, the real exchange rate is q, as well as the desires to smooth the real interest rate. Thus,

i = Lt + Qs(it-1−Lt-1) + (1 – Qs) [γq qt + γπt – Lt) + γy yt]+ us,(1)

where us, t = Qu us, t-1 + εs, t.

It is assumed that the behavior of the nominal interest rate is captured by two latent factors, the slope of the yield curve S (i.e., the difference between the long-run and the short-run interest rates), and the level of the yield curve,

i = L + S + δ.(2)

Thus L can be seen as representing inflationary expectations, and S as the real interest rate.

Replacing (2) in (1) we obtain:

S = Qs S + (1 – Qs) [γq qt + γπ πt + γy yt] + us.(3)

The parameter in (3) should indicate the relative importance of each of the objectives of the BOM. If Qs = 0, S is determined by the monetary policy adjustment. If Qs ≠ 0, monetary policy exhibits persistence, and adjustment is gradual.

The model contains a set of additional equations: an equation for the output gap, as a function of the expected output gap, lagged output, the real exchange rate, and the real interest rate; an equation for the inflationary expectations, as a function of past inflationary expectations and actual inflation, Lt = QL Lt-1 + (1-QL) πt + εLt; an equation for inflation, as a function of past inflation and the real exchange rate; and an equation for the real exchange rate, as a function of past real exchange rate and the real interest rate. The model has been estimated by maximum likelihood for the informal inflation-targeting period July 1996–March 2004, using monthly data on government bond yields (for 3-, 6-, and 12-month maturities), annualized CPI inflation, and the real exchange rate calculated by the IMF’s Information Notice System. As monthly output gap estimates are unavailable, seasonally adjusted monthly credit to the economy and to the private sector has been used as a proxy for the output gap.

The factor representing inflationary expectations Lt is extremely persistent, with QL estimated at 0.992; this means that actual inflation carries a very small weight in the formation of inflation expectations.

The BOM’s estimated monetary policy rule is

St = 0.9898 St-1 + 0.11 [0.6674 qt + 0.9301 (πt) + 0.0096 yt] + ust,

where ust = 1.17 exp(-5) us, t-1 + εs, t.

Since the estimated Qs = 0.989, St is also highly persistent, which is consistent with the inertia in the BOM monetary policy setting, as it reduces only gradually the targeted inflation. It also indicates that the BOM has only a small influence on the short-term interest rate, with 1– Qs = 0.0011. The relative importance of the inflationary deviations, the output gap, and the real exchange rate are indicated by the estimates of γπ, γq, γy,, respectively 0.9301, 0.6474, and 0.0096. These estimates suggest that multiple objectives influence monetary policy, implying a flexible inflation targeting in Mauritius. It must also be noted that these parameters are estimated imprecisely, with none actually being significant. Finally, the estimate suggests there is no autocorrelation in unanticipated monetary shocks ust.

31The bank rate remained fixed between July 1970 and December 1976 at 6 percent.
32The Mauritian rupee was linked to the pound sterling until 1976, when the rupee was pegged to the SDR in the face of continued depreciation of the British currency.
33The bank rate was not used as a primary instrument of monetary policy during the program. For instance, the bank rate stayed fixed between 1983 and 1986 at 11 percent.
34This automatic access was eliminated in mid-1996, and in the following years commercial banks’ borrowing from the central bank was negligible.
35In fact, the BOM maintained high domestic interest rates even as international interest rates declined in the 1990s.
36For a comprehensive discussion on the difference between countries with fully fledged inflation targeting and countries that announce inflation objectives but are not able to maintain inflation as the foremost policy objective, see Schaechter, Stone, and Zelmer (2000) and Stone (2003).
37There was a brief inflation spike to 6.3 percent in the year ending June 2002, as a result of increases in VAT rates and a cyclone.

    Other Resources Citing This Publication