rao and singh (peb)

1

Economic surveyPACIFIC ECONOMIC BULLETIN

Pacific Economic Bulletin Volume 21 Number 2 2006 © Asia Pacific Press

Monetary policy in Fiji

B. Bhaskara Rao and Rup Singh

B. Bhaskara Rao isProfessor of Economicswith the Department ofEconomics at theUniversity of the SouthPacific, Suva.

Rup Singh is a Lecturerin the Department ofEconomics, University ofthe South Pacific, Suva.

In order to achieve its central policy objectives of maintaininglow inflation and a stable currency, this article finds that theReserve Bank of Fiji should switch back from using the bankrate as its main policy instrument to use of the money supply.It is also recommended that the bank rate and deposit ratesbe raised to levels comparable with Australia and NewZealand in order to reduce capital outflows and increasesavings. Further, the Reserve Bank of Fiji should considerwidening its bounds for variations in the exchange rate toreduce the need for occasional large devaluations.

Monetary stability makes developmentpossible. But no tricks of monetarypolicy—certainly not artificially lowinterest or exchange rates, despite theclaims that are sometimes made forthem—can truncate the long, slowgrind of accumulating human andother capital and building economic,political and social institutions thatlead to sustained increases in incomeGarnaut (2005:103).

It is usual to begin a country survey onmonetary policy with a list of the objectivesof its central bank and then examine theirsuccesses and failures. However, as theInternational Monetary Fund has noted, inthe past central banks were notorious forbeing inscrutable (International MonetaryFund 2006).1 Following the active involve-

ment of the international agencies in goodgovernance practices, many central banksnow prize clarity in explaining their objectivesand decisions to the public. The ReserveBank of Fiji follows such good governancepractices and its statutory responsibilities arestated as follows• to regulate the issue of currency, and the

supply, availability and internationalexchange of money

• to promote monetary stability• to promote a sound financial structure, and• to foster credit and exchange conditions

conducive to the orderly and balancedeconomic development of the country(International Monetary Fund 2006).2

The first objective coincides with twofrequently stated objectives of central banks:

2

PACIFIC ECONOMIC BULLETIN

Economic survey


to regulate the issue of currency (money) witha view to maintaining its internal andexternal value through price and exchangerate stability. Waqabaca and Morling (1999)3

give a similar interpretation to the mainobjective of the Reserve Bank of Fiji, althoughthey do not highlight the importance of thestability of the exchange rate. They note that

[m]onetary policy in Fiji is conductedto achieve the Reserve Bank’s objectiveof low inflation. The Reserve Bank doesnot have a formal inflation target, butis generally comfortable with inflationrates of around 2–3 per cent (Waqabacaand Morling 1999:5).

Frequent policy interventions may benecessary to achieve this objective throughliquidity management to promote monetarystability (the second objective).

The Reserve Bank of Fiji has a limitedrole in the next two objectives: to develop asound financial structure and improve theefficiency of transactions as a foundation forthe balanced development of the economy.These objectives may not need interventionson a frequent basis. Furthermore, a centralbank’s role is limited if the banking andcorporate laws have weaknesses. At best acentral bank can only ensure that the lawsare implemented effectively . To achieve theselarger objectives, timely commissions ofenquiry, with a view to reforming bankingand corporate laws, are necessary. It is alsodoubtful if a central bank can play a directand a larger role in improving the rate ofgrowth of output. As Garnaut has noted,monetary policy can only provide monetarystability to make development possible. Lowinterest and exchange rate regimes, whichcan be mistakenly imposed by central banksto increase investment rates and net exports,cannot ‘truncate the long, slow grind ofaccumulating human and other capital andbuilding economic, political and socialinstitutions that lead to sustained increasesin income’ (Garnaut 2005:103).

Issues concerning changes to thebanking and corporate laws, policies toaccumulate human and non-human capital,and institutional reforms are beyond thescope of this survey of monetary policy inFiji. Similarly, ongoing debates on whetherFiji and other Pacific island countries shouldform currency boards or currency unions,have floating rates or fix their exchange ratesand so on, are put to one side.4 As Garnaut(1995) has observed, the need for such majorchanges arises mainly due to the failure ofthe monetary authorities to maintain a stablevalue of the currency and the failure of thegovernments to discourage rent-seekinginstitutional practices. We examine the scopefor monetary policy to achieve stable monetaryconditions by maintaining the internal andexternal value of money. In other words, thissurvey looks at whether monetary policy inFiji can be, and is, used effectively to maintainstable inflation and exchange rates.

Fiji’s growth experience

The growth experience of Fiji has been mixed,with short episodes of both rapid and sharplyreduced growth. Real GDP growth has beenmoderate, averaging around 3 per cent perannum during 1971 to 2002. Fluctuations inoutput have been influenced by both domesticand external influences. Furthermore, thepolitical coups of 1987 and 2000 reduced theaverage growth rate. During 1971 to 1986, theaverage growth rate was 3.6 per cent and afterthe first coups in 1987 and until the secondcoup in 2000, it declined to 2.6 per cent. In theperiod from 2000 to 2002, the average growthrate declined further to 1.3 per cent (Figure 1).Political instability appears to have reducedthe trend in the growth trajectory as itsnegative impacts are noted in almost allspheres of economic activity. These eventshave been responsible for the loss of consumerand investor confidence, massive declines intourist arrivals, and rapid exodus of human

3



Figure 1 Fiji: growth of real GDP, 1971–2004 (per cent per annum)

Source: Data to 2002 from International Monetary Fund, 2004. International Financial Statistics, InternationalMonetary Fund, Washington, DC; data for 2003–2004 estimated from Reserve Bank of Fiji, 2006. QuarterlyReview, March, Reserve Bank of Fiji, Suva.

Figure 2 Fiji: capital formation (private), 1985–2004 (per cent per annum)

Note: Capital series are estimated with the perpetual inventory method. The assumed depreciation rate is 4per cent. Data for 2003 and 2004 are estimates.Source: Reserve Bank of Fiji and authors’ calculations.

-0.1

-0.05

0

0.05

0.1

0.15

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003G

row

th ra

te o

f out

put (

per c

ent p

er a

nnum

)

-0.1

-0.05

0

0.05

0.1

0.15

0.2

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

Cap

ital f

orm

atio

n (p

er c

ent p

er a

nnum

)

4


Economic survey


and financial capital to neighbouringindustrialised economies. Massive migrationof skilled workers must affect productivityand the trend growth of Fiji’s output. Withunbalanced rural–urban growth, both skilledand semi-skilled populations have migratedfrom rural areas to urban centres, creatingintense competition for employment,housing and other social infrastructure.

Capital formation in the private sectorhas declined from an average annual growthrate of 5 per cent during the period prior to1987 to about 0.5 per cent afterwards (Figure2). However, employment has grown at anaverage annual rate of 2 per cent, although ithas remained static since the second coup(Figure 3). The low investment rate remainsa major problem. The ratio of public andprivate investment to output (the investmentratio) has declined from above 20 per cent inthe 1970s to around 12 per cent in the mid1980s, and has remained low since. Theprivate investment ratio declined from 15 per

cent in 1971 to below 10 per cent in 1987 dueto the political events. However, recently, newtourism and commercial building projectsand investment in residential propertieshave taken place. This is perhaps due to thelow lending rates and improvements inprocessing of investment proposals by theFiji Trade and Investment Bureau. Nonetheless,generally, private investment has been on adownward trend. Various incentives since1987—such as the Tax Free Zone scheme,Schedule 5 Export Tax Incentive, depreciationallowances, hotel investment incentives, andthe forward loss carrying option—havefailed to induce private investment to thelevels of the early 1970s and 1980s.

The inflation rate and the rate of growthof money supply (M1) are shown in Figure 4.The average inflation rate from 1971 to 2002was 6.7 per cent. In the pre and post 1987coup periods it was 10.1 per cent and 3.7 percent, respectively. Inflation was higher in thepre 1987 coup period mainly due to the oil-

Figure 3 Fiji: rate of growth of employment, 1971–2004 (per cent per annum)

Note: Data for 2003 and 2004 are estimates.Source: Reserve Bank of Fiji and Fiji Bureau of Statistics.

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

1971

1974

1977

1980

1983

1986

1989

1992

1995

1998

2001

2004

Em

ploy

men

t gro

wth

(per

cen

t per

ann

um)

5



price hikes of the 1970s. With the decline indomestic demand following the 1987 coups,inflation has generally been low. The ReserveBank’s measure of inflation (CPI, excludingvolatile items) is even more stable. Wagepressures remain subdued and thereforethere are no serious threats for inflation. Thereal money supply (M1) grew at an averagerate of 4.2 per cent over the 1971–2004 period.However, there were sharp increases in 1988and 1999. While the former was the result ofthe RBF misreading market signals followingthe first coup, in 1999 growth in moneysupply was mainly to accommodate the highgrowth in output recorded in that year. Incomparison to the growth in output, it canbe said that money growth has been slightlyhigher.

Interest rates have been moderate (Figure5). While the average lending rate has beenaround 7.8 per cent, the average two to threeyear time-deposit rate has been just over 2.5per cent, which gives a margin as high as 5.3per cent. The Reserve Bank’s policy rate has

been manoeuvred to implement monetarypolicy. It was low for a brief period in the 1990sand then tightened after 2000. Following adecline in the period from 2001 to 2003, ithas again been increased.

The nominal and real exchange rateshave been fairly stable, except during thedevaluations in 1988 and 1998. The first twodevaluations in 1988 were to contain theexcessive capital outflows that even the highinterest rates and capital controls failed tocapture adequately. The 1998 devaluationwas necessary to maintain Fiji’s internationalcompetitiveness in order to avoid negativeimplications of the deterioration of the termsof trade resulting from an overvaluedcurrency. Vulnerability to terms of tradeshocks together with a narrow export base isharmful for the economy. Fiji is a price takerin the world market and therefore inter-national forces have the potential to affect tradeprices significantly. Therefore, low productioncosts of exports and competitive export qualityare the key to turning the tables in Fiji’s

-30

-20

-10

0

10

20

30

40

50

60

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Rat

e of

gro

wth

(per

cen

t)

In fla tion RateGrowth Rate of M1

Figure 4 Fiji: growth rate of real money supply and prices, 1972–2004 (per cent)

Source: International Monetary Fund, 2004. International Financial Statistics, International Monetary Fund,Washington, DC.

Inflation rateGrowth rate of M1

6


Economic survey


Figure 5 Fiji: interest rates, 1971–2004 (per cent)

0

5

10

15

20

25

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

Inte

rest

rate

(per

cen

t)

Savings Dep RateCommercial Bank RateAverage M/Term Bonds

Figure 6 Fiji: exports and imports, 1970–2003 (F$ million)



Savings deposit rateCommercial bank rateAverage medium-term bonds

-

200.00

400.00

600.00

800.00

1,000.00

1,200.00

1,400.00

1,600.00

1,800.00

1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003

F$ m

illio

n

ExportsImports

Exp•ÿ`•````•ÿ`

7



favour. Nonetheless, Fiji has benefited fromthe bullish global growth, as well as the tradearrangements and export promotion policiesimplemented in the early 1990s. However,erosion of the major trade concessions nowpresent serious consequences for output,exports, employment and the livelihoods ofa significant proportion of the populationwho depend on the declining industries ofsugar and textiles.

The trade deficit has been widening—while exports have grown on average byaround 3 per cent since the 1987 coups,imports have increased by 7.4 per cent(Figure 6). In light of this serious increase inthe trade deficit, one is alarmed by thedeteriorating foreign exchange reservesposition (Figure 7). Note the trends inreserves in the years of devaluation—therewere sharp but temporary pick-ups inreserves in 1987 and 1998. Also note thetrends in the four decades since 1970. Thereserves are lowest from 2000 and this is a

cause of concern for the Reserve Bank of Fiji.But it should be noted that yet anotherdevaluation may only provide temporaryrelief from a potential balance of paymentscrisis. Unless exports increase strongly, it ishard to vaccinate the economy with temporaryexpenditure-switching policies such asdevaluations.

Government finance figures indicatethat the underlying deficit is high and it hasfrequently resorted to short-term borrowing,mostly from domestic sources. Domestically,the Fiji National Provident Fund is thelargest lender. The government is expectingthat the public debt which is over 50 percent of GDP will decline, although it isnot clear from the 2005 Budget how will thiseventuate in light of the lower-than-expectedgrowth projections. Therefore, it is hard toanticipate an improvement in the govern-ment’s financial position. Against thisbackdrop, we anticipate a low growthscenario in the short to medium term. With a

Figure 7 Fiji: reserves as proportion of imports, 1970–2004

0.00

0.10

0.20

0.30

0.40

0.50

0.60

1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003

Rat

io

Source: International Monetary Fund, 2004. International Financial Statistics, International Monetary Fund,Washington, DC and Reserve Bank of Fiji.

8


Economic survey


depressed investment climate, the decline inthe productivity of the major sectors of theeconomy should not be unexpected. Thesetranslate into gloomy long-term growthprospects for Fiji.

Monetary policy developments

The single most important objective of manycentral banks is the stability of the value ofmoney, that is, a low and steady inflationrate and a stable exchange rate. To avoidconfusion, we call these objectives targets,and the policies used to achieve these targets,instruments. Therefore, the desired inflationrate of 2–3 per cent is a target and the moneysupply and the bank rate are instruments.Similarly, we call the desired holdings offoreign exchange a target, and the exchangerate and bank rate instruments.5 Bycomparison, in countries with currencyboards (for example, Argentina and HongKong) the exchange rate is the target. Theexchange rate is also a target in somecountries with flexible exchange rates (forexample, Singapore) and in some countrieswith managed exchange rates (for example,Malaysia). The instruments are their largeholdings of foreign exchange reserves andthe bank rate.

However, many countries have managedexchange rate systems (for example, Australia,New Zealand, India, Israel, Fiji and Sweden)and their monetary authorities target bothinflation and the exchange rate, but allow theexchange rate to vary within predeterminedupper and lower bands.6 Since the early1990s, in virtually all of these countries withmanaged exchange rate systems the monetarypolicy instrument is the bank rate. Prior to the1990s, the instrument was the money supply.A small change in the bank rate may beadequate to achieve both internal and externalstability of the currency, provided domesticexpenditures and capital flows are sensitiveto changes in the bank rate. If there is little

sensitivity, which seems to be the case in manydeveloping economies, the effectiveness of thebank rate as the instrument of monetary policyis weak and uncertain. According to Jha(2003), in India it is difficult to reduce theinflation rate through changes to the bank ratebecause aggregate demand is not sensitive tointerest rate changes and the bank rate doesnot have significant effects on the structure ofmarket interest rates. One may also add thatcapital flows are unlikely to respond to thebank rate in India and other developingeconomies.7 Therefore, according to Jha, indeveloping economies it is difficult to stabilisethe internal and external value of moneythrough adjustments to the bank rate only.

Our brief review of the objectives ofmonetary policy indicates that in the post-war period there has been no change in themain objective of many central banks. Centralbanks are committed to maintaining theinternal and external value of money,although they may not state explicitly the rateof inflation and the exchange rate they wishto maintain. However, since the 1990svirtually all countries have switched fromusing monetary aggregates (for example,money supply) as their main instrument ofmonetary policy to the bank rate. Therefore,we may ask whether the bank rate is aneffective monetary policy instrument forachieving a stable value for money.

After the high rates of inflation of the 1970sand 1980s and the subsequent downturns ineconomic activity, there have been no majorepisodes of stagflation (see Cecchetti et al.2006). There is also a claim that policy rules—such as increasing the bank rate by more thanthe inflation rate to contain inflationarypressures—are better than discretionarypolicy adjustments. Consequently, it isdifficult to evaluate whether the aforesaidshift to the bank rate as an instrument has orhasn’t made monetary policy more effective.Justifications for policy rules and choice ofthe bank rate as the monetary policy

9



instrument, therefore, are generally based ona priori arguments, indirect evidence, andsimulation exercises.

Poole (1970) has argued that to minimisefluctuations in the level of economic activity,money supply should be used as aninstrument when the LM (money demand)relation is stable whereas the rate of interestis the appropriate instrument when the LMrelation is unstable. The switch from moneysupply to interest rate as the monetary policyinstrument has taken place in industrialeconomies due to the instability of theirmoney demand functions. The instability inturn was due to the financial innovationsfollowing the liberalisation policies of the late1980s. However, many developing economiessuch as India and Fiji have also switched tousing the bank rate as their monetary policyinstrument, although there is no evidencethat their liberalisation policies have hadlarge enough effects to make their demandfor money functions unstable. Rao and Singh(2005a, 2005b) found that demand for moneyfunctions in Fiji and India are stable. SimilarlySingh and Kumar (2006a, 2006b) found thatdemand for money functions in 12 developingeconomies—Fiji, Vanuatu, Samoa, SolomonIslands, India, Indonesia, Philippines,Thailand, Kenya, Malawi, Jamaica andRwanda—are temporally stable. Never-theless, these countries have switched tousing the bank rate as their instrument ofmonetary policy.

Cecchetti et al. (2006) have used datafrom 24 (mainly industrial) economies tocompare the effectiveness of monetary policyin the pre and post 1990 periods: 1983QI to1990QIV and 1991QI to 1998QIV, respectively.This research is of interest because centralbanks in these countries have shifted fromusing money supply to the bank rate as theirmonetary policy instrument from the 1990s.The researchers found that monetary policyhas become more effective (in 21 out of 24countries) since the 1990s in stabilising the

inflation rate.8 These findings are consistentwith Poole’s theoretical results (Poole 1970).In contrast, as noted earlier, Jha (2003) hasfound that in developing economies, suchas India, it is difficult to reduce the inflationrate by increasing the bank rate.9

Monetary policy

It is useful to review monetary policy in Fijiwith these international developments asbackground. It is also important to payattention to Garnaut’s observation, quotedat the outset, as to whether monetary policyhas succeeded in providing stable monetaryconditions. In the applied growth literature,mainly based on cross-country data, there isweak evidence that financial variables havesmall and direct effects on the growth rate.Such effects seem to be significant only whencertain policies, collectively known as theInternational Monetary Fund and WorldBank conditionalities, have beenimplemented—these are similar to thepolicies suggested by Garnaut (2005).

There are other issues that need answers.How effectively has monetary policy beenused in Fiji to achieve its objectives? Why isthe Reserve Bank using the bank rate insteadof money supply (since 1997) as its instrumentof monetary policy? Has monetary policybecome more effective since this switch? Canmonetary policy be used to increase theeconomic growth rate? Answers to thesequestions are useful not only to review theperformance of monetary policy but also todevelop a pragmatic vision for its futurecourse. We hope to provide, albeit tentatively,answers to a few of these questions.

From the stated statutory objectives andthe recent policy statements it is clear thatthe main objective of the Reserve Bank is toachieve a stable monetary environment bymaintaining the internal and external valueof the currency. Therefore, a low and steady

10


Economic survey


rate of inflation and a stable exchange rateare its two most important monetary policytargets. The instruments used are smalldiscretionary adjustments, on a daily basis,to the exchange rate and the sale andrepurchase of Reserve Bank’s short-termpaper to manage liquidity conditions, andless frequent adjustments to the bank rate toreduce anticipated inflationary pressures.

For the bank rate to serve as anappropriate policy instrument, it is firstnecessary to verify, as pointed out by Poole(1970), that fluctuations in economic activityare due to fluctuations in the LM relationship,which in turn are due to instability in thedemand for money. Furthermore, there is aneed to establish that aggregate demandadequately responds to changes in the bankrate through its effects on other interest rates.To maintain a stable exchange rate—permitting only variations within apredetermined range—the Reserve Bankneeds either adequate foreign exchangereserves and/or a significant response ofcapital inflows to changes in the bank rate.

According to some unpublished andpublished research at the RBF, these require-ments to achieve internal price stability havebeen adequately satisfied. Katafono (2001)found that the demand for money (M1) inFiji is unstable. This would justify the switchfrom money supply to the bank rate as themonetary policy instrument. Waqabaca andMorling (1999) found that a 1 per cent rise inthe (real) bank rate reduces the growth rateof output by about 0.3 percentage points.They also found that a 1 per cent increase inthe output gap (actual less potential output)leads to a 0.2 per cent increase in the inflationrate. These results support the conclusionthat changes to the bank rate have theappropriate effects on domestic demand andthe inflation rate. What is missing in thesestudies is an answer to the question ofwhether the bank rate has any effects oncapital inflows and/or imports. But this is

not a serious gap because until recently, theReserve Bank’s holdings of foreign exchangereserves seemed to be satisfactory (see Figure7). Foreign exchange reserves as a proportionof imports (goods and non-factor services)increased in 1997 and 1998 and were equalto about five months of imports.10 Thisposition remained stable until the coup in2000. By the end of 2005, however, foreignexchange reserves had declined to about 3.5months of imports. Three devaluations, twoin 1987 by 33 per cent and another in 1998by 20 per cent, did not seem to have hadlasting effects on reserves and the currentaccount balance. The effects of thesedevaluations, after an initial positive impact,seemed to diminish slowly; the gap betweenimports and exports has been widening,especially since 1999.

Therefore, it is first necessary to examinethe adequacy of research at the Reserve Bank,which seems to have influenced its choice ofthe bank rate, instead of a monetary aggregate,as its instrument of monetary policy. Weexamine Katafono’s findings on the demandfor money, albeit briefly, since some of itsweaknesses have already been pointed outin Rao and Singh (2005a). The findings ofRao and Singh are supported by the recentworks by Singh and Kumar (2006a, 2006b).We then examine, in some detail, Waqabacaand Morling’s (1999) research on the effectsof the bank rate on output and inflation.

Demand for money

Undoubtedly, estimates of the demand formoney in Fiji by Katafono (2001) were animprovement on the ad hoc empirical work thathad been done up to that time. Nevertheless,there was persistent confusion in Katafono’sstudy between estimates based on theJohansen maximum likelihood procedureand those based on the general-to-specificapproach of Hendry. Three price variables—the rate of interest on saving deposits, thetreasury bill rate and the real exchange rate—

11



were introduced into the specification.Consequently, perhaps due to some co-linearity between the two interest rates, thecoefficient of the treasury bill rate was positive.A major weakness in Katafono’s results wasthat the income elasticity of demand for moneywas about 0.5, contrary to expectationsand repeated claims that it is unity. Theseempirical anomalies are an indication thateither there were some specificationweaknesses and or the methods of estimationwere inaccurate. Consequently, Katafono’sfinding that the demand for money in Fiji istemporally unstable is difficult to accept. Asstated earlier, Rao and Singh (2005a) andSingh and Kumar (2006a, 2006b) have foundthat demand for money in Fiji is stable.11

Effects of the bank rate

Waqabaca and Morling (1999) argued thatalthough it might be expected that in adeveloping economy like Fiji, real interest rateeffects on aggregate demand (and thereforeon output) would be weak, surprisingly theirempirical results show that these effects aresignificant. From their regression estimates ofa relationship between output and itsdeterminants, they conclude that a one percent increase in the real short-term rate ofinterest would reduce output growth by 0.34percentage points.12 While this finding seemsimpressive, there are weaknesses in theirspecification and possibly in the estimationmethod. Their specification, which is a variantof the general-to-specific method (GETS) of theLSE-Hendry approach, with an adjustmentprocess based on the error correction model(ECM), and with minor changes in notationto highlight the ECM part, is as follows13

1i 2j

3k 4n

1 1

*ΔlnY = α + β ΔlnY + β ΔlnYt 0 t-i t- j

+ β Δr + β ΔZt-k t-n

c *2-c lnY - lnY + ε (1)t- t-1 tc1

∑ ∑

∑ ∑

⎛ ⎞⎜ ⎟⎝ ⎠

where Y is real GDP, Y* is GDP of tradingpartners, r is the real short-term interest rate,Z is a vector of other short-term influencessuch as shocks in agricultural supplies, theterms of trade, real exchange rate, and budgetdeficits, and e is the error term.

Although Waqabaca and Morling (1999)have many useful insights into the implement-ation of monetary policy in Fiji and use arelatively modern econometric approach,there are some worrying specificationweaknesses. They have used, albeit in an adhoc manner, the GETS approach.14 In theGETS specification there is a clear distinctionbetween the long-run equilibrium relationshipand the short-term dynamic adjustment. Inequilibrium, dynamic adjustments cease andtherefore all the lagged changes in thevariables become zero. (For an exposition ofthe GETS methodology, see Rao 2006a,2006b). Therefore, it is necessary to start witha more general specification in which all thelagged values of the hypothesised variablesaffecting output are included in the ECMpart and their current and lagged changesappear in the dynamic adjustment part ofthe specification

1i 2j

3k 4n

-

1

*α + β ΔlnY + β ΔlnY0 t-i t-j

+ β Δr + β ΔZt-k t-n

1 *ln ln1 1 2 1 3 1 4 1

ln

(2)c Y c Y c r c Z tct t t t

Yt

ε

∑ ∑

∑ ∑

− + + +− − − −

Δ =

⎡ ⎤⎢ ⎥⎛ ⎞

⎜ ⎟⎢ ⎥⎝ ⎠⎢ ⎥⎣ ⎦

It is not clear why the one period laggedvalues of the real rate of interest and thevariables in the Z vector (for example,agricultural supply and terms of tradeshocks) were not included in the ECM partof their equation. It would have been usefulnot only to know if these variables havepermanent long-run effects on the level ofoutput, but also to determine if thecoefficients of the changes in the rate of

12


Economic survey


interest and its significance are free fromspecification biases in the ECM part of theequation.

A more serious specification weaknessin Waqabaca and Morling’s output equation(which is common in ad hoc specifications inapplied growth studies with time-series data)is that they assume output is demanddetermined in Fiji. In other words, they haveignored supply-side variables in their outputequation. At the least, a few supply-sidevariables, known as the conditioningvariables in the applied growth literature—employment, capital (or investment ratio)and a time trend to capture technicalprogress—should have been included intheir reduced-form output equation. Giventhese specification weaknesses, the validityof their estimated coefficients and their t-ratios are doubtful. It is likely that, ascommonly expected, interest rates may nothave significant effects, neither in the shortor long run, on the level of output and itsgrowth rate. Therefore, until further work iscarried out on the effects of the rate of intereston output, it is difficult to justify the use ofthe Waqabaca and Morling findings as abasis for formulating monetary policies.

In comparison to their output equation,the Waqabaca and Morling specification ofthe inflation equation is perhaps one of thebest for the determination of inflation in Fiji.They assume that there is a long-runrelationship between the price level (CPI),output gap, unit costs of production, andimport prices. Therefore, their GETS-basedspecification is

1 1 2 1 3 1 4 11

ln ln1 2

ln l3 4

1ln ln( ) ln ln

ln 0

n( )

(3)

P wt i t ji j

mPt kk n

mc P c Y YP c w c P tct t t t

Pt

Y YP t n

α α

α α

ν

α ∑ ∑Δ + Δ− −

∑ ∑+ Δ + Δ−

− − − + + +− − − −

Δ = +

− −

⎡ ⎤⎢ ⎥⎛ ⎞

⎜ ⎟⎢ ⎥⎝ ⎠⎢ ⎥⎣ ⎦

where Y is output, YP is potential or capacityoutput, P is an index of CPI, w is nominalunit labour costs, and Pm is an index ofimport prices in domestic currency.15 It is notknown from the reported results in Table 5whether the estimated coefficient of 0.176 isthat of the level of the output gap or its change.However, their subsequent discussionimplies that it is the coefficient of the level ofthe output gap. Therefore, it may be said thata 1 per cent increase in the output gapincreases the price level by 0.176 per cent.This effect is increased to 0.35 if the unit costsvariable is dropped, because some of theeffects of the output gap on inflation arethrough increased wage costs. It is necessaryto validate the Waqabaca and Morling (1999)findings with further research before they areused as a basis for formulating real worldeconomic policies.

Monetary policy in Fiji

Our review of Fiji’s economy reveals thatthere are some fundamental economicproblems that cannot be corrected bymonetary policy alone. Being an insignificantplayer in world markets, Fiji is heavilydependent on trade and is vulnerable toexternal shocks. The importance of suchfactors is reflected in the large variations inthe growth rate of output (see Figure 1).Williams and Morling (2000) and Waqabacaand Morling (1999) suggest that much of theshort-term fluctuations in output and currentaccount deficits are beyond monetarypolicymakers’ direct control. Similarly, inspite of various investment incentivesprovided by the government and artificiallylow interest rates maintained by the ReserveBank, the investment ratio in Fiji did notimprove. This is because investors placehigher priority on good governance, effectiveproperty rights, and of course macroeconomicstability, in comparison to investment

13



incentives and low interest rates—a pointalso emphasised by Garnaut (2005). In sucha situation, the best policy option for monetaryauthorities is to establish broad and stablemacroeconomic conditions by maintainingthe domestic and external value of thecurrency.

However, on another front, it is fair tosay that the Reserve Bank has succeeded inpreventing major economic crises precipitatedby the financial panic due to the two coups.According to Siwatibau (1993), during thefirst coup of 1987 Fiji lost massive reservesthrough capital outflows. Both the externalloan facility and domestic credit dried up inthe two weeks following the political events.The demand for money (cash) and other liquidassets rose sharply but the Reserve Bankinitially failed to respond appropriately. Itmisread the market signals and accommodatedincreased demand and fuelled further capitaloutflows. However, after realising themistake, it quickly tightened capital outflowsand credit through capital controls andhigher interest rates. The statutory reservedeposit ratio was increased, the overnightlending facility was withdrawn, an interestpenalty was levied on overdue advances bybanks, and the discount rate, inter-bank rateand treasury bill rate were all increased.While the treasury bill rate increased from 2per cent to 20 per cent, the discount rate onpromissory notes reached 24 per cent.Siwatibau noted that the higher interest ratesfailed to reduce capital outflows, whichnecessitated tighter capital controls. Whilethe increased restrictions curtailed capitalflight, it was still possible to transfer largeamounts across national boarders. With adeteriorating external position and massivecapital outflows, the Reserve Bank wasforced to devalue the domestic currency twicein 1987—initially by 17.75 per cent and thenagain by 15.25 per cent. Following thesemeasures and application of tight fiscaldiscipline, there was a turnaround in

economic activity by the end of 1987. Reservesrose, commercial bank liquidity increased,and market interest rates declined. Sincethen, monetary policy has been generallyaccommodative, especially after the exportincentives were adopted in 1989.

With the experience of the first coupsbehind them, the Reserve Bank seems to havesuccessfully managed the crisis of the coupin 2000 and avoided devaluation. Credit wasquickly tightened and capital controls wereenforced by reducing transaction limits andincreased documentary requirements. Creditceilings were also imposed on individualbanks. Minimum lending, rediscount andrepurchase rates were increased up to 15 percent. The Reserve Bank policy rate (bank rate)was increased to 5 per cent in anticipationthat the public would hold moneydomestically.

Since 2001, monetary policy has beenmore accommodative. High levels of liquidityhave been maintained and the interest ratesgradually reduced through reductions in thebank rate (perhaps on the basis of theWaqabaca and Morling research) to encouragegrowth of output and improve investmentactivity. However, these interest ratereductions did not succeed. Although GDPgrowth rate has recovered from –3.0 per centin 2000 to 2.6 per cent in 2001, it declined to1.7 per cent in 2002. Since then the Bureau ofStatistics has stopped publishing nationalaccounts data. Many commentators conjecturethat the rate of growth of GDP during 2003to 2005 might have been worse because ofdeclining exports, the low investment ratio,increased imports, and almost neutralbudgets.16 This suspicion is supported byrecent Reserve Bank concerns about theincreased imports. As a result, the bank ratewas increased in mid 2004 to slow down theeconomy (which perhaps had alreadyslowed), to dampen consumption and reduceimports. The suspicion of a low level of recentoutput growth is further corroborated by the

14


Economic survey


modest rate of inflation. The average inflationrate (CPI) in the post 2000 coup years has been2.9 per cent.17 The Reserve Bank might haveoverreacted to the rise in the rate of inflationfrom 0.8 per cent in 2002 to 4.1 per cent in2003. In 2004, the inflation rate declined to2.8 per cent and it is doubtful if this declinewas the consequence of the rise in the bankrate from 1.25 per cent in 2003 to 1.75 per centin 2004. Subsequently, towards the end of2005, the bank rate was increased to 2.25 percent and in early 2006 to 3.25 per cent, with afurther warning that the bank rate willincrease if consumption expenditure does notslow down. Perhaps these changes to the bankrate were based on Katafono’s finding thatthe demand for money in Fiji is unstable andWaqabaca and Morling’s findings thatchanges in the bank rate have significant anddesirable effects on output and price levels.However, as noted earlier, there is no supportfor Katafono’s findings and it is doubtful thatWaqabaca and Morling’s results are robust.Furthermore, there is no evidence that theinflation rate is on the rise due to anoverheated economy. Therefore, it is hard tojustify these upward adjustments to the bankrate to slow down inflationary pressures.

Do increases in the bank rate reduceconsumption and imports, as expected by theReserve Bank of Fiji? Rao (2005) and Rao andSingh (2004) have found that consumptionexpenditure in Fiji does not respond tochanges in interest rates. However, they alsofound that consumption may respond mildlyto a reduction in the interest rate spread,which is the difference between the long andshort-term interest rates. In other words,consumption expenditure can be modestlyreduced if the availability of credit declines.18

However, it is doubtful that a modest declinein consumption will reduce imports by asignificant amount. Therefore, the recentmonetary policy measures of the ReserveBank may not reduce consumption andimports, as expected.

This is not to say that the upwardadjustments to the bank rate are not necessaryfor a different reason. Given that the bankrates (and deposit rates) in Australia and NewZealand have been recently increased to 5.75per cent and have always been well above thebank rate (deposit rate) in Fiji, it is amazingthat Fiji has managed so far to prevent largeoutflows of capital. We suspect that, in recenttimes, these outflows might have considerablyincreased due to the ease with which overseasbank accounts can be opened and manageddue to improvements in informationtechnology. The low deposit rates in Fiji arealso likely to encourage exporters andimporters to invest their funds for short or longperiods overseas. Needless to say, all suchtransactions, legal or illegal, would adverselyaffect foreign exchange reserves.

The low bank rate policy did not lead tolower lending rates in Fiji because of thehigher margins set by the banks and otherfinancial institutions. Currently, for example,the margin is 5.3 per cent; in contrast, inAustralia it is only 2 per cent. Although thelending rates are similar, there is a significantdifference in the deposit rates. If the bank ratein Fiji were to be raised to a rate comparableto Australia and New Zealand, say to 5.75per cent, and the deposit rate to say 6 percent, the Reserve Bank could negotiate withthe banks to cut their margins significantly.In return, the Reserve Bank may considerreducing the statutory deposit ratio to reducethe costs imposed on the banks of itsstabilisation policies. This is in line with thestated objective of implementing moremarket-oriented policies in the financialsector. A further gradual increase in the bankrate in Fiji by 2.5 per cent may eventuallyincrease the lending rates to 9.8 per cent to10 per cent if banks reduce their margins by1.5 per cent. The increase in the deposit rateswould be likely to reduce capital outflowsand may even give a one-shot boost to thesaving rate. It is difficult to argue that the

15



increased lending rate would reduce furtherthe already low investment rate. As pointedout earlier, lack of the investor confidenceseems to be related to changes in the bankrate and may be the main reason for thecurrent low investment rate.

A vision for monetary policy

It is hard to say which are the best indicatorsof economic stability: political stability, goodgovernance, a stable rate of growth of output,stability of the financial sector, and secureproperty rights are all potential indicators ofstability. While not disagreeing with thisview, it may be said that for the monetaryauthorities of a country, stability essentiallymeans maintaining the domestic andexternal value of its main liability, money.

Globally, inflationary pressures havebeen kept low since the 1990s. Cecchetti etal. (2006) attribute this success to responsiblemonetary policy, which essentially wasthrough restricting the rate of growth ofmoney supply so as not to exceed demandfor money excessively. From this perspective,it can be said that the Reserve Bank deservesfull credit for achieving and maintainingdomestic price stability through itssuccessful control of monetary aggregates likeM1. In the post 1987 coup period, exceptduring 1988, the rate of growth of real moneysupply (M1) has been restrained. In the threeyears following the large increase in moneysupply during 1988, the real rate of growthof money was –7.0 per cent per year to mopup the excess liquidity created to meet thepanic demand for cash caused by the coup.On average, in the post 1987 period theaverage real money growth was 4.4 per centand output growth was 1.7 per cent. The rateof inflation was 4.5 per cent, close to theunderlying trend rate of growth of 4 per cent.It is interesting to note that even afterswitching to the bank rate as the policyinstrument in 1997, the Reserve Bank did not

lose control of the rate of growth of moneysupply. The average rate of growth of realmoney supply from 1997 to 2002 was 4.3 percent. Output and inflation grew at an averagerate of 2.4 per cent and 3.1 per cent,respectively. The decline in the inflation ratemay be due to the improved growth rate ofoutput. Therefore, it may be said that theReserve Bank of Fiji has successfullymaintained the internal value of money inthe post 1987 period and avoided financialcrises.

On the exchange rate front, after two largedevaluations in 1987 and another in 1998,the Reserve Bank seems to have successfullymaintained the external value of the currencythrough a managed exchange rate system.However, in spite of the three devaluations,the gap between imports and exports hasbeen widening. If these trends continue, itmay be difficult to maintain the externalvalue of the dollar.

It is tempting to suggest that Fiji shouldabandon the managed exchange rate policyand float the dollar; see for example Chand(1998). Besides arguments made against afloating exchange rate system for Fiji byJayaraman (1999), it is worth noting someimportant observations of Masson et al.(1997) and Chang and Velasco (2000a). If thecentral bank is targeting inflation, any fixedexchange rate system (for example, currencyboards) should be given up. Furthermore,central banks should respond to exchangerate variations because they affect indirectlythe inflation rate through their effects on therelative prices of domestic and importedgoods and directly impact on inflationthrough the prices of imported goods forwhich there are no domestic substitutes.Therefore, they take the view that the fixedversus flexible exchange rate debate is onlyacademic because participants in thesedebates often define fixed and flexible inways for which there are no clear correlatesin practice. A corollary of these observations

16


Economic survey


is that a dirty float (a managed exchange ratesystem) is often an optimal choice.

With a managed exchange rate systemand a target of domestic price stability, theReserve Bank of Fiji should aim atmaintaining a stable exchange rate. Tomaintain a stable exchange rate, it will needadequate foreign exchange reserves tointervene in the currency market to smoothexchange rate fluctuations. However, foreignexchange reserves may dwindle if there arepersistent current account deficits and/orcapital outflows due to higher returns inother countries. Since domestic interest ratesare far below the rates in nearby Australiaand New Zealand, the Reserve Bank uses itspowers to control capital outflows to someextent. We anticipate that in spite of thesecontrols, a significant amount of capitaloutflow takes place. However, as far ascorrecting persistent current account deficitsis concerned, central banks have limitedpowers. If the deficits are due to fundamentaleconomic problems, devaluations may notcorrect the problem even if the Marshall-Lerner condition is satisfied. The success ofdevaluations depends, therefore, not just onthe validity of the Marshall-Lerner conditionbut also on whether appropriate policieshave been implemented by the governmentto correct the fundamental problems.

A similar reason was given by Garnaut(1995) for the failure of devaluations to correctPapua New Guinea’s current accountdeficits. Although Reddy (1997) and Singh(2006b) found that the Marshall-Lernercondition is satisfied for Fiji, the currentaccount balance worsened very quicklyfollowing the three devaluations. This is anindication that some fundamental problemsneed attention and it is doubtful that suchproblems can be rectified through monetarypolicy instruments, including adjustmentsto the exchange rate.

Maintaining the external stability ofmoney is a challenging task for many central

banks in developing economies that have notpaid much attention to the structuralproblems of their economies. In a changingworld of globalisation and economicintegration, many unexpected financialtremors are likely. With the financial crisesof the East Asian countries behind us, wenow know that the consequences of shockswill be more serious for countries thatembrace globalisation policies withoutimplementing adequate institutional reformsby modifying their existing banking andcorporate laws.

However, financial crises and runawayinflation are not currently seen to be a seriousthreat. Masson, Svastano and Sharma (1997)argue that even if a low inflation rate is theultimate objective of a central bank, monetaryauthorities need to respond to exchange ratemovements because fixed and peggedexchange rates are not viable in the long rununless a country has accumulated substantialamounts of foreign exchange reserves.Similarly, Chang and Velasco (2000b) arguethat the question for most emerging marketeconomies is no longer ‘to float or not to float?’but ‘how to float?’ There are three choices:hard pegs (for example, Argentina and HongKong), managed floats (many countriesincluding Fiji), and fully flexible exchangerates (for example, Brazil, Mexico, Peru andSingapore). Chang and Velasco found thatgrowth performance during the period 1997to 2000 was better in countries with flexibleexchange rates than in Argentina and HongKong with their currency boards and hardpegs. They recommended dirty floats foremerging countries, with wider margins forvariations in the exchange rate, because theirgoods markets are exposed to frequent externalshocks. Fiji, like several other developingeconomies, has opted for a managed float,although the Reserve Bank seems to bemaintaining a smaller range of fluctuationsfor the dollar against a weighted average ofthe currencies of its five main trading partners:

17



Australia, New Zealand, the United States,Japan and the United Kingdom.

An alternative suggestion by Garnaut(2005) is to form a currency board: meaning ahard peg, perhaps anchored to the Australiandollar. Under this regime Fiji may derive someof the benefits of a flexible exchange ratesystem, depending on the extent to which thevalue of the Australian dollar is determinedby market forces. This arrangement would bebeneficial if Australia and Fiji experiencecommon external shocks, especially given thatthe bulk of Fiji’s tourism earnings are fromAustralia and New Zealand. However, thereare also disadvantages in currency boardsbecause domestic banks are left without alender of last resort. In a world of fractionalreserves, inadequate regulations and theabsence of deposit insurance, minor financialtremors may lead to large deposit runs. Somecountries, therefore, abandoned hard pegs infavour of managed exchange rates; seeKaminisky and Reinhart (1999). If, in acurrency board system, the monetaryauthority retains its role as the lender of lastresort, financial crises may become worsebecause the central bank has to print currencyas well as satisfy the demand for foreigncurrency by domestic depositors.

If a managed float is justified as pragmatic,an important question is whether the presentrange set by the Reserve Bank for exchangerate variations is adequate? A narrow rangemay be adequate if worsening trade balancesare transient. The wider is the acceptablerange of variations for the exchange rate, thecloser will be a managed float to a flexiblefloat. Therefore, if a country prefers amanaged float, it is appropriate to considerif widening the current limits on exchangerate variations improves the trade balance,even for a few years. This avoids the pressureon domestic prices caused by largedevaluations once every five or ten years.

A downside to widening the exchangerate band is that when the exchange rate

depreciates there is the possibility of largerincreases in the inflation rate. However, someof the increased costs of imported goods maybe absorbed by reduced profit margins andsubstituting cheaper imported goods fromother sources. If the margins for exchangerate variation are not widened, eventually itmight be necessary to devalue the currencyby larger amounts, as happened in 1987 and1998. This may lead to hardship for householdsbecause it is difficult for importers and retailersto absorb large increases in costs.

Some of the adverse effects of smalldevaluations can be minimised by creatingthe conditions for investment in sectors thatcan profitably produce domestic substitutes.19

This calls for a joint effort by the central bankand the government. The Reserve Bankcannot shoulder full responsibility. At best itmay give the necessary time for the govern-ment to implement policies to encourage theproduction of domestic substitutes.

Our preference for widening the limitsfor exchange rate variations is also partlybased on the fact that it is hard to revalue thecurrency after large devaluations. Wideningthe margins at least leaves scope for the dollarto find its long-run equilibrium value andgives adequate time to the government toimplement policies to solve fundamentalstructural weaknesses.

Summary and conclusions

This survey has identified the main policyobjective of the Reserve Bank of Fiji asmaintaining the internal and external valueof money. Although it has other objectives,either they do not need frequent policyinterventions or are somewhat difficult toachieve with monetary policy instrumentsalone. Our review of the stylised facts of theeconomy shows that Fiji’s short to medium-term growth outlook is gloomy, due topolitical instability, depressed investmentclimate, exodus of human and financial

18


Economic survey


capital and widening trade deficits.Nevertheless, the Reserve Bank should becredited with preventing major financialcollapses during the two political coups,restraining inflationary pressures, andmaintaining the external value of thecurrency. These successes, in our view, aredue to judiciously controlling the rate ofgrowth of the money supply.

Our review of international trends in theuse of monetary policy shows that centralbanks in developed economies have switchedfrom using the money supply as theirmonetary policy instrument to the bank rate.This was necessary because of the instabilityin their demand for money functions.Furthermore, domestic expenditures andcapital flows are reasonably sensitive tointerest rate changes. However, althoughReserve Bank of Fiji researchers have foundthat the demand for money in Fiji is unstableand aggregate demand and the inflation raterespond to changes in the rate of interest, therobustness of these findings is doubtful.Therefore, it is difficult to justify a switch fromusing the money supply to the bank rate asits instrument of monetary policy.

The Reserve Bank’s power to stabilise theexchange rate seems to be limited at best.Although it has devalued the currency threetimes, there was no significant improvementin Fiji’s trade or current account balances. Thisimplies that trade deficits are mainly due tostructural and institutional weaknesses in theeconomy. Needless to say, monetary policymeasures alone are inadequate to correct theseproblems. Therefore, unless the governmentimplements appropriate policies to boostexport competitiveness and productivity, theReserve Bank will be forced to periodicallydevalue the currency by large amounts.

We proposed an alternative vision formonetary policy to avoid large devaluations,but take the view that unless the structuralweaknesses are alleviated, devaluations areinevitable. The key aspects of our alternative

vision are that the bank and deposit rates beraised to those comparable to Australia andNew Zealand and reduce the statutorydeposit ratio to persuade the banks to reducetheir margins so that the lending rates donot increase significantly. Implicit in ourvision is also the suggestion that the ReserveBank should switch back to using moneysupply as its monetary policy instrument sothat interest rates are market-determined.

A final point in our vision for monetarypolicy is that the Reserve Bank may considerwidening the upper and lower limits forvariations in the exchange rate. The advantageof this measure is that it would give some timeto the government to implement policies toreduce structural problems. Furthermore, anyresulting inflationary effects of fluctuationsin the exchange rate are likely to be smallerthan from infrequent large devaluations.

We hope that our survey of the course ofmonetary policy will encourage furtherresearch on the demand for money in Fiji andon the effects of the bank rate on output andinflation. Furthermore, we hope that ourwork, taken in the context of the judiciousand insightful observations by Garnaut(2005), will make policymakers think aboutthe need for some hard measures to reducethe structural problems of the Fijian economy.

Notes

1 The problem of clearly identifying objectives,instruments, and their success and failures iscomplicated by the high propensity of centralbanks and commentators to use a variety ofacronyms and clichés.

2 The problem of clearly identifying objectives,instruments, and their success and failures iscomplicated by the high propensity of centralbanks and commentators to use a variety ofacronyms and clichés.

3 Waqabaca and Morling (1999) also have agood description of the institutional detailsand how monetary policy is formulated andimplemented in Fiji.

19



4 The need for these arrangements are justifiedor rejected by examining their effects mainlyon trade flows between the membercountries. For a general discussion see Rose(1999). Jayaraman (2003) examined theseissues in the context of the Pacific islandcountries.

5 In the post war period, beginning from 1944,and until the breakdown of the BrettonWoods convention of fixed exchange ratesof the 1980s, maintaining a pre-determinedexchange rate was the target. The bank rateand/or foreign exchange reserves were usedas instruments to maintain the target.

6 Among Fiji’s main trading partners, Australiaand New Zealand had targeted exchangerates until 1983 and from 1993 switched toinflation targeting. From 1983 to 1993, theywere also targeting inflation, albeit implicitly,through controlling monetary aggregates.Since 1993 they have been using the bankrate as their main instrument of monetarypolicy. Countries that still target the exchangerate are Hong Kong, Malaysia, Singapore,Argentina, Costa Rica, Denmark, Norway,Tunisia, Turkey, Uruguay and Venezuela.Hong Kong and Argentina target theexchange rate because they have currencyboards. India, Israel and Sweden seem to betargeting a weighted average of exchangerate and inflation. In contrast, the UnitedStates and Japan target a combination ofmacro variables. For details see Fatas andRose (2004) and Mathew (2006).

7 There are claims, mainly by Reserve Bank ofIndia sources, that capital flows in India havebecome increasingly responsive to changesin the bank rate since the financial reforms in1997.

8 Countries in which there is no evidence thatmonetary policy has become more effectivein the 1990s are Austria, Germany andSwitzerland. This is understandable becausein these countries monetary policy wasequally effective in both periods.

9 It would be interesting to use the frameworkof Cecchetti et al. (2006) to analyse the efficacyof monetary policy in countries with stablemoney demand functions. One would expectthat monetary policy has become less efficientbecause of an inappropriate choice of

monetary policy instrument. Jha’s findingsfor India are a preliminary indication of thisexpectation.

1 0 During this period foreign exchange reservesin India reached a low point, sufficient onlyfor imports for a few weeks. This caused amajor crisis and the International MonetaryFund forced India to implement marketliberalisation policies. Needless to say, Indiabenefited enormously by liberalising itseconomy. Its current foreign exchangereserves exceed 6 months of imports.

1 1 A limitation in all these studies is that theyhave not used cointegration tests based onstructural breaks in the variables. It is difficultto say what would be the findings in thatcase. Although it is relatively easy to modifyunit root tests for structural breaks, there aretwo problems. First, the sample size for Fijiis small. Second, a rigorous application ofcointegration techniques based on structuralbreaks (not just unit root tests) is difficult andcomputationally demanding.

1 2 The average real bank rate during theirsample period of 1975 to 1998 was about 1.24per cent and the average rate of inflationmeasured by the CPI was 6.1 per cent.

1 3 There is confusion over whether the authorshave included in their equation the laggedvalues of the changes in the real rate ofinterest (Δr) or the lagged values of its level(r). Since in their subsequent discussion theysay that interest rate has only short-termeffects, we presume that there was atypographical error and we make thiscorrection in Equation 1 and include laggedchanges in the real rate of interest (Δr).

1 4 That they have used the GETS approach issupported by their use of the critical valuesfor cointegration between the levels of thevariables from Kremers, Ericsson andDolado (1992).

1 5 It would have been more appropriate to useln(Y/YP) instead of ln(Y–YP) because oftenYP may exceed Y and it is meaningless totake the log of a negative variable. We hopethat Waqabaca and Morling have used (Y/YP) and not (Y–YP) in their empirical work.

1 6 Therefore, only unreliable estimates (in ourview) of growth of GDP are available. TheRBF estimates of GDP growth rates are 3 per

20


Economic survey


cent (2003), 4.1 per cent (2004), 1.7 per cent(2005) and 2 per cent (2006), implying anaverage growth of 2.7 per cent for these years.

1 7 In the four years after the coup the inflationrates were 4.2 per cent, 0.8 per cent, 4.1 percent and 2.8 per cent respectively.

1 8 In many years the real interest rate on savingdeposits in Fiji has been negative. It is likelythat interest rate effects on consumption havebeen masked by these low and often negativeinterest rates. Therefore, it is difficult to denythat a substantial increase in the rate of interestto keep the real rate of interest positive wouldnot increase the saving rate through inter-temporal consumption substitution.

1 9 Casual observation of supermarket shelvesshows that Fiji is still importing commonconsumption goods like tomatoes, potatoes,capsicum, cauliflower, carrots, milk, lamb,soda water and a range of other processedfood items. Although expenditure on suchitems would be much smaller than on high-priced consumer durables—which do nothave any scope for domestic production—domestic production of many consumergoods is viable and will reduce imports.

References

Cecchetti, S.G., Flores-Lagunes, A. and Krause,S., 2006. ‘Has monetary policy becomemore efficient? A cross-country analysis’,The Economic Journal, 116(511):408–33.

Chand, S., 1998. ‘Exchange rate andfinancial management: some lessonslearnt for and from Fiji’, Pacific EconomicBulletin, 13(1):36–46.

Chang, R. and Velasco, A., 2000a. ‘Financialfragility and the exchange rate regime’,Journal of Economic Theory, 92(1):1–34.

——, 2000b. ‘Exchange rate policy fordeveloping countries’, AmericanEconomic Review, 90(2):71–75.

Fatas, A., Mihov, I. and Rose. A.K., 2004.Quantitative goals for monetary policy,Working Paper No. 10846, NationalBureau of Economic Research,Cambridge, MA.

Garnaut, R., 2005. ‘Monetary stability ineconomic development’, PacificEconomic Bulletin, 20(3):103–10.

——, 1995. ‘Monetary stability or moredevaluation?’, Pacific Economic Bulletin,10(1):19–23.

International Monetary Fund, 2006.Transparency in monetary and financialpolicies, Factsheet (April), InternationalMonetary Fund Washington, DC.Available online at http://www.imf.org/external/np/exr/facts/mtransp.htm.

Jayaraman, T.K., 1999. ‘Is there a case for afreely floating exchange rate in Fiji?’,Pacific Economic Bulletin, 14(1):93–100.

——, 2003. ‘Is there a case for a singlecurrency for the South Pacific islands?’,Pacific Economic Bulletin, 18(1):41–53.

Jha, R., 2005. Inflation targeting in India:issues and prospects, Working Paper No.2005/04, Australia South AsiaResearch Centre, The AustralianNational University, Canberra.

Kaminisky, G. and Reinhart, C.M., 1999.‘The twin crises: The causes of bankingand balance of payments problems’,American Economic Review, 89(3):473–500.

Katafono, R., 2001. Demand for money in Fiji,Staff Working Paper No. 03/2001,Reserve Bank of Fiji, Suva.

Kremers, J.J.M., Ericsson, N.R. and Dolado,J.J., 1992. ‘The power of cointegrationtests’, Oxford Bulletin of Economics andStatistics, 54(3):325–48.

Masson, P., Savastano, M. and Sharma, S.,1997. The scope for inflation targeting indeveloping countries, Working Paper No.97/130, International Monetary Fund,Washington, DC.

Mathew, J., 2006. ‘Institutional structure formonetary policy: A comparativeassessment of ten central banks’, ICFAIJournal of Monetary Economics, 4(1):6–18.

21



McConnell, M.M. and Perez-Quiros, G.,2000. ‘Output fluctuations in the UnitedStates: what has changed since theearly 1980s?’, American EconomicReview, 90:1,464–76.

Poole, W., 1970. ‘The optimal choice ofmonetary policy instruments in asimple macro model’, Quarterly Journalof Economics, 84(1):192–216.

Rao, B.B., 2005. ‘Testing Hall’s permanentincome hypothesis for a developingcountry: the case of Fiji’, AppliedEconomics Letters, 12(4):245–48.

——, 2006a. ‘Estimating short and long-run relationships: a guide to appliedeconomists’, Applied Economics(forthcoming).

——, 2006b. ‘Time series econometrics andapplied economics: a methodologicalperspective’, Indian Journal of Economicsand Business (forthcoming).

Rao. B.B. and Singh, R., 2004. ‘Aconsumption function for Fiji’, ICFAIJournal of Applied Economics, 4(3):7–14.

——, 2005a. ‘Cointegration and errorcorrection approach to the demand formoney in Fiji’, Pacific Economic Bulletin,20(2):72–86.

——, 2005b. ‘Unit roots, cointegration andthe demand for money in India’, AppliedEconomics, 38(11):1,319–26.

Reddy, M., 1997. ‘Devaluation andeconomic simulation: the Fiji economypost-coup’, Pacific Economic Bulletin,12(2):85–94.

Rose, A.K., 1999. Does a currency union boostinternational trade?, NBER WorkingPaper No. 7432, National Bureau ofEconomic Research, Cambridge, MA.

Singh, R., 2006a. An investment equation forFiji, Staff Working Paper No. 01/2006,University of the South Pacific, Suva.

——, 2006b. Cointegration tests on tradeequations: is devaluation an option for

Fiji?, Staff Working Paper No. 13/2006,University of the South Pacific, Suva.

Singh, R. and Kumar, S., 2006a.Cointegration and demand for money inselected Pacific island countries, StaffWorking Paper No. 03/2006, Universityof the South Pacific, Suva.

——, 2006b. Demand for money in developingcountries: alternative estimates and policyimplications, Staff Working Paper No.05/2006, University of the SouthPacific, Suva.

——, 2006c. Private investment in selectedAsian countries, Staff Working Paper No.14/2006, University of the SouthPacific, Suva.

Siwatibau, S., 1993. ‘Macroeconomicmanagement of small island economies’,in R. Cole and S. Tambunlertchai (eds),The Future of Asia Pacific Economies:Pacific islands at the crossroads?, NationalCentre for Development Studies,Research School of Pacific and AsianStudies, The Australian NationalUniversity, Canberra.

Waqabaca, C. and Morling, S., 1999. Theconduct of monetary policy in Fiji, StaffWorking Paper No. 01/1999, ReserveBank of Fiji, Suva.

Williams, G. and Morling, S., 2000.Modeling output fluctuations in Fiji, StaffWorking Paper No. 01/2000, ReserveBank of Fiji, Suva.

Zentler, R., 1986. ‘Monetary policy anddevelopment in Fiji’, Pacific EconomicBulletin, 1(2):18–21.

Acknowledgments

We are grateful to Professor Ron Duncan forhis encouragement and suggestions forimprovement and to the University of theSouth Pacific for a research grant (Vote Code:6599-1421) which was partly used for thispaper.

rao and singh (peb)

Documents