Chapter

V Mozambique

Author(s):
Andrew Berg, Mumtaz Hussain, Shaun Roache, Amber Mahone, Tokhir Mirzoev, and Shekhar Aiyar
Published Date:
March 2007
Share
  • ShareShare
Show Summary Details
Author(s)
Shekhar Aiyar

Mozambique received high levels of aid throughout the late 1990s, with a substantial increase beginning in 2000, a year when the country was hit by torrential rains and flooding. The aid inflows were accompanied by increased government expenditures, especially in priority social sectors. GDP and export growth rates were high, and social indicators improved.

However, during the aid surge period (2000–03), aid absorption lagged far behind spending out of aid. While about 65 percent of the increment in aid was absorbed by higher net imports, government expenditure increased by well over 100 percent of the increment in aid. Consequently there was a large injection of domestic liquidity. Despite some attempts at sterilization through treasury bill sales, the result was high inflation, nominal depreciation, and substantial currency substitution, especially in 2001.

A policy of selling foreign exchange to curb domestic liquidity could have helped contain inflation. The feasibility of such a strategy is indicated by the large depreciation that occurred in both nominal and real exchange rates, giving room for foreign exchange sterilization without a large risk of crowding out exports, and by the comfortable reserve position throughout the period.

Pattern of Aid Inflows

Gross inflows of aid were high throughout 1998–2003, but spiked in 2000, following the floods that affected the country that year. Although aid inflows decreased as a percentage of GDP after 2000, they remained at a higher level than the pre-2000 average (Table 5.1). The same pattern held for net public inflows, which rose sharply in 2000 and then remained at an elevated level. Debt relief under the Heavily Indebted Poor Countries (HIPC) Initiative accounted for a large part of the inflows, averaging more than $400 million annually from 2000 to 2002.

Table 5.1.Mozambique: Aid and Other Inflows(Percent of GDP, unless specified otherwise)
199819992000200120022003
Gross public inflows13.413.420.016.718.517.4
Net public inflows11.611.420.415.416.415.0
Net private inflows5.915.810.76.315.17.7
Net total inflows17.627.231.121.831.522.7
Memorandum item:
GDP growth rate11.97.51.513.07.47.1
Note: Figures in bold represent the aid-surge period.

Private inflows also were substantial. Foreign direct investment and private sector net borrowing accounted for less than 6 percent of GDP until 1998, but then rose sharply to almost 16 percent in 1999, mainly to finance imports for an aluminum smelting plant. After 1999, private inflows exhibited considerable volatility, but always remained higher than the pre-1999 period.

Aid Absorption

The higher aid inflows in the post-2000 period were partly absorbed by a higher average non-aid current account deficit (Table 5.2). But a substantial part of the increase in aid was also channeled into increasing international reserves, which rose from 16.4 percent of GDP at end-1999 to 21.9 percent at end-2003.1 Finally, there was also a cumulative deterioration of the non-aid capital account. This may have been partly due to the currency substitution during the period, with depositors switching from local currency deposits to dollar deposits.

Table 5.2.Mozambique: Was Aid Absorbed?(In percent of GDP)
Pre-Aid-Surge Average, 1998–99Aid-Surge Average, 2000–03Difference
Net aid inflows11.516.85.3
Non-aid current account balance–19.7–23.1–3.4
Non-aid capital account balance8.77.7–0.9
Change in reserves (increase –)–0.5–1.4–0.9
Note: Errors and omissions included in capital account.

Spending Out of Aid

Despite the less than full absorption of aid noted above, government expenditures increased by more than twice the increment in aid. Allowing for the fact that government revenues also increased over the surge period, there was still an increase in the fiscal deficit before aid of almost twice the increment in aid (Table 5.3). Thus, almost 200 percent of the additional aid was spent, despite only about 65 percent of incremental aid being absorbed. This implied an injection of liquidity into the domestic economy and determined the challenge for monetary policy.

Table 5.3.Mozambique: Was Aid Spent?(In percent of GDP)
Pre-Aid-Surge Average, 1998–99Aid-Surge Average, 2000–03Difference
Net budgetary aid13.116.53.4
Revenue (excluding grants)11.713.82.1
Expenditure (excluding external interest)22.431.18.7
Overall fiscal balance before aid–10.8–17.3–6.6

Starting from a low base, revenues maintained a steady upward path under the Poverty Reduction and Growth Facility (PRGF) program, rising from about 11 percent in 1998 to 14 percent in 2003. Revenues increased even in 2000, when flood damage shrunk the tax base. Although some quarterly revenue targets under the program were missed, these episodes did not appear to be the result of moral hazard arising from increased aid inflows.2

Government expenditures increased at a much faster rate than revenues (Table 5.4). It appears that the increased government expenditures were generally well directed. Growth was strong throughout the period, except for the flood year of 2000. Under the National Action Plan for Reduction of Absolute Poverty (PARPA), expenditures were targeted at priority social sectors, which included education, health, infrastructure development, governance, and the judicial system. PARPA priority sector expenditure rose from 13.3 percent of GDP in 1999 to 18 percent in 2002, and from 55.2 to 65.3 percent of total expenditure.3 Gross enrollment rates in primary school increased from 69 percent in 1996 to 91 percent in 2000, and from 18 to 24 percent for secondary education. The poverty rate declined from 69.4 percent in 1996 to 54.1 percent in 2002.

Table 5.4.Mozambique: Central Government Budgetary Operations(In percent of GDP)
199819992000200120022003
Revenue and grants19.423.721.228.126.024.9
Revenue11.312.013.213.314.214.3
Grants8.111.78.014.811.810.6
Expenditure and net lending21.624.727.334.634.129.4
Overall balance–2.4–1.5–6.0–6.6–7.9–4.9
Note: Figures in bold represent the aid-surge period.

With the aid surge in the form of grants and debt relief, there is little indication of an unsustainable debt accumulation over the period. The net present value (NPV) of debt-to-exports declined sharply, mainly due to HIPC debt relief, from 212 percent in 1998 to 88 percent in 2002. The debt-service-to-exports ratio also declined considerably, from 20 to 4 percent.

Real Exchange Rate and Terms of Trade

The terms of trade remained fairly stable, apart from a sharp fall in 1999 due to a decline in world agricultural prices.4 Over 2000–03, when aid inflows increased considerably, the terms of trade actually strengthened mildly. Both exports and imports grew rapidly, with exports rising especially sharply in 2000 and 2001 due to the completion of an aluminum smelting plant (Table 5.5).

Table 5.5.Mozambique: Exchange Rates, Net Exports, and the Terms of Trade(In millions of U.S. dollars, unless otherwise specified)
199819992000200120022003
Exports f.o.b.244.6283.7364.0703.0682.01,044.0
As percent of GDP6.26.910.020.418.924.2
Imports c.i.f.817.01,199.81,162.31,063.01,263.01,767.0
As percent of GDP20.629.432.030.935.040.9
Export price index117.0100.099.4100.1101.6
Import price index100.3100.0102.397.598.3
Terms of trade (percent change)–3.4–14.2–2.85.70.6
Nominal effective exchange rate (percent change; depreciation –)6.1–2.9–9.6–19.5–13.1–10.9
Real effective exchange rate (percent change; depreciation –)3.5–3.5–1.5–15.2–2.6–2.8
Notes: c.i.f. = cost, insurance, and freight; f.o.b. = free on board. Figures in bold represent the aid-surge period.

Despite elevated aid inflows, a rapid growth of exports and a stable terms of trade, the real effective exchange rate declined sharply, especially in 2001. This is consistent with the pattern of aid absorption lagging behind aid expenditure, and the consequent large injection of domestic liquidity. The real depreciation remained much less than the nominal depreciation due to the high level of inflation (Figure 5.1). Because of the failure of the real exchange rate to appreciate, there was no ex-post evidence of Dutch disease. However, to the extent that the lack of aid absorption may have been driven by the authorities’ reluctance to sell foreign exchange and allow nominal appreciation, a fear of Dutch disease effects may have led to the policy responses observed.

Figure 5.1.Mozambique: Exchange Rate, Money, and Inflation

Source: IMF staff reports, country authorities.

Note: REER = real effective exchange rate; ER = exchange rate; CPI = consumer price index. For REER and ER index, January 2000 = 100. Data cover period January 1999 to December 2003.

Monetary Policy Response

This section chronologically examines the monetary policy response, beginning in 2000, the first year when there was a substantial increase in net inflows of aid. Although total inflows jumped during 1999, this was due to an increase in private inflows, as shown in Table 5.1. In particular, the climb in private inflows was due to foreign direct investment in an aluminum smelting plant. Because the inflow was used to finance capital imports for the project, it presented no new challenges for macroeconomic policy.

Table 5.6 shows the evolution of basic monetary aggregates and macroeconomic indicators. The loosening of monetary policy is apparent from the rapid growth in reserve money from 2000 onward, accompanied by high inflation, currency depreciation, and a substantial shift within domestic portfolios toward dollar deposits (Figure 5.1).

Table 5.6.Mozambique: Selected Macroeconomic Variables
June 1999Dec. 1999June 1999Dec. 2000June 2001Dec. 2001June 2002Dec. 2002June 2003Dec. 2003
Central bank assets(Billions of meticais)
Net foreign assets–1,274–1,384–798–991,4511,7641,7273,9353,1087,104
Net domestic assets3,7744,5014,0094,0393,2464,2934,6403,1973,8621,578
Change in net foreign assets699–1105866991,550313–372,208–8273,996
Change in net domestic assets–891727–49230–7931,047347–1,443665–2,284
Reserve money2,5003,1173,2113,9404,6976,0576,3677,1326,9708,682
Percent change (12-month)15.828.426.446.353.735.617.79.521.7
Percent change (6-month)–7.124.7322.719.2295.112–2.324.6
Commercial bank interest rates(6-month backward averages)
Deposit rate7.97.99.210.213.516.719.418.114.510.5
Lending rate22.722.622.422.425.232.236.63732.529.3
Inflation(End-of-period)
Consumer price index
12-month percent change4613.511.35.12218.59.714.313.7
6-month percent change41.911.305.116.12.17.56.36.9
Dollar/meticais exchange rate(End-of-period; depreciation –)
Nominal exchange rate (6-month percent change)–1.9–5.6–15.4–11.1–25.4–8.7–2.1-0.80.20.2

2000: Year of Floods

From early February to March 2000, Mozambique was buffeted by torrential rains and a cyclone, causing floods and heavy damage to housing, crops, and infrastructure. GDP growth fell to 1.5 percent, from an average of more than 10 percent during the three preceding years. Inflation doubled from the previous year due mainly to supply shortages, and the currency depreciated. The natural disaster also elicited a sharp increase in aid inflows, much of it as humanitarian assistance but also to finance reconstruction activities.

Because of lags in government spending on reconstruction, the aid inflows were effectively sterilized during the first half of the year through an accumulation of reserves in the government’s account at the central bank. With the increase in net foreign assets (NFA) offset by a large fall in net domestic assets (NDA), reserve money grew by only 3 percent from January through June (Table 5.7). As reconstruction gathered pace in the second half of the year, NDA stopped declining, leading to a spurt in reserve money, which grew by almost 23 percent over the last six months of 2000. Despite the rise in aid inflows, the currency depreciated sharply. This may have been caused partly by the central bank’s reluctance to supply dollars to the foreign exchange market, perhaps due to a desire to keep reserves high. Evidence from the parallel market supports this view: a dollar premium on the parallel market, which had first emerged in November 1999, widened to about 10 percent by March before slowly narrowing over the rest of the year.

Table 5.7.Mozambique: Poverty Reduction and Growth Facility Benchmarks and Performance Criteria, 2000(In billions of meticais, unless otherwise specified)
March 2000June 2000December 2000
ProgrammedActualProgrammedActualProgrammedActual
Central government domestic primary deficit (ceiling)7033691,1911,0264,0143,923
Central government revenue (floor)1,6141,6203,2703,3217,4717,463
Net domestic assets of Bank of Mozambique (ceiling)5,0084,7495,1273,9344,1264,039
Reserve money (ceiling)3,2563,0763,2043,211
Net international reserves of Bank of Mozambique (floor) (millions of U.S. dollars)434451424480508526
Memorandum items:
Money and quasi money (M2)11,41812,53911,64613,62415,79116,779
Of which: Foreign currency deposits4,0494,7364,1195,5355,8917,107
Foreign currency deposits/total deposits (percent)42.544.742.247.545.049.5

The nominal depreciation exacerbated the effect of supply shortages on the price level, causing inflation to jump to over 11 percent during the first half of the year.5 Subsequently, however, inflation was kept in check, as supply shortages eased and the demand for money picked up along with reconstruction activities. The currency depreciation and inflation caused an erosion in the returns on deposits in domestic currency, leading to a strong currency substitution effect. Foreign currency deposits as a percentage of total deposits in the banking system rose from 33 percent at the end of 1999 to almost 50 percent by the end of 2000.

The spurt in reserve money over the year could have been avoided to some extent by making more dollars available to the foreign exchange market. Net international reserves remained comfortably above the program floor throughout the year, so selling more foreign exchange would have been compatible with the PRGF program.

2001: Loose Monetary Policy

In 2001, aid inflows remained high, although they came down from the elevated level of 2000. Reconstruction continued and the real economy rebounded strongly, with GDP growth of 13 percent. There was a large increase in government expenditure, which rose to almost 35 percent of GDP from about 27 percent in the previous year. These expenditures were accommodated by continuing with the monetary loosening that started in the second half of 2000, leading to large increases in reserve money (breaching the program target in every quarter), high inflation, and nominal depreciation.

During the first half of the year, reserve money increased by almost 20 percent and the consumer price index (CPI) increased by over 5 percent. This was accompanied by a nominal depreciation of more than 25 percent. Thus, the erosion of returns on local currency deposits continued from the previous year, with the proportion of foreign currency deposits in total deposits rising to 55 percent.

During the second half of the year, the authorities, with the IMF’s encouragement, tried to reign in the growth of domestic liquidity to some extent. Treasury bill sales increased, reserve requirements were increased, and enforcement of reserve requirements was tightened. Despite these measures, however, reserve money grew by a further 29 percent during this period, the CPI increased by 16 percent, and the currency depreciated by 9 percent (Table 5.8).6 A better outcome over the whole year could have been achieved by increasing net central bank sales of foreign exchange. Throughout the year, the central bank’s net international reserves remained comfortably above the program floor, suggesting that the opportunity for foreign exchange sales always existed. Foreign exchange sales would have reduced the growth in reserve money by reducing NFA, thereby abating the steep depreciation, attenuating the high rate of inflation, and possibly having a salutary effect on inflationary expectations.7

Table 5.8.Mozambique: Poverty Reduction and Growth Facility Benchmarks and Performance Criteria, 2001(In billions of meticais, unless otherwise specified)
March 2001June 2001September 2001December 2001
ProgrammedActualProgrammedActualProgrammedActualProgrammedActual
Central government domestic primary deficit (ceiling)1,4171,2272,6131,7113,7713,5894,3404,207
Central government revenue (floor)1,8932,0433,9024,1406,0396,1868,6709,616
Net domestic assets of Bank of Mozambique (ceiling)5,3653,9874,9803,2465,1624,7156,2154,293
Reserve money (ceiling)3,6423,7294,1334,6974,8265,5185,1226,056
Net international reserves of Bank of Mozambique (floor) (millions of U.S. dollars)425518385525451496439533
Memorandum items:
Money and quasi money (M2)15,91717,63716,23819,54719,12821,10419,96721,763
Of which: Foreign currency deposits5,8347,8765,7679,3619,6509,7999,6149,897
Foreign currency deposits/total deposits (percent)44.051.842.455.358.153.955.652.7

The argument against foreign exchange sales as a sterilization instrument usually rests on the crowding of net exports. But in the case of Mozambique, especially in 2001, the facts suggest that such a strategy would have been justified to curb surging domestic liquidity. Over the year, not only was there an extremely large nominal depreciation, but the real effective exchange rate (REER) depreciated by more than 15 percent. Moreover, this occurred against a backdrop of surging exports and high aid inflows during a period of rebounding economic activity. This suggests that in the absence of a policy geared toward building up international reserves, the currency could have been more stable, and that the central bank had ample scope for selling foreign exchange.8

2002: Fighting Inflation

In 2002, aid inflows remained high, increasing by 1 percent of GDP compared with the previous year. Agriculture continued its recovery from the floods, and the construction sector continued to boom, leading to high GDP growth of more than 7 percent. Government expenditure remained at a high level of 34 percent of GDP. However, the effort to tighten monetary conditions that began in the middle of the previous year continued, leading to a fall in the growth of reserve money to about 18 percent, compared with 54 percent the previous year. NDA and reserve money were contained below program targets throughout, and both foreign exchange sales and treasury bill sales were also used to keep reserve money in check, especially during the first half of the year (Table 5.9).

Table 5.9.Mozambique: Poverty Reduction and Growth Facility Benchmarks and Performance Criteria, 2002(In billions of meticais, unless otherwise specified)
June 2002September 2002December 2002
ProgrammedActualProgrammedActualProgrammedActual
Central government domestic primary deficit (ceiling)1,9712,0313,5793,2733,2243,068
Central government revenue (floor)5,3835,3218,4098,20012,40612,057
Net domestic assets of Bank of Mozambique (ceiling)4,8964,6405,6004,9625,1452,953
Reserve money (ceiling)6,4126,3666,9316,6837,4507,134
Net international reserves of Bank of Mozambique (floor) (millions of U.S. dollars)516533506530546638
Memorandum items:
Money and quasi money (M2)23,86124,57524,90126,09325,94226,145
Of which: Foreign currency deposits10,56710,71411,10711,03411,55411,544
Foreign currency deposits/total deposits (percent)50.649.651.448.151.850.9

Tighter monetary conditions yielded lower inflation, with the rise in the CPI kept under 10 percent, compared with 16 percent in 2001.9 The currency depreciation was also largely halted, with both nominal and real effective exchange rate depreciation kept under 3 percent. Currency substitution was consequently curtailed, with foreign currency deposits stabilizing at about 50 percent of total deposits.

2003: A Tale of Two Halves

In 2003, the effort to control inflation through tighter monetary policy continued through the first half of the year. Sterilization through foreign exchange sales continued, with NFA falling considerably. Consequently, the nominal exchange rate appreciated slightly, and reserve money fell over the first half of the year. Although the PRGF program had expired, the IMF continued to monitor the authorities’ performance against their own macroeconomic targets. All the monetary targets for June 2003 were met (Table 5.10).

Table 5.10.Mozambique: Poverty Reduction and Growth Facility Benchmarks and Performance Criteria, 2003(In billions of meticais, unless otherwise specified)
June 2003December 2003
ProgrammedActualProgrammedActual
Central government domestic primary deficit (ceiling)2,5092,1793,8214,108
Central government revenue (floor)6,3756,37414,70314,717
Net domestic assets of Bank of Mozambique (ceiling)4,5683,6723,3161,309
Reserve money (ceiling)7,3996,9708,2388,682
Net international reserves of Bank of Mozambique (floor) (millions U.S. dollars)581611670797
Memorandum items:
Money and quasi money (M2)28,56326,97529,67432,257
Of which: Foreign currency deposits12,92411,41812,75112,716
Foreign currency deposits/total deposits (percent)51.348.849.745.4

In the second half of 2003, the previous pattern of money growth driven by a reluctance to sell foreign exchange reappeared. International reserves increased by $186 million, a level well above the indicative target. Reserve money increased by almost 25 percent over the period, breaching the indicative ceiling, and inflation remained well above single digits for the year.

The rand and the euro strengthened significantly against the U.S. dollar in 2003. Hence, despite the mild nominal appreciation over the year, there was a significant depreciation of the nominal effective exchange rate. Coupled with high inflation, this led to a mild depreciation of the REER.

Conclusions

In Mozambique, there was a significant increase in aid inflows starting around 2000, the year that floods hit the country. Over a four-year period, aid absorption, though considerable, lagged well behind spending out of aid. In fact, government expenditures increased by well over 100 percent of the increment in aid. Hence, there was a large injection of domestic liquidity into the economy.

Sterilization through treasury bill sales proved insufficient to check inflationary pressures. Inflation remained in the 10 to 15 percent range for most of the period. The authorities proved reluctant to use foreign exchange sales as a means of sterilizing the liquidity injection, except in 2002, when the strategy did yield some lowering of inflation. In terms of program targets, this policy response implied that net international reserves remained well above program targets, while reserve money growth exceeded target growth for most of the aid-surge period. Implicitly, the PRGF program implied greater aid absorption via more sales of foreign exchange by the central bank. This would have reduced the money supply and helped check nominal depreciation and curb inflationary pressures.

International reserves in months of imports (excluding imports for large projects, notably the aluminum smelting plant) rose from 6 months at end-1999 to 6.7 months at end-2003.

In particular, in 2002 the revenue target was breached in every quarter. However, these breaches occurred against the backdrop of an ambitious revenue target (in fact, actual revenues increased by almost 1 percentage point of GDP over the year), and were caused mainly by a drop in excise and import taxes arising from an increase in world oil prices.

Excluding bank restructuring costs and interest payments.

Excluding cotton, sugar, and copra.

With imported food accounting for a considerable share of the consumer price index, the nominal depreciation contributed significantly to the jump in inflation. Due to the high level of inflation, the real effective exchange rate depreciated only mildly, by 1.5 percent over the year, despite the sharp nominal depreciation. Also, the depreciation of the nominal effective exchange rate was less sharp than the nominal depreciation against the dollar, because of the weakness of the rand and the euro. (South Africa and the European Union are Mozambique’s major trading partners.)

It is almost certainly the case that part of the high inflation was attributable to supply shortages in the wake of the natural disaster, rather than to loose monetary policy. But the floods occurred at the beginning of 2000, and supply shortages were most acute in that year. Yet inflation in 2001 exceeded inflation in 2000, suggesting that the looser monetary policy of the latter year played a role.

In this context, it is noteworthy that the program’s performance criterion—the NDA ceiling—proved to be of limited usefulness in containing inflation, given the authorities’ response to the aid inflows. Thus, reserve money—which was only a benchmark, not a performance criterion—increased rapidly with the rise in international reserves, and inflation took off. Accordingly, the current PRGF program has adopted reserve money as the performance criterion.

Gross international reserves increased from 7.2 to 7.8 months of imports from 2000 to 2001, excluding imports for large projects (which were mainly foreign-financed).

Inflation would have been lower in the absence of a drought that caused some increase in agricultural prices late in the year.

    Other Resources Citing This Publication