a. Traded goods

Traded goods other than the primary commodity are treated as a composite commodity. For simplicity, the foreign currency price of the composite good is set equal to one, so that the domestic price is equal to the fixed nominal exchange rate. Supply of the traded good depends on the price of the traded good relative to the nontraded good, while demand for the traded good depends on the nominal prices of the traded and nontraded goods, nominal money balances, and the proportion of nominal government expenditures spent on the traded good (assumed to be parametric). Demand for traded good is homogeneous of degree one in its arguments. Thus, supply and demand for the traded good are specified as follows:

where

Ts (.)	=	supply of the traded good,
Td (.)	=	demand for the traded good,
E	=	nominal exchange rate defined as units of domestic currency per unit of foreign currency,
PH	=	nominal price of the nontraded good,
e = E/PH	=	the price of the traded good relative to the nontraded good,
M	=	nominal money balances,
m = M/PH	=	the real stock of money balances,
β	=	proportion of government expenditures spent on traded goods,
G	=	nominal government expenditures,
g	=	volume of government expenditures,

View Table

Ts (.)	=	supply of the traded good,
Td (.)	=	demand for the traded good,
E	=	nominal exchange rate defined as units of domestic currency per unit of foreign currency,
PH	=	nominal price of the nontraded good,
e = E/PH	=	the price of the traded good relative to the nontraded good,
M	=	nominal money balances,
m = M/PH	=	the real stock of money balances,
β	=	proportion of government expenditures spent on traded goods,
G	=	nominal government expenditures,
g	=	volume of government expenditures,

and where

∂T^S/∂e > 0, ∂T^d/∂e < 0, ∂T^d/∂m > 0, and 0 ≥ β ≥ 1.

b. Booming primary export good

National output of the booming commodity, measured in foreign currency terms, is denoted:

where

Z	=	foreign currency value of the national output of the primary commodity,
PB	=	the world price of the commodity,
B	=	the volume of exports of the commodity.

View Table

Z	=	foreign currency value of the national output of the primary commodity,
PB	=	the world price of the commodity,
B	=	the volume of exports of the commodity.

As noted earlier, B is fixed by past plantings or past discoveries, and it is assumed that all output is exported. ^1/ The disturbance to be analyzed is a change ΔZ resulting from an increase in the world price (ΔP_B), where Δ refers to the difference operator.

c. Trade balance and international reserves

In the absence of capital flows, changes in the stock of international reserves held by the Central Bank are determined by the trade balance:

where IR = the stock, of international reserves held by the Central Bank

TB = the trade balance,

and the time subscript 0 on Z indicates the pre-boom value of the primary commodity export.

From equation (4), it is easy to see that a real depreciation of the exchange rate improves the trade balance. ^1/ On the other hand, a higher stock of real cash balances causes a deterioration in the trade balance. Setting ΔIR = 0 and for the given value of foreign exchange earnings Z₀, combinations of the real exchange rate and real money balances for which external balance is achieved can be found. This schedule is positively sloped in (e, m)-space. Intuitively, a decline in the real exchange rate e (i.e., an appreciation) results in a smaller traded goods supply (as producers shift to relatively more remunerative nontraded goods production) and a larger traded goods demand (as consumers shift toward relatively less expensive traded goods). Given Z₀, an exces demand for foreign exchange exists. To dampen this demand, a decline in real cash balances m is necessary. Thus, the schedule along which ΔIR = 0 (i.e., the external balance schedule) is positively sloped in (e, m)-space.

a. Fiscal revenue linkage

It is assumed that a proportion of any foreign exchange windfall from a transitory commodity boom is captured by the government: ^1/

where R	=	government revenues in domestic currency terms,
γ	=	the proportion of the foreign exchange windfall captured directly by the government, with 0 ≤ γ ≤ l. 2/

View Table

where R	=	government revenues in domestic currency terms,
γ	=	the proportion of the foreign exchange windfall captured directly by the government, with 0 ≤ γ ≤ l. 2/

The parameter γ is likely to be equal to 1, in cases where the government, through a marketing board or parastatal, fixes the domestic procurement price of the traditional export good, and does not pass on to producers changes in world prices, or if an explicit export tax is imposed to capture the full windfall. It will be less than 1 if part of the increase in price is passed on to the producers, or if the export tax captures only part of the windfall.

b. Government expenditure policy response

On the expenditure side, the government is assumed to have a target or desired 1evel of expenditures that varies with its revenue position. Actual expenditures, however, are posited to adjust only partially to revenue changes. In order to reflect the asymmetric behavior of the government, the adjustment parameter is assumed to be greater during a revenue boom than during a nonboom period. Thus, expenditures are assumed to behave according to:

where

G*	=	the desired or targeted level of government expenditure,
G	=	actual government expenditure,
η	=	partial adjustment parameter,
λ, η′, η″	=	constants with η′ and η″ > 0,
DUM	=	a dummy variable (1 during a fiscal boom period and 0 during a nonboom period). 1/

View Table

G*	=	the desired or targeted level of government expenditure,
G	=	actual government expenditure,
η	=	partial adjustment parameter,
λ, η′, η″	=	constants with η′ and η″ > 0,
DUM	=	a dummy variable (1 during a fiscal boom period and 0 during a nonboom period). 1/

Assuming that σ=1 ^2/ and combining equations (9), (10), and (11) yields:

The first two terms of the right-hand side of equation (12) represent the trend in government expenditures. Given the assumptions contained in equations (9), (10), and (11), and assuming that λ ≥ 0 ^3/, equation (12) suggests that the government raises the trend Level of expenditures during a boom period. Equation (12) further indicates that government expenditure policy responds to the fiscal imbalance, as measured by the difference between current period revenue and last period’s expenditure level. The strength of this response (or speed of adjustment) depends in part on whether or not the current period is characterized by a revenue windfall. This term indicates that if the fiscal imbalance, as measured by R_t - G_t _₁ is positive (negative), then the government will respond by raising (cutting) expenditures by (η′ + η″) of the imbalance during a boom period and by η′ of the imbalance during a nonboom period.

a. No revenue windfall (γ = 0)

Figure 3 illustrates the response of the system to a commodity price boom occurring at t = 1 that results in a new (higher) level of foreign exchange earnings from the traditional export good (Z₁ = Z₀ + ΔZ₀ = Z₀ + ΔP_B B₀ > Z₀), assuming that γ = 0 (i.e., that the government does not capture any of the windfall in its revenues). Since earnings from the traditional export good have risen, the real exchange rate for which external balance is achieved is now lower (i.e., has appreciated) and is denoted ē₁ · Intuitively, the reason for this is straightforward: with higher foreign exchange earnings from the traditional export good, the excess of traded goods demanded over traded goods supplied can be increased and financed by the increased foreign exchange earnings occasioned by the rise in the price of the traditional export good (Figure 1). ^2/ This increase in the excess demand for traded goods is brought about by a decline (appreciation) of the equilibrium real exchange rate (Figure 3). Thus, the new external balance schedule lies to the left of the old one. Assuming that λ = 0 ^3/, the new monetary equilibrium schedule (internal balance) also lies to the left of the old one and is again coincidental with the (new) external balance schedule. ^4/

System Response to a Commodity Boom, γ = 0

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

System Response to a Commodity Boom, γ = 0

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

System Response to a Commodity Boom, γ = 0

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

The dynamics of the transition to the new steady state occur as follows: upon the price boom, the old equilibrium point A lies to the right of the new external and internal balance schedules. Thus, at A, a trade balance surplus exists (ΔIR > 0), and hence the nominal stock of money is increasing (ΔM > 0). As the nominal stock of money increases, the price of nontraded goods rises. Together with a fixed nominal exchange rate, this implies that the real exchange rate is appreciating. In addition, real balances are increasing, since the proportionate increase in the nominal stock of money is greater than the proportionate increase in P_H. Thus, the economy moves from A along the NN schedule, and comes to rest eventually at point A′. At point A′, ē₁ < ē₀ and ${\overline{\mathrm{m}}}_1>{\overline{\mathrm{m}}}_0$; trade is again balanced; the nominal money stock is again fixed; and the nontraded goods market is in equilibrium.

At t = 2, when the boom subsides, a reverse process ensues_- When foreign exchange receipts fall back to the level prevailing at t = 0, then the real exchange rate for which external balance is achieved is again ē₀; the equilibrium level of real money balances is also the same as at t = 0, that is, ${\overline{\mathrm{m}}}_0$. In essence, when the boom wanes at t = 2, the economy retraces its path back to the initial equilibrium.

b. Windfall spent on traded goods

Figures 4 and 5 illustrate the policy trade-off which can occur when the government responds to a commodity boom with an expenditure program intensive in its use of traded goods, but is then faced with a sudden decline in its revenue position when the boom wanes. Before the boom occurs (t = 0), the economy is operating at a point like A in Figure 2, where the nontraded goods market and the traded goods market are in equilibrium and where both internal (monetary and price stability) and external balance (trade balance) are achieved. It is assumed that G₀ = R₀.

Internal and External Balance: Post-boom Policy Conflict

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Internal and External Balance: Post-boom Policy Conflict

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

Internal and External Balance: Post-boom Policy Conflict

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

Internal and External Balance: Effects of Fiscal Adjustment

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Internal and External Balance: Effects of Fiscal Adjustment

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

Internal and External Balance: Effects of Fiscal Adjustment

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

Following the primary commodity export price boom occurring at t = 1, government revenues and, hence, expenditures rise. Assuming that λ = 0 and η = 1, ^1/ that any incremental government revenue (expenditure) is spent on traded goods, and that γ = 1 (thus, the entire foreign exchange windfall which ends up in government coffers as tax revenue is recycled into purchases of traded goods), then the real exchange rate for which external balance is achieved remains unchanged. In such an instance, the external balance schedule is defined by:

where Δg is the incremental value of government purchases of traded goods, by assumption equal to ΔZ. Since the increased supply of foreign exchange resulting from the commodity boom ΔZ is absorbed by the expansion of government spending on traded goods (Δg), an excess supply of foreign exchange does not emerge. As a result, the equilibrium real exchange rate and hence the location of external balance schedule remains unchanged. In addition, since fiscal balance is also maintained, the internal balance schedule remains coincident with the external balance schedule. As a result, the small open economy remains at point A (Figure 2) while the boom is occurring. ^1/

At t = 2, when the boom subsides, the internal and external balance schedules will diverge from one another. If foreign exchange receipts fall back to the level prevailing at t = 0 (i.e., Z₂ = Z₀ < Z₁) then the loss in fiscal revenues will result in a reduced level of expenditures; however, because of the asymmetry assumption, revenues fall by more than expenditures can be cut. As a result, a fiscal deficit emerges. ^2/ The size of this deficit will depend on the value of the adjustment parameter η′. The closer η′ is to 0, the larger will be the deficit. The closer η′ is to 1, the smaller will be the deficit. ^3/ Other things equal, η′ will be closer to zero the more committed a government is to a development or capital program; the greater the political resistance to expenditure cuts; the greater the economic costs of cutting programs in midcourse; the longer the gestation period of new programs or projects initiated during the boom period; the more inflexible the expenditure structure, etc.

When foreign exchange receipts and hence fiscal receipts fall back to their initial levels, the government cannot fully cut back its expenditures on traded goods, a new (more depreciated) real exchange rate is required to maintain external balance. The new external balance schedule is to the right of the old one (Figure 4); at this higher real exchange rate ${\mathrm{e}}_1^{*}$ an increased (private) supply of traded goods plus a reduced (private) demand for traded goods allows the import requirements of the government to be met and trade balance equilibrium to be achieved.

The internal balance schedule, however, does not move, because price stability (ΔM = 0) requires the same real exchange rate. When the commodity boom wanes, at the old equilibrium real exchange rate ē₀, the trade balance is in deficit by an amount equal to the new level or (incremental) government import demand ${\mathrm{g}}_2^{*}$. This places a deflationary pressure on the money supply which is exactly offset by the flow of credit to the government to finance its fiscal deficit. In other words, at ē₀, the monetization of the fiscal deficit increases the money stock at the same rate that the drawdown of international reserves due to the trade balance deficit reduces the money stock. As such, at ē₀, the money stock and hence prices are stable.

Figure 4 thus illustrates the dilemma faced by policymakers. Given that an immediate fiscal correction may be difficult, a trade-off between internal and external balances exists. At point A, prices are stable, but the trade balance is in deficit. At point B, trade is balanced, but at the expense of inflation.

So long as the government successfully works to reduce its fiscal deficit to zero by cutting expenditure in each time period, the internal and external balance schedules will eventually coincide ^1/ (Figure 5). However, the exact policy choice to attain these balances would depend on, among other things, the central bank’s reserve position in the aftermath of the boom. When the boom wanes, the authorities may have a large enough reserve position to stay at a point like A in Figure 4, where price stability prevails but a trade balance deficit results in a rundown in reserves. For this to be feasible, the central bank’s reserve position must be large enough to finance a sequence of (diminishing) trade balance deficits.

On the other hand, the central bank’s reserve position may be inadequate to sustain the series of trade balance deficits. In this case, it may choose to devalue in order to move to a point like B in Figure 4. As a result, the government maintains trade balance equilibrium, but at the expense of inflation, with the rate of price acceleration depending on the size of the fiscal deficit. Each successive period, as the fiscal deficit contracts, the government can choose to maintain external balance by changing its exchange rate, for example, from ${\mathrm{e}}_1^{*}$ to ${\mathrm{e}}_2^{*}$ and then to ${\mathrm{e}}_3^{*}$, etc. in Figure 5. Or, given the cut in its fiscal deficit, the government may choose to maintain the real exchange rate, despite the leftward shift of the external balance schedule, in order to rebuild its reserve position.

Of course, the government may choose to operate on neither the internal nor the external balance schedule. By operating at a point like C in Figure 4, the government can choose a real exchange rate which yields both inflation (though less than if the government were to maintain external balance at point B) and a trade balance deficit (though smaller than if the government were to maintain internal balance at point A). Such a choice might be opted for if reserves were just high enough to spread the cut in government expenditures over a period of time, so as to balance the costs of inflation with the costs of a real devaluation, or if expectations existed that the commodity price bust might reverse itself.

c. Windfall spent on nontraded goods

Figure 6 considers the case when the government responds to a commodity boom with an expenditure program intensive in its use of nontraded goods. Under the same assumptions as in the previous case (i.e., C₀=R₀, λ = 0, η=l, and γ = 1), a primary commodity export boom at t = 1 results in a leftward shift of the external balance schedule; that is, since foreign exchange earnings from the primary export commodities have risen, the real exchange rate for which external balance is achieved is now lower, that is, has appreciated. Compared to the previous case, the exchange rate appreciates because the windfall foreign exchange is not absorbed (or “sterilized”) by government purchases of traded goods. The internal balance schedule also shifts leftward, and remains coincident with the external balance schedule, reflecting the fact that because η = 1 the budget remains in balance (since η = 1, then ΔD = 0) during t = 1. The revenue windfall captured by the government and the government’s utilization of this windfall on programs or projects intensive in their use of nontraded goods results, however, in an inward movement of the NN schedule. Because of the increased demand for nontraded goods, the price of nontraded goods must rise, resulting in this inward movement. Thus, a new equilibrium is established at a point like B in Figure 6 at time t = 1.

Revenue Windfall Spent on Nontraded Goods

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Revenue Windfall Spent on Nontraded Goods

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

Revenue Windfall Spent on Nontraded Goods

Citation: IMF Working Papers 1991, 044; 10.5089/9781451846362.001.A001

• Download Figure
• Download figure as PowerPoint slide

When the boom subsides at t = 2, the external balance schedule returns to its initial position, but the internal balance schedule will be located to the left of the external balance schedule. As in the previous case, the loss in fiscal revenue results in a reduced level of expenditure; but, because of the asymmetry assumption, revenues fall by more than expenditures can be cut. As a result, a fiscal deficit emerges. The internal balance schedule is thus located at an exchange rate at which the monetization of the fiscal deficit increases the money stock at the same rate that the drawdown of international reserves due to exchange rate overvaluation would decrease the money stock, so that ΔM = 0. The cut in government expenditures and the subsequent reduction in overall demand for nontraded goods result in an outward shift of the NN schedule.

Because of the divergence of the internal and external balance schedules, government policymakers will again be faced with a trade-off between price stability and trade balance. However, so long as the government gradually reduces its expenditures on nontraded goods and hence the fiscal deficit, the internal and external balance schedules will eventually coincide, and the NN schedule will return to its pre-boom position. In this case, however, the cuts in the fiscal deficit result in a rightward movement of the internal balance schedule until it coincides with the external balance schedule (as opposed to the leftward movement of the external balance schedule in the previous case). Eventually, the pre-boom equilibrium will be restored.

a. Episode 1

The high level of international prices for agricultural exports during 1977-78, coupled with an adjustment of export duties in 1977 to capture the windfall gains, caused export duty revenues to surge to 10.3 percent of GDP in 1978 from 2.0 percent a year before (Table 2). ^1/ Reflecting higher export duty rates as well as the impact of the trade liberalization of 1977 on customs duties and expansionary policies partly engendered by the boom, total revenues rose by 9.2 percentage points to 26.4 percent of GDP in 1978. ^2/ In line with the easing of the commodity boom, export duty revenues returned to the pre-boom level in 1982 with a concomitant reduction in total revenues. ^3/

Table 2.

Sri Lanka: Fiscal Finances

(In percent of GDP)

Source: Government Finance Statistics and International Financial Statistics.

Table 2.

Sri Lanka: Fiscal Finances

(In percent of GDP)

			1976	1977	1978	1979	1980	1981	1982	1983	1984	1985	1986
Total Revenue			17.7	17.2	26.4	23.2	20.2	17.4	16.3	19.2	22.2	22.3	20.8
Export taxes			1.6	2.0	10.3	8.3	5.8	4.4	2.7	2.8	4.2	1.8	0.9
Total Expenditure and net Lending			27.3	23.2	40.3	37.9	42.4	33.0	33.8	32.6	31.1	34.0	33.0
Of which:
	Current Expenditure		18.7	17.5	28.0	23.4	24.7	17.2	18.5	18.1	16.0	20.1	18.9
	Capital Expenditure		8.1	5.5	11.0	13.7	16.6	13.2	15.6	13.1	13.0	13.0	13.0
Overall deficit (before grants)			−9.6	−6.0	−13.9	−14.7	−22.2	−15.6	−17.5	−13.4	−8.9	−1.1.7	−12.2
	Domestic financing Of which:		6.4	2.5	4.7	7.5	13.8	6.8	9.3	5.6	2.6	5.1	5.1
		Banking, system	…	…	…	…	…	1.8	3.7	—	−1.4	…	1.7
	External financing		3.2	3.5	9.2	7.2	8.4	8.8	8.2	7.8	6.3	6.6	7.1
Memorandum items:
External current account, excluding grants (in percent of GDP)			−1.8	2.0	−4.5	−11.0	−19.7	−13.7	−14.9	−12.3	−3.3	−10.0	−9.3
Nominal exchange rate change (depreciation -)			−20.0	−6.0	−75.3	−	−5.8	−16.4	−8.3	−13.0	−8.1	−7.1	−2.9
Average inflation rate			1.4	1.1	12.2	10.8	26.0	18.0	10.8	14.0	16.6	1.5	8.0

Source: Government Finance Statistics and International Financial Statistics.

View Table

Table 2.

Sri Lanka: Fiscal Finances

(In percent of GDP)

			1976	1977	1978	1979	1980	1981	1982	1983	1984	1985	1986
Total Revenue			17.7	17.2	26.4	23.2	20.2	17.4	16.3	19.2	22.2	22.3	20.8
Export taxes			1.6	2.0	10.3	8.3	5.8	4.4	2.7	2.8	4.2	1.8	0.9
Total Expenditure and net Lending			27.3	23.2	40.3	37.9	42.4	33.0	33.8	32.6	31.1	34.0	33.0
Of which:
	Current Expenditure		18.7	17.5	28.0	23.4	24.7	17.2	18.5	18.1	16.0	20.1	18.9
	Capital Expenditure		8.1	5.5	11.0	13.7	16.6	13.2	15.6	13.1	13.0	13.0	13.0
Overall deficit (before grants)			−9.6	−6.0	−13.9	−14.7	−22.2	−15.6	−17.5	−13.4	−8.9	−1.1.7	−12.2
	Domestic financing Of which:		6.4	2.5	4.7	7.5	13.8	6.8	9.3	5.6	2.6	5.1	5.1
		Banking, system	…	…	…	…	…	1.8	3.7	—	−1.4	…	1.7
	External financing		3.2	3.5	9.2	7.2	8.4	8.8	8.2	7.8	6.3	6.6	7.1
Memorandum items:
External current account, excluding grants (in percent of GDP)			−1.8	2.0	−4.5	−11.0	−19.7	−13.7	−14.9	−12.3	−3.3	−10.0	−9.3
Nominal exchange rate change (depreciation -)			−20.0	−6.0	−75.3	−	−5.8	−16.4	−8.3	−13.0	−8.1	−7.1	−2.9
Average inflation rate			1.4	1.1	12.2	10.8	26.0	18.0	10.8	14.0	16.6	1.5	8.0

Source: Government Finance Statistics and International Financial Statistics.

A pronounced expenditure response followed the revenue increase. In 1978, total expenditure and net lending soared, rising by 17.1 percentage points to 40.3 percent; by 1982, this figure reached 42.4 percent, even as revenue performance had begun to slacken. The bulk of the increase was concentrated in current spending which rose by 10.5 percent of GDP in 1978 to 28.0 percent, spread almost equally between goods and services (in the form of higher wages and salaries) and current transfers and subsidies (by way of increased fertilizer and petroleum subsidies). Capital expenditures also increased, partly influenced by the new policies initiated in late 1977 to support ambitious development plans. Thus, the undertaking of new projects was facilitated by the increased availability of foreign aid and the fiscal windfall from the commodity boom. As a result, capital outlays rose, by 5.5 percent of GDP in 1978 to 11.0 percent, and continued to rise, reaching 16.6 percent of GDP by 1980. Within capital spending, a large proportion of the increase was traceable to capital transfers which tripled in relation to GDP to 6.1 percent in 1978. ^4/

Because the expenditure response of the government, particularly pertaining to current expenditures, outpaced the revenue windfall, the central government budgetary deficit rose by 8.0 percent of GDP in 1978 to 13.9 percent, and then further to 22.2 percent in 1980. ^1/ Although dependence on domestic financing of the deficit remained low until 1978 in view of available foreign financing, reliance on domestic sources reached a record level of 13.8 percent of GDP in 1980. The resulting inflation of 26.0 percent offset a part of the 1977 exchange rate devaluation and caused the external current account to worsen to 19.7 percent of GDP in 1980 from 4.5 percent in 1978.

By 1981, a process of fiscal adjustment to restore macro-balances had begun, via the containment of current expenditures. Expenditures on goods and services, current transfers, and subsidies fell to levels prevailing during the pre-boom period. The fiscal deficit declined to 17.5 percent of GDP in 1982, with a substantial reduction in domestic borrowing by the government. However, unlike current spending, capital expenditures did not contract in the post-boom period and remained at twice the level of the pre-boom period, reflecting the continued availability of foreign funding for capital projects. Thus, the investment projects initiated in 1977/78 and facilitated partly by the revenue gains from commodity price increases could not be phased out with the ending of the commodity boom, leading to a long-term shift in favor of development expenditures. Nonetheless, the improvement of fiscal balances during 1982 contributed to the strengthening of the external current account and to lowering the rate of inflation.

b. Episode 2

Once again in 1984, there was an all-around increase in international prices of the three main agricultural commodities exported by Sri Lanka (Table 1). As a result, export duty revenues rose to 4.2 percent of GDP in 1984 from 2.8 percent in the preceding year; all of the export duty increase came from tea exports. The boom-led expansion in economic activity also increased revenues from other direct and indirect taxes. Consequently, total revenues rose by 3.0 percentage points in 1984 to 22.2 percent of GDP. Both the external and the internal balances improved; the external current account deficit fell to 3.3 percent of GDP from 12.3 percent a year before, and the overall fiscal deficit declined by 4.4 percentage points of GDP to 9.0 percent. A large net repayment by the government was made to the banking system, and inflation also eased.

The effect on expenditure was felt in the following year (1985), when total expenditures and net lending increased by 2.9 percentage points of GDP. Current expenditure rose by 4.0 percentage points of GDP as outlays on goods and services more than doubled with a near quadrupling of expenditures on defense and security, and the granting of a substantial increase in wages and salaries to the civil service. On this occasion, capital expenditure did not grow and exhibited relative stability. As the commodity boom began to wane in 1985, export duties declined substantially; however, this revenue loss was offset by higher taxation of imports (including raising of the turnover tax). With the growth of expenditures, the overall deficit widened to 11.7 percent of GDP.

In 1986, total revenues declined by 1.5 percent of GDP, as export duties on tea dropped by 0.9 percent of GDP following a 26 percent decline in tea prices. However, the cut in total expenditures and net lending, by 1.0 percent of GDP, suggests that the expenditure program initiated in response to the temporary revenue gain was relatively flexible. Thus, in this instance, the structure of expenditures did not undergo any long-lasting change. By 1986, the overall fiscal and external current account deficits stood at 12.2 percent and 9.3 percent of GDP, respectively, lower than or in the range of the levels prevailing in 1983.

Two broad conclusions emerge from the Sri Lankan experience. First, current expenditures exhibited much greater variation in response to exogenous movements in export prices. Second, capital expenditures were rigid downward in the short run, attaining a higher level during the first episode-a shift that was sustained even after the boom had subs ided.