IV. Output, Expenditure, and Prices
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund
  • | 2 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund


The ultimate objective of financial programming is to establish an economic framework for sustainable growth. This goal cannot be achieved without a realistic forecast of output, expenditure, and prices in the program period. The forecast serves as the basis for subsequent sectoral projections of the fiscal, monetary, and balance of payments accounts. This section, then, discusses the conceptual framework for determining output, expenditure, and prices and suggests some practical projection techniques. During the financial programming exercise at the end of the workshops, a consistency check is conducted to ensure that projections for the budgetary, monetary, and balance of payments accounts forecasting workshops are consistent with the initial projections established in this workshop.

1. Introduction

The ultimate objective of financial programming is to establish an economic framework for sustainable growth. This goal cannot be achieved without a realistic forecast of output, expenditure, and prices in the program period. The forecast serves as the basis for subsequent sectoral projections of the fiscal, monetary, and balance of payments accounts. This section, then, discusses the conceptual framework for determining output, expenditure, and prices and suggests some practical projection techniques. During the financial programming exercise at the end of the workshops, a consistency check is conducted to ensure that projections for the budgetary, monetary, and balance of payments accounts forecasting workshops are consistent with the initial projections established in this workshop.

2. Determining Output: A Conceptual Framework

a. Potential vs. actual output

Targeting the real growth in output is one of the most important elements of a financial program. To be credible, the output target has to be compatible not only with the other elements of a program but also with external realities and the special features of the domestic economy.

The level of output produced in a given period is a function not only of the inputs used during the production process but also of the way they are used and includes technological, managerial, and environmental factors. Thus, output depends in part on how efficiently inputs are combined, and the volume of the inputs themselves is a function of past output. In industrial countries in particular, these relationships are often analyzed using an aggregate production function that indicates the relationship between output and inputs, or factors of production. This function can be written as follows:


Equation 4.1 states that real output (Q) rises with increases in the availability of capital (K) and labor (N), and with technological and other improvements (A) that enable existing factor supplies to be used more efficiently.

Increases in output over time must reflect either increased utilization or more efficient uses of the factors of production. In the long run, technological improvements and other contributions to factor productivity help explain growth rates. In the short run, the economy is working with a given stock of capital, labor, and other resources, but short-run fluctuations in output occur as a result of changes in demand and the availability of raw materials such as petroleum. Short-term changes in real GDP reflect changes in the extent to which labor and capital are fully employed, or capacity utilization rates. The level of output that would be realized if the labor force were fully employed and labor and capital used at normal intensity is referred to as potential or capacity output.1 Capacity utilization is measured by the ratio of actual to potential output; alternatively, unused capacity or a GDP gap may be measured as the difference between actual and potential output (expressed as a ratio of the latter).

Short-run changes in real aggregate supply often reflect changes in final output prices relative to factor prices. Most important in this context is the relationship between increases in output prices and labor costs. In the short term, the combination of a growing work force and a given stock of other factors results in a decline in marginal labor productivity (the extra output generated by additional workers). Because employers increase their work force only if the value of additional workers’ marginal product exceeds their wages, real wages must decline for employers to engage additional labor. To the extent that nominal wages remain relatively fixed in the short run, real wages would fall only if output prices increase. This development is reflected in an upward-sloping aggregate supply curve, as higher levels of employment and output are associated with higher price levels. Concomitantly, a shift in the aggregate (downward-sloping) demand curve caused (for example) by a more expansionary fiscal and monetary policy results in a movement along the supply curve and an increase in both output and prices.2

An important policy issue is the amount of “excess capacity” in the economy, as it is easier to expand production if there are slack inputs available. Once inputs become more scarce, attempts to expand output will have a significantly greater impact on prices. A frequently used measure of capacity utilization is the unemployment rate, which can be expected to be inversely related to the level of capacity utilization. It should be noted that in addition to cyclical unemployment, there is a more permanent category that is sometimes referred to as the “natural rate of unemployment” or, more recently, as the “nonaccelerating inflation rate of unemployment” (NAIRU). The NAIRU concept embodies the fact that even when the economy is operating at full capacity, people are entering and leaving the labor force, and thus some unemployment exists.3 As this type of unemployment may also change over time, the same unemployment rate observed at two different times may reflect different capacity utilization rates.

b. Factors affecting output determination in developing countries

(1) Short-term factors

The preceding analysis of short-run supply responsiveness is most relevant to industrial economies or the industrial sectors of developing countries. In contrast, aggregate supply in primary-producing developing countries may often be relatively unresponsive to changes in prices in the short run but more elastic in the long term. The supply of most agricultural commodities is relatively inelastic over periods of less than a year; other products, such as rubber and coffee, have an even longer gestation period.4 Even the price elasticity of mineral production in developing countries may be low in the short run, owing to the limited availability of the necessary skilled labor and specialized equipment.

Measuring short-run changes in output on the basis of capacity utilization or unemployment rates is difficult in developing countries. First of all, the concepts of unemployment and the labor force are more difficult to define in these countries than in industrial economies. Open unemployment in developing countries may be confined to the urban sector and to those members of the labor force who can afford to engage in a job search. A more complete measurement of unemployment in developing countries needs to allow for underemployment and discouraged workers. These concepts are especially difficult to define for the rural sector, which dominates most of these countries.

Measuring the growth in aggregate capital stock and labor input is also problematic. To solve this technical problem, growth in the labor force can be used as a proxy for an increase in aggregate employment. The implicit assumption is that no significant involuntary unemployment exists in a competitive labor market. Again, however, such an approach is not useful in many developing economies. In this regard, equation 4.1 is a highly aggregative form of the production function, and attention must be paid to the structural composition of the economy in question. In many countries, a significant share of the population lives in rural areas, and agriculture absorbs a dominant proportion of the total labor force but accounts for a relatively small share of total output. As a result, the marginal labor productivity in the agricultural sector is low, and there is often a significant labor surplus. Moreover, although official urban unemployment may be reported at a relatively low level, in many countries, a large floating population has migrated from the rural areas to the cities in search of work. Hence, growth in the labor force is a poor proxy for increases in aggregate employment, and an alternative approach may have to be sought.

(2) Long-term factors

As mentioned earlier, in addition to changes in labor (N) and capital (K), growth in output is dependent on technological improvements, advances in knowledge, the education and know-how of workers, and shifts in input use into high-productivity sectors. The determination of output in developing countries is complicated by price distortions that render economic decisionmaking more complex and inefficient and by heavy dependence on an agricultural sector that is subject to the vagaries of weather and fluctuations in world prices. Predicting the impact of possible changes in government rules responsible for price distortions or of a change in climate on agriculture and agro-based industries is difficult and makes forecasting output in developing countries exceedingly uncertain.

(3) Price distortions

Empirical studies suggest that the efficiency with which available factors of production are used accounts for a substantial portion of the growth rate of output and for differences in growth rates among countries with similar natural endowments.5 Efficient resource allocation maximizes the value of output from a given level of input. A key factor in determining efficiency is the pricing of inputs and outputs to reflect relative scarcities. Prices that deviate significantly from the scarcity value of goods (or opportunity cost) may be regarded as distorted. Such distortions often reflect the authorities’ skepticism about the importance of price signals or government efforts to achieve equity objectives, even at the cost of efficiency.6

Price distortions have been emphasized in analysis because of their impact on both the balance of payments and growth. Measures affecting the exchange rate and trade restrictions have been particularly important in this context. An exchange rate that has been allowed to become overvalued in relation to the purchasing power of the currency reduces the incentive to export or to produce import substitutes.7 In the absence of corrective measures, an overvalued exchange rate can be expected to result in a foreign exchange constraint and non-price rationing of imports. Such rationing can limit the availability of intermediate inputs, seriously constrain the development of new industries, and undermine capacity utilization levels in existing sectors.

Substantial distortions are often involved in the pricing of factors of production. Interest rates in some countries have been kept at artificially low levels, with a negative real interest rate that discourages savings and inflows of foreign capital. Non-price rationing of available financial savings leads to substantial administrative problems and allocative inefficiencies. For those with access to funds, low interest rates encourage an excessively capital-intensive mode of production, which is further encouraged if unrealistic minimum wage laws and social security taxes raise the price of labor.

(4) Relative sectoral roles

In developing countries, the often dominant role of agriculture (in terms of its contribution to the GDP and employment) accounts for the strong relationship between developments in the agricultural sector and overall economic growth.8 Agriculture is not only the source of food for the population but the source of labor for industry, saving for capital formation, and exports for foreign exchange. In order to perform functions in all these capacities, the agricultural sector must itself have a satisfactory growth rate. Variations in weather conditions, unfavorable relative price developments, and inefficient processing and marketing agencies work against strong agricultural growth.

Industrialization may be constrained in many developing countries because their small markets, low population densities, and poor transportation systems limit economies of scale and reduce the choice of industries. These constraints may be exacerbated by domestic industrial policies. Governments often establish public enterprises to lead the industrialization process, but these enterprises are frequently plagued by managerial inefficiencies and by reliance on noneconomic considerations in setting priorities. In addition, prolonged protection of domestic production fosters the development of inefficient industries with limited growth potential, leading to inefficient allocation of resources and unfavorable effects on overall growth.

3. Determining Expenditures: A Conceptual Framework

Because the forecasting of public expenditure is to be developed as part of the workshop on budget forecasting, this section discusses techniques for forecasting private sector expenditure, which, together with public consumption and investment, determines domestic absorption.9

Chapter II showed that GDP or GNP can be measured by expenditure on final products. The expenditure components of GNP can be expressed as follows:10



Y = GNP;

CP = private consumption;

IP = private investment;

CG = government consumption;

IG = public investment;

X’ = exports of goods and services; and

M’ = imports of goods and services.

In most countries, the private expenditure components (CP + IP) of demand account for a substantial share of GNP. Therefore, the behavior of private expenditure is crucial in determining overall stability in the domestic economy. Private spending also has an impact on the amount of private saving and the balance of payments position.

a. Determinants of private consumption

Early theoretical and empirical analysis emphasized the importance of current disposable income as a determinant of consumption. Disposable income can be derived by starting from the identity for GNP, which can be written as:



SP = private saving; and

T = taxes less government transfers.

Private disposable income (YPD) can be obtained by deducting net taxes (T) from gross national disposable income (GDI), which is GDP plus net factor income from abroad and net transfers from abroad (see Chapter II). Thus, YPD can be expressed as:


From this accounting identity, a possible behavioral relationship between private consumption and private disposable income can be postulated as:




This specification of the consumption function implies that the marginal propensity to consume, ∆CP/YPD, is constant and equal to b, while the average propensity to consume, CP/YPD, falls as YPD increases because of the declining relative importance of the constant term. These properties suggest that developments in consumption and saving help to stabilize the economy as the saving-to-GDP ratio rises during booms and declines during slumps. For instance, as income rises, the average propensity to consume falls, exercising a restraining influence on demand-induced output growth. A further implication is that as income grows over time, the saving ratio is likely to rise.

Other approaches to estimating consumption functions, such as the “life-cycle theory” and the “permanent income hypothesis,” recognize that consumers are unlikely to be influenced only by income in a particular period. Rather, consumers consider some longer-run or normal income stream in determining their consumption behavior. For example, during cyclical booms transitory income is positive and the ratio of consumption to actual (as opposed to long-run) income declines. Similarly, cyclical slumps are associated with negative transitory income, and the ratio of consumption to measured income tends to rise. This relationship also suggests that the marginal propensity to consume will be different in the short and long run. The life-cycle theory and permanent income hypothesis represent two essentially complementary attempts to make long-term concepts operational.

The permanent income and life-cycle hypotheses of consumption behavior can both be empirically approximated as follows:



CPR = private consumption in real terms; and

YPDR = disposable personal income in real terms.

It can be shown that the coefficients β1 and β1/(1-β2) are the short- and long-run marginal propensities to consume, respectively.

b. Determinants of private investment

(1) The conceptual framework

Private investment spending exhibits substantial instability in most countries. The simple accelerator theory of investment states that the desired stock of capital (KRPt*) depends on the expected level of output (YRte) as follows:


where k represents the capital-output ratio and is assumed to be fixed.

Since investment represents the change in capital stock, the above equation can be written as follows, assuming that the actual and desired stock are kept equal:


where IRPt = real private investment. This function implies that a constant increase in output is required to maintain an unchanged level of investment spending; expectations of a slower change in output imply a reduction in private investment.

Because of the time it takes to design, build, and install new capital equipment, lags occur in the adjustment of actual capital investment to the desired level. A partial adjustment mechanism is used in explaining the accompanying investment (IRPt) behavior:


where 0 ≤ β ≤ 1.

The coefficient of adjustment (β) is assumed to vary systematically with economic factors that influence the ability of private investors to achieve the desired level of investment. Since ΔKRPt represents net private investment, gross investment, including depreciation on the existing capital stock, can be defined as:


where δ is the rate of depreciation.

The simple accelerator theory of investment assumes fixed capital-output ratios, which imply a production function with technologically fixed proportions among factor inputs, so that factor prices do not enter into the specification. The neoclassical theory of investment provides for some substitutability between capital and labor and includes the user cost of capital as a determinant of investment.

(2) Private investment in developing countries

In developing countries, the standard optimizing investment models (based on the user cost and marginal productivity of capital) are weakened by institutional and structural factors such as the absence of well-developed financial markets, the relatively large role of government in capital formation, distortions created by foreign exchange constraints, and other market imperfections. Moreover, severe data constraints usually hinder attempts to apply such models; data on the capital stock, labor force, wages, and user cost of capital are often limited or nonexistent.

Studies on developing countries have emphasized that the availability rather than the cost of finance represents the major constraint to private investment.11 This situation reflects the fact that real interest rates are being maintained at below-market equilibrium levels and that financial resources are allocated partly by non-price rationing. As a result, bank credit, foreign capital inflows, and retained profits represent the major determinants of private investment, emphasizing the importance of monetary policy in these countries. In addition, there is the possibility of financial crowding out if the government’s recourse to the banking system limits credit to the private sector. Government investment may also cause real crowding out by preempting scarce supplies of domestic and imported physical resources and limiting market opportunities. In contrast to this competitive relationship between private and public investment, the counterargument claims a complementary relationship based on the provision to the private sector of publicly financed essential infrastructure and support services. Empirical research has provided support for this complementary relationship between government and private investment.12

Among other factors that should be considered in projecting private investment are fiscal incentives, the amount of excess capacity in the economy, the exchange rate level and expectations, and import constraints. Introducing tax credits and more attractive depreciation allowances can enhance incentives to invest. With respect to capacity utilization, as the gap narrows between potential and actual output, producers—recognizing rising demand—are likely to try to expand their plants and buy new equipment. Of course, there can be problems with absorptive capacity and shortages of key inputs and/or technical expertise. An overvalued exchange rate tends to discourage foreign direct investment and to make the climate less attractive for domestic investors as well. Exchange rate expectations also play an important role; for example, if a depreciation is expected, capital flight may become significant. Investment can also be constrained by external factors, such as import shortages stemming from a scarcity of foreign exchange quotas, or donor shortfalls.

Another possible approach to forecasting private investment uses the incremental capital-output ratio (ICOR). One difficulty with this approach is the “lumpiness” and “bunching” of investments—that is, very large investments may occur in certain years, while output may be more evenly spread out; such behavior may not be characteristic of either past or future developments. Another shortcoming of the ICOR approach is that it implicitly assumes constant factor ratios or fixed technological coefficients. Clearly, this assumption is not realistic; for example, changes in the price of capital (interest rates) relative to labor have an impact on desired investment and hence should be taken into account.

At its simplest, the ICOR method entails the use of the equation below. The ICOR may be based on data from selected earlier years and is defined as the ratio of total real investment to the change in output during the period in question, or:



INV = total real investment;

ΔY = the first difference in real GDP; and

ICOR = the incremental capital-output ratio.

Private investment can be derived by subtracting government investment from the estimate for total investment yielded by equation 4.9.

4. Prices and Inflation: A Conceptual Framework

Inflation is one of the most complex and difficult economic variables to analyze. Technically, inflation occurs when the demand for goods and services expands more rapidly than total output, exerting pressure on the purchasing power of money. But inflation is often caused by a complicated interaction of several factors, which may in turn be obscured by varying lags in their effects. Expectations of future inflation can be a primary determinant to current inflation.

There are three primary measures of inflation: the consumer price index (CPI), the wholesale price index (WPI), and the gross domestic product deflator (GDPD). Each provides valuable insight into pricing pressures in the economy. The CPI is based on the actual prices of a typical “basket” of domestic and imported goods and services consumed by residents and thus measures price changes felt by the “average” consumer. The WPI measures price movements at the wholesale transactions level and therefore provides some advance information on prospective CPI movements. The GDPD is a price index for domestically produced goods and services and thus provides a measure of trends in international competitiveness. Clearly, a key difference between the CPI and the GDPD are the movements in the prices of traded goods, and if import prices rise faster than domestically produced goods, changes in the CPI can outpace those in the GDPD.

The causes of inflation are often categorized as being cost-push or demand-pull, depending on the primacy of supply or demand factors. On the demand side, it is often argued that large swings in monetary growth strongly influence inflation rates. Rapid monetary growth, often to finance large fiscal deficits, results in an increase in domestic demand, but a slowdown in monetary growth exerts a dampening effect on domestic prices.

A more structural approach to inflation emphasizes the role of bottlenecks, particularly in developing countries, where shortages of skilled labor and capital are important issues. Sometimes the problem can arise from labor agreements mandating wages and benefits that exceed productivity gains. Such agreements can have secondary “demonstration” effects on wage demands in other sectors—for instance, when excessive wage increases granted to government workers are carried over to nonfinancial public enterprise employees. Import prices and foreign interest rates can also play a major role in establishing inflationary trends. Many developing countries, particularly small, open economies, are considered particularly vulnerable to movements in the prices of traded commodities. (The oil price shocks of the 1970s had a sharp impact on inflation rates in many countries.) Finally, international capital mobility has forced interest rates worldwide to become more competitive, seriously eroding the abilities of countries to maintain independent interest rate structures.

5. Output and Expenditure: The Case of Sri Lanka

a. A historical and comparative perspective

(1) Comparative growth performance

In the early 1960s, Sri Lanka was at the same stage of economic development as many of its neighbors in the region, with a per capita income that ranked favorably among these countries.13 By the early 1990s, however, Sri Lanka had fallen well behind many of the Asian countries, in particular Korea, Malaysia, and Thailand. Economic growth in Sri Lanka averaged only 4.4 percent between 1961 and 1990, substantially below the 6.5 percent, 7.6 percent, and 8.7 percent average annual growth rates achieved by Malaysia, Thailand, and Korea, respectively. Even after the Sri Lankan government embarked on a far-reaching economic liberalization program in 1977, the annual average growth rate edged up only modestly, from 3.6 percent (in 1961–72) to 5.5 percent during 1978–90—still well below the 7 percent average growth rate achieved by the three Asian countries and Pakistan, a country at approximately the same stage of development as Sri Lanka.

Table 4.5 illustrates the patterns of growth, saving, and investment in Sri Lanka, Korea, Malaysia, Thailand, and Pakistan from 1961 to 1990. The growth rates in these countries are closely and positively related to their saving and investment rates: Korea, the country with the highest growth rate, also has the highest saving and investment rates; Pakistan and Sri Lanka, the countries with the lowest growth rates, also have the lowest saving- and investment-to-GDP ratios. Differing degrees of price distortions among the countries may also have contributed to the differences in ICORs.

Table 4.5.

Sri Lanka: Growth, Domestic Saving, and Investment, 1961–90

(In percent)

article image
Sources: IMF, International Finance Statistics, and World Bank, World Tables.

l/GDP = Investment/GDP ratio.

S/GDP = Saving/GDP ratio.

(2) Pattern of saving and investment

Low capital accumulation and productivity improvement have been major contributing factors to Sri Lanka’s weak growth performance. These factors are in turn influenced by the pattern of saving and investment.

(i) Saving

Saving rates in Sri Lanka are among the lowest in developing countries. Between 1971 and 1990, the gross domestic saving rate averaged 13 percent of GDP. Taking into account net factor income and private transfers from abroad, gross national saving, on average, amounted to no more than 14.5 percent of GDP per annum (Box 4.1), significantly below the 22–23 percent average saving rate for all developing countries and the 22–26 percent saving rates in Korea, Malaysia, and Thailand during 1971–90.14

Sri Lanka: National Saving/GDP Ratios, 1971–90

article image
Source: IMF, International Financial Statistics.

There are many plausible explanations for the low saving rates in Sri Lanka, such as cultural factors, the level and distribution of income, and large fiscal deficits. During 1971–90, for example, government expenditures and net lending averaged about 32 percent of GDP per annum, while total revenue, including grants, averaged only about 22 percent of GDP. This large fiscal deficit affected saving in several important ways. First, the dominant form of savings available for households was low-interest deposits, which were channeled through the National Savings Bank (NSB) to finance the budget deficit. Second, large fiscal deficits tended to put pressure on the government to monetize the public debt. In fact, large budget deficits have been the main cause of the rapid expansion of both domestic credit and high inflation in Sri Lanka over the last 30 years. In an inflationary environment, households with only limited access to equity markets and real assets have had little incentive to save because of low or even substantially negative returns.15 Third, high current expenditures by the government directly reduced public saving. Moreover, Sri Lanka’s extensive welfare programs served as an extra disincentive for households to save.16

To a certain extent, Sri Lanka, as a small economy, can substitute foreign for domestic saving to increase investment. However, relying too heavily on foreign saving is unrealistic. Given the uncertainty occasioned by the civil conflict and the economy’s relatively early stage of development, Sri Lanka’s ability to absorb foreign direct investment is limited, even if the external sector is further liberalized. In almost all of the high-growth industrializing economies, only modest amounts of foreign direct investment were available in the early stages of development.17 The scope for heavy external borrowing is also limited. With external debt already exceeding 70 percent of GDP in Sri Lanka, relying on foreign borrowing to continue covering an investment-saving gap equivalent to 10 percent of GDP is an unsustainable proposition.

(ii) Investment

In the successful Asian economies used here for purposes of comparison (Korea, Malaysia, and Thailand), public investment plays a largely supporting role in productive economic activities. In Sri Lanka, however, the public sector has always played the dominant role. From the mid-1950s to 1977, the government followed a wide-ranging substitution policy that emphasized the public sector as the driving force in economic development. Private sector development was constrained by price controls, limited access to financial resources and foreign reserves, and other administrative and legal barriers.

Having realized the limits of this inward-looking policy, in 1977, the government began to dismantle administrative controls on resource allocation and to promote market- and export-oriented economic policies. These policies led to rapid private sector growth, and private investment jumped from 8 percent of GDP in 1974–77 to 13 percent of GDP in 1978–84. However, owing to the rapid expansion of government investment in a number of large public projects, public investment continued to exceed private investment until 1988 (Box 4.2).

This investment pattern contrasts sharply with those in the three other Asian countries, where private investment exceeded public investment even during the early stages of development. In Malaysia, for instance, the private investment/GNP ratio reached about 16 percent during 1973–80, while the public investment/GNP ratio remained constant at about 9 percent.

Aside from excessive public investment, the composition of corporate investment also appears to have contributed to low productivity growth in Sri Lanka during the 1980s. Although investment in equipment has been recognized as one of the potentially most effective ways to enhance productive capacity in Sri Lanka, machinery and equipment investment as a share of total fixed investment declined steadily from 50 percent in the early 1980s to only 30 percent in 1989–90. This decline was the result of the continued expansion of construction (especially of buildings); the share of construction in total corporate fixed investment rose from 38 percent in 1981–82 to 58 percent in 1989–90 (Box 4.3).

Sri Lanka: Private and Public Investment, 1978–90

(As percent of GDP)

article image
Source: Central Bank of Sri Lanka.

Budgetary capital transfers to public corporations were excluded from private investment and included in public investment.

Sri Lanka: Composition of Fixed Investment in Private and Public Corporations, 1981–90

(As percentage of total)

article image
Source: Central Bank of Sri Lanka.

(3) Other issues related to total factor productivity

(i) Macroeconomic stability

One common characteristic of the successful growth experiences of Korea, Malaysia, and Thailand has been their ability to maintain a relatively stable macroeconomic environment characterized by fairly low and predictable inflation, a sustainable fiscal balance, and a viable balance of payments position.

The macroeconomic environment in Sri Lanka was not generally stable during 1970–90, with inflation fluctuating between 10 and 20 percent in most years. Although the civil conflicts were an important contributing factor in the ongoing macroeconomic instability, the main problem seems to have been the large and persistent fiscal deficits (including grants), which averaged 9.4 percent during 1971–90. The government’s overly ambitious social welfare and capital expenditure programs also put continual pressure on domestic demand and external balances, causing irregular patterns of inflation and exchange rate changes.

(ii) The labor market and human capital accumulation

A relatively efficient labor market has been an important element in the success of Korea, Malaysia, and Thailand. In contrast, Sri Lanka’s labor market appears highly segmented and inefficient. Although only limited data are available on employment and real wages, periodic surveys indicate that urban and rural unemployment rates have been persistently high. For instance, the labor survey taken in 1985/86 showed an unemployment rate of 19.5 percent in urban areas and 13.2 percent in rural areas and did not include the large number of underemployed workers. On the other hand, pressures on wages were persistent, pushing up the average yearly rate of wage growth by 2 percentage points more than the rate of total factor productivity growth during 1970–90.18

Several factors have contributed to the inefficiency of the labor market in Sri Lanka. The most obvious is the strict labor legislation. Of particular concern has been the Termination of Employment and Industrial Disputes Act, which provides extensive protection to employees and practically eliminates the right of firms employing more than 15 workers to lay off employees.

A second contributing factor to labor market inefficiency is the government’s role as a major employer. The government sector in Sri Lanka accounts for nearly half of all nonagricultural employment and employs two-thirds of all college graduates. Traditionally, public sector salaries have been much higher than those in the private sector, even after adjusting for differences in educational levels.

A third factor is the generally adverse climate for private business that has existed in Sri Lanka for the last three decades. Experiences in both industrial and developing countries suggest that one of the most effective ways to expand employment is through the development of small and medium-sized private businesses. Unfortunately, as mentioned above, Sri Lanka’s government has not played a significant positive role in promoting private sector development.

(iii) Foreign direct investment and technology transfer

Foreign technology transfer is essential to productivity improvement in a small economy such as Sri Lanka, where the domestic market and resources for research and development are limited.19 Technology transfer can take many forms, including foreign direct investment, technology licensing, capital goods imports, and overseas training. Despite government efforts to promote foreign investment in Sri Lanka after 1978, the amount of foreign direct investment has remained modest (Box 4.4).

Sri Lanka: Foreign Direct Investment, 1981–90

article image
Source: International Monetary Fund, International Financial Statistics.
(iv) Financial system efficiency

Despite its diversity, the financial system suffers from serious problems that have impeded efficient intermediation between households and private business.20 Owing to various restrictions, the market share of private commercial banks has been limited, and the commercial banking system has been dominated by two state-owned banks—the Bank of Ceylon and the People’s Bank—which together account for two-thirds of the total financial assets in the commercial banking system.21 In addition to managerial and operational weaknesses (such as high employee-related costs), the two state banks have suffered from serious financial problems, owing in part to pressure to provide credit for purposes accorded high priority by the government, such as supporting public enterprises and promoting rural development. These problems have undermined the banking system’s efficiency in allocating resources to the most productive investments.

Another aspect of the inefficiency of the financial system stems from the roles of competing government financial institutions, especially the NSB, in raising funds to meet the government’s financial needs.22

b. Recent developments

(1) Performance in the late 1980s

Sri Lanka is essentially an agricultural economy, although the share of the agricultural contribution to the economy has been steadily declining, mostly because of the very substantial drop in real output in plantation agriculture during 1987–88 and significant downturns in rubber and coconut production after 1987. Nonplantation agricultural output has also declined. The share of paddy (rice) declined from 6 percent of GDP in 1986 to 5 percent of GDP in 1990; production was adversely affected by the drought conditions of 1989, which hindered planting and fertilizing.

The mining and quarrying sector, which has performed marginally better, saw its share in GDP increase from 2.3 percent in 1986 to over 3 percent in 1990. Much of this improvement has come from a substantial increase in the production of gemstones to meet export demand.

The share of the nascent manufacturing sector in GDP grew from 15 percent in 1986 to over 17 percent in 1990, largely due to the improved performance of subsectors other than tree crop processing. Food processing (54 percent), textiles/garments (19 percent) and chemical/petroleum/plastic products (10 percent) account for over 80 percent of total value added in manufacturing. The textile and garment subsectors have generally performed quite well in recent years. After a change of management in the 1980s, government-owned textile mills have been able to reduce their losses sharply, and the creation of export processing zones in the suburbs of Colombo has led to a rapid expansion of exports. The government has actively sought to foster such zones by providing incentives such as infrastructure development and reduced taxes. Other subsectors, such as wood and chemical products, have tended to perform less well following liberalization, largely because of increased competition from imported products.

The share of the construction sector in GDP has remained constant at about 7 percent, partly owing to the effects of cuts in government fixed investment, especially in 1990. The bulk of the large infrastructural projects were undertaken in the first half of the 1980s.23 The slowdown in construction sector growth also reflects the effects of a decline in private investment in the latter half of the 1980s, which may have been caused by the continuing civil strife in the north and new civil disturbances in the rest of the country.

In recent years, the share of the services sector in GDP has remained at about 50 percent. The most important subsectors in this group are wholesale and retail trade (including hotels and restaurants), transport and communications, banking, insurance, real estate, and government services. The continuing civil strife in the north and east and the disturbances in the west and central provinces have adversely affected tourist inflows since the early 1980s.24 The shares of financial and government services in GDP increased after 1986, the latter in response to the civil disturbances and the related increase in government defense expenditures.

(2) Growth in 1990

A number of factors contributed to a strong growth performance in 1990, including the introduction in mid-1989 of a new economic stabilization and structural adjustment program, improvements in the security situation in the south, and good weather conditions during the cultivation seasons. As a result, real GDP increased by over 6 percent. Several sectors experienced rapid expansion, including the plantation sector, which recorded a strong recovery after a weak performance in 1986–88, and the manufacturing sector, which had a growth rate of over 9.5 percent. Despite growing domestic demand, the expansion of nontraditional exports (including garments) lay behind much of this growth. The relative “normalization” of conditions in most of the country helped to stabilize trade, commerce, tourism, and related sectors, contributing to a growth rate of over 4 percent in the services sector.

6. Recent Developments in Prices and Wages in Sri Lanka

Until the liberalization program was launched in 1978, inflation in Sri Lanka (as measured by the CPI) was relatively low. This apparent stability was in part the result of price controls on major food items and the output of public enterprises, as the government maintained a food distribution system and imposed import and exchange controls. Because of the nature of the country’s interventionist, inward-looking policy regime, repressed inflation was substantial prior to 1978. Until 1985, the reform process initiated in 1978 was associated with double-digit inflation attributable in part to reduced government intervention in commodity prices. During 1985–86, domestic inflation fell to single-digit levels as world inflation eased and the “catch-up” period for domestic prices ended. However, in spite of the government’s ongoing efforts to influence some strategic prices, inflation since 1988 has accelerated, in part because of emerging fundamental disequilibria.25 In particular, inflation has reflected the persistence of high government budget deficits fueled by mounting defense expenditures and the resulting pressure for monetary financing.

In 1990, inflation measured by the CPI index increased to over 21 percent, almost double the rate recorded in 1989, and the GDP deflator rose nearly as fast. A substantial depreciation of the exchange rate in late 1989 was the proximate cause of this sudden acceleration, but adjustments in the administered prices of wheat, fertilizer, and petroleum products were also important.

Trends in private sector real wages moved inversely to trends in inflationary pressures during the 1980s, rising in 1984 when inflation slowed, then falling gradually until 1987, when they generally recovered from the erosion of the prior two years. Real wages for central government employees declined steadily until 1988, the year of the last major revision of the government salary structure. By 1990, the erosion in wages had progressed to the point that cost-of-living increases were granted despite the absence of any provision for such increases in the government salary structure. Nevertheless, the upsurge in inflation that year eroded real wages in the public sector as well as throughout the rest of the economy.

7. Practical Steps to Forecasting Output, Expenditure, and Prices in Sri Lanka

a. Forecasting medium-term growth

(1) Extrapolating the trend

The simplest approach to forecasting GDP is to estimate its historical growth rate and assume that the same rate will hold in the future. A variant of this approach involves disaggregating the GDP into its major sectors (Table 4.1); the projection is then essentially a weighted sum of the sectoral forecasts derived from the historical growth rates for each of the sectors.

Table 4.1.

Sri Lanka: Gross Domestic Product by Industrial Origin at Constant Prices, 1986–91

(In millions of Sri Lanka rupees at 1982 factor cost)

article image
Source: IMF Institute database.


Including forestry and fishing.

Tea, rubber, and coconuts.

The historical growth rates used in the extrapolation method can be estimated using either a regression or the compound growth formula. The latter estimates an average annual growth rate (r) from the data for the first and the last year of the period under review and can be written as follows:26



GDPRt = GDP at constant prices at time t;

r = constant rate of growth; and

t = time(0, 1, 2, …).

Thus, the average annual growth rate for the period under review would be


Note that if r were expressed in percent, the expression on the right-hand side of equation 4.11 would need to be multiplied by 100.

Alternatively, a regression equation can be fitted to find the growth rate over time, as follows:



T = time (0, 1, 2, …); and

ln = natural logarithm.

(2)The “trend through peaks” method

The trend through peaks method of estimating capacity output over time makes it possible to identify years when an economy has operated below capacity and to measure the extent of capacity utilization. The explicit assumptions are that past peak levels of output—which should be identifiable—represent potential output, since resources were being utilized at full capacity, and that potential output increases at a constant rate between two peaks. The implicit assumption is that economic policies will be sufficient to support this growth rate. The growth target is then derived by extrapolating the trend through the last two peaks into the forecasting period. This estimate may be adjusted for exceptional factors, including an acceleration in the pace of development.27

(3) Sectoral projections

When detailed information is available on likely developments in individual sectors, projecting each sector individually and combining the separate estimates provides the most accurate projection of aggregate output.28 For example, if it is known that certain new hotels will be operating in the target year, this information can be utilized in projections for the tourism sector. Because this methodology provides increasingly accurate projections as more information becomes available, it is significantly superior to econometric techniques that cannot incorporate revised data directly from updated estimates during the program year.

b. Forecasting short-term growth

In forecasting short-term output (that is, for one year), the objective is to project actual rather than potential output. We are interested both in determining the extent to which capacity will be fully utilized and in turning points or cycles, so trend projection methods are not particularly useful.

One way to start, then, is to analyze leading economic indicators that can predict changes in economic activity—for example, changes in orders, building permits, inventories, stock prices, profit margins, cash flows, capacity utilization, and the number of new enterprises. Because such data are not available in countries such as Sri Lanka, IMF missions have developed an alternative starting point, contacting representatives from different economic sectors who can provide forecasts of real sectoral growth in the coming year. Sri Lanka’s economy has five key sectors: agriculture, mining and quarrying, manufacturing, construction, and services (Table 4.1). Assume that, in contacts with sectoral representatives, the following projections are made: agriculture will grow by 3–5 percent, mining and quarrying by 5–6 percent, manufacturing by 7–9 percent, construction by 2–3 percent, and services by 3–5 percent. Assuming that policies remain unchanged, a weighted sum of the various components will produce a forecast for real GDP at factor cost.

The next step is to project the GDP deflator (see section 7c). This forecast can then be used to project nominal GDP. Multiplying the GDP deflator by the forecasted value of real GDP provides an estimate of nominal GDP at factor cost. Adding a forecast for indirect taxes less subsidies provides an estimate of nominal GDP at market prices.

These projections should then be cross-checked using a projection of GDP based on the expenditure approach. The forecast for government consumption and investment expenditure can be obtained from the budget forecasting workshop; real magnitudes are derived using deflators for government consumption and investment. Projections for exports and imports of goods and nonfactor services can be taken from the balance of payments forecasting workshop. Forecasts for export and import unit value changes can be used to deflate the nominal values.

At this point, only private consumption and private investment remain, and both can be derived as residuals. Equation 4.14, which has been estimated for real private consumption, can be used to make the consumption forecasts. This projection may need to be modified in light of the likely effect of changes in both inflationary expectations and confidence in the political and economic outlook—factors that have played a key role in influencing consumption in Sri Lanka. The deflator forecasts for the various expenditure components must be consistent with the aggregate GDP deflator forecast.

The standard optimizing investment model is not easily applied to Sri Lanka, owing to the rudimentary nature of financial markets, the large role played by government investment, other market imperfections, and difficulties obtaining information. The nature and role of public sector investments need to be carefully considered in order to prevent real crowding out. Substantial public sector investments have been made, especially in the post liberalization era following 1978, with a view to providing the infrastructure and support services essential to private sector development. Consequently, public investment may have complemented rather than competed with private investment during this period. However, continued recourse to the banking system to finance budget deficits suggests that credit to the private sector has been financially crowded out. Thus, the level of bank credit, real interest rates, and retained profits are important determinants of private investment in the country.

For the reasons given above, forecasting private sector investment is not straightforward and often requires making subjective judgments. In this regard, several factors that have significantly affected the level of private investment in recent years should be taken into account, especially national security. Ethnic conflicts and civil disturbances have dampened the domestic private sector’s incentive to invest. The surge of direct foreign investment that followed liberalization in 1977 has also been severely curtailed by the continuing instability in the country. In addition, concerns about the adequacy, credibility, and durability of the economic reforms have had a negative influence on private investment.

Finally, the ICOR defined in equation 4.9 can play an important role in projecting investment. One possibility is to use it as a device for projecting investment based on estimates of recent ICORs. Alternatively, it can be used as a consistency or “reality” check to see whether the estimates yielded by another projection technique appear plausible relative to past values and current information about the state of the economy.

c. Projecting prices

The key inflation variables to forecast in a financial programming exercise are the GDP deflator and the consumer price index. We can forecast the GDP deflator using the main factors that affect it: money growth, fiscal deficits, capacity utilization, wage growth, inflation in import prices, and expectations. A disaggregated approach that forecasts the components making up the GDP deflator is perhaps the most useful, and we can begin by forecasting consumer prices from either the supply or demand side.

One approach to forecasting consumer price inflation involves preparing an estimate based on available cost information. In the absence of any new developments, a convenient starting point is the assumption that inflation in the forecasting period remains unchanged from that of the preceding period (either the last year or the annualized rate for the last 3–6 months). The costs that should be reviewed for their possible impact on inflation in the forecast year include changes in the rate of inflation for imports and wages and adjustments that can affect disposable income, such as discretionary tax increases or curtailed subsidies. Box 4.5 outlines this procedure, while Box 4.6 shows the details of the individual calculations. That part of any increase resulting from changes in import prices, for example, is calculated as the difference between import price increases for 1990 and those for 1989 (25.76–15.44), weighted by the share of imports of goods and nonfactor services in GDP (0.382). Assuming that the ratio of imports to consumption is the same as that for GDP, the effect on the CPI is 3.9 percent. Wages in Sri Lanka constitute about 40 percent of the cost of production; therefore, 0.4 × [–13.8 – (0)] gives us the reduction in the rate of consumer price increase in 1990, or 5.5 percent.

In 1990, a reduction in subsidies on certain imported basic commodities, including wheat, sugar, and kerosene, resulted in upward adjustments in the administered prices of these goods. Because of their relatively heavy weight in the CPI, the consequence was an increase in the inflation rate of about 5 percentage points. Following these adjustments, which in the past had occurred infrequently, the prices of these items moved close to free market levels.

Illustrative Estimates of Inflation, 1990 (Period Average)

(In percent)

article image

This approach overlooks several aspects of the inflationary process, most notably the interaction of prices and wage pressures and the entire role of demand in determining prices. One of the fundamental goals of the financial programming exercise is to relate demand to supply, something that is ignored in this approach. Another problem with the cost-based technique is that the lagged effect of cost increases in prior years can produce secondary inflation.

To rectify these shortcomings, inflation forecasts derived from a cost perspective can be modified to take into account the influence of demand factors and the “ripple effect” of cost changes in previous years. To accommodate demand, the forecast for actual real GDP is compared with the projection for potential GDP. A smaller real GDP and a wider gap between the two measurements point to a reduction in inflationary pressures. If the gap has narrowed, however, inflationary pressures have increased.

In addition, among demand factors, monetary expansion appears to have been important in Sri Lanka. During 1978–90, an often rapid increase in domestic credit and the incomplete sterilization of foreign capital inflows and workers’ remittances resulted in relatively rapid monetary expansion. These factors led to an increase in the stock of money and greater aggregate demand, with inflationary consequences. In terms of lagged effects, the sharp increase in consumer prices in early 1990 (Table 4.9) may reflect the sizeable depreciation of the exchange rate in late 1989. The importance of money and demand in the inflation process in Sri Lanka is reflected in the forecasting equation 4.15 for the CPI.29 This equation, which does not take into account the adjustment in administered prices, projects an increase in the CPI of nearly 14 percent in 1990 (Table 4.8). Thus, a comprehensive approach to forecasting consumer prices accurately must take into account cost-push factors, demand elements, and administrative actions.

Table 4.8.

Sri Lanka: Equations 4.13 to 4.15, Actual and Predicted Values, 1978–90

article image
Source: IMF Institute database.

GDP at constant 1982 market prices.

Private consumption at constant 1982 market prices.

Consumer price index.

Table 4.9.

Sri Lanka: Consumer Prices, 1988–90

(Percent change over corresponding period of previous year!

article image
Source: IMF Institute database.

Details of Price Calculations

(In percent, unless otherwise noted)

article image

Average percentage change in real wages (Table 4.7) for workers in agriculture, trade and industry, services, central government, and teachers.

IBRD, Report No. 10079-CE on Sri Lanka.

Table 4.2.

Sri Lanka: National Income and Product Accounts by Industrial Origin at Current Prices, 1986–91

(In millions of Sri Lanka rupees)

article image
Source: IMF Institute database.


Including forestry and fishing.

Tea, rubber, and coconuts.

Table 4.7.

Sri Lanka: Index of Real Wage Rates, 1983–90

article image
Source: IMF Institute database.

On the basis of the forecast for consumer prices and information about export and import unit values and expected changes in the exchange rate, we can derive a projection for the GDP deflator. Forecasting the government consumption deflator requires weighting the expected wage increase for government workers on the basis of the share of the government wage bill in government consumption expenditure, using the increase in the CPI as a proxy for inflation for the nonwage component of government expenditure. Both private and government investment expenditure contain a large import component. Thus, to forecast an investment deflator for Sri Lanka, we can rely on the historical correlation between increases in import prices in rupees and the cost of investment goods. To arrive at a projection of the GDP deflator, then, we add the deflators for the various components of GDP, weighted by their relative shares in real GDP.

8. Selected Statistical Estimations for Output, Expenditure, and Prices for Sri Lanka

Econometric estimations for 1978–90 that can be used to project output, expenditures, and prices for Sri Lanka are presented below.

Output (real GDP)


Private consumption (real)


Prices (CPI)



GDPR = real GDP, in millions of 1982 Sri Lanka rupees;

T = time (1977, 1978, …);

GNIR = real gross national income, in millions of 1982 Sri Lanka rupees;

CPR = real private consumption, in millions of Sri Lanka rupees;

CPI = CPI (1982 = 100);

M2AVG = average of end-of-year broad money,(M2t-1 + M2t)/2;

MUV = index of import unit value in SDRs (1985 = 100);

ENSA = exchange rate, Sri Lanka rupees per SDR, annual average; and

LN = logarithm operator (natural log).

9. Exercises and Issues for Discussion

a. Exercises

The projections of output, expenditure, and prices are developed in two stages. Structural problems and policy responses can be identified in tandem with these projections.

  • (1) In the first stage, you will produce preliminary working estimates for real GDP growth, the GDP deflator, and the CPI. These estimates will be used as assumptions in the subsequent sectoral workshops on the balance of payments, the budget, and the monetary sector. Base the real growth assumption on both trends exhibited in Table 4.1 and on your assessment of the sustainability of the 1990 growth rate, taking into consideration:

    • recent trends in output growth and special short-term factors relevant to Sri Lanka;

    • factors affecting medium- to long-term growth prospects;

    • trade-offs between growth and social policy objectives; and

    • the general stance of economic policies in recent years.

    The econometric equation for consumer prices (4.15) can be explored after you complete other workshops in Chapters V to VII. At this stage, project the CPI on the basis of estimated cost pressures (as in Box 4.5), or on the trend indicated in Table 4.9.

    Base your estimate of the GDP deflator on available price information for key components of GDP.

    Given the assumptions for growth and the deflator, make your initial projection of total nominal GDP in 1991. A final projection of nominal GDP by expenditure category (Table 4.4) cannot be prepared until forecasts in the other macroeconomic sectors are available (with the exception of preliminary estimates of total consumption and investment).

Table 4.4.

Sri Lanka: Gross Domestic Product by Expenditure, 1986–91

(At current market prices)

article image
Source: IMF Institute database.


Includes autonomous public corporations and the Mahaweli Authority.

Central government capital expenditure less capital transfers to corporations, plus local government investment.

Using equation 4.14 to project private consumption requires a forecast of gross national income. For a preliminary estimate, you can assume that net factor payments abroad and unrequited transfers (in SDRs) are unchanged from 1990. You will need an initial working assumption for the exchange rate: for instance, either the rupee/SDR rate does not change from the end of 1990, or the currency depreciates at the same average rate as in 1986–90.

Project total investment using the ICOR (as in equation 4.9) or other factors you judge to be essential.

Total consumption plus total investment constitutes total domestic demand (TDD). Compare total domestic demand derived from the above steps—making an assumption for general government consumption—with GDP. How does the ratio TDD/GDP in 1991 compare with previous years?

  • (2) Identify the major structural problems and issues pertaining to output, expenditure, and prices. List these, together with possible policy responses, in Table 4.10.

  • (3) After the other sectoral workshops have been completed, it will be necessary to review the forecasts made in this workshop in order to complete the baseline scenario. To this end, in the second stage you will make forecasts of output, expenditure, and prices (including net exports of goods and nonfactor services from the balance of payments forecast) that are as precise as possible. Only at this stage can Tables 4.1 to 4.4 be completed in detail.

Table 4.10.

Sri Lanka: Problems of and Possible Policy Responses to Stabilization and Structural Adjustment (Real Sector)

article image

b. Issues for discussion

  • (1) Consider how the projections of output and prices are affected by the following factors:

    • poor weather conditions;

    • a depreciation of the nominal exchange rate;

    • changes in the external environment (such as the outcome of GATT trade negotiations and changes in world energy prices); and

    • an easing of monetary and fiscal policies.

  • (2) Discuss the relationship between export growth and import growth and the rate of growth of output. What special factors affect Sri Lanka’s export performance?

  • (3) How do high rates of inflation and real interest rates affect output growth?

  • (4) Identify the different channels through which fiscal policy can affect output growth, both positively and negatively. How effectively have these channels operated in Sri Lanka, especially in relation to the enterprise sector?

  • (5) In your opinion, what are the major determinants of private investment in developing countries? Do real interest rates really matter in determining saving/investment levels in underdeveloped financial markets? What is the role of foreign direct investment in these countries? What role does official foreign aid play, especially in relation to the public investment program?

Table 4.3.

Sri Lanka: Gross Domestic Product by Expenditure, 1986–91

(At constant 1982 market prices)

article image
Source: IMF Institute database.


Includes autonomous public corporations and the Mahaweli Authority.

Central government capital expenditure less capital transfers to corporations, plus local government investment.

Table 4.6.

Sri Lanka: Colombo Consumer Price Index, 1970–90

(Percentage changes)

article image
Source: IMF Institute database.

See Jacques R. Artus, “Potential and Actual Output in Industrial Countries,” Finance and Development (June 1979), pp. 25–28; and Lawrence J. Christiano, “A Survey of Measures of Capacity Utilization,” IMF Staff Papers (March 1981), pp. 144–198.


The aggregate demand curve indicates a negative relationship between demand and the price level. Essentially, this relationship results from the impact of changes in the price level on the real purchasing power of money balances and other financial assets and consequent effects.


Estimates of such unemployment rates have been made by using as a benchmark a period when the economy was operating at a high level of capacity utilization with reasonably stable prices. The rate is then adjusted for changes in the composition of the labor force since that time.


However, to the extent that plantations are neglected because of policy shortcomings (such as unrealistic producer prices), correcting these distortions can increase agricultural output even in the short run.


See World Bank, World Development Report 1983 (Washington, D.C: 1983), Part II; and Ramgopal Agarwala, “Price Distortions and Growth,” Finance and Development (March 1984).


Certain types of taxation, price control, and other regulatory measures may be designed to reduce the indirect economic costs (externalities) of various private economic activities. For example, subsidies for education, health care, and mass transit, as well as taxes on road use, energy, and pollution, are often justified on these grounds.


Overvaluation often results from the maintenance over prolonged periods of fixed nominal exchange rates in the face of rates of domestic inflation that substantially exceed those in trading partner countries.


In a number of developing countries, mining is the dominant sector and is therefore in a position similar to that of agriculture in other economies. Sharp price movements and shifts in global demand for metals and minerals contribute to the volatility of this sector.


The amount of direct government influence over expenditure by nonfinancial public enterprises varies from country to country as well as over time within one country, and the decision to include these expenditures either with the private or the government sector depends on individual circumstances. One practical factor that is important in such decisions is the availability of data.


This identity holds in both current and constant price terms.


See U Tun Wai and Chorng-Huey Wong, “Determinants of Private Investment in Developing Countries,” The Journal of Development Studies (October 1982), pp. 19–36; and Mario I. Blejer and Mohsin S. Khan, “Government Policy and Private Investment in Developing Countries,” IMF Staff Papers (June 1984), pp. 379–403.


See S. Khalid, “Determination of Private Investment in Pakistan,” IMF Working Paper WP/93/30 (Washington, D.C: 1993); Joshua Greene and Delano Villanueva, “Private Investment in Developing Countries: An Empirical Analysis,” IMF Staff Papers (March 1991); and Wai and Wong, “Determinants of Private Investment.”


Based on past annual growth rates and World Bank estimates, per capita GDP in Sri Lanka in 1964, adjusted for purchasing power parity (PPP), was $950 (measured in 1990 U.S. dollars), as opposed to $810 in Korea, $860 in Thailand, and $870 in Pakistan.


IMF, World Economic Outlook (Washington, D.C.: 1993).


Real interest rates in Sri Lanka were substantially negative until the mid-1980s.


The government has implemented a variety of large-scale welfare programs during the past 30 years. The most recent, the Jana Saviya Program, has been in place since 1989.


Foreign direct investment amounted to no more than 1 percent of GDP in both Korea and Thailand during 1971–80.


See World Bank, World Development Report (Washington, D.C: 1991).


All successful Asian countries, including Korea, Malaysia, and Thailand, have been highly receptive to foreign technology. Singapore is a particularly good example, especially during the early period of its economic development.


Sri Lanka has long had a relatively sophisticated and diverse financial system. As well as a stock market, the country has a well-established banking system that includes the Central Bank, 2 state-owned and 22 private commercial banks, the NSB, and over a dozen other specialized and regional financial institutions. There are also other financial institutions, such as pension funds and insurance companies, that mobilize long-term resources.


The market share of private domestic banks has risen from 10 percent to 16 percent in recent years, mainly at the expense of foreign banks.


The NSB is 100 percent owned by the Sri Lanka Government and operates under the National Savings Bank Act of 1971. While the NSB has recently made efforts to expand into rural lending activity, its two main functions are to mobilize saving in urban and rural areas and utilize that saving to finance the government deficit. The NSB’s share in total saving has hovered between 32 percent and 40 percent in the past few years, and over 90 percent of its deposits are invested in government securities. Because of its special role in financing government deficits, the NSB has received various government subsidies, including tax exemptions on interest income.


The so-called “construction boom” in Sri Lanka took place immediately after the liberalization of the economy in 1978 and the launching of such gigantic projects as the Mahaweli Development scheme for irrigation and hydroelectric power generation. However, sizable construction work on the Mahaweli project continued during 1986–90, especially the Samanala Wewa project and downstream developments. In addition, construction work on the Million Houses Programme (MHP) initiated in 1984 continued during this period.


Prior to 1982, tourism was one of the most dynamic sectors of the Sri Lankan economy, recording substantial gains in foreign exchange earnings. After the tea sector, it was the second largest foreign exchange earner in 1982.


During 1989–90, the government eliminated price controls on rice, fertilizer, sugar, and several other products regulated by the Fair Trading Commission. However, its direct influence on the general pricing policies of state enterprises (including rail and road transportation) and on petroleum products and food imports (through the state-run Cooperative Wholesale Establishment) has continued.


If the first and the last observations are out of line with other observations, the compound growth method may yield misleading results. One way to remedy this problem is to substitute two-year averages for the first and last observations; another is to delete “abnormal” years from the series.


See Mohsin S. Khan and Malcolm D. Knight, “Fund-Supported Adjustment Programs and Economic Growth,” Occasional Paper 41 (Washington, D.C.: IMF, 1985) for an approach that ties the growth rate of output to the ratio of real investment to real income, thus avoiding the problem of estimating capital stock in developing countries.


An overall growth rate, R, can be calculated as the weighted average of the sectoral growth rates:R=i=1nwiri,

where wi and rj are the sectoral weights and growth rates, respectively.


The relationship between prices and monetary aggregates may be reinforced to the extent that the velocity of circulation of money increases as the rate of inflation accelerates.