Hidden subsidies.

Pasha, Hafiz A. ; Ghaus-pasha, Aisha ; Aamir, Naveed 等

The objective of this paper is to quantify the magnitude of subsidies provided from budgetary sources for major economic and social services including irrigation, roads, education, health, and water supply and sanitation. This exercise is complicated by the limited availability of data, especially on costs of provision. The research demonstrates that the budgetary subsidy on major economic and social services in Pakistan is large and there is less of it for 'merit goods'. Therefore, considerable scope exists for improving cost recovery and raising the level of non-tax revenues, with much of the benefit going to provincial governments. It needs to be recognised that resource mobilisation efforts, which have hitherto been concentrated in the taxation area, need to focus more on non-taxes, by raising charges on services which are not in the nature of 'merit goods' if the overall equity consequences of public budgeting are to be enhanced.

INTRODUCTION

Many governments use price subsidisation (total costs less total revenues from user charges) to meet social protection objectives in lieu of, or in addition to, direct income transfers. Such subsidies may be perceived as influencing behaviour to further other socially desirable policies. For example, the price response induced by lowering the price of schooling will both lower the cost of living for the beneficiaries and also increase the investment in education more than a similar income transfer would achieve.

The incidences of benefits from a general price subsidy are proportional to purchases and can be deduced from the pattern of expenditures. Some goods are inappropriate vehicles for redistribution since subsidies on them will not only accrue mainly to the rich they will actually increase inequality in welfare. It is therefore important to ensure that commodities chosen for price subsidisation are largely consumed by the lower income groups. Also, detailed data on such commodities should be made public to make the extent of subsidy easily tractable.

In the case of Pakistan, the problem of lack of transparency of federal and provincial budgets is vividly demonstrated by the inability of such budgets to readily highlight the subsidy on the various economic and social services, which are essentially in the nature of 'private' goods, provided by such governments. This is not only a reflection of the problem of the nature of budgeting practices whereby, first, revenues and expenditures on different heads are shown separately and no account is made either of depreciation of assets or the costs of capital used to finance the acquisition of assets which yield a stream of services. Second, to the extent that the subsidies largely benefit the upper income groups, political compulsions dictate that such subsidies largely remain hidden.

It is clear, therefore, that one of the major areas of reform of public expenditure must be first to reveal the magnitude of service-related subsidies for subsequent scrutiny by public opinion and then to implement a policy of enhancing and improving the efficiency of collection of user charges to minimise the subsidy on services which are consumed by the section of the population that has greater ability to pay. The higher revenues generated can then be used to bring down the fiscal deficit or to cross-subsidise the provision of services which are pro-poor.

The objective of this paper is to quantify the magnitude of subsidies provided from budgetary sources on major economic and social services including irrigation, roads, education, health and water supply and sanitation. This exercise is complicated by the limited availability of data, especially on costs of provision. Therefore, various approaches are adopted to estimate the subsidies. It needs to be emphasised that in some cases the resulting magnitudes must be considered as approximate in character.

The paper is organised as follows: Section I presents the general methodology for deriving the magnitude of subsidy on a particular service. Sections II to VI describe the procedure used for estimation of costs of provision and the magnitude of the subsidy on irrigation, roads, education, health, water supply and sanitation respectively. Section VII brings all the individual service subsidies together to obtain an estimate of the overall subsidy bill. Finally, in Section VIII are highlighted the key policy implications of the findings.

I. METHODOLOGY

The costs of provision of a particular service have two components-----capital and recurrent costs. The former consists of depreciation of assets, which provide the service and the interest costs on the capital used to acquire the assets. The latter are composed of the salary costs of employees engaged in the delivery of the service and the overhead staff plus the operations and maintenance costs.

Therefore, the total costs are derived as:

total costs=depreciation + interest costs + salary bill + operations and maintenance costs .... .... .... ... (1)

Then, the subsidy can be computed as:

subsidy = total costs--total revenues from user charges ... ... (2)

The analysis in the paper has been conducted throughout at constant prices (of 1997-98). The base year corresponds to the latest year for which more reliable estimates are available of expenditures. The real rate of interest is computed as the nominal rate of interest minus the rate of inflation. Since development expenditure is generally financed by borrowing, the former is taken to correspond to the average cost of domestic borrowing, which has averaged about 9 percent. The long term rate of inflation of Pakistan is about 7 percent. Therefore, the real rate of interest is taken at 2 percent. However, given the annual variation in these rates, we generate a range of estimates. The first can be referred to as a relatively high estimate in which the real rate of interest is taken as 2 percent. In the second estimate, interest costs are ignored.

We turn now to the estimation of the subsidy on individual services.

II. IRRIGATION

This is by far the most important publicly provided service, benefiting the agricultural sector of Pakistan. Estimation of the capital stock embodied in the irrigation infrastructure is a complex exercise. Irrigation systems (including the major dams and water storage reservoirs) are developed mostly by the Water and Power Development Authority (WAPDA) and then handed over for operations and maintenance to the respective provincial governments. WAPDA's annual investments are not apportioned into the water and power components respectively. Therefore, annual outlays for developing the irrigation system are not available directly.

Fortunately, the Federal Bureau of Statistics gives sectoral estimates of the gross and net national product. In the agricultural sector, the difference corresponds to the depreciation of the capital stock in the sector. In agriculture, bulk of the capital stock consists of the irrigation system, tractors, tubewells and agricultural implements. 80 percent of the depreciation is assumed to take place in the irrigation system. Recurrent costs correspond to the current expenditures of the irrigation departments of the provincial governments, which have been extracted from the four provincial budget documents and aggregated to arrive at the national estimate. Similarly, revenues from irrigation charges (abiana) have also been obtained from those documents.

Resulting estimates of the irrigation subsidy are presented in Table 1. This table reveals that depreciation has increased rapidly in the 90s as the wear and tear on assets created in the 60s and 70s has risen. There is a, more or less, corresponding increase in interest costs. Recurrent costs, which rose rapidly in the 80s, have been held down in the 90s probably by largely postponing operations and maintenance expenditures. In real terms, revenue collections have actually declined, especially in the 90s.

The overall subsidy on irrigation is estimated at between Rs 28 billion and Rs 36 billion in 1997-98. In real terms, it has increased annually during the 25 year period, 1972-73 to 1997-98, by 2 to 3 percent.

III. ROADS

For roads a different methodology has been adopted for estimation of the capital stock. For this sector, information is available directly from the federal and provincial budget documents on development expenditures (corresponding to the investment). An investment deflator has been used to generate the investment, [I.sub.t], series at constant prices of 1997-98. Based on this, the capital stock, [K.sub.t], has been derived as follows:

[K.sub.t] = [K.sub.t] - 1(1 - [delta]) + [I.sub.t] ... ... ... ... ... (3)

Therefore,

depreciation = [K.sub.t - i]. [sup.[delta]] ... ... ... ... (4)

The rate of depreciation, [delta], has been taken as 2.5 percent annually.

Beyond this, the recurring expenditure and revenues (in the form of tolls and other charges) have been obtained directly from the budget documents. It needs to be emphasised that only revenues from user charges have been included. Revenues in the form of taxes on petroleum products consumed by transport vehicles have not been considered in the analysis.

Estimates of the overall national subsidy on roads and bridges are presented in Table 2. The major component of costs is recurrent costs, especially on operations and maintenance. In 1997-98, recurrent costs accounted for 36 percent of total costs (inclusive of interest costs). Revenues from tolls and other levies are extremely low, at below Rs 1 billion. Consequently, the subsidy bill is high, ranging from Rs 16 to Rs 25 billion in 1997-98. It has also grown rapidly, at almost 4 percent per annum.

IV. EDUCATION

Education financing is a shared responsibility of the federal and provincial governments. The former makes grants primarily to universities (through the University Grants Commission) while recurrent and capital costs of primary, secondary and college education are financed out of provincial budgets. We analyse the subsidy at each level of education.

Primary Education

The major involvement of governments in education is in the area of primary education. Since this is pre-dominantly a 'merit good' catering mostly to poor households, there is a strong case for subsidising this level of education. It is not surprising, therefore, that the subsidy bill for primary education is large. Table 3 gives estimates of the government subsidy on primary education.

Depreciation has been estimated from the series of capital stock, by application of Equation (3). However, bulk of the costs are recurring in nature, consisting primarily of teachers' salaries and allowances. Recurrent costs have shown fast growth after the mid-80s, following the launching initially of the Five Point Programme and then the Social Action Programme. These programmes have led to a major expansion in the number of primary schools and, consequently, in school teachers. As expected, since primary education is essentially free, revenues are negligible.

Table 3 reveals that in 1997-98 the subsidy on primary education had approached Rs 30 billion. In real terms it has demonstrated very rapid growth at almost 8 percent per annum.

Secondary Education

The same methodology has been used to estimate the public subsidy on secondary education. The derived magnitudes are presented in Table 4. Here again, recurrent costs dominate, showing fast growth upto the early 80s and then some moderation as resources got diverted to primary education. Total subsidy on this level of education is estimated at about Rs 14 billion, with a relatively high growth rate in real terms of almost 8 percent.

Higher Education

During the 70s and 80s the higher education sector expanded rapidly. Many new universities and colleges were established. This was an expression not only of the rising political influence of the urban middle class but in response also to the growing demand for educated manpower caused by the emergence of employment opportunities in the Middle East and in the expanding public sector of Pakistan. During the 90s, development expenditure on higher education has fallen sharply. Consequently, the capital stock embodied in higher education institutions, more or less, doubled by the end of the decade of the 80s. But since then it has actually tended to decline, implying that the level of investment is not even adequate for replacement purposes.

Table 5 presents estimates of the subsidy on higher education. The subsidy may be somewhat overstated because, especially in the case of universities, revenues from fees and other charges are retained and not reverted back to the provincial budgets. But these collections are limited and unlikely to make a major difference to the quantum of subsidy.

The table reveals that the subsidy in real terms has actually started falling after 1987-88. For the year, 1997-98, it is estimated at between Rs 10 billion and Rs 11 billion. For the twenty five year period as a whole, the annual growth rate is estimated at about 5 percent.

V. CURATIVE HEALTH

Curative health services provided primarily by provincial governments consist of the network of basic health units and rural health centres in the rural areas and hospitals and clinics in the major towns and cities. Development expenditures on these services peaked in the second half of the 80s during the tenure of the Five Point Programme. Studies reveal that the major portion of beneficiaries from government curative health services belong to relatively poor households, while private facilities essentially cater for the needs of upper income households.

The capital stock in public curative health facilities has been derived from the stream of development expenditures, by application of Equation (3). The value of this capital stock exceeded Rs 100 billion in 1997-98. Therefore, the depreciation and interest costs are significant, although recurrent costs are the single largest item of cost in this service also.

Table 6 derives the subsidy bill on curative health. Costs of provision increased rapidly upto 1987-88. Revenues are marginal, due to the low level of fees charged. Altogether, the subsidy bill is estimated at between Rs 15 billion and Rs 17 billion in 1997-98, with an annual growth rate in real terms of about 7 percent.

VI. WATER SUPPLY AND SANITATION

The public health engineering departments of the provincial governments supply drinking water, primarily to rural households, by implementing schemes generally with piped connections and levy fixed charges for this service. The basic issue is whether these charges recover the costs of provision.

Table 7 gives the costs of provision, revenue from tariffs and the resulting subsidy. The major component of costs in this service are capital costs. These have peaked during the 90s, because of the high priority attached by the Social Action Programme to the expansion of rural water supply and sanitation coverage.

The table reveals the disappointingly low level of revenues. Consequently, bulk of the costs have to be subsidised. The subsidy is estimated at between Rs 4 billion and Rs 7 billion in 1997-98, with an annual growth rate in real terms of about 5 percent.

VII. OVERALL BUDGETARY SUBSIDY

We are now in a position to quantify the overall subsidy on major services, economic and social, which is financed by budgetary resources of the federal and provincial governments. Estimates are presented in Table 8. The high estimate includes interest cost.

The analysis reveals that the aggregate 'hidden' subsidy was between Rs 115 to Rs 140 billion in 1997-98. The latter is as much as 5 percent of the GDP. The subsidy bill on 'merit goods' (that is, services which are basic and pro-poor in nature), which includes primary education, curative health and drinking water supply, is Rs 53 billion, representing 38 percent of the aggregate subsidy bill. Therefore, the remainder, almost Rs 87 billion in unlikely to be justified on equity grounds.

However, two positive trends are visible. The share in total subsidy of 'merit goods' has almost doubled in the last twenty five years, from about 20 percent in 1972-73. Second, the aggregate subsidy bill has declined somewhat from over 6 percent of the GDP in the early 70s to about 5 percent of the GDP currently.

There exists a strong case for enhancing user charges for services that are not in the nature of 'merit goods'. Initially if efforts are made to recover recurrent costs only on such services (irrigation, roads, secondary and higher education) then potentially additional revenues equivalent to over 1 percent of the GDP can be generated. In the long run, if full cost recovery is targeted for then the contribution to revenues could be as much as 3 percent of the GDP. Provincial governments' fiscal position, in particular, could improve substantially as most of the additional revenue would accrue in the form of non-tax revenues to these governments. This would enable not only a significant reduction in the overall fiscal deficit but would also increase the availability of resources for financing the expanded and improved provision of pro-poor services.

VIII. POLICY IMPLICATIONS

In conclusion, the research demonstrates that the budgetary subsidy on major economic and social services is large in Pakistan at about 5 percent of the GDP. Less than 40 percent of this subsidy goes to 'merit goods'. Therefore, considerable scope exists for improving cost recovery and raising the level of non-tax revenues, with much of the benefit going to provincial governments. It needs to be recognised that resource mobilisation efforts, which have hitherto been concentrated in the taxation area, need to focus more on non-taxes, by raising charges on services which are not in the nature of 'merit goods', if the overall equity consequences of public budgeting are to be enhanced.

Comments

The objective of the paper is to quantify the magnitude of subsidies provided from budgetary sources on major economic and social services including irrigation, roads, education, health, potable water supply and sanitation. It is difficult to find data to carry out such studies, since hardly any data is available on the subject matter. Thus, authors must have made tremendous efforts to compile the data and carry out the study. Besides, there is limited literature on the subject matter which may have analysed the matter in so depth. In this respect, the efforts of the authors need to be appreciated. Irrespective of the quality of quantification of the results of the study, the authors have contributed to initiate debate on Varity of subsidies in a broader prospect. Thus, the paper will initiate debate to rationalise subsidies, which have been granted in the past without economic rationale.

Notwithstanding the above, the study creates an impression that huge subsidy is provided for the sectors mentioned above. It is claimed that removal of such subsidies could provide a good support to solve budget deficit problem and therefore it may also help to resolve other related budgetary issues. However, considering the definition and methodology used by the authors to derive the results, it appears that there is a need to pay due attention to properly define subsidies. No rational has been provided therein for the definition used to calculate the results. The study suffers from such definitional and methodological limitations. It could be improved further if the following comments are given due importance to revise the study.

1. There is a need to have some space allocated to review the literature on the subject matter so that subsidy is properly understood and it is not mixed up with state welfare works and services. It will help to analyse the issue within the domain of debate on subsidies. There is not a single study referred there in which may highlight the debate on subsidies. Thus the importance of the issue is not established. Besides, the review of literature is totally neglected which has led to create ambiguous definition of subsidies.

2. A major problem with the paper is that it has no theoretical foundation spelled out therein. If the theory of subsidies is made a part of the paper, it will help to address the issue in a proper framework. Thus, many social services which have been treated as subsidies may be eliminated. As a result, the claims and quantification may become insignificant. Surprisingly, no support from theory or literature has been derived. I think public goods and public services which are responsibility of the state have been mixed up with private goods for example, roads, basic health, primary education, potable water and sanitation expenditure etc., which may not fall in the range of the definition of subsidies. These expenditures are state welfare expenditures and need not be treated as subsidies. The provision of such services is a responsibility of the state under constitution. How these can be treated subsidies on private goods is not understandable.

3. Moreover, while discussing the results, the authors themselves ignore the figures for expenditures on public services. If so, then why these estimates are a part of the study anyway?

4. For a proper identification of distortions like subsidies, there is also a need to look at other distortions in the market. Take an example of irrigation, for sure; cost of the same may be greater than revenue recovery from irrigation. But, what about the controlled prices of agricultural products (state monoposony), like cotton, rice and other cash crops. State corporations bought these crops at half prices and then exported these commodities at double prices. The difference (surplus) is taken away by the state. It had been the practice of Cotton Trading Corporate and Rice Export Corporation. By taking care of these aspects, do the subsidies still exceed the cost?, or revenue collected from the sector. Such distortions have not been a part of the methodology. Thus, the methodology use for calculation of subsidies can hardly be justified. Therefore, no attempt is made to justify these aspects to reach to net results of subsidies. Besides, no rationale for the formula used to calculate subsidies is provided. Such neglects led to overestimate the volume of subsidies.

5. On the one hand, the authors claim that objective of the paper is to quantify the volume of subsidies. At the end of the paper, they claim that Rs 87 billion out of (Rs 115-140 billion) is unlikely to be justified as subsidies on equity grounds. All ends up with net subsidies of 1 percent of GDP, which the authors claim that it can be raised by elimination of subsidies. If the figures are analysed by separating public services and net subsidies, the figures may be insignificant. Besides, keeping in view wide spread of poverty, the position of small farmers and over 65 percent rural population deprived of basic public services, hardly any claim of the authors is justified. Thus, there is a need to properly review the literature, justification of methodology need to be spelled out and there is also a need to capture price distortions, hidden taxes on agriculture and separate public services from subsidies.

M. Aslam Chaudhary

Quaid-i-Azam University, Islamabad.

Hafiz A. Pasha, Aisha Ghaus-Pasha and Naveed Aamir are former Managing Director, Senior Technical Adviser, and Research Officer, respectively, at the Social Policy and Development Centre, Karachi.

Table 1

Irrigation Subsidy (Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 12.2 9.8 4.3
1977-78 11.5 9.2 3.7
1982-83 13.5 10.8 5.9
1987-88 14.8 11.8 7.9
1992-93 16.5 13.2 6.3
1997-98 23.5 18.0 6.9
ACGR (%) -- -- --

 Total Costs Subsidy

Year Low High Revenues Low High

1972-73 16.5 26.3 3.8 12.3 22.0
1977-78 15.2 24.4 2.8 12.4 21.0
1982-83 19.4 30.2 3.6 15.8 26.0
1987-88 22.7 34.5 3.1 19.6 31.1
1992-93 22.8 36.0 2.0 20.8 34.0
1997-98 30.4 38.4 2.7 27.7 35.7
ACGR (%) -- -- -- 3.3 1.9

Source: Pakistan Economic Survey.

Federal and Provincial Budget Documents.

Table 2

Subsidy on Roads (Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 2.2 1.8 4.4
1977-78 3.3 2.6 2.1
1982-83 4.1 3.3 5.7
1987-88 4.8 3.8 7.8
1992-93 6.0 4.8 8.5
01 1997-98 7.5 6.0 9.3
ACGR (%) -- -- --

 Total Costs Subsidy

Year Low High Revenues Low High

1972-73 6.6 11.0 0.3 63.0 10.7
1977-78 5.4 7.5 0.3 5.1 7.2
1982-83 9.8 15.7 0.3 9.5 15.4
1987-88 13.6 21.4 0.5 13.1 20.9
1992-93 14.5 23.0 0.6 13.9 22.4
01 1997-98 16.8 26.1 0.7 16.1 25.4
ACGR (%) -- -- -- 3.8 3.5

Source: Pakistan Economic Survey.

Federal and Provincial Budget Documents.

Table 3

Subsidy on Primary Education (Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 0.2 0.2 4.6
1977-78 0.2 0.2 4.5
1982-83 0.4 0.3 8.1
1987-88 0.7 0.6 15.4
1992-93 0.9 0.7 19.4
1997-98 1.3 1.0 27.5
ACGR (%) -- -- --

 Total Costs Subsidy

Year Low High Revenues Low High

1972-73 4.8 5.0 0.2 4.6 4.8
1977-78 4.7 4.9 0.0 4.7 4.9
1982-83 8.5 8.8 0.0 8.5 8.8
1987-88 16.1 6.7 0.2 15.9 16.5
1992-93 20.3 21.0 0.2 20.1 20.8
1997-98 28.8 29.8 0.3 28.5 29.5
ACGR (%) -- -- -- 7.6 7.5

Source: Pakistan Economic Survey.
Federal and Provincial Budget Documents.

Table 4

Subsidy on Secondary Education (Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 0.2 0.2 1.8
1977-78 0.3 0.3 2.9
1982-83 0.4 0.3 5.0
1987-88 0.7 0.6 10.0
1992-93 0.8 0.6 10.0
1997-98 0.9 0.7 13.0
ACGR (%) -- -- --

 Total Costs Subsidy

Year Low High Revenues Low High

1972-73 2.0 2.2 0.1 1.9 2.1
1977-78 3.2 3.5 0.0 3.2 3.5
1982-83 5.4 5.7 0.0 5.4 5.7
1987-88 10.7 11.3 0.3 10.4 11.0
1992-93 10.8 11.4 0.3 10.5 11.1
1997-98 13.9 14.6 0.4 13.5 14.2
ACGR (%) -- -- -- 8.1 7.9

Source: Pakistan Economic Survey.
Federal and Provincial Budget Documents.

Table 5

Subsidy on Higher Education (Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 0.6 0.5 2.5
1977-78 0.8 0.6 5.9
1982-83 0.9 0.7 5.6
1987-88 1.4 1.1 9.8
1992-93 1.5 1.2 8.0
1997-98 1.5 1.2 8.5
ACGR (%) -- -- --

 Total Costs Subsidy
Year Low High Revenues Low High

1972-73 3.1 3.6 0.3 2.8 3.3
1977-78 6.7 7.3 0.2 6.5 7.1
1982-83 6.5 7.2 0.2 6.3 7.0
1987-88 11.2 12.3 0.3 10.9 12.0
1992-93 9.5 10.7 0.2 9.3 10.5
1997-98 10.0 11.2 0.4 9.6 10.8
ACGR (%) -- -- -- 5.5 4.9

Source: Pakistan Economic Survey.
Federal and Provincial Budget Documents.

Table 6

Subsidy on Curative Health (Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 0.4 0.3 2.4
1977-78 0.5 0.4 3.0
1982-83 0.9 0.7 4.6
1987-88 1.7 1.4 9.9
1992-93 2.2 1.8 10.9
1997-98 2.6 2.1 13.0
ACGR (%) -- -- --

 Total Costs Subsidy

Year Low High Revenues Low High

1972-73 2.8 3.1 0.2 2.6 2.9
1977-78 3.5 3.9 0.1 3.4 3.8
1982-83 5.5 6.2 0.2 5.3 6.0
1987-88 11.6 13.0 0.3 11.3 12.7
1992-93 13.1 14.9 0.3 12.8 14.6
1997-98 15.6 17.7 0.4 15.2 17.3
ACGR (%) -- -- -- 7.3 7.4

Source: Pakistan Economic Survey.
Federal and Provincial Budget Documents

Table 7

Subsidy on Water Supply and Sanitation
(Rs in Billion at 1997-98 Prices)

 Interest Recurrent
Year Depreciation Costs Costs

1972-73 1.0 0.8 0.4
1977-78 1.1 0.9 0.2
1982-83 1.3 1.0 0.4
1987-88 1.7 1.4 0.9
1992-93 2.1 1.7 1.6
1997-98 2.7 2.2 1.9
ACGR(%) -- -- --

 Total Costs Subsidy

Year Low High Revenues Low High

1972-73 1.4 2.2 0.1 1.3 2.1
1977-78 1.3 2.2 0.1 1.2 2.1
1982-83 1.7 2.7 0.1 1.6 2.6
1987-88 2.6 4.0 0.1 2.5 3.9
1992-93 3.7 5.4 0.1 3.6 5.3
1997-98 4.6 6.8 0.2 4.4 6.6
ACGR(%) -- -- -- 5.0 4.7

Source: Pakistan Economic Survey.
Federal and Provincial Budget Documents.

Table 8

Overall Budgetary Subsidy on Major Economic and Social Services
(Rs in Billion at 1997-98 Prices)

 Economic Services Social Services

 Irrigation Roads Primary Secondary
Year Education Education

High Estimate
 1972-73 22.5 10.7 4.8 2.1
 1977-78 21.6 7.2 4.9 3.5
 1982-83 26.6 15.4 8.8 5.7
 1987-88 31.4 20.9 16.5 11.0
 1992-93 34.0 22.4 20.8 11.1
 1997-98 35.7 25.4 29.5 14.2
 ACGR (%) -- -- -- --
Low Estimate
 1972-73 12.3 6.3 4.6 1.9
 1977-78 12.4 5.1 4.7 3.2
 1982-83 15.8 9.5 8.5 5.4
 1987-88 19.6 13.1 15.9 10.4
 1992-93 20.8 13.9 20.1 10.5
 1997-98 27.7 16.1 28.5 13.5
 ACGR (%) -- -- -- --

 Social Services
 Overall
 Higher Curative Water Budgetary
Year Education Health Supply Subsidy

High Estimate
 1972-73 3.3 2.9 2.1 48.4
 1977-78 7.1 3.8 2.l 50.2
 1982-83 7.0 6.0 2.6 72.1
 1987-88 12.0 12.7 3.9 108.4
 1992-93 10.5 14.6 5.3 118.7
 1997-98 10.8 17.3 6.6 139.5
 ACGR (%) -- -- -- 4.3
Low Estimate
 1972-73 2.8 2.6 1.3 31.8
 1977-78 6.5 3.4 1.2 36.5
 1982-83 6.3 5.3 1.6 52.4
 1987-88 10.9 11.3 2.5 83.7
 1992-93 9.3 12.8 3.6 91.0
 1997-98 9.6 15.2 4.4 115.0
 ACGR (%) -- -- -- 5.3

Source: Pakistan Economic Survey.
Federal and Provincial Budget Documents.