This Selected Issues paper analyzes Pakistan's tax reform and revenue performance. The paper assesses the nature and magnitude of Pakistan's actual twin debt problem in a historical context, and reviews the policy options. The study highlights the factors explaining the recent stagnation in merchandise exports, and analyzes the country's export performance with regard to the linkages between performance and the structure of exports. The paper reviews a survey of poverty in Pakistan, and also provides a statistical appendix report of the country.


This Selected Issues paper analyzes Pakistan's tax reform and revenue performance. The paper assesses the nature and magnitude of Pakistan's actual twin debt problem in a historical context, and reviews the policy options. The study highlights the factors explaining the recent stagnation in merchandise exports, and analyzes the country's export performance with regard to the linkages between performance and the structure of exports. The paper reviews a survey of poverty in Pakistan, and also provides a statistical appendix report of the country.

I. Tax Reform and Revenue Performance1

A. Introduction and Background

1. Efforts to reduce Pakistan’s chronically large fiscal deficit tend to focus on improving revenue performance. The poor fiscal position is the achilles heel of macroeconomic stability and it severely limits the government’s ability to support poverty alleviation, sustainable growth, and economic recovery. While there is little doubt that the full arsenal of revenue augmentation, rationalization of government spending, and more efficient allocation of government resources will need to be employed for bringing about lasting betterment, highest hopes are usually placed on mobilizing additional revenue.

2. This Section tries to put the quest for additional revenue into perspective. Pakistan’s tax system has indeed undergone profound changes over the last decade. The next subsection analyzes how they have affected die composition and over all performance of revenue over time. It is followed by subsections that take a closer look at the reform and performance of the main sources of revenue: the sales tax, the income tax, customs duties, and petroleum taxes. Pakistan’s revenue performance is then contrasted with that of a sample of thirteen other countries from around the world at varying degrees of development. A final subsection summarizes the conclusions.

B. Revenue Performance in Historical Perspective

3. Notwithstanding profound changes of Pakistan’s tax system, tax revenue in relation to GDP has remained remarkably stable over the last two decades (Chart I-1 and Table I-1). It averaged 13 percent of GDP, never exceeded 14.5 percent of GDP, and never dipped below 12 percent of GDP. Although 1999/2000, the final year of the observation period, registered the weakest revenue performance, the time series does not exhibit a downward trend. However, there certainly is no upward trend either. This analysis focuses on the tax revenue of the central government, comprising five components: sales taxes, excises, direct taxes, customs duties, and surcharges on petroleum products and natural gas. The remaining sources of government revenue, not covered here, are provincial tax revenue, central government nontax revenues, and provincial government nontax revenues. They typically account for around 0.5, 2.5, and 0.5 percent of GDP, respectively.

Chart I-1.

Pakistan: Tax Revenue, 1979/80-1999/2000 1/

(In percent of GDP)

Source: Pakistani authorities, and staff estimates.1/ Revised actual data for 1993/94-1998/99 and preliminary actual data for 1999/2000. Unrevised data prior to 1993/94 have been adjusted proportionately with revisions during 1993/94-1998/99.
Table I-1.

Pakistan: Tax Revenue, 1979/80-1999/2000 1/

article image
Sources: Pakistan authorities and Fund staff estimates.

Revised actual data for 1993/94-1998/99 and preliminary data for 1999/2000. Unrevised data prior to 1993/94 were adjusted in proportion to revisions during 1993/94-1998/99.

4. The composition of tax revenue has shifted towards more reliance on sales tax and direct tax collections with the contribution of customs duties and excises shrinking. The most striking feature is the declining trend of customs duty collections, which fell more or less continuously from over 6 percent of GDP in 1989/90 to just under 2 percent of GDP in 1999/2000. While they used to account for well over 40 percent of tax revenue in the 1980s, they now contribute a mere 15 percent. Equally striking is the relatively recent surge of sales tax collections, which culminated at the equivalent of 3.7 percent of GDP in 1999/2000 compared to just 1 percent of GDP two decades earlier. Direct tax collection improved over the observation period with most of the headway occurring in the first half of the 1990s. The importance of excise duties diminished over time while surcharges remained volatile throughout.

5. While revenue stagnated relative to GDP during the observation period, substantial headway was made in increasing revenues earlier on. During the 1960s, revenue increased from about 9 percent of GDP to over 12 percent of GDP and then flattened out.2

C. Sales Tax Reform

6. Over the past ten years, Pakistan has successfully reformed the sales tax regime but important reform elements remain to be implemented. The introduction of GST in VAT mode in November 1990 together with successive broad basing and movement toward accounts-based assessment has meant that revenues increased by about 2 percentage points of GDP and that revenues are raised in a less distortionary way. Key items still on the reform agenda include: (a) extension to the retail sector; (b) firmly establishing the GST in the service sector; and (c) better enforcement in general. It will also be important to consolidate the progress already achieved and guard against policy reversals, which have frequently marked the reform process in the past.

7. The Sales Tax Act of 1990 replaced a single-stage general tax on manufacturers and importers, which was highly distortionary, difficult to administer, and unsuccessful in raising much revenue. It was beset by a large number of exemptions, cascading (despite a complex credit mechanism),3 and tax liabilities determined on the basis of notional rather than actual prices. The credit mechanism allowed for an adjustment in the amount of the sales tax embedded in the cost of materials and supplies used up in the production process. At least in theory, it required the tracing of material and supplies through the production process. Levied at a rate of 12.5 percent, it raised on average 1.2 percent of GDP in the 1980s.

8. The GST is currently levied at a single positive rate of 15 percent—exports are zero-rated—which minimizes distortions, simplifies tax administration, and helps keep compliance costs down. Although the GST was introduced at a single positive rate of 12.5 percent, multiple rates were introduced later on. The initial rate was hiked to 15 percent in July 1993. In July 1995, the standard rate was further increased to 18 percent and two additional bands at 10 and 20 percent were introduced. Yet another band at 23 percent was added in July 1996 meaning that a total of four positive rates was in effect in 1996/97. A simplified rate structure came into effect in July 1997 when the standard rate was dropped to 12.5 percent and only one nonstandard rate of 10 percent was retained. The current rate structure was adopted in December 1999.

9. While the tax base was broadened considerably over the past ten years, coverage is not yet satisfactory. The initially narrow base of the GST was a feature that carried over from the tax it replaced, under which 8 goods accounted for about 80 percent of revenues. Outright exemptions, exclusion of large parts of the retail and service sectors, and enforcement problems all impinged on coverage.

10. GST exemptions have been reduced drastically. In its early years, the GST was riddled with exemptions which resulted in all supplies being exempted except for supplies made by a manufacturer, supplies of furniture, and supplies of certain imported or locally produced consumer durables. In addition, a government order exempted a further 120 locally produced goods, including agricultural products, many raw products and semi-manufactured goods, petroleum, electricity, pharmaceuticals, fertilizers, motor vehicles, dolls, toys, etc. Major headway in broadening the tax base was made in the context of the 1994/95 budget when 266 exemptions were eliminated. Likewise, the GST extension to petroleum products, natural gas, and electricity in August 1999 constituted an important milestone. Remaining major non-standard GST exemptions relate to agricultural inputs (including fertilizers, pesticides, and animal feeds), edible oils (excluding import of palm and soybean oil), and computer hard and software.

11. In the past, the integrity of the GST was also undermined by geographical exemptions, which were granted to promote the setting up of new industries in backward regions. In the mid-1990s, a plethora of Special Industrialized Zones, the region of Gadoon Amazai, and large parts of the North West Frontier Province and the Province of Balochistan enjoyed such preferential treatment. These incentive schemes have meanwhile been grandfathered and have, for the most part, expired. However, the Provincial and Federal Administered Tribal Areas as well as the Northern Areas remain outside the coverage of the Sales Tax Act.

12. Despite several attempts, the GST has not yet been successfully extended to the retail sector. Mindful of administrative constraints, collection of the GST was initially deliberately confined to the manufacturing and import stages, although this entailed structural weaknesses and a considerable revenue loss. As experience with the new tax grew, the extension to the retail stage was seriously contemplated from the mid-1990s. However, all moves in this direction were met with fierce taxpayer resistance. In the event, the government resorted to imposing a turnover tax of 3 percent in the retail sector in June 1997, although this was not much of an improvement from a structural point of view. In the face of renewed taxpayer resistance and implementation problems, the government agreed in April 1998 to replace the turnover tax by a system of fixed fees, so-called trade enrollment certificates, to be administered by the retail associations themselves. The new arrangement yielded little revenue and was abandoned soon afterwards. Although the GST was legally extended to the retail sector in July 1998, enforcement did not take place in earnest. In May 2000, the government started a new attempt to bring the retail sector into the tax net. Enforcement was stepped up in the context of a tax registration drive and at the same time retailers were granted the option to pay a 2 percent turnover tax in lieu of the GST up to end-June 2001. Although this new arrangement also triggered disruptive strikes and violent protests across the country, the government has stood its ground.

13. The GST has been extended to a positive list of services in the context of the 2000/01 budgets. The authority for levying sales tax on goods was transferred from provinces to the federal government in the years following the partition of colonial India. The taxation of services was not explicitly considered at the time and remained under provincial authority. This legal situation made it difficult to extend the GST to services without going against the conventional wisdom that the VAT is not a suitable instrument for lower-level jurisdictions in a federation.4 In the event, the government opted for imposing the GST on services through identical provincial legislation, which adopts most rules and regulation of the Sales Tax Act through a blanket reference and delegates the administration to the federal government.5 While this ensures that goods and services are taxed in an integrated fashion for now, it might create problems in the future should some, but not all, provinces decide to modify their legislation. It is also likely to complicate expanding GST coverage to include additional services. Currently, only six categories of services are covered, excluding, construction, legal, consultancy, accounting, and many transportation services. The current legal framework could be strengthened through a constitutional amendment, which would empower the federal government to levy sales taxes on services, in the context of a wider review of intergovernmental relations.

14. Accounts-based assessment is now the norm in Pakistan, although remnants of simplified assessment methods might still be practiced in some cases. Soon after the introduction of the GST in VAT mode, simplified assessment methods emerged in the textile sector and spread to as many as 31 sectors in 1994/95. In that year, almost 40 percent of the taxpayers were under so-called fixed tax schemes. Under these schemes the tax liability was typically determined as a function of some characteristics of the taxpayer, such as the capacity of the enterprise, rather than actual taxable sales net of input tax credit. While the schemes were partly motivated by breaks in the VAT chain, which meant that tax on inputs could not fully be credited, they exacerbated the problem and started to seriously undermine the viability of the VAT. The government attempted to terminate all fixed tax schemes first in July 1995, then in July 1996. But in the face of taxpayer noncompliance, an agreement was struck in November 1997 allowing for the extension of the schemes not beyond 1998. In 1998/99 all fixed tax schemes were eliminated and audits of all affected taxpayers were initiated.

15. Important progress has also been made in other areas but the small number of taxpayers remains a concern. Over the years, GST administration has been separated from the excise department, has been organized along functional lines, and is now fully computerized. A turnover threshold for GST registration has been put in place, replacing earlier thresholds that were defined in terms of capital employed. The working of the refund mechanism has been improved, although remaining weaknesses have prompted unwarranted calls for exempting sales to exporters and their suppliers from the GST. Despite these improvements, the number of taxpayers remains low (about 85,000). Systematic enforcement efforts therefore remain paramount in the period ahead.

16. Since the introduction of the GST in VAT mode, revenues have increased steadily with the largest improvement occurring in 1999/2000. Initially, the GST yielded around 1.6 percent of GDP about the same as the tax it replaced. The ensuing eight years brought a steady increase to 2.4 percent of GDP in 1998/99, although a temporary setback was experienced in 1997/98 when rates were lowered. The large revenue gain of 1999/2000 reflects primarily the extension of the GST to petroleum products, natural gas, and electricity. It should be noted, however, that these GST extensions were designed in an overall revenue neutral way: the gains on account of the GST are offset by revenue losses elsewhere or additional expenditure. The GST extension to petroleum products was accompanied by a commensurate lowering of petroleum surcharges; the GST on electricity increased after-tax tariffs of only those consumers which were in a position to claim input tax credit, thus necessitating subsidies to electricity producers. In the case of natural gas, the GST extension coincided with the elimination of excises but also with an increase of retail prices. The 2000/01 GST extension to services follows the same pattern; most services brought under the GST were previously subject to excises which were abolished in the process.

D. Income Tax Reform

17. Transformation of the income tax over the last two decades was successful in mobilizing additional revenue, but tax collection continues to rely on outmoded methods that will likely constrain future buoyancy. During the 1980s, revenues from direct taxes6 declined from 2–2.5 percent of GDP to just under 2 percent of GDP, reflecting the reduction of tax rates and the growing use of tax concessions. The aggressive development of the withholding tax regime in the 1990s meant that revenues surged ahead reaching 3.8 percent of GDP in 1995/96, followed by a slight decline in the second half of the decade, which left direct tax collections at 3.5 percent of GDP in 1999/2000. Throughout the last two decades, little progress was made in establishing a system built around modern methods of self-assessment underpinned by the threat of audit, increasing the number of taxpayers, or putting in place clear and easily accessible legislation. Moreover, with the withholding tax system already overextended, there is little hope for augmenting income tax collection further through the strategy of the 1990s.

18. The income tax has become unusually dependent on revenues raised through a complex pattern of withholding taxes. They accounted for around 70 percent of net income tax collections in the late 1990s and for almost the entire revenue augmentation over the last decade. In 1979, when the current Income Tax Ordinance was promulgated, only six kinds of payments/transactions were subject to withholding taxes, which was increased to 12 in 1989/90 and further to 24 in 1999/2000. Eight withholding taxes accounted for over 90 percent of withholding tax revenue in 1999/2000. This includes, in declining order of importance: (a) a withholding tax on payments for the supply of goods, services rendered, and contracts executed at rates generally ranging between 3.5 percent and 6 percent; (b) a 5 percent withholding tax on imports, which was raised to 6 percent in the 2000/01 budget; (c) a 30 percent withholding tax on the interest income from holding securities; (d) a withholding tax on salaries and prerequisites; (e) a 10 percent withholding tax on interest or profit on an account maintained with banks or other financial institutions; (f) a withholding tax on export proceeds of between 0.75 percent and 1.25 percent, which was raised by 0.25 percentage points in the 2000/01 budget; (g) graduated withholding taxes on electricity bills; and (h) graduated withholding taxes on telephone bills. While the four top yielding withholding taxes were already provided for in the original Income Tax Ordinance of 1979, the ones pertaining to exports and electricity bills were added in 1992/93, the one on bank account interest was introduced in 1994/95, and the one on telephone bills was first levied in 1996/97. In addition, tax rates of the two most important withholding taxes were pushed up in several rounds. Withholding taxes on supplies, contracts, and services were raised from 2 percent, 3 percent, and 3 percent at the beginning of the decade to 3.5 percent, 5 percent and 5 percent in 1999/2000, respectively.7 The withholding tax on imports was temporarily abolished and reintroduced in 1989/90 at a rate of 1.5 percent, and increased several times to reach 6 percent in 2000/01.8 As a result of the above policies, the share of withholding taxes increased sharply from around 50 percent to 70 percent of income tax collection during the 1990s. Moreover, collections other than those related to withholding taxes stagnated at under 1 percent of GDP throughout that time.

19. Many withholding taxes are indeed presumptive taxes, more akin to indirect taxes than to direct taxes. Their payment constitutes a final discharge of tax liability. This includes the withholding tax on supplies/contracts/services, on imports by commercial importers, and on bank account interest. Revenue from presumptive taxes as a share of withholding tax collections has remained roughly constant throughout the 1990s at around 50 percent. However, it has made up an increasing proportion of total income tax collections. Withholding taxes under the presumptive regime tend to work like indirect taxes, especially when broadbased.9 For instance, a noncreditable income withholding tax levied on imports is equivalent to an import tariff.

20. The extensive reliance on withholding was a likely key factor behind the failure of the tax administration to adopt the procedures needed to develop a modern income tax system. The current system remains based on largely manual procedures and either routine face-to-face contact between taxpayers and assessing officers or arrangements that enable the taxpayer to escape risk of audit. A modern system, in contrast, would be built around voluntary compliance supported by an effective, trusted, and well-targeted audit.

21. The Income Tax Ordinance of 1979 provided for self-assessment from the outset, but the scheme was continuously liberalized and the scope of audit minimized. This trend was reversed in late 1999/2000 and in 2000/01. Available to all, except companies,10 self-assessment, like elsewhere, requires taxpayers to apply the income tax rules to compute their taxable incomes and exposes them to the risk of audit. It authorizes the tax authorities to specify the operational details. These details were flawed in that they granted taxpayers immunity from audit provided that their tax payment rose by a specified percentage. Moreover, taxpayers were not legally required to keep any records or books of account which could be used to substantiate their self-assessed income. As a result, very little revenue was collected through the self-assessment scheme, about 7 percent of income tax revenues other than from withholding in 1997/98. In February 2000, the government inserted general book/record keeping requirements into the Income Tax Ordinance and the 2000/01 budget subjects filings under the self-assessment scheme to a 15 to 20 percent audit probability.

22. The number of taxpayers remains low, although progress has been made over the last ten years. In August 2000, there were 1.8 million taxpayers registered under the income tax, compared to just over one million in 1990/91. Moreover, a large proportion of enrolled taxpayers does not actually file a return (about ⅓ in 1997/98).

23. Income tax rates have been lowered considerably over the past two decades but rates remain high and highly differentiated. In 1979/80, income was highly taxed with effective tax rates in the corporate sector of between 55 percent and 71.5 percent and an effective top personal income tax rate of 66 percent.11 Ten years later these rates had been brought down for companies to the range of 44 to 66 percent and for individuals to 49.5 percent.12 Rates were further reduced in the ensuing decade and in 2000/01 they were 58 percent for banks, 47.3 percent for private companies, 45.2 percent for public companies, 38.5 percent for individuals, and 30 percent for individuals whose income consists mainly of salary.13 Moreover, the tax structure was simplified over the years by merging the separate so-called supertax and its rebate schedule into the income tax schedule and reducing reliance on surcharges. Notwithstanding the general trend toward lower rates and a simplified tax structure, the process was not continuous; e.g., the 2000/01 budget reintroduced surcharges on corporate income, which had been abolished with effect to 1992/93.

24. Progress has been made lately in broadening the coverage of the income tax. However, there is a need to rationalize remaining exemptions and guard against expectations of recurring tax amnesties. Tax concessions of various forms that were widely granted until the mid-1990s to promote priority industries and development of backward areas have been used more sparingly in the second half of the decade. So-called tax-whitener schemes, which guaranteed immunity from tax probe into the source of monies invested in certain instruments, were grandfathered in December 1999. Enabling legislation for taxing agricultural income, which is not taxable under the federal income tax, has been promulgated with the provincial finance bills of 2000. Employees’ benefits in kind have been brought within the ambit of the income tax; they became taxable in 1998/99 as a separate block and are taxed together with any other income beginning 2000/01. Nonetheless, the list of exemptions and concessions still comprises several hundred items, many of them relating to interest income and allowances of the civil service and the armed forces, In March 2000, a tax amnesty was launched, which allowed whitening of evaded assets against payment of a 10 percent tax. While the scheme was relatively successful in raising revenue (0.3 percent of GDP), there is the distinct possibility that taxpayers perceive it as part of a long history of frequent tax amnesties in Pakistan, thus undermining their willingness to make regular tax payments in the future. In this context, it is problematic that the government has launched a new tax amnesty scheme in August 2000, which levies a reduced rate of 2 percent on newly declared assets provided they are in the form of inventories of traders/retailers.

25. Recognizing that a more fundamental reform of the income tax is needed to ensure its buoyancy over the medium term and to improve its structure, the government has established a reform committee. It is planned that a reformed income tax will go into effect in 2001/02.

E. Tariff and Trade Reform

26. While certainly a significant step forward from a structural point of view, trade reform over the past 20 years has taken a heavy toll on fiscal revenues.14 While customs duties equivalent to 5.7 percent of GDP were collected in 1979/80, tariff reductions reduced this yield to under 2 percent of GDP in 1999/2000.

27. Twenty years ago, Pakistan’s import system was highly restrictive. Not only were tariff rates high, but the number and amount of permitted imports was also closely regulated. Items not explicitly listed on the positive list of imports could generally not be imported. Moreover, imports were permitted only against licenses issued by the Chief Controller of Imports and Exports and a subset of items on the positive list were subject to monetary licensing ceilings. Certain items not on the positive list were importable under special authorization, subject to complex regulations and value limits.

28. The first phase of import liberalization, which was undertaken in the 1980s, did not involve any revenue loss. Indeed, customs duties collections increased somewhat to 6.4 percent of GDP in 1989/90 from 5.7 percent of GDP ten years earlier. The first phase of liberalization was initiated in 1983/84 and basically aimed at converting reliance on the positive list of permitted imports into reliance on a negative list of prohibited/restricted imports followed by a gradual shortening of the negative list and making monetary licensing ceilings less binding. This did not involve any revenue loss as items removed from the negative list became subject to their previously notional statutory tariff rates. Moreover, during this period revenue benefited from the imposition of paratariffs, in the form of a 10 percent general surcharge, a 5 percent Iqra surcharge, and a 6 percent licensing fee. As a result, the effective tariff rate increased from 38 percent in 1979/80 to 42 percent in 1987/88.15

29. In the late 1980s, the government embarked on the second phase of the liberalization program, which soon started to impinge upon custom revenues. While the negative list continued to be shortened, the second phase also involved the reduction of the maximum tariff rate and the rationalization of the tariff structure. The maximum tariff rate, which stood at 225 percent in June 1988, was down to 90 percent in July 1991, further lowered to 80 percent in July 1992, reduced to 45 percent in July 1997, and reached its current level of 35 percent in March 1999. Moreover, the paratariffs were integrated into the statutory tariff schedule without raising the maximum rate in 1992/93 and 1994/95. Regulatory duties were imposed between October 1995 and March 1997 and also during part of 1999, albeit on a much smaller range of goods. At end-1999/2000, no significant regulatory duties were levied.16 As a result of these policy changes the effective tariff rate declined to 22 percent in 1999/2000 from 42 percent in 1987/88.

F. Petroleum Taxation

30. Budgetary revenue from the taxation of petroleum products was exposed to the vagaries of international oil prices throughout the past two decades. Collections surged as high as 2 percent of GDP in the face of declining international prices and dipped below 0.5 percent of GDP when international prices increased (Chart I-2).

Chart I-2.

Pakistan: Oil Prices and Petroleum Surcharge Revenue, 1979/80-1999/2000 1/

Source: Pakistani authorities, IFS, and staff estimates.1/ The GST extension to petroleum products of August 1999 led to a reclassification of the petroleum revenue. The 1999/2000 data adjusted according to the old classification.

31. Petroleum taxes were an unstable source of budgetary revenue because they essentially took the form of differential taxes. Retail prices, as well as margins for the distributors, transporters, and dealers are regulated. As a result, petroleum surcharges get compressed or expanded with fluctuations in the landed costs of fuel, unless the latter are fully and instantaneously passed through to retail prices, which was generally not done.

32. Several attempts have been made over the years to better insulate budgetary revenues from the volatility of international oil prices, but such efforts were insufficiently sustained. In the early 1980s, the government adopted a policy of passing through to consumers higher costs of petroleum products. However, this policy was not adopted uniformly across different types of petroleum products and gave way to a continuing program of price increases through the mid-1980s. More importantly, the government decided not to lower retail prices when international prices declined substantially in the period 1984/85-1987/88. The increase of international prices in the late-1980s and early-1990s eroded petroleum revenues despite discretionary, mostly upward, price adjustments. A renewed attempt to better link retail prices to international price movements was made in December 1995; the Fund-supported program envisaged monthly price adjustments triggered by any 3 percent change of the landed cost of fuel. However, the mechanism was suspended for need of revenue a year later when international prices started to decline. It was revived in a modified form in late-1998 when a ± 5 percent band on the petroleum tax was established. The tax, however, veered outside the band in the period January-March 1999. An automatic price adjustment formula was put in place in August 1999, which envisaged the one-to-one pass through of changes in the landed cost of fuel to retail prices on a quarterly basis. However, the government failed to implement the mechanism at the first test date in September 1999 because it would have implied significant price hikes. After the current government took office, it increased prices in December 1999 and in subsequent quarters, largely in line with the price adjustment formula. The formula underwent a technical modification in September 2000, and retail prices are now being reset every quarter as the sum of lagged landed costs of fuel, fixed margins, petroleum surcharges fixed in PRs/liter, and all other taxes, including the 15 percent GST.

33. More fundamental change in petroleum pricing will likely be needed to better ensure that budgetary revenues are stabilized in the face of fluctuating international oil prices. While the current government has shown resolve in implementing the pricing formula, the overall track record of formula driven adjustment of regulated retail prices has been poor. As long as retail prices are regulated likely adverse customer reactions will always argue against price hikes and Pakistan’s chronically weak fiscal position will always argue against price cuts. Deregulation of retail prices would mute these pressures by depoliticizing the price setting, In this context, it is encouraging that the government deregulated the prices of fuel oil in July 2000.

G. Revenue Performance in a Cross-Country Comparison

34. This section compares Pakistan’s ability to and efficiency in raising public revenue with the experience in other countries. The set of comparators consists of thirteen countries from around the world at different stages of development. It includes countries from the region such as Bangladesh, India, and Sri Lanka, as well as countries from Africa, Asia, the Middle East, the Western Hemisphere, and a transition economy. Some of the countries, such as Tanzania and Uganda, exhibit a per capita GDP below US$300, while the sample also covers countries as rich as Mexico or even South Korea. Data cover the last two available years and are drawn from Fund documents with supplementary information provided by the respective country teams. An attempt has been made to use fiscal aggregates with similar coverage and definitions across countries.

35. Pakistan’s ability to raise revenue is not out of line with what one would expect from a country with a per capita GDP of around US$450 per year (Table I-2). This confirms findings of earlier cross-country studies.17 Pakistan raises about 15 percent of GDP in general government revenue compared to an average of around 18 percent of GDP in the sample. The wedge is not surprising given that average per capita GDP is significantly higher at around US$1,700. None of the countries with per capita GDP below that of Pakistan, i.e., Bangladesh, Tanzania, and Uganda, manage to collect as much as Pakistan. India, with a very similar GDP per capita, also collects less at the central government level, but it surpasses Pakistan’s revenue collection by about 1 percentage point of GDP when revenues of provinces are taken into account. Countries with per capita GDP in the range of US$800-1,000 (Sri Lanka, Cote d’Ivoire, the Philippines, and Albania) are doing significantly better than Pakistan. They are raising additional revenue in excess of 3 percent of GDP.

Table I-2.

Pakistan: Revenue Performance in a Cross-Country Comparison 1/

article image
Source: Fund staff estimates.

Ratios are calculated on the basis of data for the last two fiscal years, except in the cases of Mexico and South Korea where only 1999 data has been used. Data pertaining to the last fiscal year are mostly preliminary actuals and sometimes estimates.

Unweighted averages of values of comparator countries.

36. Basically the same picture emerges with regard to central government tax revenue. Pakistan mobilizes about 12.5 percent of GDP compared to 14.4 percent of GDP in comparator countries on average. It collects, in other words, 13 percent less tax revenue. The shortfall for total central government revenue is somewhat larger (16 percent), indicating that it disproportionately falls on nontax revenues.

37. The revenue gap with comparator countries shrinks further when focusing on collections under the three main tax heads: sales tax, income tax, and customs duties. While Pakistan collected almost 9.1 percent of GDP in 1999/2000, the average comparator country collected 9.9 percent of GDP. This corresponds to a shortfall of around 8 percent, compared to a shortfall of 16 percent for revenue from all taxes. Pakistan seems to be less at a disadvantage in raising revenue through the three main taxes than through other taxes.

38. In terms of revenue from the three main taxes, Pakistan outperforms a number of countries with higher income. It collects more than Mexico as well as Sri Lanka, and it is almost at par with Egypt. Overall these countries seem to fare better than Pakistan only because they have access to other sources of revenue: oil revenue in the case of Mexico, the National Security Levy in the case of Sri Lanka, and oil revenue, as well as revenue from the operations of the Suez Canal in the case of Egypt.

39. Pakistan is somewhat less effective in raising revenues through the three main taxes than the much richer comparator country group. If Pakistan were as effective in raising revenue from the three main taxes as the countries in the comparator group, it would have collected an additional 1 percent of GDP in 1999/2000. Effectiveness is measured by efficiency ratios. They indicate how well countries are doing in mobilizing revenue relative to the tax rates they are imposing.18 Pakistan’s tax rates are similar to those prevailing in the other countries; its standard sales tax rate is slightly lower, its top income tax rate (averaged across businesses and households) is higher, while its tariff rate (average of tariff bands) is almost the same.

40. The adjusted sales tax efficiency ratio comes to 0.28 in 1999/2000 compared to 0.33 in the other countries. Although this corresponds to a shortfall of 15 percent, it still means that Pakistan does better than some countries with higher per capita GDP (Cote d’Ivoire and Mexico) and is not outperformed by any country with a lower per capita GDP.

41. Pakistan’s income tax efficiency ratio of 0.09 compared to an average of 0.1 suggests an underperformance of 15 percent. Again, Pakistan surpasses richer countries like Egypt, Jordan, and Albania. However, it is also outperformed by one country with a lower per capita GDP (Tanzania). One needs to caution though that the income tax efficiency ratio likely paints too positive a picture of Pakistan’s income tax performance. As discussed above, the bulk of income tax revenue is raised through nonstandard withholding taxes and is therefore largely unrelated to the rates of the income tax schedule.

42. Pakistan’s effectiveness in raising customs duties exceeds those of comparator countries. The adjusted efficiency ratio comes to 0.68 in 1999/2000 compared to 0.6.19 It does better than many richer countries including the Philippines, Morocco, Sri Lanka, and even South Korea and Mexico. The outperformance of South Korea and Mexico reflects these countries having entered into extensive free trade arrangements rather than a superior Pakistani customs administration.

43. Overall, there is little in the data presented in this section that would support the claim that Pakistan’s revenue performance is out of line with experience in other countries. While Pakistan does not match the revenue mobilization of the other countries in the sample, this appears to reflect primarily Pakistan’s lower level of overall development and the absence of access to certain forms of revenue. Moreover, weaknesses appear to be more pronounced in nontax revenues and provincial revenues and less pronounced in the area of the three main taxes: sales taxes, income taxes, and customs duties. However, it should be noted that many of the comparator countries are beset by severe fiscal imbalances and therefore should not be considered a wholly satisfactory benchmark.

H. Conclusions

44. Enhancing the buoyancy of the tax system is generally perceived as key for improving Pakistan’s fiscal position. As current tax rates leave only limited room for further hikes, efforts tend to focus on better tax enforcement, bringing more taxpayers into the tax net, and other improvements in the realm of tax administration. The analysis in this Section suggests that these efforts will only be successful if Pakistan breaks with its historical record and surpasses the performance of comparator countries at a similar, and sometimes even higher, level of development.

45. While Pakistan’s historical record of revenue performance looks dismal at first sight, some positive signs emerge upon closer inspection. True, revenue as a percentage of GDP has not changed significantly over the last 20 years. However, this masks the significant progress that has been made in boosting non-trade taxes which now yield revenue of 10.2 percent of GDP compared to 7.4 percent of GDP only ten years ago, implying an annual buoyancy coefficient of 1.27.20 Much of the additional revenue has been mobilized in a structural sound way, especially through sales tax reform. Revenue enhancement under the income tax, however, have relied excessively on a mushrooming net of withholding taxes, with more fundamental improvements left for the period ahead.

46. The cross-country comparison suggests that Pakistan’s tax system is, by and large, an average performer given the overall level of the country’s development. While this might come as a pleasant surprise to the many observers with less favorable expectations, it also means that the upside potential, at least in the near term, is probably limited. It should also be noted that the cross-country comparison is subject to several caveats. First, the choice of comparator countries might have inadvertently tilted the sample too much towards countries which are either too similar to Pakistan (and therefore not suitable benchmarks) or too different from Pakistan (and therefore not comparable). Second, the exercise focuses exclusively on the revenue yield relative to GDP and relative to tax rates. It ignores other critical elements, such as the complexity of the tax system and governance issues in tax administration, which are much more difficult and judgmental to compare across countries.

47. The inherent difficulties in quickly mobilizing significant amounts of revenue through better tax enforcement means that revenue-losing measures and expenditure overruns are to be avoided if fiscal consolidation is to be achieved. Indeed, supplementary revenue measures could help secure the consolidation effort, although a pure enforcement strategy remains preferable from a structural point of view.


Prepared by Christoph Klingen (FAD).


See Ahmad, Ehtisham, and Nicholas Stern, The Theory and Practice of Tax Reform in Developing Countries, Cambridge University Press, 1991.


A tax is called cascading if a commodity/service is taxed more than once as it passes through the various stages of the production-distribution chain and the effective tax burden therefore exceeds the nominal tax rate.


For a discussion of the problems involved and possible solutions see Keen, Michael, 2000, “VIVAT, CVAT and All That: New Forms of Value-Added Tax for Federal Systems,” Canadian Tax Journal, 48, 409–24.


Rules and regulations still need to be modified to take into account the special requirements of service taxation. This includes practices related to the import of services, the complex issue of taxing air travel and communication services, and the formulation of rules that apportion inputs of companies supplying goods as well as services.


Income tax accounts for almost 95 percent of direct tax revenues. The balance is made up by the wealth tax, workers welfare fund tax, capital value tax, and corporate assets tax.


The rate pertaining to supplies was increased from 2 to 2.5 percent in 1991/92 and from 2.5 to 3.5 percent in 1997/98; the one pertaining to contracts was increased from 3 to 5 percent in 1995/96; and the one pertaining to services was raised from 3 to 5 percent in 1991/92.


It was raised in 1991/92 to 2 percent, in 1995/96 to 4 percent, in 1997/98 to 5 percent, and in 2000/01 to6 percent.


The same is true of withholding taxes that are, in principle, creditable but not actually credited. With holding taxpayers might chose not to file a return, in particular if that would involve having to pay taxes over and above the ones already withheld. The withholding tax then becomes, de facto, a final tax.


The scope of the regime was extended to include part of the corporate sector in 2000/01. It was also available to companies in 1980/81 and 1981/82.


Banks were subject to a statutory tax rate of 30 percent, a supertax of 35 percent and a surcharge of10 percent. Companies were faced with a statutory tax rate of 30 percent, a supertax of 25 percent, and a surcharge of 10 percent Public companies, i.e., companies that are listed on the stock exchange, were eligible for a 5 percent rebate on the supertax. The top marginal personal income tax rate was 60 percent and a 10 percent surcharge applied to income above a certain threshold.


Banks were subject to a statutory tax rate of 30 percent, a supertax of 30 percent, and a surcharge of10 percent. Companies were faced with a statutory tax rate of 30 percent, a supertax of 15 percent, and a surcharge of 10 percent. Public companies were eligible for a 5 percent rebate on the supertax. The top marginal personal income tax rate was 45 percent and a 10 percent surcharge applied.


Banks are subject to a statutory tax rate of 58 percent. Companies are faced with a statutory tax rate of 43 percent, and a surcharge of 5 percent. The statutory rate for public companies is 33 percent and they are also subject to a 5 percent surcharge.


For a broader discussion of the revenue implications of trade liberalization see Ebrill, Liam, Stotsky, and Gropp, Revenue Implications of Trade Liberalization, IMF Occasional Paper No. 180, 1999.


The effective tariff rate is defined here as customs duty collections (including revenue from paratariffs) in percent of the value of dutiable imports.


Regulatory duties were reintroduced in the context of the unification of excise duties between imported and domestically produced goods. They will be removed once antidumping legislation goes into effect.


See Tanzi, Vito, “Quantitative Characteristics of the Tax Systems of Developing Countries,” in Newbery, David and Nicholas Stern (eds.), The Theory of Taxation for Developing Countries, Oxford University Press, 1987.


The efficiency ratio of a tax is defined as its yield in percent of GDP divided by the tax rate. Adjusted efficiency ratios are normalized by the share of the tax base in GDP. For instance, a sales tax yielding 5 percent of GDP at a standard rate of 10 percent and consumption accounting for 80 percent of GDP has an efficiency ratio of 0.5 and an adjusted efficiency ratio of 0.625. A low efficiency ratio would be indicative of weak tax administration, widespread exemptions, or many sales being taxed at a reduced rate.


Calculation of the efficiency ratios is based on the average tariff rate (simple average of tariff bands).


Defined as the average growth rate of (nontrade) tax revenue relative to the average growth rate of nominal GDP.