The emphasis on fiscal reforms encompassing documentation, transparency, and improving tax compliance, initiated by the present government, continued during FY02. However, unlike FY01, which saw a notable fiscal consolidation, the efforts to reduce the budget deficit during FY02 to the targeted 4.9 percent of GDP were not successful as the overall deficit rose to 6.6 percent of GDP. The specific factors driving this outcome include the negative impact of international political developments, unanticipated defense expenditures, and some one-off adjustments. Importantly, adjusting for the latter expenses of Rs 168.4 billion, the "baseline" deficit for FY02 falls to 4.6 percent of GDP. This is well below the FY02 target, as well as the FY01 deficit of 5.3 percent of GDP (see Table 4.1).

A compositional breakdown of the overall FY02 deficit is instructive; while revenue collections were admittedly lower than the target, the primary contribution to the above-target deficit was from a sharp rise in expenditures. The decline in revenues was, to a degree, explainable by a slowdown in domestic economic activity, a consequent decline in imports, as well as the negative impact of two unexpected trends that altered key tax bases, i.e. a continuing decline in domestic inflation and the appreciation of the Rupee against US Dollar, both of which contributed to a decline in ad-valorem tax receipts. However, the fact that the tax-to-GDP ratio remains weak suggests that the tax base needs to be broadened.

On the expenditure side, while higher defense spending was to be expected given the escalated tensions on borders with India (following the December 13 incident),' the larger drain was on account of three additional one-off expenditures; a grant to the CBR allowing it to clear accumulated income tax arrears owed to banks, a substantial investment in KESC to restructure its finances before privatization, and a settlement of WAPDA arrears.

As no single indicator can properly assess the overall fiscal performance of a country, a set of indicators has been presented to gauge the consolidated fiscal operations of the federal and provincial governments during FY02.

The unadjusted overall budgetary deficit has clearly deteriorated during FY02; not only is it higher than the actual budgetary deficit in FY01, the rise clearly disturbs the overall downward trend visible since FY99 (see Figure 4.1). However, the increase in the deficit over the FY02 target is almost entirely attributable to exceptional expenditures;

excluding these gives an "adjusted" deficit of 4.6 percent of GDP (see Table 4.2), i.e. maintaining the recent downtrend.

billion Rupees

Source : Ministry Of Finance

The performance of the overall budget deficit is similar to that of another indicator, the revenue deficit, i.e. the gap between revenues and current expenditures, which indicates the portion of the government's revenues consumed by non-development expenditures.3

Borrowings undertaken to meet current expenditures are a matter of some concern, as the expenditures do not add to the repayment capacity of the country, but they increase the debt burden. Pakistan has been experiencing a revenue deficit for last two decades, but this has declined significantly in FY02 to 2.4 percent of GDP (see Figure 4.1).

To its credit, the government appears to recognize that even the FY02 revenue deficit figure is high. This concern is explicitly accepted in the government's proposed Fiscal Responsibility and Debt Limitation Ordinance, 2002 that aims to gradually eliminate the revenue deficit by FY07.

A troubling development for FY02 is the abrupt reduction in the primary surplus to 0.1 percent of GDP from 2.0 percent of GDP in the previous year (see Figure 4.1). This indicator is arrived at by deducting interest expenses from the overall budgetary deficit, and is used to gauge a country's ability to finance its expenditures in the absence of a debt burden. Pakistan has been running primary surpluses (or almost negligible deficits) since FY94, but it was only in the last few years that substantial primary surpluses had been generated, reflecting the increased emphasis on containing the debt burden.

The abrupt loss of this surplus in FY02 is largely attributable to the exceptional expenditures witnessed during the year (see Table 4.2). The absence of these in future years could therefore lead to the re-appearance of large primary surpluses.

FY02 has seen a surprisingly strong jump in the revenue-to-GDP ratio, which rose by 1.1 percentage points even as the tax revenue to GDP ratio remained almost unchanged (see Figure 4.2). In other words, the buoyant 15.4 percent growth in revenues is primarily driven by rising non-tax collections, which is based on three main sources: (1) interest income on government loans, (2) dividends from corporations, and (3) profits from other organizations such as the SBP.4 None of these has traditionally been a stable income source, as evident from the high variance in these receipts over the last 12 years (see Table 4.3). As seen in Figure 4.2, tax revenues as a percent of GDP showed some growth from FY92 to FY97 and afterward remained worryingly stagnant despite the considerable efforts over the years to broaden the tax base of the economy. Clearly, the efforts to capture the tax potential of the economy have not succeeded so for.

To gain more insight, the buoyancy of revenues has also been computed for a 12-year period to assess the overall growth in revenue receipts. The estimated buoyancy of tax revenues averaged 0.9 as compared to 1.3 for non-tax revenue (see Table 4.3). In other words, the average growth and volatility in non-tax revenues remained significantly higher than for tax revenues over the estimation period.

Table 4.2: Adjusted Budget Deficit for FY02

During FY02, the share of non-tax revenues and its buoyancy estimates were higher than the average for the 1990s. However, the concern due to the historical volatility may be misplaced, as the restructuring of public sector enterprises in recent years may finally permit them to regularly service their debt to the government. If so, this would add greatly to the stability of non-tax revenues going forward.

The total expenditure to GDP ratio approximates the government's share in the overall economy and provides information on the fiscal stance of the government. As evident from Figure 4.3, after falling sharply in the first half of the 1990s, it depicts a gradual downtrend in succeeding years. Unfortunately, this gradual slide was almost exclusively at the cost of development expenditures.

The average buoyancy of total (1.0), current (1.1), and development (0.5) expenditures computed for the preceding 12 years reinforces the conclusions drawn above (see Table 4.3). While expenditures kept pace with GDP on average, it is current expenditures that grew at the faster pace, driven mainly by a sharp rise in interest payments, and fiscal adjustments were largely made by cutting development expenditures.

During FY02, the total expenditures to GDP ratio has jumped sharply from 21.3 to 23.7 percent, largely on account of current

expenditures (see Figure 4.3).5 However, for once, this rise is not at the expense of

development expenditures, which has also rose by 0.7 percent of GDP. Moreover, incorporating the one-off spending, the "adjusted" FY02 expenditures to GDP ratio falls to 21.7 percent.

A compositional breakdown of current expenditures over the years also provides interesting insights. As a ratio to GDP, expenditure on public sector development programs and defense recorded considerable declines since the early 1990s. However, the realized reductions did not lower the overall budget deficit, as the interest payments picked up over the same period (see Figure 4.4). On the positive side, over the last three years, the interest payments to GDP ratio has also witnessed a notable decline, reflecting the government's emphasis on better debt management.

The ratio of budgetary targets to actual performance, for various revenue and expenditure heads, provides interesting insights on the government's ability to set realistic targets and perform within

these budgetary parameters. It is pertinent to note here that both large positive or negative deviations from the target are not desirable. Table 4.3 shows that revenue receipts remained persistently lower than the target, and strikingly, there has been very little deviation from the mean, i.e. there is a very consistent upward bias in the government's budgetary revenue estimates.

The expenditure ratio shows greater variability; current spending has seen both positive and negative deviations from targets, but on average, positive deviations (realized figures greater than budget targets) have dominated (see Figure 4.5). On average development expenditures are over-estimated (see Table 4.3).

The revised FY02 federal government revenue receipts stand at Rs 632.8 billion. While below budget target, this figure is still 18.3 percent higher than the collections during FYO 1. The improvement is almost entirely based on higher non-tax revenues.

The overall performance of the taxation system depends on its revenue generation, which is generally gauged by the tax to GDP ratio. Historically, the Federal tax to GDP ratio fluctuated in a band of about 2 percentage points around a mean of 11.4 percent (see Figure 4.6). The fact that post-1996, this ratio has generally remained below the average, despite lower GDP growth, indicates a continuing weakness in the taxation system. Within total taxes, there is a structural shift from indirect taxes to direct taxes as reflected

in the rising direct tax to GDP ratio (see Figure 4.7). During first half of 1990s, a sharp rise in this ratio was largely explained by the

increasing resort to withholding taxes. However, recent tax reforms reduced the role of these taxes, and therefore the continuing marginal up

trend in direct tax collections over the past three years is quite encouraging.

As can be seen from Figure 4.6, the variation in the indirect taxes to GDP ratio explains most of the variation in total tax collections. This co-movement also highlights the government's heavy dependence on indirect tax revenues despite the regressive nature of such taxes.

The present government has initiated wide- ranging taxation reforms to correct structural weaknesses, which include a Tax Survey and Registration Scheme supplemented with Tax Amnesty Schemes to ease public concerns, introduction of new Income Tax Law, a new Self-Assessment Scheme, etc. Broadening the tax base and improving efficiency in tax administration remained the main planks of taxation policy. These were implemented by reducing the number of taxes, rationalizing tax rates and penalties, and simplifying collection procedures. All these measures helped the government realize double- digit growth in tax collections during FY00 and FY01, as also reflected in the rising tax to GDP ratio (see Figure 4.6). However, lower growth in FY02 primarily due to slowdown in economic activity, lower imports and higher tax refunds undermined the government efforts to record buoyant growth for a third year in a row.

Buoyancy is used to measure the relative increase in revenue collection (including the effects of discretionary changes in tax system) compared with the relative increase in the base. The buoyancy estimates for various federal taxes with respect to GDP, as well as with alternative bases, are reported in Table 4.4.7 The buoyancy for total tax receipts is less than one for both GDP and non-agricultural GDP. This means that the CBR tax collections could not keep pace with growing national income over the period of estimation, despite discretionary tax measures.

Within total taxes, buoyancy estimates for direct taxes are significantly greater than one, both with respect to GDP and non-agricultural GDP.8 This implies that the growth in direct tax collection remained healthy over the period of estimation, reflecting the success of the government policy to increasingly shift the tax burden to direct taxes.

In contrast, indirect taxes could not grow in proportion to either overall GDP or private consumption;

only sales tax witnessed buoyant growth, while the central excise duty and customs duty could not record any significant increase over the period of estimation. This behavior of indirect taxes is largely explainable by the taxation policies.

Table 4.4 : Buoyancy Estimates

Customs duty collections fell as the maximum tariff rates were slashed from over 100 in FY91 to just 30 percent by FY02, driving the share of custom duties in total tax revenues from 48.6 percent in FY91 to 11.9 percent in FY02. The government attempted to offset this loss by imposing sales tax, but this effort was not too successful until FY99 (see Figure 4.8).  Thereafter, the sales tax receipts have risen strongly amidst the broadening of the tax base, through elimination of exemptions, and shifting of CED collections. The central excise duty (CED), which had a 25.8 percent share in indirect taxes during FY91, was also gradually phased out for most commodities, being replaced by sales tax. As a result, the CED collection fellfrom 2.1 percent of GDP in FY99 to just 1.3 percent during FY02.

The buoyancy estimates also reflect these developments, being greater than one for sales tax and less than one for customs duties and CED. The low buoyancy for indirect taxes thus simply reflects that the government failed to offset revenue losses from tariff cuts during the 1990s. However, the situation has witnessed tremendous improvement during FY00 and FY01.

Compared to buoyant growth in CBR tax collections during FY01, receipts increased by only 3.0 percent during FY02 (see Table 4.5). This marginal net growth was weaker than real GDP growth and the inflation rate, mainly due to an unexpected decline in import-based taxes, and exceptionally high tax refunds (that dragged down the net collections). Both factors primarily pertain to indirect tax collections, which consequently saw a decline of 2.4 percent. Thus, in sharp contrast to low growth in total tax collections during FY02, revenues from direct taxes recorded a healthy growth of 14.5 percent over the preceding year (see Table 4.5).

A compositional break down of direct taxes showed that the growth was largely driven by higher collections on account of both normal returns and on demand payments (arrears).9 The compositional change is more encouraging given that the collections from withholding taxes have remained almost at the FY01 level. This is due to abolition of five types of withholding taxes, lower collections from interest/profit income on securities (largely driven by lower T-bill rates during the year), as well as a fall in receipts of import-based withholding taxes. The double-digit growth in direct taxes is thus quite creditable and indicates that the government taxation reforms have started paying dividends.10

Table 4.5: Actual Tax Collections (Net) by CBR

Source : Central Board Of Revenue

The breakup of indirect taxes showed that only sales tax collections recorded positive growth as compared to FY01, while the customs collections were the hardest hit with a steep decline of 26.1 percent (see Table 4.5). In fact, the sales tax collections maintained moderate positive growth for the year, both on domestic and import related items. Specifically, while FY02 sales tax collections on domestic goods and services registered healthy growth of 13.1 percent over a year ago, sales tax collections on import related items rose only 4.8 percent, which lowered the overall growth rate to 8.3 percent.

Higher sales tax collections on domestic goods and services are largely attributable to the broadened tax base, both due to new sales tax payers and the government's efforts to substitute CED with sales tax. Also, it must be noted that relatively low growth in import-related taxes is not too discouraging, in the presence of dwindling dutiable imports and appreciating local currency." Additionally, a one-time distortion caused by exceptional GST refunds was a key FY02-specific dampener on net collections. In absolute terms, FY02 sales tax refunds were Rs 6.1 billion higher than in FY01.

FY02 custom revenues too, were a major drag on CBR net tax collections. In absolute terms, these alone contributed Rs 21.5 billion to revenue shortfall of Rs 53.8 billion from the budget target. In fact, in proportion to its 11.9 percent share in actual tax collections, the contribution to overall shortfall is exceptionally large at 40.0 percent. The reasons for the fall are the same as for sales tax.

Although overall growth in net tax collections was disappointing during FY02, trends in monthly tax collections are insightful. Despite exceptional developments during the year under review, the seasonably remained unchanged, with peaks in revenue collections coinciding with quarter ends (see Figure 4.9). Another important point is the clear improvement in revenue collections during the last quarter of the year (see Figure 4.10). This suggests a waning of the negative September 11 impacts, with the passage of time. Specifically, revenue collection during Q4- FY02 witnessed impressive growth of 16 percent over the same period a year ago, with both direct and indirect taxes contributing to the improvement.

As in the previous year, the CBR tax targets saw three revisions during FY02, but the actual collections ofRs 403.9 billion were still quite low compared to the final revised target ofRs 414.2 billion, as well as the budget target ofRs 457.7 billion (see Table 4.6).

The initial budget target ofRs 457.7 billion was revised downward to Rs 444.7 billion in August 2001 to account for the shortfall realized in actual tax collections during FYO 1.12 However, the impact of the September 11 shocks, and realized shortfall during Q1-FY02 forced CBR to resort to another downward revision of the FY02 target to Rs 429.9 billion by October 2001. based on preliminary projections of further revenue losses.

However, the mid-year collections again fell short of the revised target, as the economic assumptions proved too optimistic. To adjust for the shortfall accumulated upto January 2002. and to incorporate revisions in economic assumptions, the annual target was again adjusted downward to 414.2 billion, which was only 5.6 percent higher than the actual collections during FYO I.13 Unfortunately, the actual collections still fell short of this final revised target (see Table 4.6).

During FY02, gross tax collections rose by 6.3 percent year-on-year, but net collections rose by only 3.0 percent due to a sharp increase in

the payment of tax refunds/rebate. Specifically, refunds rose to Rs 79.3 billion

during FY02 compared to Rs 62.1 billion a year before (see Table 4.7). Consequently, refunds/rebate as apercentage of gross collections jumped from 13.7 percent inFYOl to 16.4 percent during FY02.

The massive growth in refunds/rebates was the upshot of government efforts to substantially reduce the accumulated arrears of refunds/rebates. The government also streamlined the refund/rebate claims process, in a bid to help exporters remain competitive.

Table 4.6 : CBR Revenue Targets

Table 4.7 : Quarterly Collections And Refunds by CBR

The monthly data on the disbursement of refunds shows that the bulk of the incremental refunds were concentrated in Q2-FY02 when business confidence was probably at its lowest ebb (see Figure 4.11).14 During H2-FY02, refunds disbursements were in line with the corresponding period last year. However, substantial incremental payments are also visible in August (Rs 3.8 billion), March (Rs 1.4 billion) and May (Rs 2.1 billion), which suggest that this was not merely an effort to help out the exporters, but was also a genuine effort to eliminate the payments arrear overhang.15 If the latter view is correct, this would be reflected in a relative increase in net collections in the next year.

Revised estimates of surcharges stood at Rs 53.9 billion for FY02 compared to the budget target ofRs 47.0 billion. The entire increase is attributed to higher development surcharges on petroleum products (see Figure 4.12).

The rise is the result of a Rs 0.75 per litre increase in petroleum development surcharge on diesel and Rs 0.25 per litre on other oil products.16 This revenue generation measure was specifically designed to improve overall revenue receipts in the presence of a binding fiscal deficit target under the PRGF. Unlike tax revenues, surcharges remained a volatile source of government revenue receipts in the past. The development surcharges are the difference between production costs and the fixed sales price of the commodities, and the latter were not readily adjusted as international prices changed.

However, the volatility in surcharges is attributable entirely to the fluctuations in petroleum surcharges given the smaller share of gas surcharges. With the deregulation of petroleum prices from the FY02, it is envisaged that surcharges will emerge as a more consistent source for the exchequer.

In sharp contrast to the tax revenue picture, revised non-tax receipts stood at Rs 164.7 billion, which are Rs 25.6 billion higher than the budget target and over Rs 50.0 billion higher than the actual receipts during FY01 (see Table 4.8). The breakup of non-tax revenue showed that all three heads posted a healthy growth over a year ago, but the major driving force was higher receipts from civil administration.

Receipts from property and enterprise were almost in line with the budget target, but Rs 13.4 billion higher as compared to the previous year. This improvement was shared by interest income from provinces and institutions (the biggest component was the higher dividend income received from OGDCL).'7

Source : Annual Budget Statement

The increase in interest income was primarily driven by higher loan recoveries from financial and non-financial institutions. The increase in the former was largely on account of interest received on foreign loans ofNDFC (which has now been merged with National Bank of Pakistan), while the surge in latter was due to loan recoveries from KESC, WAPDA and NHA.18

Civil administration receipts witnessed remarkable improvement during FY02 as compared to FY01. This increase of over Rs 30.0 billion was attributed to two factors: (1) a Rs 5.5 billion increase in transfers from SBP, and (2) a Rs 23.0 billion increase in defense receipts, largely on account of the logistic support to US forces operating in Afghanistan. The SBP profits largely depend on interest income and exchange gains or losses. Due to net exchange gains, the SBP was able to transfer an amount ofRs 28.5 billion to the government during FY02. As can be seen from Table 4.8, miscellaneous receipts also posted an increase ofRs7.6 billion over a year before, largely on account of higher income from sale proceeds of oil and gas (on account of the government share in joint ventures).

The transfers to provincial governments are revised downwards to Rs 175.1 billion compared to Rs 190.0 billion in budget estimates (see Table 4.8). This downward revision is entirely on account of lower receipts in the divisible pool.19 The underlying reasons for lower collections of these taxes are discussed in Section 4.3.1.

Revised FY02 federal government expenditures on the revenue account stand at Rs 697.7 billion, which is Rs 35.1 billion higher than the FY02 budget estimates and Rs 85.0 billion higher than the actual expenditures during FY01 (see Table 4.9). Both, current and development expenditures, shared in this increase, as the former surged by Rs 55.0 billion and the rest was absorbed by the latter.

Table : Federal Government Expenditures