A Framework for Reducing the Lebanese Budget Deficit


Part II- Analysis of the 1998 Budget Proposals



Background

The struggle to contain the budget deficit in Lebanon is an endeavour that has been on-going for a number of years. Very large deficits were unavoidably incurred during the war years, but in 1993 when the governmanaged to regain significant control over revenue sources, the deficit-to-GDP ratio fell to less than nine per cent. (The deficit discussed in this study refers to the deficit incurred in Lebanese pounds in the State Budget; it does not include the deficit incurred by the Council for Development and Reconstruction (CDR) and financed by foreign loans. If this component of public spending is consolidated within the State Budget, the overall budget deficit would be a few percentage points higher.)

Another factor that helped to hold the deficit down in 1993 was the lag that is usually experienced between initiating spending plans and implementing them. This lag was overcome in 1994 when total expenditures in the budget rose by 72 per cent while revenues increased by 20 per cent. The consequence was that the deficit exceeded the equivalent of 19 per cent of the GDP in 1994. In the circumstances efforts were made in 1995 to reduce the deficit through enhancing revenue generation while the expansion in spending was held down. The result was that the deficit contracted by nearly four percentage points to the equivalent of 15.6 per cent of the GDP.

The 1996 budget foresaw further progress in deficit reduction and it projected a deficit equivalent to 12 per cent of the GDP. This result was to be attained by a 33 per cent growth in revenues while the increase in spending was to be held at ten per cent. In the event, revenue collections were LL492bn less than projected in the budget, while expenditures exceeded the budgeted allocations by LL767bn. The consequence was that the overall budget deficit, rather than being reduced significantly, rose by 2.5 percentage points, to the equivalent of 18 per cent of the GDP.

A summary of Government Finances, 1996-1998 is shown in Table 6.


Developments in 1997

In the 1997 budget another attempt was made to cut back the budget deficit, this time drastically by almost eight percentage points. The strategy rested on a decision to freeze total spending in 1997 at the level of the approved budget allocations for 1996 and not to allow any additional expenditures unmatched by new revenue resources. This meant holding aggregate spending in 1997 significantly below the actual level expected for 1996. At the time of the formulation of the budget, total outlays in 1997 were already forecast to attain LL7200bn, and in fact turned out to be only marginally higher than that estimate. Therefore, to achieve the 1997 spending reduction target of 12 per cent relative to the expected actual spending in 1996, great efforts in exercising expenditure controls were required. The revenue forecast in the 1997 budget was founded on realistic estimates.

Preliminary actual data for the January-September 1997 period indicate total expenditures in the budget at LL5,883bn (92 per cent of allocations for the year as a whole). Clearly the intention to exercise strict restraint on spending had not been effective. For the year as a whole aggregate spending is projected at close to LL8000bn - as an order of magnitude - but it could be some LL100-200bn less than this forecast. In any case spending in excess of the budgeted allocations seems to amount to about LL1,500bn of which additional interest payments borne by the budget are estimated at LL500bn. The cost of higher salaries for teachers and uncovered Treasury advances apparently also contributed to excess spending.

With regard to revenue collections, in the first nine months of the year receipts totaled LL2,719bn (66 per cent of the budget forecast). For the year as a whole revenue generation is projected - also as an order of magnitude - at around LL4,000bn. As was noted above about expenditures, receipts in the year as a whole could also be some LL100-200bn less than presently estimated. Considering a slower rate of economic expansion in 1997 than was anticipated a year ago (that is, at the time the budget forecast was prepared) and also in view of the punitive raises in import duties on certain categories of cars, which are believed to have adversely affected collections, the revenue estimate at LL4,000bn or marginally less tends to substantiate the view that the revenue forecast in the budget was generally realistic.

With the broad magnitudes for revenues and expenditures mentioned above, the overall deficit in the budget is forecast at LL4,000bn, equivalent to almost 18 per cent of the GDP (with CDR expenditures included, the ratio would likely exceed 20 per cent of the GDP). If these estimates materialise the net debt-to-GDP ratio would be at the 90 per cent level and the gross debt ratio would exceed the 100 per cent level. A substantial use of government deposits in the second half of 1997 would draw the two ratios closer together and enable the gross debt ratio to remain below the 100 per cent mark. That would constitute an effective use of assets to reduce the stock of debt.

In addition a new debt component has recently emerged in public discussions. These are the reported delayed payments of about $125m to hospitals and an uncovered deficit in the Fund for the Development of the South of $50m. These are commonly referred to as domestic arrears. They are debts owed to the private sector which had extended this lending involuntarily. Technically, it may be difficult to allocate these expenditures to the years in which they were incurred. A practical arrangement could be to recognize these arrears - and any others that may exist - as part of the stock of debt as it stands now. When such arrears are repaid in the future from borrowed resources, the implications would be substituting one form of debt for another.

The Pre-Budget Discussions

By mid-1997 the unsatisfactory fiscal situation had become increasingly evident and subject to widespread public discussion. The Prime Minister called for a conference in July 1997 at which the problem of the fiscal deficit and the mounting debt were major topics. In these meetings there was a broad consensus that continuing large fiscal deficits could not be tolerated and that ambitious measures would have to be implemented if the overall debt situation was to be contained within manageable proportions. With that objective in mind it was strongly recommended that total spending in 1998 be reduced by LL200bn relative to their 1997 level. It was also urged that the quality of spending be scrutinised carefully in order to eliminate waste. The conference also called for redoubling efforts to enforce tax laws and to collect fees and utility charges comprehensively. It was also suggested that the authorities should look into the feasibility of new revenue measures.

In September 1997 the government considered, but did not approve, a package of taxes, spending and borrowing measures - the so-called $1bn project, later reduced to $800m. The relationship between the proposed package and the then forthcoming budget proposals for fiscal 1998 was not spelled out. In the first section of this report, it was assumed that the whole package would be part of the budget.

The revenue component of the package was ambitious in scope and very important in content. The two most important elements in the proposed revenues consisted of LL5,000 fee per 20 liters of gasoline and the introduction of a sales tax. These, with other measures, were estimated to generate new resources for the budget equivalent to about four per cent of the GDP. Apart from the resource mobilization aspect, which is very important in itself, these measures would have strengthened greatly the component of tax spending within the context of the Lebanese tax system. Lebanon still laca modern sales tax to use as an instrument in fiscal management, including replacing receipts from import duties which might need to be reduced in the context of a trade association agreement with the European Union. As for the gasoline tax, it is needed not only to generate quick and considerable revenues but also to encourage conservation in fuel use, and to support policies designed to tackle the problems of traffic congestion and environmental pollution. Retail gasoline prices in Lebanon are about half those charged in Greece.

The spending components of the package were listed as follows: $200m for the Fund for Displaced People, $125m to pay overdues to hospitals, $50m to cover the deficit in the Fund for the Development of the South, and the balance of $425m to finance social projects in remote regions of the country. The financing was to be generated by dollar long-term borrowing abroad.


The 1998 Budget Proposals

In early October 1997 the government submitted to Parliament the proposals for the 1998 budget. Revenues were estimated at LL4,956bn, and expenditures at LL7,925bn, resulting in an estimated deficit of LL2,969bn. The details for the revenue and expenditure proposals are as follows.

1. Revenues

The revenue forecast in the 1998 budget proposals suggests a rate of expansion about twice the projected rate of growth in nominal GDP (12 per cent under the assumptions of this study). To achieve the ambitious result the proposals specify in detail a large number of tax and fee adjustments including the introduction of substantive new revenue measures. Wide-ranging increases in fees are enumerated in Table 9 of the attachment to the draft budget law. The 12 pages of this attachment comprise increases in the following categories of fees:

1. Car licenses (with considerable differentiation based on horsepower)

2. Driving license fees

3. Passport fees

4. Work license fees for non-Lebanese

5. Real estate fees for a wide variety of transactions

6. A fee on cement

7. University registration and degree fees

All these fees are specific (so many Lebanese pounds per transaction) and the most important are the fees on cars, real estate transactions, and on cement. With respect to cars, compliance has been a problem in the past, therefore collecting effectively significantly higher fees could be problematic and will require very determined efforts.

Assuming strict implementation of the proposed increases in all categories of fees, a significant expansion in revenue generation could be expected. The budget proposals also suggest tapping new sources of revenue, as shown in Table 7.

A breakdown of Revenues is shown in Table 8.


Around LL387bn in receipts arise from new taxes and fees, constituting 7.8 per cent of revenues. Tax revenues are estimated at LL3.63trn, an increase of 15 per cent over the budgeted figure for 1997 (see Table 3 ). Non tax revenues are forecast to reach LL1.33 trn in 1998, reflecting a 46.5 per cent rise compared to the 1997 budget.

A more detailed analysis of tax revenue reveals the following:

* Revenue from income and profit taxes is estimated at LL440bn, representing 12 per cent of total tax revenue.

* Revenue from sales tax on goods and merchandise is a new source of revenue estimated at LL240bn, representing 6.6 per cent of total tax revenue.

* Revenue from property taxes is estimated at LL470bn, representing 13 per cent of total tax revenue.

* Revenue from indirect taxes on commodities is estimated at LL286bn, or eight per cent of total tax revenue.

* The major source of revenue stems from taxes imposed on international trade, estimated at LL1.95trn and constituting 53.8 per cent of tax revenues or 39.3 per cent of total revenues.

* Taxes from other sources stand at LL230bn, which represents 6.3 per cent of total tax revenue.

* Revenues from public boards, governmental institutions, and government property drastically increased from LL374bn to LL704bn.

* Revenues from administrative fees drastically increased from LL230bn to LL428bn, of which traffic law infringement fines increased from LL115bn to LL275bn.

* No revenues from the sale of public assets and settling of loans are expected, compared to LL40bn in 1997 budget.


2. Expenditures

On the expenditure side, budget expenditures are expected at LL7,375bn. However, after adding both of the amounts expected to be spent from amounts carried over from 1997 and the amounts expected to be allotted to municipalities, total gross expenditures estimate amounts to LL7,925bn. In the draft budget, current expenditures are estimated at LL6,588bn and investment expenditures LL703bn, representing 89 per cent and 11 per cent of proposed expenditure respectively. The breakdown of the anticipated spending allotted to municipalities and the amounts expected to be spent from amounts carried over from 1997 not available.

The proposed public spending for 1998 is shown in Table 9. A breakdown of Expenditures is shown in Table 10.

The main components of expenditure by type were sub-divided as follows:

* Expenditures on wages, salaries and indemnities are estimated at LL2,492bn, representing 37.8 per cent of current expenditures, or 31.4 per cent of total expenditures.
* Debt service constitutes LL3,200bn, or 40.4 per cent of budget expenditures, of which 91 per cent are domestic interest payment.
* Expenditure on goods and services totaled LL234bn, subdivided into LL136bn for goods and LL98bn for services. This represented three per cent of expenditures.
* Investments expenditure in Lebanese pounds totaled LL703bn of which LL386bn was allocated to infrastructure and other projects; LL66bn went to equipment and supplies; LL74bn to maintenance; and LL34bn to miscellaneous expenditures and 143bn in support of CDR.

A more detailed analysis of the breakdown of expenditures by ministries reveals that:

* Spending by the Ministry of Defense increased by three per cent, from LL738bn to LL760bn, accounting for 18 per cent of total expenditures.

* Spending by the Council of Minister rose by 16 per cent, from LL408.7bn to LL475bn, accounting for 11.4 per cent of total expenditures.

* Spending by the Ministry of Education increased by 21 per cent, from LL388bn to LL470bn, accounting for 11 per cent of total expenditures.

* Spending by the Ministry of Health increased by 69 per cent, from LL160bn to LL270bn, representing 6.5 per cent of total expenditures.




3. Overall position

There are several favourable features in the 1998 budget proposals and a few weak spots. The revenue estimate appears to be achievable, assuming a recovery in the economy, particularly in the real estate sector, and the spending forecast seems to be realistic enough to be adhered to.

The projected 25 per cent rise in revenues is expected to be underpinned by the enhancement of many revenue items and the introduction of new measures including the nucleus of a potential broad-based sales tax. It is assumed that all these measures would be in effect as of the beginning of 1998. In the aggregate they appear to be sufficient to promote revenue growth in line with the projected nominal expansion of the GDP (12 per cent), and to generate incremental receipts equivalent to about two per cent of the GDP. A main weakness in the revenue package lies in a partial dependence on raising fees, some of which are specific (i.e. specified as a fixed amount per transaction). Such fees grow with the volume of transactions and therefore tend to lag behind nominal GDP growth. These fees would need to be adjusted periodically.

On the expenditure side total spending is projected at LL7,925bn consisting of LL7375bn in the 1998 draft budget, LL350bn in unused allocations from previous years re-authorized for spending in 1998, and LL200bn authorized for outlays by the municipalities. Excluding the municipalities, the proposals appear to fulfill the recommendation of lowering the government's own expenditures by LL200bn. Including the municipalities, aggregate expenditure is essentially frozen at the estimated 1997 level. Considering that total outlays are estimated to have increased by over one third in 1996-1997, a cap on expenditures in 1998 would appear to be feasible. It would nevertheless require strict monitoring in order to assure continuing compliance. The weakness in the expenditure estimates might be in the projection for interest payments, which are capped at the expected 1997 level. To realize this estimate would require a significant decline in interest rates.

The projected overall fiscal deficit of LL3trn is about LL1trn less than the deficit estimated for 1997. In terms of ratios to GDP, the deficit is forecast to fall by six percentage points to just under the equivalent of 12 per cent of the GDP. That would represent the best fiscal performance since 1993. Impressive as this achievement would be if realized, the impact on the debt burden would be merely to stabilize the situation at an uncomfortably high level. While the gross debt ratio would decline significantly, it would be expected to remain around the 100 per cent mark; the net debt position would be expected to grow marginally to 92 per cent of the GDP. Clearly, there is no room for slippage in implementing the budget and reinforcement might well become necessary during the course of the fiscal year.




A Revised Framework for a Deficit Reduction Effort

Table 11 showing a summary of government finances provides two sets of projections for the years 1998-2001. The original projections (which were published earlier) have been revised to reflect the 1998 budget proposals. Below are listed the assumptions for the original and the revised projections.

    Original
    Revised
Objective
Fiscal balance in 2001
Same
GDP growth
12% in nominal terms
Same
Expenditures
Capped at LL 7.8 trillion annually
Capped at LL 8 trillion annually
Revenues
Grow in line with nominal GDP plus 4% of GDP additional
evenue effort in 1998.
Grow in line with nominal GDP plus 2% in 1998
and 1% additional in each of 1999, 2000, and 2001

In the first part of the LCPS report, which was published in October, it was stressed that these scenarios were merely intended as examples to clarify issues and possibilities. With this objective in mind it might be useful to compare the aggregated numbers over the four years in the two scenarios:

(In LL trillion)

Original Revised Difference (Revised minus Original)
Revenue +26.30 +25.63 -0.67
Expenditure -31.20 -31.94 -0.74
Deficit -4.90 -6.31 -1.41

Broadly, the revised scenario provides over the four years for some LL740bn in additional expenditures and generates some LL670bn less in revenues. The latter occurs despite the fact that the revenue effort in the revised scenario totals to the equivalent of five per cent of the GDP, but is spread over the four years; while the original scenario assumes four per cent of the GDP all built into the first year. The cumulative difference in the deficit of LL1.41trn gets reflected in the debt burden position in the year 2001. While the original scenario aimed for the 70-75 per cent of GDP range, the revised one projects the 75-80 per cent range.

The measures and policies required to underpin a deficit reduction effort, whether as outlined above in the original or the revised scenarios, are essentially the same. The options and possibilities were described in the first section of the report which was published earlier. It is clear that in Lebanon's circumstances the effort will have to be comprehensive, utilising a wide range of instruments, and cast in the context of the medium-term programme that is pursued with determination. Scenarios can be invented, developed, scrutinised, and adjusted. Their intent is to explore, enlighten, and stimulate discussion. But in the end the only scenario that really counts is the one articulated and implemented by the authorities.

A revised Summary of Government Finances for the years 1993- 2001 is shown in Table 11.


Back to Table of Contents | Economy Index | Research Projects