Israel: Selected Issues and Statistical Appendix

Contents include: Budget deficit - concepts, measurements, and developments; definition, accounting basis, scope, and coverage. Structural reform issues: privatization; capital market reform; exchange and trade liberalization; other government intervention - investment subsidies and price controls; labor market reform; and land reform. Current account sustainability in Israel: overview of the Israeli economy and its transformation in the 1990s; a model of the current account applied to Israel; and sustainability indicators. Statistical tables.


Contents include: Budget deficit - concepts, measurements, and developments; definition, accounting basis, scope, and coverage. Structural reform issues: privatization; capital market reform; exchange and trade liberalization; other government intervention - investment subsidies and price controls; labor market reform; and land reform. Current account sustainability in Israel: overview of the Israeli economy and its transformation in the 1990s; a model of the current account applied to Israel; and sustainability indicators. Statistical tables.


A. Introduction

To instill fiscal discipline, since 1992 the Israeli authorities have each year set a budget deficit target that was lower than that of the preceding year (though not necessarily lower than the preceding year’s actual outturn). These targets have been set, however, in terms of a specific deficit definition that is not entirely comparable to that employed in many other countries. Hence, for policy analyses, it is important to understand the nature of this definition and its implications for assessing fiscal developments.

Broadly speaking, three separate conceptual issues are involved in any analytical discussions of a deficit definition. In descending order of specificity, they are: (1) the accounting basis on which the deficit is to be measured, e.g., it could be measured on a cash or accrual basis; (2) the scope of the transactions for which the deficit is designed to capture, e.g., certain types of revenue and of expenditure could be excluded for specific reasons (such as capital revenue and expenditure, which would result in a deficit that corresponds to government/public sector dissavings; or interest revenue and expenditure, which would lead to a primary deficit); and (3) the coverage of the transactions (for a given scope) for which the deficit is to encompass, e.g., it could cover only transactions of the central government, or also those of extrabudgetary funds (such as social security), those of sub-central level governments (usually referred to, together with the central government, as the general government), and those of the nonfinancial state-owned enterprises (making up, together with the general government, the nonfinancial public sector).

The appropriate choice of a deficit definition depends to a large extent on the objective of the investigation—no one definition emerges conceptually as the best alternative under all circumstances. The following sections discuss in turn the basis, scope, and coverage of the deficit definition employed in Israel which, as it turns out, does not fall cleanly into any one of the conventional deficit definitions.2

B. Definition of the Deficit

For the purpose of setting the annual deficit targets, the deficit in Israel is defined as the outcome of a subset of transactions in the state budget. Specifically, (1) the targets for 1992-96 were set only for the domestic component of the state budget (targets for 1997-2001 would comprise both the domestic and foreign components); (2) net lending operations of the central government have been, and will continue to be, excluded from the deficit definition, although such operations are formally part of the state budget; and (3) the deficit definition also excludes proceeds from the sale of government assets.

The targets are stipulated as a percentage of GDP, which implies that the target in a given year could technically be met even if the actual deficit in nominal terms exceeded the budgeted level in that year, as long as the excess was more than compensated by higher-than-projected growth in nominal GDP. As shown in Table A24, while the (domestic) deficit target so defined declined from 6.2 percent of GDP in 1992 to 2.5 percent of GDP in 1996, both the actual outturn for 1995 and estimated outturn for 1996 exceeded their respective targets. The (total) deficit target for 1997 has been set at 2.8 percent of GDP, declining to 1.5 percent of GDP by 2001.

C. Accounting Basis of the Deficit

The accounts of the state budget are largely, but not entirely, cash-based. This ambiguity arises from the asymmetrical treatment of interest payments on indexed and nominal government debt. To fully understand the significance of the accounting convention adopted by the state budget, some conceptual discussion of the related issues of the cash versus accrual accounting for government finance, the conventional and operational concepts of budget deficit, and the interest payments on indexed and nominal debt is necessary.

Cash versus accrual accounting

In contrast to the generally accepted principles of accrual accounting as it applies to the accounts of a commercial enterprise, government accounts are almost always kept on a cash or payment basis, due to a variety of informational and data limitations. Largely for this reason, some statistical systems recommend reporting fiscal data on a cash/payment basis to avoid complexities arising from timing and valuation issues.3 An important shortcoming of cash-based accounting, however, is that it may result in data that do not reflect the true underlying economic situation in a sufficiently timely manner to be fully useful for policy analyses. There are, therefore, also statistical systems that recommend the adoption of accrual-based accounting for government accounts,4 but such an approach would typically involve spelling out a set of adjustment rules for accounts for which data on an accrual basis are not usually and readily available.5

A particularly important aspect of the issue of cash versus accrual basis as it relates to government accounts is the treatment of interest expenditure on government debt, as the distinction between interest payments (which are “above the line” transactions and, therefore, would affect the magnitude of the budget deficit) and amortization payments (which are recorded “below the line” and have no impact on the deficit) in any given period is not always straightforward, e.g., payments related to zero-coupon bonds.6 Matters become even more complicated when such payments are made under inflationary conditions, and/or when they are related to indexed debt.

Conventional versus operational deficit

Consider first the simpler case where all government bonds are nominal, i.e., nonindexed, with fixed periodic coupon payments over the life of the bonds. Under inflationary conditions, the nominal interest payment on outstanding debt overstates its service burden on the Government, as the real value of the debt stock is being eroded by inflation. Conceptually, it could, therefore, be argued that a portion of the nominal interest payment is really an implicit amortization payment (through inflation) that should be netted out from government expenditure above the line before computing the deficit. Carrying out such an adjustment would clearly result in a deficit, known as the operational deficit, that is smaller than that measured conventionally—either on a cash or accrual basis.7

While the conceptual underpinnings of the operational deficit are appealing, its application is not unequivocally meritorious under all circumstances. In particular, it may mask the true underlying budgetary imbalance and, consequently, understate the needed fiscal consolidation.8 Its use so far has not been recommended by any of the internationally recognized statistical systems, although a number of countries with high inflation do employ it.9

Nominal versus indexed bonds

An indexed bond is one whose coupon payments and/or redeemable principal value are linked to some price index, typically the CPI. If the indexation is perfect,10 the before-tax real yield of the bond is protected against inflation risk. Indexed government bonds will give rise to a different complication in measuring the budget deficit when government accounts are cash-based, however, because the intertemporal profile of cash flows from such bonds are conceptually arbitrary, depending on the chosen indexation scheme among many such schemes that are all equivalent in present value terms. In contrast, this complication does not arise with accrual-based accounting.

To see the nature of the above complication in simple analytical terms, consider first the case of no inflation (π = 0), in which case there is no difference between a nominal and an indexed bond. Let P be the bond’s redeemable principal value and r its coupon rate. For simplicity, assume that the bond matures in two periods and its current price is also P (i.e., r is the bond’s real yield). The coupon payment from such a bond in each of the two periods is simply rP9 implying a present value of P for the bond’s cash flow over its life, as expected. This is illustrated in the first row of Table 1.

Table 1.

Illustrative Cash Flows Under Alternative Indexation Schemes

article image
Source: Staff computations.

Consider now the case of positive inflation (π > 0), which is assumed to be constant over the two periods. For a nominal bond, its coupon rate (i) will rise above r to compensate for the expected inflation (πe), i.e., i = r + πe + r.πe (the discrete-time version of the Fisher equation). The coupon payment each period is now i.P, with the redeemable principal value remaining at P (second row in Table 1). The expected present value of this cash flow is Pe, which will be equal to P only if πe = π. Thus, the real yield of a nominal bond can be protected against inflation only when inflationary expectations are correctly realized. Clearly, Pe>P as π > π e.

For an indexed bond, however, there is no inflation risk, because its coupon payments and/or redeemable principal value are adjusted to actual, rather than expected, inflation.11 Two particularly illustrative indexation schemes are considered below, each guaranteeing a before-tax real yield of r in the face of positive inflation, but generating a different profile of cash flows.

Indexation of principal

The first scheme—adopted by almost all countries (including Israel) which issue indexed bonds—involves indexing the principal, with the coupon rate r remaining fixed. With a changing principal, however, the coupon payment also changes from one period to the next, i.e., an extra payment (at the rate r) is earned on the adjusted (larger) principal every period. The redeemable principal value is also adjusted upward by the compounded inflation rate (third row in Table 1). Note that, under this scheme, the actual coupon payment in each period still rises with inflation.

Indexation of coupon rate

Instead of indexing the principal, the second—and conceptually equivalent—scheme indexes the coupon rate, with the redeemable principal value P remaining fixed. This alternative essentially involves increasing the original coupon rate r by an adjustment factor ε in each period, so that the present value of the entire stream of cash flows is maintained at P. From the Fisher equation, this adjustment factor must clearly be ε = π.(1+r). Hence, the coupon payment in each period under this scheme involves the sum of two terms: the rate r on the per-period inflation-adjusted principal, and the compensation for the loss in the redeemable principal value due to inflation, i.e., π·P (fourth row in Table 1).

Comparative cashflows under alternative indexation schemes

Comparing the cash flows under the first indexation scheme with that under the second, it is easily seen that the former has a lower coupon payment in the first period but a higher redeemable principal value, than those under the latter. The relative magnitudes of the coupon payment in the second period between the two schemes are, however, ambiguous.12

As with the principal indexation scheme, the actual coupon payment in each period under the coupon rate indexation scheme also rises with inflation.13

Since the intertemporal profiles of cash flows under the two indexation schemes are different, the intertemporal profile of deficits computed on a cash basis will depend on the chosen indexation scheme, even though the schemes have equivalent present values.14 The arbitrary nature of this outcome could raise questions about the usefulness of cash-based deficits as a fiscal policy indicator when indexed bonds comprise a significant proportion of outstanding government debt. This potential shortcoming would not arise, however, with accrual-based deficits, because any change in the value of an indexed bond’s redeemable principal between the beginning and the end of a particular accounting period due to indexation is treated as interest accruing in that period, in addition to any interest actually paid.15 Hence, with accrual-based accounting, interest expenditure in a given period always corresponds to that of the coupon rate indexation scheme, regardless of the indexation scheme actually chosen, the intertemporal profile of deficits is, therefore, invariant to the choice of the indexation scheme.16

Interest expenditure in the state budget

According to the accounting convention adopted by the state budget at present, interest expenditure reflects the sum of: (1) actual interest paid on foreign debt; (2) actual interest paid on domestic indexed debt (based on the principal indexation scheme); and (3) real (i.e., nominal less inflation) interest paid on domestic nominal debt. Hence, technically, the computed deficit is partly cash-based (on account of the interest payments on foreign and domestic indexed debt) and partly operational (on account of the monetary correction on the interest payment on domestic nominal debt). In reality, however, the share of nominal debt in the total outstanding stock of government debt, while rising, is still minimal (accounting for only 4.4 percent of total domestic government debt at end-1995; see Table 2). For this reason, the measured deficit in the state budget can be regarded as largely cash-based.

Table 2.

Budget Deficits and Government Debt, 1994-95

article image
Sources: Data provided by the authorities; and staff computations.

Includes only that part of the deficit financed by debt.

Budget deficits and government debt

As shown in Table 2, the bulk (about 91 percent during 1994-95) of the domestic government debt (accounting for about three quarters of the total debt during the same period) is CPI-indexed.17 Hence, a large discrepancy inevitably arises each year between the change in the value of outstanding government debt and the amount of measured deficit in the state budget that is financed by the issuance of debt. Based on annual period-average data, the changes in the total government debt were about NIS 23.6 billion in 1994 and NIS 29.2 billion in 1995, which did not bear a meaningful relationship to the debt-financed deficit as recorded in the state budget of NIS -0.2 billion and NIS 10.4 billion in the corresponding periods.

In contrast, once the (largely) cash-based state budget deficits were converted to an accrual basis, the deficits turned out to be about NIS 23.5 billion in 1994 and NIS 27.7 billion in 1995—amounts that corresponded closely to the changes in total government debt in the same periods.18

D. Scope of the Deficit

As noted earlier, until 1996 three types of transactions in the state budget have been excluded from the definition of the deficit when setting deficit targets, although one—the foreign component—will be included beginning with the 1997 budget. The foreign component of the state budget primarily consists of foreign grants on the revenue side (foreign revenue averaged about 4.6 percent of GDP, or 12 percent of total state budget revenue, during 1992-96; see Table A25) and foreign interest and defense imports on the expenditure side (foreign expenditure averaged about 4.7 percent of GDP, or 10 percent of total state budget expenditure, during the same period; see Table A26). While the average deficit in the foreign component has been relatively low as a share of GDP during the past few years, there has been a discernible upward trend in the foreign imbalance since 1993,19 which is expected to reach 0.8 percent of GDP in the 1997 budget (Table A26). As the foreign balance has an impact on the domestic economy just like the domestic balance, excluding the former in policy formulation and implementation would produce an incomplete picture of the true stance of fiscal policy.

The appropriateness of excluding net lending operations in the state budget when setting deficit targets is less clear cut. While the GFS recommends that they be classified as expenditure (thus they would have an impact on the measured deficit), on the ground that they involve transactions in claims upon others acquired for purposes of public policy, it also recognizes the potential usefulness in grouping them with financing (thus they would have no impact on the measured deficit) to measure the Government’s net financial position.20 Regardless of the procedure adopted, the quantitative impact of their exclusion on the measured deficit has been minimal since 1994 (see Table A27).

The third type of transactions excluded from the definition of the deficit for target-setting purposes concerns the proceeds from the sale of government assets, which, under the GFS, such proceeds are classified as revenue. The magnitude of these proceeds has been somewhat erratic from year to year in the recent past, but is budgeted to be about 1 percent of GDP in both the 1996 and 1997 budgets. Given its relatively unpredictable nature, its exclusion from the deficit definition seems justified.21

E. Coverage of the Deficit

Technically the state budget is the budget only of the central government. In practice, however, the substantial degree to which the finances of the National Insurance Institute (NII), the local authorities, and other public nonprofit institutions rely on unilateral transfers from the central government renders the overall balance of the state budget not very far from that of the general government on a cash basis, even though the revenue received and expenditure undertaken by these latter units are not properly consolidated with the accounts of the state budget. While the following discussion will focus on the financial interactions between the state budget and the other aforementioned entities on a cash basis, fully consolidated revenue, expenditure, and overall balance of the general government are available (albeit with a longer time lag) on a national accounts basis (see Tables A28-A31).22 On this basis, the general government deficit was about 3.4 percent of GDP, compared to the state budget deficit of about 4.2 percent of GDP, in 1995. This discrepancy was likely to be accounted for mostly by the net deficit (i.e., deficit after unilateral transfers from the state budget) of the local authorities. Data on the net deficit of the local authorities in 1995 are not available, but during 1992-94 it averaged about 1.5 percent of GDP (Table A8).

National Insurance Institute

The NII administers a whole host of social security programs, some of which fall under its own mandate, such as the old age/survivors pensions and child allowances, while others it administers on behalf of the Government, such as a variety of welfare benefits for special population groups. Since 1995, the NII has also been required by law to collect a new health tax from employees and distribute the revenue, together with additional budget funding, to four health funds responsible for the provision of a standardized basket of health services to Israeli residents. Previously, participation in the health funds was voluntary and contributions were made on a private-billing basis.23 In 1995, about two thirds of NII’s current revenue was derived from contributions in one form or another from the state budget; contributions from the public (i.e., national insurance taxes from the employers and employees, the new health tax, and the parallel tax) accounted for the rest.

While the NII’s revenue and expenditure accounts are not consolidated with those of the state budget, the impact of the former’s current transactions on the latter is fully reflected in the overall balance of the state budget each year. This is because, in addition to explicit budgetary transfers to the NII (5.8 percent of GDP in 1995; see Table A26), the state budget includes as revenue loans from the NII (1.9 percent of GDP in 1995; see Table A25) and as expenditure repayment of loans to the NII (1.1 percent of GDP in 1995; see Table A26). Since by law cash surpluses of the NII are lent to the Government (in 10-15 years indexed bonds at preferred interest rates of 4.5-5 percent presently), the divisions among loans, repayment of loans, and budgetary transfers are somewhat arbitrary. Hence, the net burden on the state budget of NII operations, which is included in the state budget’s deficit calculations, was about 5 percent of GDP in 1995.24

Local authorities

Unlike the NII, local authorities, in addition to revenues from property taxes (known as arnona), an assortment of fees, and transfers from the state budget, have access to bank credits to finance expenditure. Hence, the impact of the operations of local authorities on the balance of the state budget is only partial—the impact being limited to the extent of explicit budgetary transfers (2.2 percent of GDP in 1995; see Table A26)25

National and nonprofit institutions

In addition to obtaining sizable transfers from the state budget (5.6 percent of GDP in 1995; see Table A26), these institutions also receive grants directly from foreign sources (0.8 percent of GDP in 1995; see Table A29). While little information is available concerning their financial operations, data on a national accounts basis indicate that the net financial imbalance (after budgetary transfers) of these institutions as a whole is relatively insignificant.

F. Concluding Remarks

The above suggests that the comprehensiveness of the definition of the deficit for target-setting purposes has up till now suffered from three potentially significant limitations: (1) cash-based accounting has excluded the accrued interest liabilities on the stock of outstanding indexed government debt; (2) focus on the domestic component of the state budget leaves the deficit definition incomplete in its scope in capturing all relevant transactions; and (3) availability of bank credits to local authorities renders the state budget incomplete in its coverage of the relevant activities of the general government that are important in assessing the stance of fiscal policy. As noted earlier, the authorities have already decided to remove the second limitation beginning with the 1997 budget and taken steps to address the third limitation. Quantitatively, however, a conversion from the cash- to accrual-based accounting—insofar as interest expenditure is concerned—to remove the first limitation would have the greatest impact among the three limitations on the measured deficit.


  • Eurostat, European System of Accounts (Luxembourg 1995).

  • Eurostat, International Monetary Fund, Organization for Economic Cooperation and Development, United Nations, and World Bank (EIOUW) System of National Accounts (Brussels/Luxembourg, New York, Paris, and Washington 1993).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund (IMF), A Manual on Government Finance Statistics (Washington 1986).

  • Levin, JonathanGovernment in the 1993 System of National Accounts” in John W. Kendrick (ed.), The New System of National Accounts (Boston: Kluwer Academic Publishers 1995).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, Mario I. Blejer, and Mario O. Teijeiro,Effects of Inflation on Measurement of Fiscal Deficits: Conventional Versus Operational MeasuresStaff Papers Vol. 34 (1987) pp. 71138.

    • Search Google Scholar
    • Export Citation

Prepared by Howell Zee.


For a general taxonomic discussion of alternative deficit concepts, see IMF (1985).


For example, this is the approach adopted by the IMF’s A Manual on Government Financial Statistics (GFS). See IMF (1986).


Examples include the new international System of National Accounts (SNA) and the European Union’s European System of Accounts (ESA). See EIOUW (1993) and Eurostat (1995), respectively.


For a comparative discussion of the GFS and SNA as they pertain to government statistics, see Levin (1995).


In a cash-based system such as the GFS, interest payments on zero-coupon bonds are recorded as the difference between the issuance and redemption prices only at the time the bonds are redeemed. In contrast, an accrual-based system such as the ESA would spread the difference over the life of the bonds as if coupon payments were actually paid and reinvested by the bond holder, thus also necessitating adjustment entries to the value of bonds outstanding in the accounts of both the issuer and holder of such bonds.


In practice, the operational deficit is usually derived by including only the “real” interest payment above the line, the real interest rate being the nominal rate minus the rate of inflation. This procedure may produce a negative real interest rate, however, if the actual nominal interest rate for some reason does not fully compensate for the effects of inflation. The conceptual significance of such an occurrence is extensively discussed in the SNA.


For an extended discussion of the merits and shortcomings of the concept of operational deficit, see Tanzi et al. (1987).


The SNA does recognize, however, the usefulness of the concept of operational deficit in computing meaningful measures of income and expenditure in “high” inflationary conditions, and, therefore, does not object to its use under such circumstances.


This will be assumed for simplicity. In reality, perfect indexation is, of course, not always achieved, as the price index to which the bond is linked could be a biased indicator of inflation, or the linkage mechanism could have a lagged structure.


The absence of inflation risk with respect to an indexed bond refers to its before-tax yield. Inflation risks continue to exist with respect to its after-tax yield, however, as long as the tax system is not entirely inflation proof.


Specifically, in the second period, the coupon payment under the first scheme would be higher (lower) than that under the second if r(l + 0) is greater (less) than unity.


Note that the cash flows under this indexation scheme are identical to that of the nominal bond, when the inflation rate is correctly anticipated. Hence, indexing the coupon rate is analytically identical to the case of the nominal bond with a fully floating coupon rate.


As a third indexation example, a zero-coupon indexed bond would generate still another intertemporal profile of deficits.


This is explicitly provided for in the ESA.


The counterpart entry below the line is increased borrowing in an amount equal to the accrued interest so calculated, i.e., the accrued interest is treated as if it had actually been paid to, and reborrowed from, the bond holder.


The rest of the domestic debt not accounted for by CPI-indexed and nominal debt comprised debt indexed (or optionally indexed) to the U. S. dollar.


The small differences between the changes in the government debt and the accrual-based deficits are probably due to the fact that the former had been calculated on the basis of annual period-average data, while the latter had been calculated on the basis of a month-by-month debt conversion.


The projected improvement in 1996 is entirely due to an unexpected delay in the receipt of a grant from the United States from the 1995 to the 1996 budget year.


Both the SNA and ESA classify net lending operations of the Government as financing items.


Proceeds from the sale of government assets are treated in the same way as net lending operations in the SNA and ESA.


No data are available on consolidated accounts of the entire nonfinancial public sector.


In contrast to employees, employers have long been required to contribute to the health funds through a special payroll tax (known as the “parallel” tax); this tax continues to exist under the new regime.


This figure somewhat understates the burden, because it excludes the interest payment on Nll loans, which amounted to about 0.9 percent of GDP in the 1995 budget.


Since 1994 the Government has continuously sought to put the public finances of local authorities on a sounder and more transparent basis, including drawing up explicit criteria for allocating transfers and determining local taxing powers. This process is still on-going.