The West Bank and Gaza faces considerable challenges over the medium term, and the fiscal policy of the Palestinian Authority (PA) can play an important role in overcoming them. For example, Chapter 2 discussed how difficult it will be to achieve and sustain economic growth rates high enough to reduce unemployment in a period of rapid labor force growth. By avoiding large debt-financed deficits and improving the composition of expenditure, fiscal policy can make a great contribution to this endeavor. Just as fiscal policy can play an active role in shaping the West Bank and Gaza’s economic prospects, however, it will in turn be affected by several medium-term developments, such as the very same demographic dynamics mentioned above. Apart from the expected effects of population growth, this chapter discusses briefly two further factors that can have particularly significant consequences for public finances: the future choices of tax and trade policy regimes; and a permanent solution to the refugee question.
With respect to the demographic dynamics, the West Bank and Gaza is projected to undergo important changes over the next 10 years. The annual rate of population growth is projected to slow to 3.2 percent by 2010 from about 3.8 percent in 1999, while the working-age population is expected to continue to grow annually at about 3.8 percent on average, and the labor force at 4.4 percent a year. Most importantly, these changes would result in a significant rise in the share of the population in the labor force with far-reaching macroeconomic implications (see Chapter 2). Should the economy be able to absorb the additional labor, the next section argues that the impact of these developments could be unambiguously positive for fiscal revenue, while it is less clear for expenditure; bearing this caveat in mind, therefore, the demographic dynamics in the West Bank and Gaza may influence public finances positively, suggesting a window of opportunity for fiscal consolidation. Moreover, if the PA were to maintain current spending levels, but opted for using the additional room for changing its relative spending priorities, the potential demographic dividend could put the PA in a position to make a greater contribution to economic development. On the revenue side, it could rely less on import duties for revenue purposes and adopt an open trade regime with a low uniform tariff across the board. On the expenditure side, higher social spending on a per capita basis and a greater contribution to infrastructure investment may become possible without increasing total outlays as a share of GDP.
There is nothing inevitable about such an outcome, however, and it hinges on high growth in the Palestinian economy so that the expected inflow to the labor market can be absorbed at reasonable real wages. Should the economy falter, the blessing may well turn into a curse. Our analysis constitutes, moreover, a comparative static exercise isolating the impact of demographics alone, while assuming all other factors to be constant. In the end, even if it were to materialize, the positive influence of a larger workforce may be more than offset by policy decisions, which have a much stronger and more direct impact on fiscal revenue and expenditure in the medium term. This highlights how important it is for the PA to form a view already now on the trade and tax regimes it would want to implement in the future, once a permanent status agreement has been reached with Israel.
Fiscal Policy and Projected Demographic Dynamics
The Impact of Demographics on Fiscal Revenue
Public finances are highly sensitive to demographic changes, as shown by the experience in both developed and developing countries. Table 5.1 summarizes the demographic projections on which the analysis in this chapter is based. The projected demographic transition in the West Bank and Gaza can influence fiscal revenue both directly and indirectly. First, a larger workforce, if employed, creates a larger tax base, boosting both direct and indirect tax receipts. Second, an increase in the workforce relative to the nontax-paying population (or a decline in the dependency ratio) could bolster per capita GDP growth, which in turn may stimulate taxable income and tax revenue (see Chapter 2).1 Finally, tax receipts could be further enhanced should demographic pressures lead to increasing urbanization and a shift away from hard-to-tax sectors such as agriculture. In particular, direct taxes may become relatively more important vis-à-vis indirect taxes (see, for example, Tanzi, 1987).
The Underlying Demographic Projections
Migration scenario.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (between 15 and 65)
The Underlying Demographic Projections
| 1999 | 2010 | 20101 | |
|---|---|---|---|
| Total population (in millions) | 2.8 | 4.0 | 4.6 |
| Share of population under 15 (in percent) | 47.0 | 44.0 | 43.0 |
| Share of population over 65 (in percent) | 3.4 | 2.9 | 1.4 |
| Share of working age population (in percent) | 49.7 | 52.9 | 53.8 |
| Age dependency ratio2 | 1.0 | 0.9 | 0.9 |
Migration scenario.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (between 15 and 65)
The Underlying Demographic Projections
| 1999 | 2010 | 20101 | |
|---|---|---|---|
| Total population (in millions) | 2.8 | 4.0 | 4.6 |
| Share of population under 15 (in percent) | 47.0 | 44.0 | 43.0 |
| Share of population over 65 (in percent) | 3.4 | 2.9 | 1.4 |
| Share of working age population (in percent) | 49.7 | 52.9 | 53.8 |
| Age dependency ratio2 | 1.0 | 0.9 | 0.9 |
Migration scenario.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (between 15 and 65)
Some Simple Accounting
A simple and intuitive way to see how demographic changes can affect revenue is to look at fiscal revenue in per capita terms. Typically, fiscal revenue is expressed as a percentage of GDP, and this provides a useful indication of the fiscal burden and the size of the government. But looking at fiscal variables in per capita terms is more relevant if the starting point is the desire to be able to spend certain amounts of fiscal resources on a per capita basis (for example, education expenditure per pupil and health expenditure per capita). The revenue to population ratio can be decomposed as follows:
The first item on the right-hand side is the revenue-GDP ratio, which can be viewed as the effective tax rate. The second item can be seen as average labor productivity, and the last item is the workforce-population ratio. This is an accounting expression that does not explain interlinkages among the various items. Nevertheless, it is immediately obvious that an increase in the workforce relative to the population at large would lead to an increase in fiscal revenue per capita for a given rate of labor productivity and a given average tax. In reality, of course, the other items are not given. As discussed in Chapter 2, an increase in the workforce-population ratio (as expected in the West Bank and Gaza) can be expected to raise per capita real GDP growth, reinforcing the effect on revenue, and as discussed below, tax revenue has typically been found to grow faster than GDP, further reinforcing the positive revenue effect.
Preliminary data put PA fiscal revenue in 2000 at NIS 3.8 billion (22 percent of GDP), which comes to about NIS 1,340 per capita and about NIS 7,207 per worker.2 Table 5.2 shows the estimates for the West Bank and Gaza using equation (1). Taking into account only the projected increase in the workforce-population ratio, PA fiscal revenue would rise from NIS 1,340 to NIS 1,581 on a real per capita basis by 2010—an 18 percent increase (2010a in Table 5.2). As was seen in Chapter 2, however, a rising workforce-population ratio can be expected to have positive effects on investment and growth, and if we assume that real wages grow by a 1.5 percent per year (the medium scenario in Table 2.9 in Chapter 2), PA fiscal revenue would rise to NIS 2,067 per capita—a 54 percent increase in real terms (2010b). This means that, under these growth scenarios, even if PA fiscal revenue were to remain constant at around 22 percent of GDP between 2000 and 2010, PA fiscal spending on a per capita basis could increase sharply. Alternatively, if the PA wanted to limit the increase in revenue (and expenditure), to say, NIS 1,581 per capita (in real terms), the revenue-GDP ratio could be reduced from 22 percent of GDP in 2000 to 17 percent of GDP in 2010 (2010c).
Demographic Dynamics and Fiscal Revenue, 2000–101
In real terms.
Demographic Dynamics and Fiscal Revenue, 2000–101
| Revenue | = | Revenue | X | GDP | X | Workforce | |
|---|---|---|---|---|---|---|---|
| Population | GDP | Workforce | Population | ||||
| 2000 | 1,340 | 0.217 | 32,972 | 0.19 | |||
| 2010a | 1,581 | 0.217 | 32,972 | 0.22 | |||
| 2010b | 2,067 | 0.217 | 43,089 | 0.22 | |||
| 2010c | 1,581 | 0.167 | 43,089 | 0.22 | |||
| 2010d | 2,673 | 0.282 | 43,089 | 0.22 |
In real terms.
Demographic Dynamics and Fiscal Revenue, 2000–101
| Revenue | = | Revenue | X | GDP | X | Workforce | |
|---|---|---|---|---|---|---|---|
| Population | GDP | Workforce | Population | ||||
| 2000 | 1,340 | 0.217 | 32,972 | 0.19 | |||
| 2010a | 1,581 | 0.217 | 32,972 | 0.22 | |||
| 2010b | 2,067 | 0.217 | 43,089 | 0.22 | |||
| 2010c | 1,581 | 0.167 | 43,089 | 0.22 | |||
| 2010d | 2,673 | 0.282 | 43,089 | 0.22 |
In real terms.
It has been assumed thus far that the revenue-GDP ratio would remain constant at 22 percent (except for 2010c), but there is international evidence that the elasticity of fiscal revenue with respect to real GDP growth is greater than one.3 This is often referred to as tax buoyancy. It is inherently difficult to estimate a future rate of tax buoyancy for any country, and in the case of the West Bank and Gaza, this exercise is complicated further by the short and multifaceted history of the PA tax administration.4 Nevertheless, with due qualifications, and purely for illustrative purposes, the computed buoyancy for 1998–99 is 1.2.5 Thus allowing for revenues to increase at a higher rate than GDP would render a projected revenue-GDP ratio of 28 percent in 2010 equivalent to NIS 2,673 per capita (2010d in Table 5.2). Such an increase is far from automatic, however, and may hardly be desirable, as the tax burden in the West Bank and Gaza is already relatively high, given its level of development. The scenarios provided just serve to illustrate the dynamic potential of demographic change and the room it creates for reshuffling expenditure priorities and redesigning the tax system. A much more clean-cut approach to isolate the effect that demographic changes may have on fiscal revenue as a percentage of GDP is to use regression analysis.
Regression Analysis
Two simple cross-country regressions are estimated to obtain the elasticity of tax revenue with respect to the age dependency ratio and the size of the working age population, while controlling for other factors.6 There are obviously several caveats to this analysis, some of which will be analyzed in greater detail below. Importantly, the overall structure and efficiency of the tax system and other key dynamic factors, which may be driving revenue in the medium term, are assumed to remain unchanged. As discussed later, this will require a major effort on the part of the PA, if and when it assumes responsibility for collecting revenue now collected by Israel on its behalf under the customs union. Furthermore, future tax revenue will depend on policy decisions about the tax regime, with respect to tax levels and exemptions. For example, the 1998 Law on the Encouragement of Investment provides rather generous tax incentives, and their possible fiscal implications are not clear. Nevertheless, bearing these caveats in mind, the results indicate that a lower dependency ratio and larger workforce could, all else being equal, have a positive effect on revenue. The exact magnitude of these effects can be computed for the West Bank and Gaza by combining the estimated elasticities from these two regressions with the population growth projections from Chapter 2. As Table 5.3 shows, the projected decline in the age dependency ratio7 and the increased working-age population could add 1.3 and 1.2 percentage points of GDP to the revenue-GDP ratio, respectively, by 2010.
Estimated Impact of Demographic Variables on Revenue, 1999–20101
As a percentage of GDP.
Dependent variable is tax revenue/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Estimated Impact of Demographic Variables on Revenue, 1999–20101
| Regressor | Sign | Base Value (1999) | Estimated Value (2010) | |
|---|---|---|---|---|
| IMF staff estimate2 | Age dependency3 | — | 20.7 | 22.0 |
| Working-age population | + | 20.7 | 21.9 |
As a percentage of GDP.
Dependent variable is tax revenue/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Estimated Impact of Demographic Variables on Revenue, 1999–20101
| Regressor | Sign | Base Value (1999) | Estimated Value (2010) | |
|---|---|---|---|---|
| IMF staff estimate2 | Age dependency3 | — | 20.7 | 22.0 |
| Working-age population | + | 20.7 | 21.9 |
As a percentage of GDP.
Dependent variable is tax revenue/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
From these exercises it becomes clear that, provided the Palestinian economy will be able to maintain the high growth rates necessary, demographic developments could have the potential to boost PA tax revenue over the medium term—certainly in per capita terms, but also as a percentage of GDP—both through its direct and indirect effects.
The Impact of Demographics on Fiscal Expenditure
As in the case of fiscal revenue, there are direct and indirect ways in which demographic dynamics may affect fiscal expenditure. A larger population will demand more public services and perhaps public sector employment. The latter might become a particular problem should private sector job creation turn out to be inadequate. On the other hand, should the proportion of retirees decrease over the medium term, as has been projected in the West Bank and Gaza, this could reduce demand for health and pension expenditure. Similarly, the expected decline in the school-age population could alleviate the pressure on education spending. Thus, in contrast to the revenue side, where demographic change appears to have a rather unambiguously positive impact, the direct effect on expenditure is less clear.
Regarding indirect influences, government consumption as a share of GDP is generally observed to increase with per capita income, a phenomenon termed Wagner’s Law. The channels here are higher costs of regulation, more generous provision of welfare services, and the need for improved infrastructure in a more advanced economy (see, for example, Payne and Ewing, 1996).8 Therefore, should a larger workforce have positive growth effects in the West Bank and Gaza, as projected in Chapter 2, it may lead to higher expenditure of the PA.9 Urbanization, population density, and infant mortality—all phenomena to some extent driven by demographic changes—have been shown to influence expenditure as well (Heller and Diamond, 1990; Rodrik, 1996). The empirical investigation of the impact of demographics on aggregate expenditure, however, focuses on a few key variables, and given the sensitivity of education, health, and pension outlays to changes in the age-structure of the population, expenditure in these sectors is analyzed in greater detail.
Again, the most direct way of quantifying the relationship between demographic changes and public expenditure is to run a regression across a large sample of developed and developing countries. Regressing total expenditure as a percentage of GDP on per capita GDP, and either the age dependency ratio or the share of the working-age population has less explanatory power than the equivalent regression for tax revenue. Nonetheless, the results, which are summarized in Table 5.4, confirm the general trends just noted. The projected decline in the age dependency ratio between 2000 and 2010 could reduce total expenditure by 1 percentage point of GDP by 2010, whereas the growing working-age population would produce a reduction of 0.1 percentage points of GDP by 2010. These results are smaller than those found in other studies. For example, using the estimates from Rodrik (1996) and Commander, Davoodi, and Lee (1997), the expected decline in the age dependency ratio in the West Bank and Gaza could knock off 1.8 to 1.3 percentage points, respectively, of the expenditure-GDP ratio by 2010. Another explanatory variable used by Heller and Diamond (1990) is the share of the population over 65. Applying their coefficient to the West Bank and Gaza reduces the projected expenditure by 0.3 percentage points of GDP by 2010. As changes in the age structure of the population, such as the share of the working-age population and the population over 65 years, are closely related to the age dependency ratio, these estimates cannot simply be added up. Therefore, we take the average of our results with respect to the age dependency ratio; it can be expected to capture most of the underlying demographic dynamics. This suggests that demographic factors could allow total expenditure to decline by about 1.4 percentage points of GDP by 2010.10
Estimated Impact of Demographic Variables on Expenditure, 1999–20101
As a percentage of GDP.
Recurrent expenditure/GDP.
Dependent variable is total expenditure/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Dependent variable is total current expenditure/GDP (for 1999 total expenditure equals current expenditure of the PA).
Estimated Impact of Demographic Variables on Expenditure, 1999–20101
| Regressor | Sign | Base Value (1999)2 | Estimated Value (2010) | |
|---|---|---|---|---|
| IMF staff estimate3 | Age dependency3 | + | 22.3 | 21.3 |
| Working-age population | – | 22.3 | 22.2 | |
| Rodrik (1996)5 | Age dependency4 | + | 22.3 | 20.5 |
| Commander, Davoodi, and Lee (1997) 5 | Age dependency4 | + | 22.3 | 21.0 |
| Heller and Diamond (1990) 5 | Population share > 65 years old | + | 22.3 | 22.0 |
As a percentage of GDP.
Recurrent expenditure/GDP.
Dependent variable is total expenditure/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Dependent variable is total current expenditure/GDP (for 1999 total expenditure equals current expenditure of the PA).
Estimated Impact of Demographic Variables on Expenditure, 1999–20101
| Regressor | Sign | Base Value (1999)2 | Estimated Value (2010) | |
|---|---|---|---|---|
| IMF staff estimate3 | Age dependency3 | + | 22.3 | 21.3 |
| Working-age population | – | 22.3 | 22.2 | |
| Rodrik (1996)5 | Age dependency4 | + | 22.3 | 20.5 |
| Commander, Davoodi, and Lee (1997) 5 | Age dependency4 | + | 22.3 | 21.0 |
| Heller and Diamond (1990) 5 | Population share > 65 years old | + | 22.3 | 22.0 |
As a percentage of GDP.
Recurrent expenditure/GDP.
Dependent variable is total expenditure/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Dependent variable is total current expenditure/GDP (for 1999 total expenditure equals current expenditure of the PA).
In the end, fiscal expenditure is determined by discretionary policy decisions, and the most the reader should take away from this analysis is that the projected demographics would—if economic conditions turn out to be favorable—at least not add further expenditure pressure on the budget. Indeed, one interpretation can be that, for a given level of aggregate fiscal expenditure, there would be room for the PA to improve the composition of fiscal expenditure. As will be elaborated later, issues that have to be addressed in the medium term include the need for the PA to contribute to public investment with its own resources, and the requirement to rebalance recurrent expenditure. Should demographic changes allow for a reduction in expenditure, at least some of this room should be used to increase spending on certain priority areas within the budget, such as social outlays and the development budget. A closer look at the education and health sectors confirms these insights, it should be emphasized, however, that medium-term expenditures may well take a different path if unemployment increases and fiscal timebombs, such as the financially unsustainable pension system, fail to be addressed.
Effects on Expenditure in the Health Sector
Two short scenarios serve to illustrate the impact of demographic changes on health expenditure in the medium term. First, a rough estimate of overall health expenditure is derived on the basis of expected population growth and current health spending per capita. Second, health outlays are estimated as a percentage of GDP for 2010 using Heller and Diamond’s (1990) coefficient estimates together with the projected developments in the age structure and infant mortality for the West Bank and Gaza.
Health spending per capita by the PA in 2000 is estimated at NIS 142 (Table 5.5). If this level of spending were to be maintained in real terms through 2010, given the projected population growth and using the medium economic growth scenario in Chapter 2 (Table 2.9), the PA health budget would decline from around 2.3 percent of GDP in 2000 to 1.7 percent of GDP in 2010.11 In reality, though, the demand for health expenditure is unlikely to be invariant to changes in, for example, the age structure, labor force participation, or GDP per capita. Specifically, per capita health spending can be expected to increase with per capita GDP and the old age dependency ratio, and decrease with a rise in female labor force participation rates (see, for example, Kornai and McHale, 2000).12
Projections for the Health Sector
Based on population projections, Chapter 2.
In 1999 prices.
Projections for the Health Sector
| 2000 | 2010 | ||
|---|---|---|---|
| Total population (millions)1 | 2.9 | 4.0 | |
| Per capita health expenditure (NIS)2 | 142.0 | 142.0 | |
| Total health expenditure (NIS millions)2 | 407.0 | 567.0 | |
| In percent of GDP | 2.3 | 1.7 | |
Based on population projections, Chapter 2.
In 1999 prices.
Projections for the Health Sector
| 2000 | 2010 | ||
|---|---|---|---|
| Total population (millions)1 | 2.9 | 4.0 | |
| Per capita health expenditure (NIS)2 | 142.0 | 142.0 | |
| Total health expenditure (NIS millions)2 | 407.0 | 567.0 | |
| In percent of GDP | 2.3 | 1.7 | |
Based on population projections, Chapter 2.
In 1999 prices.
In a study of the composition of government expenditure in developed and developing countries, Heller and Diamond (1990) regress health expenditure on various demographic variables. They find that a lower percentage of the population over 65 years could lead to a decrease in aggregate and per capita health expenditure.13 On the other hand, reductions in the share of the population under 15 years of age and the infant mortality rate are associated with higher health expenditure.14 In the West Bank and Gaza, the relative shares of each of these age groups are expected to diminish over the medium term, and the Ministry of Health projects a lower infant mortality rate by 2010. Hence, these developments may end up largely offsetting each other’s impact on health expenditure. On balance, therefore, using the coefficients in Heller and Diamond (1990), health expenditure would result in only a modest increase in from 2.2 percent of GDP in 2000 to 2.3 percent of GDP in 2010.
Effects on Expenditure in the Education Sector
For the education sector, a similar exercise can help to illustrate future spending paths. Total education expenditure in 2010 is projected on the basis of the expected number of pupils and the 2000 level of expenditure per pupil. Taking into consideration the projected population between six and seventeen years of age, planned enrollment rates, and holding the percentage of pupils attending government schools constant at 70 percent, the total number of pupils in government schools would rise from 618 thousand in 2000 to 886 thousand in 2010.15 Keeping per pupil expenditure at its 2000 level in real terms leaves a projected total education expenditure in 2010 of 4–2 percent of GDP, compared with 5.5 percent of GDP in 2000 (Table 5.6). Within these aggregates, recurrent expenditure in the education sector would decline to 3.3 percent of GDP in 2010, compared with 4.3 percent of GDP in 2000.
Projections for the Education Sector
Projected on the basis of population projections, planned enrollment rates and slight increase over current share of government schools (70 percent).
Ministry of Education five-year education plan.
Held constant at 1999 level.
At 1999 prices.
Projections for the Education Sector
| 2000 | 2010 | ||
|---|---|---|---|
| Total number of pupils in government schools (1000s)1 | 618.0 | 886.0 | |
| Enrollment rates (ages 6–14)2 | 0.9 | 0.9 | |
| Pupil-teacher ratio3 | 24.0 | 24.0 | |
| Student-classroom ratio3 | 41.0 | 41.0 | |
| Salary costs (NIS millions) | 718.0 | 1522.0 | |
| Expenditure per pupil (NIS)3 | 1,555.0 | l,555.0 | |
| Total education expenditure (NIS millions)4 | 4,754.0 | 1,097.0 | |
| In percent of GDP | 5.5 | 4.2 | |
| Of which: recurrent expenditure | 4.3 | 3.3 | |
Projected on the basis of population projections, planned enrollment rates and slight increase over current share of government schools (70 percent).
Ministry of Education five-year education plan.
Held constant at 1999 level.
At 1999 prices.
Projections for the Education Sector
| 2000 | 2010 | ||
|---|---|---|---|
| Total number of pupils in government schools (1000s)1 | 618.0 | 886.0 | |
| Enrollment rates (ages 6–14)2 | 0.9 | 0.9 | |
| Pupil-teacher ratio3 | 24.0 | 24.0 | |
| Student-classroom ratio3 | 41.0 | 41.0 | |
| Salary costs (NIS millions) | 718.0 | 1522.0 | |
| Expenditure per pupil (NIS)3 | 1,555.0 | l,555.0 | |
| Total education expenditure (NIS millions)4 | 4,754.0 | 1,097.0 | |
| In percent of GDP | 5.5 | 4.2 | |
| Of which: recurrent expenditure | 4.3 | 3.3 | |
Projected on the basis of population projections, planned enrollment rates and slight increase over current share of government schools (70 percent).
Ministry of Education five-year education plan.
Held constant at 1999 level.
At 1999 prices.
This is a somewhat stronger effect than what would be expected on the basis of the analysis in the Heller and Diamond (1990) study, which estimates that education expenditure as a percentage of GDP increases with the percentage of population under 15 years of age. Using that study’s coefficient estimates renders a decrease in recurrent education spending to 3.7 percent of GDP in 2010.16
The Effects on the Pension System
Finally, as mentioned above, demographic dynamics might also put pressure on the budget through their impact on the pension system.17 Currently, most of the workforce relies on informal, family-based transfers for old-age provision. Only public sector employees, about 19 percent of total employment currently, are covered by three centrally organized pension schemes, two for civil servants in the West Bank and Gaza and one for the PA police force.
Changes in the population’s age structure will have obvious implications for these arrangements. Among the public sector pension schemes, only the Gaza Pension Fund is financially sustainable, under existing contribution and benefit levels, while the system for civil servants in the West Bank is running an operational deficit, and the police scheme is projected to be running deficits by 2005. While there will be a certain demographic dividend resulting from a larger working population paying a higher total sum of contributions into the pay-as-you-go system, other demographic factors such as improved life expectancy may exacerbate the underlying problems produced by a disproportionate combination of low contribution rates and generous benefits.
In the absence of pension reform, the financial gap in the public system will have to be filled by increased taxes or decreased expenditure in other areas, or a mixture of both. The preferred action would, of course, be to make the pension systems viable to avoid a financial and social crisis. Ideally, the West Bank and Gaza schemes should be unified with contributions and benefits aligned at a sustainable level and ultimately include all private sector workers. The police force could be covered by a supplementary scheme, drawing on a merged administration and a reserve fund.
The Overall Impact of Demographics on Public Finances
Combining the regression results with respect to aggregate revenue and expenditure, while bearing the caveats in mind, suggests that demographic change in the West Bank and Gaza may promote fiscal consolidation, which in turn could support economic growth in an important way, as shown in Chapter 2 and Figure 5.1. Since economic growth will be needed to help the demographic dividend materialize there seem to be two possible equilibria—a bad and a good scenario. Just considering the average impact of changes in the age dependency ratio gives us a rough sense of the orders of magnitude involved. Under the good scenario, the labor market could absorb all the additional labor without a cut in real wages, and revenue could increase by 1.3 percent of GDP and expenditure decline by 1.4 percent of GDP, allowing for a reduction in the overall fiscal deficit from 3.7 percent in 1999 to 1 percent of GDP in 2010.
Summary of the Impact of Demographic Change on Fiscal Revenue, Expenditure, and Deficit of the Palestinian Authority
Source: IMF staff calculations and estimates.Nevertheless, a word of caution is warranted at this point. First, it needs to be emphasized again that this constitutes the isolated effect of demographics alone, holding all other factors constant. Clearly this does not depict reality. Policies are likely to change over the next 10 years, and pressing issues, such as the medium-term financial sustainability of the pension system, the potential cost of a pension reform, and the efficiency of the tax administration, will have to be addressed. The PA also will, at some point, have to take over the services provided by UNRWA. The actual size of the deficit in 2010 will finally be a reflection of these and other factors. Second, as mentioned above, there is nothing automatic about these positive effects on public finances and economic growth. They have to be accompanied by good economic policy, including improvements in the PA’s expenditure composition, which is currently heavily biased towards wage expenditure, and a reasonable level of aggregate fiscal expenditure. Rather than using all of the room created by the demographic dynamics to reduce the overall fiscal deficit, it might be more appropriate to use some of it to increase real social expenditure and infrastructure investment or reduce taxes. Indeed, as the separate analyses for the health and education sectors have illustrated, it could be possible to increase such real outlays significantly on a per capita and per pupil basis, while holding total spending constant as a fraction of GDP. Fiscal policy over the medium term will also have to take into account the fiscal implications of reforms to the tax and trade regimes, and of possible large-scale immigration.
Further Fiscal Policy Challenges Over the Medium Term
Apart from the demographic changes underway, a permanent status agreement with Israel would pose two additional fiscal policy challenges. This section discusses the future of the tax and trade regime and the potential fiscal implications of large-scale immigration, including refugees.
The Tax and Trade Policy Regime
The West Bank and Gaza is in a customs union with Israel—as set out in the Protocol on Economic Relations of the Interim Agreement—which offers only very limited scope for an autonomous indirect tax policy. Tariff rates can deviate from those of Israel only for a limited number of goods and quantities, and the VAT cannot be more than 2 percentage points lower than in Israel.18 Rates and coverage of direct taxes may differ, however, and do so to a great extent.19 Under the revenue clearance system established within the customs union Israel transfers 75 percent of all income tax withheld from Palestinian workers in Israel to the PA, and indirect tax revenue (from VAT, customs duties, and the purchase tax) is allocated according to the destination principle. Only duties and VAT (and recently purchase taxes) on direct imports into the West Bank and Gaza are transferred to the PA; Revenue from duties and purchase taxes on indirect imports via Israel and on purchase taxes on Israeli goods exported to the West Bank and Gaza are not. 20 On the whole, the system works rather smoothly, but it exposes the PA to the risk of revenue transfers being suspended when the political and security situation deteriorates, as happened in 1996, late 2000, and early 2001.
While it remains to be decided what trade regime will be adopted between the West Rank and Gaza and Israel in the future, the customs union seems likely to be replaced with an arrangement that invokes the establishment of a customs border between the two entities. This would provide greater scope for tax policy autonomy in the West Bank and Gaza.
To start, the revenue-GDP ratio of the PA has risen from 8 percent of GDP in 1994 to 22 percent of GDP in 2000. Israel’s tax burden is even higher at 30 percent of GDP, but if one takes into account per capita GDP and the age dependency ratio, the West Bank and Gaza lies well above the trend for 62 developed and developing countries (Figure 5.2). This raises the question to what extent PA tax policy should be and could be better tailored to meet the needs of the Palestinian economy. Moreover, currently more than 80 percent of all revenues originate from indirect taxes (60 percent from the revenue clearance system with Israel), and only 8 percent are generated by personal income taxes. A better balance between direct and indirect tax revenues would be desirable on equity grounds.21 Over the medium term, however, with an increase in the ratio of the labor force to the population at large, direct taxes will inevitably become relatively more important.
The West Bank and Gaza’s Tax Burden in a Cross-Country Comparison
Sources: World Bank Development Indicators database, and IMF staff estimates.Finally, if the revenue-GDP ratio were indeed to increase in line with the higher estimates in the previous section, it might be appropriate to consider reducing certain tax rates, rather than allowing the government sector to grow too large. A much higher revenue-GDP ratio than today could undermine the growth prospects of the Palestinian economy. The tariff changes mentioned above offer this potential. Similarly, total PA fiscal expenditure (including through the donor-financed development budget) equals roughly 27 percent of GDP. This is very high by international standards, especially considering that, for a large part of the population, public services are mainly, if not entirely, provided by UNRWA and NGOs. For the future, it will be important for the PA to plan how it can gradually assume these services. Foreign assistance is also unlikely to remain at current levels and on current favorable terms indefinitely, and the PA should seek to begin contributing to the development budget with its own resources. To these ends, and given that it would not be desirable for the revenue-GDP ratio to rise significantly higher than its current level (22 percent), the PA will have to rein in recurrent expenditure growth.22 Adopting a low uniform tariff rate across the board with a minimum of exceptions, as Chapter 4 argues, would also have the beneficial side effect of working towards this end. Such a tariff regime would be simple and transparent, and would help to minimize economic distortions, administrative problems, and potential rent seeking. Moreover, the problem of tax leakages resulting from indirect imports via Israel could finally be addressed by well enforced rules of origin.
Tax Administration
If a customs border between Israel and the West Bank and Gaza were to be established, this would not only open the door for a more autonomous and growth-oriented trade and tax policy, but the burden on the PA tax administration would also increase. An important priority for the coming years must, therefore, be to strengthen the efficiency of the tax administration to ensure that there will not be too much of a loss in revenue collection if and when the PA becomes responsible for collecting all of its revenue.23 Specifically, it will be important to improve the institutional structures and systems by unifying the separate VAT and income tax organizations in the West Bank and the Gaza Strip. This would improve efficiency and allow for better cross checking. To further boost revenue collection and fight tax evasion, it will be important to create a large taxpayer unit, alongside an enhanced enforcement system and selective audit programs. In addition, the administrative burden on the central tax authority could be relieved by introducing a self-assessment concept for income tax—a basic principle of any modern tax system—and devolving more revenue autonomy to local governments. Local property taxes are underutilized in the West Bank and Gaza relative to other countries in the region, and apart from delegating responsibility, such a move could alleviate the constraint on local governments’ revenue capacity and diversify the revenue base in general.
Fiscal Consequences of Migration
A permanent status agreement between Israelis and Palestinians may be followed by large-scale immigration to the West Bank and Gaza, including of refugees. While the numbers are surrounded by considerable uncertainty, UNRWA estimates that about 1.4 million Palestinian refugees currently live in the West Bank and Gaza, and a further 2.3 million in Jordan, Lebanon, and Syria.24 Chapter 2 highlights the potential macroeconomic consequences of an inflow of only a fraction of the Palestinians living outside the West Bank and Gaza over the next 10 years. Obviously, there are also important fiscal implications. For a start, the revenue-enhancing and expenditure-reducing effects estimated earlier in this chapter could be amplified by the higher growth in population and slight changes in age composition (see Table 5.1).25 There will also be considerable direct costs that might be incurred by the PA.
Tables 5.7 and 5.8 report the effects of demographic change using the same methods as in the previous section, but amplified by 600 thousand additional immigrants living in the West Bank and Gaza by 2010. Revenue would receive a further boost and could increase by 1.6 percentage points of GDP, in contrast to the 1.3 percentage points in the no migration scenario. At the same time, expenditure could be reduced a bit more through a slight reduction in the age-dependency ratio, leading to a reduction by 1.7 percentage points, compared with the 1.4 calculated in the earlier case. Figure 5.3, as a complement to Figure 5.1, sums up the overall results: immigration may further facilitate fiscal consolidation, as the deficit could shrink to 0.4 percent of GDP. Of course, the caveats mentioned earlier still apply. Above all, employment creation for such a large number of returnees will be even more difficult than the already formidable challenge of accommodating the natural growth rate of the workforce—especially given the current state of the economy. Hence, rather than leading to domestic employment growth at rising real wages, there might well be an increase in unemployment, a reduction in real wages, or both.
Estimated Impact of Demographic Variables on Revenue with Migration, 1999–20101
As a percentage of GDP.
Dependent variable is tax revenue/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Estimated Impact of Demographic Variables on Revenue with Migration, 1999–20101
| Regressor | Sign | Base Value (1999) | Estimated Value with Migration (2010) | |
|---|---|---|---|---|
| IMF staff estimate2 | Age dependency3 | — | 20.7 | 22.3 |
| Working-age population | + | 20.7 | 22.2 |
As a percentage of GDP.
Dependent variable is tax revenue/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Estimated Impact of Demographic Variables on Revenue with Migration, 1999–20101
| Regressor | Sign | Base Value (1999) | Estimated Value with Migration (2010) | |
|---|---|---|---|---|
| IMF staff estimate2 | Age dependency3 | — | 20.7 | 22.3 |
| Working-age population | + | 20.7 | 22.2 |
As a percentage of GDP.
Dependent variable is tax revenue/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Estimated Impact of Demographic Variables on Expenditure with Migration, 1999–20101
As a percentage of GDP.
Dependent variable is total expenditure/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Dependent variable is total current expenditure/GDP (for 1999 total expenditure equals current expenditure of the PA).
Estimated Impact of Demographic Variables on Expenditure with Migration, 1999–20101
| Regressor | Sign | Base Value (1999) | Estimated Value with Migration (2010) | |
|---|---|---|---|---|
| IMF staff estimate2 | Age dependency3 | + | 22.2 | 20.9 |
| Working-age population | – | 22.2 | 22.0 | |
| Rodrik (1996)4 | Age dependency3 | + | 22.2 | 20.0 |
| Commander, Davoodi, and Lee (1997) 4 | Age dependency3 | + | 22.2 | 20.6 |
| Heller and Diamond (1990) 4 | Population share > 65 years old | + | 22.2 | 21.8 |
As a percentage of GDP.
Dependent variable is total expenditure/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Dependent variable is total current expenditure/GDP (for 1999 total expenditure equals current expenditure of the PA).
Estimated Impact of Demographic Variables on Expenditure with Migration, 1999–20101
| Regressor | Sign | Base Value (1999) | Estimated Value with Migration (2010) | |
|---|---|---|---|---|
| IMF staff estimate2 | Age dependency3 | + | 22.2 | 20.9 |
| Working-age population | – | 22.2 | 22.0 | |
| Rodrik (1996)4 | Age dependency3 | + | 22.2 | 20.0 |
| Commander, Davoodi, and Lee (1997) 4 | Age dependency3 | + | 22.2 | 20.6 |
| Heller and Diamond (1990) 4 | Population share > 65 years old | + | 22.2 | 21.8 |
As a percentage of GDP.
Dependent variable is total expenditure/GDP.
Ratio of nonworking-age population (under 15 and over 65) to working-age population (15–65)
Dependent variable is total current expenditure/GDP (for 1999 total expenditure equals current expenditure of the PA).
Summary of the Impact of Demographic Change on Fiscal Revenue, Expenditure, and Deficit of the Palestinian Authority, with Migration
Source: IMF staff calculations and estimates.In addition to the fiscal implications stemming from returnees, a permanent settlement of the refugee question would also require the PA to (eventually) take responsibility for the services provided by UNRWA—with a budget of US$353 million in 1999—to the refugees living inside the West Bank and Gaza. Highly tentative calculations suggest that current expenditure could come to about US$200 per refugee per year, abstracting from any compensation payments. Therefore, should the total population be boosted by 600 thousand immigrants and their descendants, as in the migration scenario in Chapter 2, and supposing that they are all returning refugees, this could mean additional expenditure of US$120 million for the PA. Should the PA also assume the cost of public services to the 1.4 million refugees currently provided for by UNRWA in the West Bank and Gaza, this would add up to a total of US$394 million per year (roughly 9 percent of GDP in 2000).26 In comparison, total recurrent expenditure of the PA stood at US$943 million in 1999. Additionally, one-time costs, for example in infrastructure, of absorbing such a large number of people, would have to be taken into account. Obviously, such an expense could not be covered by an improved revenue effort alone but highlights that sustained donor support will be required together with stringent expenditure controls in the PA.
Immigration may have beneficial macroeconomic and fiscal effects over the medium term—but only if the economy is able to generate and sustain employment without undue compression in real wages. Given current unemployment rates, this raises a big question mark. In the short term, however, donor support would surely be essential to meet the needs of the greatly expanded population.
Conclusion
It is difficult to predict medium-term fiscal developments for any economy, and even more so for the West Bank and Gaza, given the many uncertainties. Thus, projections up until 2010 should not be taken too literally. They constitute little more than suggestions of the directions and the broad order of magnitude that may be involved. Nevertheless, it is clear that while the projected demographic changes have the potential of developing into a fiscal blessing for the PA, they may also turn out to be a curse. Tax revenue seems likely to increase with a larger workforce, but only a larger workforce that finds jobs. Expenditure may be reduced with a smaller number of school children and fewer people older than 65 in relation to the total population, but only it public sector employment is contained and the fiscal dangers of an unsustainable pension system are addressed.
To produce and at the same time reap any benefits of this potential demographic dividend the PA will have to make a few important policy choices. Above all, the PA should reconsider its spending priorities. On the basis of current allocations, changes in the age structure may allow social budgets to decrease as a percentage of GDP, but it may in fact be preferable to maintain the current expenditure-GDP ratio by increasing per capita expenditure in the education, health, and social sectors, and by contributing more to the capital budget.
It should be borne in mind that positive growth depends on the assumptions about participation rates (working-age population actually joining the workforce), employment creation (workers finding jobs-not that simple given the current situation), and labor productivity. Other indirect mechanisms through which demographic change can affect growth rates are through its impact on savings behavior and capital accumulation.
This phenomenon may be due to a combination of factors, including improvements in tax administration with economic development, workers moving up in the income tax brackets, and expansion of the tax base as a percentage of GDP during the transition from an agricultural and informal economy to a more industrialized and formal economy.
Since 1994, PA fiscal revenue has grown considerably faster than GDP, but this is mostly on account of improvements in tax administration, and it would be (grossly) misleading to interpret this as tax buoyancy. It is difficult to separate the tax administration effects from buoyancy, and instead we estimate buoyancy for a shorter and more recent period (1998–99) which does not include the period with the strongest improvements in tax administration. The drawback is that the estimate is based on only two years, and two years when the Palestinian economy experienced strong recovery in growth. Ideally, the buoyancy should be estimated over a time period that is sufficiently long to include both economic upswings and downturns.
The buoyancy ratio is defined as the percentage change in tax revenue over the percentage change in nominal GDP, or the elasticity of tax revenue with respect to nominal GDP. In this context, however, buoyancy refers to total revenue, not just tax revenue.
We regress tax revenue as a percentage of GDP on the age dependency ratio and the share of the working population respectively, controlling for per capita GDP and the share of agriculture. The data set included information on 60 developing and developed countries for the year 1997.
The age dependency ratio is defined as the ratio of the population over 65 and under 15 to the population between 15 and 65.
It should be noted, however, that empirical evidence investigating this relationship has been less than conclusive (Ram, 1987; Payne and Ewing, 1996).
There is also a more cynical political economy argument why expenditure as a share of GDP is higher in richer economies—as seen above, revenue tends to grow faster than GDP over time (for given tax structure and rates), and the argument would be that policymakers prefer to spend these resources rather than take the opportunity to lower tax rates.
The age dependency ratio was also chosen, as a broad range of estimates was available, lending more power to the analysis.
This discussion ignores health services provided by UNRWA and other agencies, some of which the PA might have to assume responsibility for over the medium term.
More developed countries generally have higher per capita spending on health care, and the population over 65 tends to demand relatively more health services per capita.
The direction of causality regarding infant mortality rate is unclear. It has been suggested that the negative correlation reflects the success of higher health expenditure (Heller and Diamond, 1990).
Planned enrollment rates are taken from PA’s five-year education development plan; see Palestinian Authority Ministry of Education (2000)
To be exact, total education spending should have been used in this context, as this is the basis for Heller and Diamond’s study. The PA, however, only covers current expenditure, and hence, using current expenditure was more informative.
See Annex I in the IMF study in Alonso-Gamo, Alier, Baunsgaard, and Erickson von Allmen (1999) for a detailed description of the Palestinian pension system.
See Chapter 4 for a fuller discussion of the trade regime and trade policy options for the future.
The new income tax law in West Bank and Gaza reduced tax rates and the number of personal income tax brackets to four (5, 10, 15, and 20 percent). In contrast, Israel maintains five inflation-adjustable brackets of 10, 20, 30, 45, and 50 percent on active personal income.
Under the revenue clearance system, each party recovers from the other party’s tax authority some taxes paid by persons and firms in their respective jurisdictions. For further details, see Alonso-Gamo, Alier, Baunsgaard, and Erickson von Allmen (1999).
A progressive direct tax system can achieve limited income redistribution.
These points are further elaborated in Alonso-Gamo, Alier, Baunsgaard, and Erickson von Allmen, (1999).
Even today, there are large amounts of reported tax arrears outstanding (estimated at NIS 1.6 billion in June 2000), and a significant number of taxpayers appear to not have paid any income tax for the last five years.
UNRWA.
The age dependency ratio in 2010 is slightly lower in the scenario with migration than in the one without migration.
This does not necessarily contradict the finding that overall expenditures as a percentage of GDP could decline as per capita GDP growth may be stimulated by the larger workforce.
