Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 1 Haydeeliz Carrasco Hamidou Jawara Moritz Meyer The World Bank University of The Gambia The World Bank January 2022 Abstract The overall objective of this study is to assess the impact of the fiscal system on poverty and inequality in The Gambia as of 2015. The study presents the first empirical evidence on the distributional impacts of taxes and social spending on households in The Gambia. Furthermore, it also evaluated the distributional effects of recent fiscal policy reforms in The Gambia. The assessment was based on the Commitment to Equity (CEQ) Methodology with data from the Integrated Household Survey of 2015 and fiscal administrative data from various government ministries, departments, and agencies. The analyses show that while the fiscal system in The Gambia reduces inequality by 1.2 Gini points, it increases the national poverty headcount by 5.3 percentage points as all households (including the poor) are net payers into the fiscal system. Most of the inequality reduction is due to primary education benefits, with a marginal contribution of 0.44 Gini points, and most of the poverty increase is due to custom duties and VAT with marginal contributions of -2.63 percentage points and -2.07 percentage points, respectively. Simulating the effect of changes in the structure of personal income tax (PIT) and the government’s ongoing absorption of the School Feeding Program indicate that these changes reduce inequality but do not offset the impoverishing effect of the fiscal system. Hence, more cashable transfer programs targeted to the poor are needed to offset the impoverishing effect of indirect taxes and make the fiscal system more pro-poor. Keywords: fiscal policy, social spending, taxes, fiscal incidence, inequality, poverty, The Gambia JEL classification: H22, I38, D31 1 The authors would like to thank their World Bank colleagues (World Bank Equity Policy Lab, The Gambia country team), the Government of The Gambia and development partners for facilitating the data collection and for adapting the methodology to the Gambian context, and participants of a validation workshop organized by the Ministry of Finance and Economic Affairs for their comments on the results and draft document. We declare that we have no relevant or material financial interests that relate to the research described in this paper. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the World Bank or any affiliated organizations, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. 1 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 1. Introduction Over the past two decades, The Gambia has experienced limited economic growth which undermined inclusive growth (World Bank 2020b). In effect, between 2000 and 2018, the average growth in per capita income was minimal, at less than 0.5 percent per year, significantly below the 2 percent average of countries in Sub-Saharan Africa (SSA). Consequently, Gambians’ standard of living has trailed behind that of their SSA peers: per capita gross domestic product (GDP) dropped from 45.5 percent of the SSA average in 2000 to 32.1 percent in 2017. In terms of social indicators, as of 2015, 48.6 percent of the population lived below the national poverty line, and a Gini coefficient of 0.359 pointed towards inequality across households and regions within the country.2 Moreover, high levels of monetary poverty are closely intertwined with deficits in human capital endowment and limited access to basic infrastructure, with 15.4 percent of the population being multidimensionally poor (World Bank 2020b);3 this problem is particularly acute in rural areas.4 These socioeconomic deficits in the population translate into low levels of productivity and limited resilience, as well as challenges to attempts to tackle economic and social exclusion. Following the peaceful political transition in 2016, there is hope that faster and more inclusive economic growth will reduce both inequality and poverty, especially after the government launched its national development plan. The National Development Plan (NDP) 2018–21 aims to “deliver good governance and accountability, social cohesion, and national reconciliation and a revitalized and transformed economy for the wellbeing of all Gambians.� To move towards these objectives, the plan has identified eight priority areas: (i) building good governance; (ii) stabilizing the economy; (iii) modernizing the agricultural sector; (iv) investing in education and health services; (v) building infrastructure and restoring energy services; (vi) promoting inclusive tourism; (vii) empowering youth; and (viii) strengthening private sector development for growth and job creation. Progress towards these priorities will require sizeable investments into human and physical capital, which are expected to enhance productivity, strengthen resilience, and support inclusion. Fiscal policy is a key input to achieving the policy goals in The Gambia’s NDP 2018–21. As the country prepares for a better future, growth alone will not be enough to eliminate poverty. The fiscal system can support this process, through the right mix of investments and redistribution via taxation and social expenditures. However, The Gambia still faces fiscal challenges such as a high public debt and limited tax revenue to finance recurrent expenditures and investments. As the country explores the mix of taxation and expenditure policies to ensure fiscal sustainability (World Bank 2020a), it is important that the design and implementation of reforms maintain a focus on equity. Pro-poor fiscal policies could lay the foundation for future economic growth and human development, while paying special attention to poverty reduction and social inclusion. Pro-poor policies contribute to fiscal and social sustainability and take account of differences in households’ ability to pay and need for support. More specifically, progressive fiscal policies must be designed so that relatively richer households pay a higher share of their income as taxes, while poorer households receive a relatively larger share of transfers. Yet, fiscal policy is driven by multiple objectives which often contradict each other. First, direct and indirect taxes are needed to finance public expenditure while aiming for fiscal 2 The Gini coefficient reached 0.436 in 2010, so its level has fallen in recent years. However, it is challenging to make poverty comparisons over time given that previous household surveys were not really comparable. 3 Multidimensional poverty in The Gambia reflects a combination of low consumption levels, limited educational attainment, and gaps in access to drinking water, sanitation, and electricity. Deprivations are often overlapping and contribute to the depth, complexity, and persistence of poverty. 4 Access to basic services and facilities is worse in rural areas and there is a strong divide between the capital city region and the rest of the country. 2 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia sustainability. Second, fiscal policies need to be implemented in an efficient way that minimizes distortions in the allocation of resources. Third, fiscal policies can be used to reduce negative externalities arising from harmful and undesirable behavior. And fourth, fiscal policies can also be used to improve equity through the redistribution of resources. This study focuses on this last objective with the specific aim of determining the impact of fiscal policy on poverty and inequality in The Gambia. In doing so, its analysis focuses on addressing three broader questions. First, does the fiscal system in The Gambia reduce inequality and poverty? Second, who bears the burden of taxes, and who receives the benefits from transfers? And third, which fiscal interventions are the main drivers of these results? This fiscal incidence analysis for The Gambia describes the distributional impact of selected tax interventions and public social expenditure items on poverty and inequality for the year 2015. To the best of our knowledge, this is the first fiscal incidence analysis conducted for The Gambia that aims to cover the complete fiscal system. The analysis is based on the Commitment to Equity (CEQ) Methodology with data from the Integrated Household Survey (IHS) 2015/16, which is a nationally representative survey.5 Under this approach, taxes and transfers are either directly observed or simulated from the survey data. The analysis assesses the incidence of several selected fiscal interventions in The Gambia—including direct and indirect taxes, direct transfers, indirect subsidies, and in-kind government social benefits in the form of spending on providing health and education services at subsidized rates. For the main results, we include a discussion on international comparability with other African countries. Lastly, the fiscal microsimulation model built for the analysis is used to explore simulations of recent fiscal policy developments. The main findings from the fiscal incidence analysis show that fiscal policy in The Gambia in 2015 was inequality reducing but poverty increasing. In 2015, total taxes represented 12.1 percent of GDP while total primary government expenditure (excluding debt interest payments) accounted for 16.4 percent of GDP (of which 3.8 percentage points were social expenditure).6 The combined effect of taxes and social spending reduced inequality by 1.2 Gini points, and increased poverty by 5.3 percentage points. Most of the inequality reduction came from in-kind primary education benefits and most of the poverty increase was explained by indirect taxes. An international comparison suggests that these findings resonate with findings in similar studies conducted in Africa. Specifically, the fiscal systems in most countries in the region (e.g. Ghana, Ethiopia, Uganda, and Tanzania), are also found to be inequality reducing and poverty increasing. The results from these analyses show the need to strengthen social protection systems that offset the impoverishing impacts of indirect taxes, so that poor households are net receivers and not net payers of the fiscal system. This paper is organized as follows: Section 2 presents the literature review. Section 3 describes the structure of tax revenues and public expenditure in The Gambia. Section 4 describes the data and general methodology, and Section 5 presents assumptions and allocation rules used for the modelling of the fiscal incidence analysis in The Gambia. Section 6 presents the main results and Section 7 presents international comparison. Section 8 presents recent developments and policy simulations, and Section 9 concludes. Annexes present more details as well as recommendations for data and methodological extensions. 5 At start of the study, the most recent household survey for The Gambia was collected between April 2015 and March 2016. In combination with fiscal data for 2015, this study provides an assessment of the distributional impact of fiscal policy on poverty and inequality for the year 2015. While there have been changes to taxes and social expenditure between 2015 and 2020, Section 6 shows that the main findings remain relevant for the current country context. 6 The analysis accounts for 78.6 percent of total tax revenues, but only 18.7 percent of total expenditure was included. While this is likely to provide an unbalanced picture of the effect of the fiscal system, the study includes the main social expenditures that were funded by the government in 2015 (mainly health and education). This also reflects the fact that a large share of expenditure, including public goods such as infrastructure and defense, might be considered distributional neutral. Under this assumption their omission does not bias the estimated effect on inequality, whereas poverty would be lower if the usage value would be added to household welfare. 3 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 2. Literature Review The use of the CEQ approach to analyze the distributional effect of the fiscal system is becoming popular across the globe. Although most of the earlier studies were in done in Latin America (e.g. Beneke et al. 2018; Daude et al. 2017; Lustig 2016b), evidence from Africa is also growing (see Gasior et al. 2018). In this section, we review some of the empirical evidence in Africa. In Ghana, Younger et al. (2017) find that the effect of the fiscal system on poverty and inequality was moderate. It reduces the Gini coefficient by 3.5 Gini points when going from market income plus pensions to final income. At the same time, the fiscal system increases the national poverty headcount in Ghana by 2.2 percentage points, when going from market income plus pensions to consumable income. Furthermore, they find that the effect of the fiscal system on poverty differs across different fiscal items. For instance, while pensions, direct taxes, and cash transfers have “almost no effect� on poverty, indirect subsidies to electricity and fertilizers have a positive effect on poverty reduction. They also found that indirect taxes increase poverty significantly and the effects are larger for poorer households. Hill et al. (2017) also used the CEQ approach and data from the 2010/11 Ethiopia Household Consumption Expenditure Survey and Welfare Monitoring Survey to study the redistributive effect of fiscal policy in Ethiopia. As in many developing countries (see Inchauste and Lustig 2017 for a review), they find that Ethiopia’s fiscal policy reduces inequality and increases poverty. Specifically, the fiscal system reduces inequality from a Gini coefficient of 0.322 to 0.302 when going from market income to final income (i.e., a reduction of 2.0 Gini points). As for poverty, their estimates indicate that fiscal policy increases the national poverty incidence from 31.2 percent at market income to 32.4 percent at consumable income (i.e., by 1.2 percentage points). Despite this increase in poverty incidence, they also find that the fiscal system reduces the depth and severity of poverty in Ethiopia. However, based on the fiscal impoverishment indexes, they found that the impact of the fiscal system on poverty is greater for poor households than for nonpoor households in Ethiopia. De La Fuente et al. (2017) studied the impact of fiscal policy on inequality and poverty in Zambia. The household level data they used for their analysis were from the Zambia Living Conditions Monitoring Survey of 2015, which was administered to about 12,250 households and 63,000 individuals. Regarding the distributive effect of the fiscal system, they find that fiscal policy significantly reduces inequality in Zambia in 2015 as the Gini coefficient falls by 11 points between market income and final income with the overall decrease being greater for rural areas. They find the fiscal system increases poverty as the net cash position at consumable income is worse than the net cash position at disposable income for most households. Similar results were found in Mali and Niger by Hounsa et al. (2019). In Tanzania, Younger et al. (2016) find that the fiscal system reduces the Gini coefficient by 5.1 points: from 0.382 at market income (plus pensions) to 0.331 at final income. This redistributive effect is attributed mainly to very progressive taxation as well as indirect taxes, with education and health benefits also making a difference. As for poverty, they find that the fiscal system increases the national poverty rate by about 6.5 percentage points: from 28.3 percent at market income (plus pensions) to 34.8 percent at consumable income. Hence, without taking into account in-kind benefits in poverty measurement (which is a standard practice in CEQ analysis), they find that fiscal policy increases poverty in Tanzania. Like many of the studies reported above, their results reveal that the fiscal system in Tanzania did more to increase poverty than to reduce inequality. Therefore, the fiscal system’s redistributive effects are weaker than its welfare effects. 4 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Lara Ibarra et al. (2019) also assessed the distributive effects of the fiscal system for Egypt. They find that the redistributive effect of the fiscal system in Egypt for 2015 (the same year covered by this study) is positive; the fiscal system reduces inequality by 4.4 points from a Gini coefficient of 0.325 at market income (including pensions) to 0.281 at final income. On poverty, they find that the fiscal system reduces the poverty headcount by 11 percentage points. Hence, in Egypt the fiscal system has positive effects on both inequality and poverty reduction, with a better effect on poverty than most fiscal systems in the continent; this is explained by the fact that social protection programs (mostly the Tamween program) are well targeted. Another study of fiscal incidence in North Africa using the CEQ approach was conducted in Tunisia by Jouini et al. (2018). Studying the distributional effect of the fiscal system on inequality and poverty they find that in Tunisia the fiscal system had a significant effect on inequality in 2010 as it reduces the Gini coefficient by 9 points: from 0.44 (market income) to 0.35 (final income). On poverty, they find that the effect depends on the type of poverty line used: they found a reduction in poverty when using the lower international poverty line of US$1.25 per day, but an increase in poverty of 2.7 percentage points when using the national poverty line. Their results suggest that the increase in poverty based on the national poverty line is due to the increase in the tax burden and social contributions for people in the second decile and above of income distribution. The fiscal system only reduced poverty among the poorest 10 percent because they benefit substantially from transfers despite the tax burden they face. In studying the effect of the fiscal system on inequality and poverty in South Africa, Inchauste et al. (2017) find that the fiscal system in 2010 reduces inequality by 17.4 Gini points when going from market income to final income. Despite higher levels of inequality in South Africa than in most middle-income countries, its fiscal system does well in redistributing income. On poverty, they find that the fiscal policy reduces the national poverty headcount by 2.2 percentage points. South Africa is among the few countries in Africa where the fiscal system achieves a reduction in both poverty and inequality. The authors conclude that these redistributive results in South Africa are attributed to the combination of a highly progressive tax system (that collects mostly from the rich) and a highly progressive social spending (that redirects resources to the poorest). Overall, the empirical evidence in Africa shows that for all countries the fiscal system has a positive effect on inequality, ranging from a reduction of about 2 Gini points in Ethiopia to one of about 17 Gini points for South Africa. On poverty, the results are mixed, but for most countries the fiscal system increases the national poverty at consumable income (i.e. excluding in-kind benefits such as from health and education). The notable exceptions in this literature review are South Africa and Egypt, where the fiscal systems actually reduced poverty; other authors have similarly attributed national poverty reduction to the Namibia’s fiscal system (Sulla et al. 2017). Five out of 25 countries have fiscal systems that increase poverty with respect to the extreme poverty line of US$1.25/day, rising to 15 countries when the poverty line is pegged at the US$4/day level (Lustig, 2016a). 5 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 3. Tax Revenues and Public Expenditure in The Gambia Public finance in The Gambia involves central government and public enterprises. The General Revenue Authority (GRA) is the institution responsible for tax collection (both domestic and customs).7 Social security funds are managed by government directly through the Public Service Commission and Treasury Department of the Ministry of Finance and Economic Affairs (MoFEA) and indirectly through the Social Security and Housing Finance Corporation (SSHFC). The latter is responsible for managing pensions for quasi-government and private sector workers.8 For public expenditure, MoFEA is responsible for managing the state budget as well as official development assistance. Moreover, MoFEA plays the leading role in the design and analysis of economic and fiscal policies.9 In The Gambia, most social spending is for the provision of health and education services. Hence, the main institutions involved in the execution of social expenditure are the Ministry of Basic and Secondary Education (MoBSE), the Ministry of Higher Education, Research and Technology (MoHERST), and the Ministry of Health (MoH). The Ministry of Women Affairs, Children, and Social Welfare (MoWCSW) is also now involved in social protection policies. Indirect subsidies are related to energy and agriculture. The main government institutions channeling these are the Ministry of Agriculture and the National Water & Electric Company (NAWEC). Finally, donors play a vital role in supporting government’s interventions, particularly in the areas of social protection, education, and health. 3.1. Tax Revenues The structure of the tax revenues in The Gambia is reported in Table 1 below. In 2015, tax revenues reached 12.1 percent of GDP (94 percent of total government revenue). Most of the tax revenue (75 percent) came from indirect taxes, particularly from custom duties and import value added tax (VAT). Direct tax revenues constituted about 24 percent of total taxes, with similar shares coming from personal income tax (PIT) and corporate income tax (CIT). The Gambia’s tax revenue is lower than peer low-income countries in Africa, which average 14 percent of GDP,10 and the gap is even larger compared to the SSA average of 17 percent of GDP.11 The World Bank (2020a) estimates that The Gambia has a structural tax gap of 4–6 percent of GDP, which suggests that its tax ratio could reach 15–17 percent of GDP. The Gambia’s only contributory social security contribution (SSC) scheme is the National Provident Fund (NPF) for private sector employees, which is managed by the SSHFC. The SSC scheme for public servants is the Public Service Pension Scheme (PSPS) while quasi- public employees use the Federated Pension Scheme (FPS); both are non-contributory and fully funded by the government. More details about the benefits from social security schemes can be found in the Social Expenditure section (Box 1). 7 More details about GRA: http://www.gra.gm/ 8 The Department of Labor is also involved through the Injuries Compensation Fund. 9 More details about MoFEA: https://mofea.gm/about-us 10 African low-income average tax revenue (2013) based on: (Beegle & Luc, 2019). 11 World Bank (2020a). The Gambia Public Expenditure Review: Creating Fiscal Space. 6 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Direct Taxes Personal Income Tax In 2015, PIT amounted to 1.2 percent of GDP, and contributed about 10 percent of The Gambia’s tax revenues. Personal income tax is payable on annual gross income using a pay as you earn (PAYE) system.12 PIT is calculated based on marginal rates, depending on annual gross income. Gross income includes employment income, business income, property income, and any other income,13 reduced by the annual deductions allowed to the taxpayer (PwC 2013; p. 9). In 2015, PIT was based on six progressive brackets: 0 percent for annual gross income between GMD 0 and GMD 18,000; 5 percent for the excess of income between GMD 18,001 and GMD 28,000; 10 percent for the excess income between GMD 28,001 and GMD 38,000; 15 percent for the excess of income between GMD 38,001 and GMD 48,000; 20 percent for the excess of income between 48,001 to GMD 58,000; and the maximum tax rate of 30 percent, for the excess of annual income above GMD 58,000. Table 1. The Gambia: General Government Revenue, 2015 GOVERNMENT (in millions of Share of tax Included in Share of GDP REVENUES GMD) revenue analysis? Total revenue & grants 8,369.85 14.12 Revenue 7,647.55 12.91 Tax revenue 7,168.88 100.00 12.10 Direct taxes 1,719.91 23.99 2.90 Personal income tax 730.26 10.19 1.23 Yes Corporate income tax 857.03 11.95 1.45 No Payroll tax 43.90 0.61 0.07 No Taxes on property 63.69 0.89 0.11 No Other direct taxes 25.03 0.35 0.04 No Total contributions to - - - No social insurance Indirect taxes 5,352.24 74.66 9.03 Customs duties 2,108.99 29.42 3.56 Yes Import VAT 1,345.68 18.77 2.27 Yes Domestic VAT 820.60 11.45 1.38 Yes Excise taxes 627.11 8.75 1.06 Yes Taxes on exports 0.09 0.00 0.00 No Stamp duties 96.73 1.35 0.16 No Other indirect taxes 353.04 4.92 0.60 No Grants 722.30 10.08 1.22 No Source: Authors’ elaboration based on official data from MoFEA 12 PIT in The Gambia functions as a withholding tax since employers must withhold the tax from wages of their employees. Source: The Gambia’s Income and Value Added Tax Act 2012, Art. 89. 13 There are some forms of income that are exempt or excluded from the income tax base: certain pensions, income from education scholarships. Some individuals are also exempt (e.g. the President, local authorities, diplomats). Source: Gambia’s Income and Value Added Tax Act 2012, Art. 28. 7 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Indirect Taxes Custom Duties In 2015, total custom duty collections reached 3.6 percent of GDP and 29 percent of total tax revenue, which was close to the combined contribution of VAT on imports and domestic sales. The Gambia has four custom duty rates (5 percent, 10 percent, 20 percent, and 35 percent) applicable to imported goods.14 Based on consultations with Custom Duty officials at GRA, the study was able to define the custom duty rates for the main products purchased by Gambian households, listed in Annex III.15 Most food and household products attracted the 20 percent rate but a few were taxed at the lower rates, such as some milk products (5 percent), and fuel and some other food products (10 percent). The 35 percent rate covers some specified food, beverage, and confectionary items, along with a few clothing items. Excises In 2015, total tax revenue collected from excises amounted to 1.1 percent of GDP, or 8 percent of total tax revenue. Excises in The Gambia are mostly levied in absolute terms based on the amount of the goods in question rather than their value. The main excises are associated with alcoholic drinks and tobacco products. Specifically, there were the following excise taxes as of 2015: soft drinks and mineral water (5 GMD per liter), beer and spirits (between 175 and 300 GMD per liter), tobacco products (495 GMD per kg, including environmental tax of 165 per kg), wheelbarrows (5 percent) and imported motor cars (15 percent plus an environmental tax of GMD 1,000). Value Added Tax In 2015, total VAT (domestic and imports) amounted to 3.7 percent of GDP and contributed about 30 percent of total tax revenues; about 60 percent of total VAT came from imports and the rest from domestic sales. VAT is a consumption tax16 at a standard rate of 15 percent of the taxable value of imported and domestic products (with the value including custom duties and excises); only exempted and zero-rated goods are excluded. The main VAT exemptions are listed in Annex III and include basic foods; educational services; prescription drugs; medical, dental veterinary and optical services; agriculture and aquaculture inputs and equipment; unprocessed agriculture and aquaculture products; life and health insurance; financial services not rendered for a fee or commission; domestic transportation and ferry services; residential properties; monthly domestic electricity consumption below 1,000 kw and water below 250 cubic meters. In line with international best practices, the VAT zero rate also applies to exports. VAT is collected either at the point of import (for imported goods and services) or at the point of sale (for domestically produced goods and services). Registered businesses file monthly with the GRA and VAT registration is mandatory for businesses with annual turnover of over GMD 1 million. In terms of VAT 14 According to Harvard-CID-Economic Complexity Atlas (https://atlas.cid.harvard.edu/), as of 2015, total imports in The Gambia were estimated at US$1.05 bn, and more than half of total imports came from agricultural products (34.33%) and textiles (25.08%). See Annex I 15 Based on the list of products in the Integrated Household Survey 2015/16 that are used in the fiscal incidence analysis. 16 VAT is a consumption tax because consumers are the net payers. Even though producers and sellers bear the legal incidence of VAT filing to the government, they can claim a credit for the VAT paid on their inputs. Some businesses (if exempt) could be net VAT payers since they are not eligible to receive VAT credit for their inputs, however, we assume that they indirectly translate their input VAT (hidden) into higher final prices to the consumers 8 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia administration, there is a significant difference between zero rated and exempt. The sellers of zero rated goods and services (e.g. exports) can still claim credit for the VAT paid on their inputs (similar to the sellers of products subject to standard VAT); in contrast, the sellers of exempted goods are not eligible for the input VAT credit, which could result in indirect effects that increase final prices to consumers.17 We could not find any previous estimates on the scope of VAT informality,18 but qualitatively, the Gambian authorities acknowledge that: “The Gambia’s tax structure is complicated and difficult to comply with, resulting in very low compliance and low levels of revenue collection which affects the performance of the economy� (MOTIE 2018).19 For the year 2015, the World Bank (2020a) estimates that the VAT C-efficiency rate in The Gambia reached 25.1 percent, while the VAT gross compliance ratio was 28.6 percent,20 meaning that about 76–80 percent of potential VAT revenues are uncollected. The forgone VAT revenue or efficiency gap is due to a combination of exemptions (policy gap) and evasion (compliance gap) (Ueda, 2017). 3.2. Social Expenditure Table 2 displays government expenditure in The Gambia in 2015. Total government expenditure was 22.9 percent of GDP in 2015, while primary government expenditure (excluding debt interest payments of 6.7 percent of GDP) was 16.4 percent of GDP. Within primary government expenditure, the largest outlays were on indirect subsidies and social expenditure. The subsidies go mainly to agriculture and electricity (64 percent of total primary expenditure)21 while social expenditure is mostly driven by education and health (23.1 percent of total primary expenditure). Table 2. The Gambia: General Government Expenditure, 2015 (in millions of Share of primary Share of Included in GMD) expenditure GDP analysis? Total expenditure 13,592.60 22.94 Primary government expenditure 9,687.67 100.0 16.35 Defense expenditure 33.98 0.4 0.06 No Social expenditure 2,240.81 23.1 3.78 Social protection 128.41 1.3 0.22 Social assistance 47.54 0.5 0.08 Non-contributory pensions 47.54 0.5 0.08 No Near cash transfers - 0.0 No Social insurance 80.87 0.8 0.14 Yes Old-age pensions 80.87 0.8 0.14 Yes Education 1,289.20 13.3 2.18 Basic education 966.90 10.0 1.63 Yes Secondary 167.60 1.7 0.28 Yes 17 The indirect effect of VAT is also known as the cascading effect (i.e. the tax on tax effect). In the case of VAT exemptions, since producers and distributors cannot claim VAT credits on inputs, there is a cascading effect as they will try to recoup the hidden VAT (part of their costs) into higher final prices to be paid by the consumer. 18 IMF (2017) included comparative estimations of the informal economy for most countries in Sub-Saharan Africa. However, The Gambia was one of the few countries not included in the analysis due to data limitations. 19 To tackle this issue, the GRA has implemented an Informal Tax for small taxpayers. However, the tax works as a simplified CIT rather than a simplified VAT. 20 VAT C-efficiency is defined as the ratio between actual VAT revenue and final consumption (including households, nonprofit organizations, and government) with respect to the standard VAT rate. The VAT gross compliance ratio is similar to the VAT efficiency ratio but only includes household consumption in the denominator. 21 Although indirect subsidies are recorded as “net government losses�, in effect they are equivalent to government expenditure and represent the highest share of the budget. 9 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Post-secondary or tertiary 154.70 1.6 0.26 Yes Health 684.70 7.1 1.16 Inpatient 166.70 1.7 0.28 Yes Outpatient 155.80 1.6 0.26 Yes Other non-contributory health 362.20 3.7 0.61 No expenditures Housing and urban 138.50 1.4 0.23 No Subsidies 6,216.81 64.2 10.49 Electricity - 0.0 0.00 No Fuel 122.49 1.3 0.21 Yes Nonfuel 6,094.33 62.9 10.28 No Infrastructure 308.30 3.2 0.52 Other non-social expenditure 806.90 8.3 1.36 Debt servicing 3,904.94 6.59 Source: Authors’ elaboration based on official data from MoFEA; MoBSE and MoHERST for education data; and MoH for health data. Data based on executed budget. Indirect Subsidies In the Gambia, indirect subsidies amounted to 10.5 percent of GDP as of 2015, with about 98 percent going on non-fuel subsidies (import and VAT duty waivers and agricultural subsidies22) and the remainder on fuel subsidies. Agricultural subsidies are mainly in the form of subsidized prices for seeds and fertilizers; usually, the Government of The Gambia (GoTG) buys these items at market price and sells them to farmers at subsidized prices, mostly via intermediaries.23 Energy and fuel subsidies are price subsidies on the consumption of light fuel (diesel, petrol, and kerosene), which is used mainly for transport (diesel and petrol) and home lighting (kerosene). The rationale of these subsidies is to provide consumers with final pump prices below market prices so that they are not affected by international price fluctuations. In 2015, the fuel subsidies budget was minimal as international prices of oil were historically low; in effect, market prices were below subsidized fuel prices for the year of analysis, making these subsidies work as implicit taxes (see Annex II). Another indirect subsidy is a long-standing duty waiver on heavy fuel oil, which is imported mainly for NAWEC, and so the waiver can be seen as a subsidy on electricity production.24 In 2017, the duty waiver was replaced by an electricity subsidy program given to NAWEC directly but since this only started in 2017, it is out of the scope of this study. Social Protection As of 2015, public expenditure on social protection was only 0.2 percent of GDP. Even when donor-funded programs are considered, overall social protection expenditure in The Gambia only reaches about 0.9 percent of GDP, still lower than the average among low-income African countries of around 1.5 percent of GDP (Beegle and Luc 2019). Social protection (SP) includes all public policies that help households with socioeconomic vulnerability or shocks, and it can be subdivided into: (i) preventive SP (or social insurance), (ii) protective SP (cash transfers or in-kind benefits to poor households); (iii) promotion SP (supporting livelihoods and access to jobs among vulnerable groups); and (iv) targeted social services (World Bank 2018; Rawlings 2015). 22 Data on non-fuel subsidies were not disaggregated. 23 Given that middlemen are heavily involved in the trading of these items, it is not clear how the farmers benefit directly from such subsidies. 24 The duty waiver on heavy fuel oil for NAWEC can be considered as a truly forgone tax revenue (i.e. revenue that should have been received but was not), which do not appear in the government budget (at least using most common accounting standards). 10 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia In The Gambia, the overall social protection system was underdeveloped in 2015. In terms of preventive SP, there was no universal health insurance nor unemployment insurance and only 10 percent of formal employees had access to social security. In terms of protective SP (or social assistance), there were some programs supporting maternal and child nutrition, but these were small-scale, fragmented and funded by different donors.25 In effect, the only social assistance programs targeting nutrition that reached 1 percent of the population (about 100,000 beneficiaries) were financed outside the fiscal system: the School Feeding Program is funded and implemented by the World Food Programme (WFP) and the Micronutrient- Deficiency Program is funded by the European Union (EU), the United Nations Children’s Fund (UNICEF), and the WFP and implemented by the National Nutritional Agency (NaNA) – a government agency.26 27 Within the social assistance category (non-contributory pensions) the Public Service Pension Scheme is the main source of government-funded transfers, which accounts for the total social protection expenditure in Table 2. Box 1 describes the pension schemes available in The Gambia. Box 2 describes the School Feeding Program (the flagship social protection program), which has also been receiving government funding since 2018. Box 1. Benefits from Social Security Schemes in The Gambia Social insurance schemes are currently not well developed in The Gambia. Only 10 percent of formal employees have access to social insurance, including sick pay, work-related injury cover, and pensions (World Bank 2017a). As of 2015, there were three types of pension schemes in The Gambia: - The National Provident Fund (NPF) is the only contributory pension scheme in the country, and it is mandatory for private employees. The social security contribution (SSC) rate is equivalent to 15 percent of employees’ income (5 percent paid by employees, and 10 percent paid by employers). In this scheme, individuals at retirement age (60 years old) receive a lump sum equivalent to their saved contributions plus interest. According to the SSHFC, as of 2015, total NPF claims paid reached 1,134 and total benefits paid reached GMD 107.3 million. - The Public Service Pension Scheme (PSPS) is the civil servants’ pension scheme, including the military and the police. It is still non-contributory (e.g. fully funded by the government) and it is equivalent to a defined benefit for public servants. Under the PSPS, individuals can retire as early as 45 years old. Individuals can receive up to 25 percent of their annual retirement benefit up front as a lump sum and then receive the remainder as a reduced pension throughout their retirement lifetime. Based on World Bank (2010), the average pension benefits under the PSPS were assessed as low (about 31 percent of average pensionable wage) and unpredictable (e.g. not covering all the retirement period). According to MoFEA, as of 2015, the scheme was paying a total of 6,612 pensioners, and total benefits paid reached GMD 80.9 million. The Federated Pension Scheme (FPS) is available for quasi-government employees or those working for parastatals and, like the PSPS, it is also non-contributory. In this scheme individuals receive a (low) defined benefit. According to the SSHFC, as of 2015, total pensions paid by the FPS reached GMD 74.2 million and the total number of pensioners was 2,812. The NPF and FPS are managed by the SSHFC while the administration of the PSPS is divided between the Personnel Management Office (PMO), the Public Service Commission, the Treasury, and the Auditor General. Self-employed workers may also voluntarily enroll in any of the SSHFC pension schemes. Source: Authors’ elaboration based on data from the World Bank and official sources (SSHFC and MoFEA). 25 World Bank (2018) presents in Table 2 the list of the main SP programs in The Gambia along with characteristics. 26 The Micronutrient Deficiency Program’s goal is to tackle micronutrient deficiency by providing Vitamin A supplements; the target population is children below 5 years old and post-partum mothers (less than 8 weeks post- delivery. Sources: World Bank 2018, and the NaNA website http://nana.gm/2020/06/10/micro-nutrients-deficiency- control-programme/. 27 The Gambia also had a disaster-relief program called Targeted Food and Nutrition Assistance, Protracted Relief Recovery�, that provided supplemental feeding for households affected by floods. However, it was funded by WFP . 11 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Box 2. The School Feeding Program in The Gambia The School Feeding Program (SFP) is the largest social protection program (near-cash transfer) in The Gambia. The program’s origins date to the 1970s and 80s, but it was discontinued and restarted in 2012, as a partnership between the World Food Programme (WFP) and the Government of The Gambia (GoTG). As of 2015, the program was fully funded and implemented by the WFP in four regions (Banjul City, Central River Region, Upper River Region, and North Bank Region), meaning it was not part of the fiscal system at that time. The SFP aims to improve nutrition and education access among The Gambia’s most vulnerable population. It targets the geographic areas with the lowest school enrolment, and high levels of malnutrition, food insecurity and poverty. Within these regions, it has selected certain early childhood development (ECD) and primary schools. The beneficiary schools are from the public system, including some Madrassas (religious schools) accredited by MoBSE for providing the conventional school curriculum. The SFP provides each child attending the beneficiary schools a daily meal consisting of: 80 grams of rice, 30g of beans or tinned fish, 5g of vegetable oil, 50g of corn soya blend and 10g of sugar. The SFP also strengthens local procurement capacity by linking the program with local agricultural producers and by employing vulnerable women in the community as cooks. As of 2015, the SFP reached a total of 99,603 children in ECD and primary schools: 11,104 children under 5 and 88,499 5-18 year-olds. It also hired 822 adult women as cooks. The total executed budget was US$11.87 million (GMD 504.72 million), and about 51 percent of this budget was destined to the food component. Based on MoBSE data, in 2015 the schools benefitting from the SFP were concentrated in four regions (Banjul City, Central River Region, Upper River Region, and North Bank Region). Starting in 2018, the GoTG extended the program to the other two regions (Lower River Region and West Coast region) not fully covered by the WFP with the plan of gradually taking over the entire running of the program. Source: Authors’ elaboration based on data from the World Food Programme. Education The Gambian formal education system is divided into four main levels: early childhood development (ECD) for 3-6 year-olds, basic education (6 levels of lower basic education and 3 levels of upper basic education), senior secondary education (3 levels), and post-secondary education (tertiary, non-tertiary, and higher education28). In terms of governance, MoBSE is responsible for ECD and basic and secondary education, including Madrassas (religious schools following the official curriculum29). MoHERST is responsible for tertiary, vocational, and higher education. As of 2015, 23.8 percent of children attending ECD were enrolled in public centers, so the private sector was the main provider.30 For lower basic, upper basic and secondary education, over 70 percent of total enrollment was in public schools. In the case of post-secondary education, about 72 percent of total enrollment was in public schools.31 To analyze education public expenditure in The Gambia from a variety of data sources,32 this study re- grouped education levels using the UNESCO international standard classifications: basic education (ECD plus grades 1-6); secondary education (grades 7-12), and post-secondary education (post grade 12). Based 28 Non-tertiary and tertiary education provide certificates and diplomas; the latter can lead to higher education. Higher education provides degrees (World Bank 2017b). 29 The GoTG also provides grants to Madrassas to facilitate their integration to the conventional education system (World Bank 2017b). 30 Source: World Bank 2017b 31 The leading public university is the University of The Gambia, which includes The Gambia College and the School of Nursing. 32 Data sources: MoBSE, MoHERST and World Bank (2017b). 12 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia on these definitions, as of 2015, the GoTG spent about 2.2 percent of GDP on education.33 Basic education accounted for 75 percent of this expenditure, secondary education 13 percent, and post-secondary education 12 percent. The Gambia’s education expenditure is lower than the average for low-income countries in Africa (4 percent of GDP34). Despite the free tuition policy in public schools at the basic and secondary education level, there are still significant indirect costs (e.g. uniform, transport, school meals), that hinder education access among the poor (World Bank 2017b, p.98; World Bank 2011). Government support for higher and tertiary education is usually in the form of universal subventions to institutions and the provision of scholarships to high- performing students. However, financial constraints and lower school completion rates, especially for students from deprived backgrounds, mean students from poor households are less likely to access higher education. Health In The Gambia, the public sector is the main provider of healthcare with facilities in both rural and urban areas. Public facilities are particularly important for covering rural areas. In urban areas (Greater Banjul and West Coast regions), there has been a recent increase of private providers for secondary and tertiary levels of care, but the public sector still provides over 75 percent of health facilities (Sine et al. 2019). Access to health services is mixed compared to peer countries in the region, with both relative strengths and relative gaps.35 There is no national health insurance; only 4 percent of the population had access to health insurance in 2015, through formal employment. Patients pay user fees, but these are very minimal (Sine et al. 2019). Based on the data sources used in this study,36 public health expenditure in The Gambia amounted to 1.2 percent of GDP in 2015. About 24 percent of this expenditure was driven by inpatient services (e.g. tertiary care), 23 percent by outpatient services (e.g. primary and secondary care), and the remaining 53 percent corresponded to other health non-contributory expenditure.37 The GoTG health expenditure is lower than the average for low-income countries in Africa of 2 percent of GDP (Beegle and Luc 2019, p.254). However, when external funding is also included, total health expenditure in The Gambia reached 4.7 percent of GDP, which shows that the health sector is very dependent on donor funding.38 33 A significant part of government expenditure is financed through the School Development Fund, which might not be reflected in budget accounts. 34 The average for African low-income countries is based on Beegle and Luc (2019, p. 253). 35 “Areas where The Gambia outperforms compared to the Africa regional average include prenatal care, childhood immunization, and access to improved water and sanitation. Areas where The Gambia underperforms relative to the Africa regional average include maternal mortality and antiretroviral therapy coverage among HIV-positive individuals� (Sine et al. 2019, p. 14) 36 Ministry of Health and World Bank (2020b). 37 Other health expenditure includes public expenditure on water, sanitation and hygiene (WASH), which is managed by the Ministry of Health. 38 Sine et al. (2019), based on National Health Accounts 2015. 13 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 4. Data and Methodology 4.1. Concepts and Methodology The design and implementation of fiscal policies should consider different dimensions: (i) fiscal sustainability; (ii) economic efficiency, (iii) externalities and (iv) equity (Figure 1). The objective of this fiscal incidence analysis is to analyze the impact of the fiscal system on poverty and inequality in The Gambia. The motivation for performing the analysis is threefold: (i) there has not been a comprehensive distributional assessment of the country’s fiscal system; (ii) the NDP 2018–21 has prioritized social expenditure but the country has limited fiscal space to finance these expenditures; and (iii) given the above, having a fiscal incidence analysis will be an important input for promoting tax and expenditure reforms that are consistent with both fiscal sustainability and fiscal equity. Fiscal incidence analysis aims to measure how government tax collection and social expenditure affect households’ income distribution. The motivation of such analysis is to answer questions such as: What is the impact of taxes and spending on poverty and inequality? Which taxes and transfers are progressive or regressive? Who bears the burden of taxes and receives the benefits of social expenditures? Which households are net payers or net receivers of the fiscal system? Moreover, by developing detailed and parametrized microsimulation models for fiscal incidence analysis, it creates a workhorse which can be used not only to assess the distributional impacts of existing fiscal policies but also to simulate the distributional impact of potential policy reforms. These models can therefore be particularly useful to inform evidence- based fiscal policy design, taking into account equity considerations before and after implementation. Figure 1. Key Considerations of Fiscal Policy Source: Authors’ elaboration. The Fiscal Incidence Analysis in The Gambia (2015) is developed using the Commitment to Equity (CEQ) Methodology,39 which provides a systematized framework to determine the distributional impacts of the fiscal system on outcomes such as poverty and inequality. The standard CEQ model covers several fiscal interventions (direct taxes, indirect taxes, direct transfers and indirect subsidies, and in-kind benefits from 39 The Commitment to Equity project (CEQ) has been led by Nora Lustig at Tulane University. The latest CEQ Handbook was published in 2018 and is available online. More information at: http://commitmentoequity.org/publications-ceq-handbook 14 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia health and education) and aims to model how fiscal systems work in practice.40 The CEQ Methodology has already been implemented in over 70 countries, which facilitates the production of results that are internationally comparable. Figure 2. Definitions of Income Underpinning the CEQ Fiscal Incidence Analysis Market income (gross wages and salaries, income from private capital, private transfers, income from own-production and imputed rent) Market income plus pensions Direct transfers Direct taxes (e.g. social assistance in cash or in-kind) (e.g. Personal Income Tax) Disposable income Indirect taxes Indirect subsidies (e.g. VAT, custom duties, (e.g. fuel, food, other) excises) ( Consumable income In-kind benefits Payment for public (e.g. monetized value of public services services such as education and health) (e.g. co-payment, fees) Final income Source: Adapted from Lustig (2018b). Commitment to Equity Handbook. Building a fiscal incidence model under the CEQ Methodology requires understanding how the country’s fiscal system works (based on legislation and administrative data) and then modelling how taxes and transfers are allocated across households and individuals, using micro data from a representative socioeconomic household survey. Once all taxes and transfers are modelled, the CEQ Methodology calculates different income concepts for each household to assess how fiscal policy affects households’ income at various stages (Figure 2). For each household, the analysis starts with “market income� or “market income plus pensions�as the pre-fiscal income and then subtracts taxes and adds transfers to obtain households’ final income (post-fiscal income). 40Other fiscal incidence methodologies such as EUROMOD simulate the different fiscal interventions de jure, or based on the design rules, which is perhaps a reasonable approach in developed countries. However, in developing countries, facing tax informality and social program implementation challenges, it makes more sense to aim for de facto modelling, such as that embraced by the CEQ Methodology. 15 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Fiscal incidence analysis is “both methodology and art�. Every country poses its own challenges, depending on the complexity of the fiscal system and the data availability. Some underlying assumptions used when building the standard fiscal incidence model under the CEQ Methodology imply certain limitations in terms of scope or methodology: (i) it uses standard incidence analysis without behavioral, lifecycle or general equilibrium effects; (ii) it does not consider the quality of services delivered by the government; and (iii) it does not include taxes such as corporate income tax, which is better modelled with tax administrative data, nor investments in infrastructure and public goods, since the methodologies are yet to be developed. 4.2. Data Sources National Household Survey IHS 2015/16 For this fiscal incidence analysis, the master microeconomic dataset is the Integrated Household Survey (2015/16), which is a nationally representative socioeconomic survey.41 The IHS is among a series of household surveys regularly conducted by the Gambia Bureau of Statistics (GBoS) in collaboration with its international partners the World Bank, the United Nations Development Programme (UNDP) and UNICEF. The IHS collects information on households’ income and employment, along with a detailed consumption module with expenditure data collected at the product level. It also collects information on the use of health and education services, household demographic characteristics, and other themes.42 The 2015/16 IHS surveyed 13,281 households,43 the first time that a survey of this magnitude had been conducted in The Gambia.44 Although the IHS was conducted between April 1st 2015 and March 1st 2016, most of the survey was carried out in 2015.45 Hence, this study took 2015 as the reference year for the analysis of the fiscal system. Administrative Data Administrative data were collected from different sources, broken down as follows: - The GRA provided data on actual tax collections (2015), disaggregated by tax type. GRA officials also shared feedback on indirect tax rates by products. - MoFEA provided data on executed government expenditure. - Data on pensions (budget and pensioners) were provided by the SSHFC for the NPF and FPS and by MoFEA for the PSPS. - MoFEA provided data on indirect subsidies for fuel (market prices and subsidized prices). - MoBSE provided data on the education sector (executed budget and students enrolled) for the primary and secondary level of education. This was complemented with official data previously collected by the World Bank (2017b). - The MoH provided data on the health sector (itemized budget and health cases). This was complemented by expenditure analysis performed by the World Bank (2020a). - The WFP provided data from the School Feeding Program (executed budget and beneficiaries). MoBSE provided additional data on beneficiary school districts and budget for the government-run programs. 41 The survey is statistically representative on the level of districts. 42 Some modules were available at the individual level (income and employment, education, and health) while other modules were available at the household level (expenditure data by product). 43 According to GBoS, 13,340 households were sampled, but 13,281 households were interviewed. 44 The expanded population in the sample (applying weights) was equivalent to 1.872 million individuals, close to the official population estimate of the country in 2015 (1.978 million). 45 More information about the IHS 2015/16 available at: https://microdata.worldbank.org/index.php/catalog/3323 16 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 5. Application of the Fiscal Incidence Analysis in The Gambia 5.1. Coverage The Fiscal Incidence Analysis in The Gambia (2015) aims to model the core fiscal interventions from the country’s tax system and social public expenditure. On the tax side, the study modeled PIT, total VAT (from domestic and imported goods), excises, and custom duties. Together, the total direct and indirect taxes included in this study represent approximately 78.6 percent of total government tax revenues in 2015. On the social expenditure side, the analysis covered the non-contributory PSPS scheme (social protection), indirect subsidies from fuel, health expenditure (inpatients and outpatients), and education expenditure (basic, secondary, and post-secondary). Altogether, the total social expenditure covered in the analysis represents approximately 18.7 percent of total primary expenditure or 75.5 percent of total social expenditure. Due to limitations of the data and methodology, the analysis excluded taxes such as CIT, property tax46 and payroll tax.47 The analysis also excluded indirect agriculture subsidies and infrastructure spending. 5.2. Assumptions The analysis follows the CEQ Framework (Figure 2) for the definition of income concepts. Given that consumption data are more reliable than income data in the IHS 2015/16, the calculation of income concepts starts with disposable income equated to the official consumption aggregate available in the IHS 2015/16.48 The CEQ income concepts above disposable income were then calculated backwards: net market income equals disposable income minus direct transfers; market income plus pensions equals net market income plus direct taxes. The CEQ income concepts below disposable income are calculated in the standard way: consumable income equals disposable income minus indirect taxes plus indirect subsidies; final income equals consumable income plus in-kind benefits from health and education. The income concepts described above were calculated using a combination of a direct identification approach from survey data, simulation, and imputation techniques using survey and administrative data as inputs. Lastly, regarding the treatment of SSC and pensions in The Gambia, the study worked with the CEQ “pensions as deferred income� scenario, where market income plus pensions is the pre-fiscal income. Applying this scenario implies that: (i) the contributions to the NPF (contributory scheme) are considered a forced savings rather than a tax, hence they are not included in direct taxes; (ii) income from the NPF (contributory) is considered part of “market income plus pensions�; and (iii) income from the PSPS (non- contributory) is considered a social protection transfer. A description of the main assumptions used to model each fiscal intervention is found below; more detailed data on parameters are presented in Annex III. 46 CIT amounted to 1.4 percent of GDP in 2015. In The Gambia, the CIT paid is the higher of 27 percent of chargeable profit or 1 percent of turnover for the tax year. However, given the data limitations of business income variables in the household survey, CIT it is not included in the analysis. For a similar reason, the main property tax or residential tax (0.1 percent of GDP), which is equivalent to a 10 percent annual rate paid by individuals that collect taxable rental income, is not included in the analysis. 47 There is also payroll tax (0.1 percent of GDP), which is a lump sum of GMD 40,000 expatriate tax collected for hiring non-Gambia nationals. This tax is not modelled since the household survey used is based on the national population. 48 Using consumption as the welfare measure (instead of income) is typical in low-income countries, because the former is easier to report or less subject to fluctuations. (Burtles, G., cited in Lustig, 2018b). 17 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Direct Taxes The main direct tax modelled was PIT on employees, with the analysis focusing on employment income.49 Since the IHS 2015/16 does not have information on direct taxes paid by the household, PIT was simulated based on the income and employment data available for each household member. PIT was estimated based on a variable of annualized employment income (own calculations) and using the 6-bracket PIT schedule available in the official legislation (see Annex III). As a proxy for informality, the study assumed that 100 percent of government employees were formal;50 for all other employees, the study assumed that individuals were formal only if they self-reported in the survey that they were entitled to social security in their jobs. Based on these assumptions, we estimated that about 12.3 percent of employed individuals were formal,51 which is close to previous estimates of labor formality in The Gambia.52 Indirect Taxes Similar to direct taxes, indirect taxes were calculated based on simulations. The indirect taxes modelled were custom duties, excises, and VAT. The IHS 2015/16 has household expenditure data (at the product level53) and this information was combined with statutory rates by product obtained from Customs Department of the GRA. As a proxy for informality, which is prevalent in The Gambia, the study classified products according to the likelihood of formality as the IHS does not have information on place of purchase (see below). - Custom duties: Given that compliance is higher at customs, we assume that all goods that were subject to custom duties and imported in 2015 (based on COMTRADE data) paid custom duties. - Excises: We assume that liable goods were only formal if they paid custom duties. To calculate quantities of excisable products, we divided household expenditure by product price data.54 - VAT: We assume that all goods that paid custom duties in the model also paid VAT; non-tradable services were divided between formal (e.g. telecommunications) and informal (e.g. tailoring, beauty, home repair), based on local technical insights.55 The analysis of indirect taxes focuses on their direct effects. Products (goods and services) that were considered formal were assumed to pay the statutory rate and products that were considered informal were assumed to pay zero tax. Based on these classifications, 43 percent of food products were classified as informal (but since most food items are tax exempt, only 10 percent of VAT-liable food items were classified as informal), and 29 percent of the non-food products were classified as informal.56 The final estimates of taxes for customs duties and VAT were scaled to match their effective tax rates with respect to consumption (see Section 5.3). To improve the distributional approach of informality in indirect taxes, a potential 49 As noted in Section 2, the PIT taxable base also includes other non-labor income (e.g. business income, property income and any other income). However, the IHS 2015/16 did not detail non-labor income data at the individual level. 50 According to GBoS, based on the Labor Force Survey (2018), less than 1 percent of Gambians working in “public administration� are informal. 51 Estimates based on main occupation. Source: Calculations based on IHS 2015/16. 52 Gavrilovic and Dibba (2013) estimated that only 10 percent of the Gambian labor force was formal, based on social security coverage. 53 Approximately 500 items were available in the households’ expenditure module, including both goods and services. Note: The classification of product codes was not under the COICOP classification. 54 The IHS 2015/16 had a module with price data for food and beverages. In addition, we used tobacco price data from secondary sources (Chisha et al. 2019). Some products were excluded due to lack of price data: tobacco wrapped in paper; snuff; washing soap; toilet soap; and international phone calls. 55 This includes the business classification of the Informal Tax Brochure from the GRA. 56 Percentages based on the total products not the value of products. 18 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia extension to this study could be to make an out-of-sample prediction57 of Gambian households’ consumption informality based on evidence available in Bachas et al. (2020). Data limitations such as the lack of an input-output matrix for The Gambia (or for a similar country58) prevented the study from estimating the indirect effects of indirect taxes. This means that the model will not capture the embedded VAT (higher final prices due to unrecovered VAT paid on formal inputs in the production chain) for exempt or informal goods purchased by households. Social Protection For the fiscal year 2015, the main government-funded social protection program was the Public Service Pension Scheme (PSPS). To model the PSPS, the study combined inference and imputation. First, in the household-income module, households that self-reported having received income from “public pensions� were identified. Second, in the individual-level module we inferred how many individuals per household were potential retirees based on the information on age and employment status; we realized that in most households the number of potential retirees was one. Third, for the eligible households that self-reported receiving income form “public pensions,� we imputed the median annual pension for 2015, based on MoFEA’s PSPS pension dataset.59 While the social protection system was very underdeveloped in The Gambia—in terms of government- funded programs—the study recognized that the WFP School Feeding Program was the flagship scheme in the country. To analyze the potential effects of the fiscal system including the SFP in the government’s budget (a policy that started in 2018), we present results from simulations in Section 8. Indirect Subsidies The analysis of the study focused on the modelling of indirect subsidies on light fuel products (kerosene, gasoline, and diesel) based on 2015 price data received from MoFEA. In 2015, these products had fixed prices: gasoline was fixed at GMD 58.77/liter), diesel at GMD 56.59/l, and kerosene at GMD 53.59/l. First, subsidies per unit were calculated based on the difference between the fixed price and the market price (both available in MoFEA data). Given the international drop in oil prices in 2015 the implied price subsidies for fuel products were negative in The Gambia (implicit tax). Second, in the IHS 2015/16, the amount of fuel consumed by households were calculated by dividing total expenditure on each fuel product by its fixed price. Third, for each household, the direct effect of indirect subsidies was calculated by multiplying the average subsidy per unit times the quantity consumed. Due to data limitations,60 the analysis excluded the indirect effects of indirect subsidies. Education In-kind education benefits were modelled for the basic (ECD and grades 1–6), secondary (grades 7–12), and post-secondary (post grade 12) education levels. Following the standard CEQ Methodology, the 57 The estimation of consumption informality based on Bachas et al. (2020) would be “out of sample� because The Gambia was not included in the list of countries studied. 58 There was no input-output matrix for any of The Gambia’s structural peers, nor any country with similar characteristics (including size of population/economy, structure of economy), which could be used. 59 The PSPS dataset received from MoFEA was de-identified at the individual level with monthly pensions data. For each individual (retiree), the 2015 annual pension was calculated. Since there were outliers with high pensions at the top of the distribution, we preferred to impute the median annual pension instead of the average. 60 As with indirect taxes, the lack of an input-output matrix for calculating indirect effects. 19 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia allocation of education in-kind benefits was based on imputation using the average cost approach (at the national level) as follows: 61 - First, for each education level, an annual average transfer per student was calculated for 2015 by dividing the total expenditure executed in each education level by the number of students enrolled in public schools for that education level. 62 In The Gambia, this assumption adapts well to the country’s reality given that MoBSE centralizes the education budget and then distributes it to regional directorates (schools) based on a “per student� grant (World Bank 2017, p34). - Second, potential beneficiaries were identified from the household survey based on individual characteristics: attending a public school, at the basic, secondary, or post-secondary education level. - Third, the annual average transfer was imputed to the public-school students identified in the household survey, according to education level. Health In-kind health benefits were modelled for outpatient services (health consultations) and inpatient services (hospitalizations). The allocation of health benefits was also based on imputation under the average cost approach (at the national level) as follows: - First, based on administrative health data, the average benefit per patient was computed. For each type of health service (inpatients or outpatients), in-kind benefits per patient were calculated by dividing total executed budget by the total number of cases treated in public facilities.63 - Second, the potential beneficiaries in the household survey were identified based on individual-level health variables. For outpatient services, the IHS 2015/16 data allowed the identification of (sick) individuals who attended a health consultation with a public practitioner. For inpatient services, survey data allowed the identification of individuals who had been hospitalized in the last 12 months; households’ out-of-pocket expenditure on hospitalization was used as a proxy for public service.64 - Third, the average annual transfer was imputed to the patients identified in the household survey, according to public health service. One limitation of the “average cost� approach is that it makes households with sick individuals look better off after fiscal transfers compared to households who were eligible to access to these services but did not get sick.65 5.3. Macro-validation After modelling each fiscal intervention from a combination of administrative and household survey information, one key step prior to analyzing the results is to do a macro-validation (i.e., to compare estimates from the study with actual values from official administrative data on taxes and transfers in the year of analysis). Table 3 presents the main results from the macro-validation of this fiscal incidence analysis. The 61 National averages were calculated because it was not possible to get information on the education budget and enrollment disaggregated by region. 62 Based on own calculations, the average annual transfers per student (2015) in The Gambia were GMD 3,253.8 for students enrolled in public primary schools and GMD 3,699.9 for those in public secondary . 63 Based on own calculations, the average annual transfers per patient (2015) in The Gambia were: GMD 7,148.1 and 168.16 for patients that received inpatient and outpatient services in public hospitals, respectively. The lower transfer size for outpatient services was due to a combination of low public expenditure estimated at this level (GMD 155.84 million) and the higher reported number of patients (926,720). In contrast, for inpatient services, total public expenditure was estimated at GMD 166.7 million and the reported number of patients was 23,315. 64 For more details on the assumptions see Annex III. 65 One alternative approach for modelling health benefits is the “insurance value� methodology. Under this methodology, all individuals who potentially have access to public health services are imputed the average in-kind benefit. This is considered under future model extensions (see Annex VIII). 20 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia highlighted rows show the estimates used in the final specifications of the model. The first column presents the micro estimates in GMD (using the household survey data) and the second using the official administrative data. The third column presents the micro estimate (column 1) as a share of column 2. Columns 4 and 5 present the results of columns 1 and 2 as a share of their respective total consumption (based on the IHS and administrative data, respectively). The estimations of taxpayers and beneficiaries by deciles are presented in Annex V. The main takeaways from the macro-validation are the following: - Total household consumption: The total household consumption provided in the IHS 2015/16 covers 88 percent of the aggregate private consumption in the national accounts (NA).66 This means that the comparison of fiscal estimates should be done in relative terms, i.e., as a share of consumption (as in columns 4 and 5). One reason why household consumption in the survey is lower than in the NA is that high-income households, which usually consume more, are typically not well covered in household surveys. - Tax coverage: When comparing the ratios of tax estimates to consumption (columns 4 and 5), the effective PIT rate in our model is 0.8 percent, compared 1.4 percent in the administrative data. This gap could be due to the household survey underestimating income or due to it under-covering high- income earners, who are typically the larger payers of PIT.67 For indirect taxes, the original estimates (unscaled) overestimated custom duties, while VAT estimates had good coverage. In the final model, we applied scaling factors for custom duties and VAT, so that the effective tax rate relative to consumption in the household survey is close to that observed in the macro data. After scaling, the model’s VAT effective rate was close to the macro level (4.0 percent versus 4.1 percent), as it was for custom duties (3.9 percent vs. 4.0 percent). The study found a large underestimation of excise rate (0.2 percent vs 1.2 percent) but upscaling was not done as the gap was so wide. One potential reason for this underestimation could be due to under-reporting of “sin-goods� (alcohol, tobacco) in the household survey; Annex VII includes a potential method to tackle this in future model extensions. - Social expenditure coverage: The model has good coverage of education in-kind benefits, when comparing the ratios with respect to consumption (1.7 percent vs 2.4 percent). In the case of health in-kind benefits (inpatient and outpatient services), the original estimates overestimated admin expenditure due to an overestimation of outpatient benefits, so the latter was downscaled. After downscaling health in-kind benefits, the model’s estimates relative to consumption matched the macro data (0.6 percent vs 0.6 percent).68 In the case of the only social protection expenditure examined (the PSPS), the model used covered only 10 percent of the benefits reported in the administrative data received from MoFEA. The underestimation of PSPS benefits was due to underestimation of the number of PSPS pensioners, given data limitations in the IHS 2015/16. Hence, upscaling was not pursued in this case.69 Lastly, for indirect subsidies, the model’s estimates accounted for -0.2 percent of total household consumption; it is not possible to macro-validate this estimate given that the negative subsidies (implied tax) in 2015 were not recorded in the government accounts.70 66 The gap is not large and is usually expected since household surveys and national accounts have different methodologies for calculating private consumption. For instance, in The Gambia, private consumption in national accounts is calculated as a residual variable. 67 Understanding sources of PIT underestimation would require getting access to further tax admin data from PIT taxpayers and collections by income bracket. This was a data limitation for macro-validation. 68 The source of overestimation was that the number of outpatient beneficiaries in the household survey was higher than the admin value (215 percent). One reason is that the household survey asked for health consultations in a recall period of 2 weeks, and in annualization these consultations (multiplying by 26) we assumed that these visits would occur with the same frequency throughout the year. 69 According to IHS 2015/16, only 1,339 households responded that they received income from public pensions. There was no income from pensioners at the individual level, so it was assumed that each household had one pensioner, based on analysis of proxy variables at the individual level (age and employment status). Hence, only 1,339 pensioners were considered whereas the administrative data from MoFEA recorded 6,612 pensioners. 70 The government recorded a small positive fuel subsidy budget (0.2 percent of GDP) in 2015. 21 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Table 3. Macro-validation of Fiscal Incidence Analysis in The Gambia Values in GMD millions IHS Estimates Admin IHS estimate Admin estimate from IHS value (% of (% of / 2015/16 2015 consumption) consumption) Admin WELFARE AGGREGATES Total household consumption 46,855 53,334.0 87.9% DIRECT TAXES - PIT on formal employees 355 730.3 49% 0.8% 1.4% DIRECT TRANSFERS - - Public Service Pension Scheme* 8.3 80.9 10% 0.0% 0.2% INDIRECT TAXES VAT, unscaled 1,546 2,166.3 71% 3.3% 4.1% Custom duties, unscaled 2,789 2,109.0 132% 4.5% 4.0% Excises, unscaled 89 627.1 14% 0.2% 1.2% VAT, scaled 1,870 2,166.3 86% 4.0% 4.1% Custom duties, scaled 1,832 2,109.0 87% 3.9% 4.0% Excises, unscaled* 89 627.1 14% 0.2% 1.2% INDIRECT SUBSIDIES Fuel subsidies (112) 122.5 n/a -0.2% 0.2% IN-KIND BENEFITS In-kind education benefits, 782 1,289.2 61% 1.7% 2.4% unscaled In-kind health benefits, unscaled 410 322.5 127% 0.9% 0.6% All in-kind education benefits, 782 1,289.2 61% 1.7% 2.4% unscaled All in-kind health benefits, scaled 271 322.5 84% 0.6% 0.6% Source: Authors’ elaboration based on IHS 2015/16 and GMB administrative data. Note: (1) Public Service Pension Scheme and excises are underestimated, but unscaled given the large gap. Suggestions for improving these estimates in terms of data and methodology are given in Annex V. (2) In-kind health benefits are benchmarked against health benefits from inpatient and outpatient services (47 percent of health expenditure in 2015). 22 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 6. Main Findings This section presents the main findings from the fiscal incidence analysis. It aims to answer three questions: - Who are the net payers and net receivers of the fiscal system? - What are the aggregate impacts of the fiscal system on poverty and inequality? - Which fiscal interventions are driving the main results? 6.1. Net Payers and Net Receivers To examine how the 2015 fiscal system affected households across the income distribution, the study calculated the net cash position across different groups across the welfare distribution (Figure 3). The analysis divided the population into 10 deciles ranked by market income plus pensions (pre-fiscal income).71 For each decile, the staked bars show the incidence of the fiscal intervention with respect to market income plus pensions. All the fiscal interventions that represent an income gain to the household are above the zero axis (direct transfers and in-kind education and health benefits); and all the fiscal interventions that represent an income loss for the household are below the zero axis (direct taxes, indirect taxes and negative indirect subsidies). The net cash position (red line) shows the aggregate sum of all cashable interventions (all taxes, direct transfers and indirect subsidies) for each decile; the total cash position (green line) includes all cashable interventions plus in-kind benefits. Figure 3. Net Cash Position of Households After the Fiscal System 4.0 2.0 Incidence by fiscal intervention 0.0 1 2 3 4 5 6 7 8 9 10 (% of mktypp) -2.0 (Poorest) (Richest) -4.0 -6.0 -8.0 -10.0 Deciles by Market Income plus Pensions -12.0 Direct Taxes Direct Transfers Indirect Taxes Indirect Subsidies Education Benefits Health Benefits Net Cash Position Total Cash Position Source: Authors’ calculations based on IHS 2015/16 and administrative data following the CEQ Methodology. When looking at the red line or net cash position (how much cash is left in households’ pockets after paying all taxes and receiving all cash transfers) the results show that all deciles were net payers of the fiscal system in 2015. The net cash loss (with respect to pre-fiscal income) was -9.4 percent for the poorest decile and - 10.1 percent for the richest; the net cash loss was similar for the 2nd–9th deciles (average -8.6 percent). Moreover, the green line shows that, even when looking at the total cash position (adding in the monetized value of in-kind benefits from public health and public education), all deciles were still net payers of the 71The size of the effects represents the overall picture by decile, but there could be further heterogeneity of the effects within deciles (Lustig 2018b, p. 36) 23 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia fiscal system in 2015. The net cash position connects to the impact on poverty and inequality shown in Figure 4 and Figure 5. The fact that all deciles were net payers of the fiscal system is consistent with finding in the next section that the fiscal system increased poverty. Similarly, the fact that the total cash loss was slightly higher for the richer deciles than the poorer deciles is consistent with the finding that the fiscal system (slightly) reduced inequality. 6.2. Overall Impacts on Poverty and Inequality Figure 4. Poverty Impacts of the Fiscal System Figure 5. Inequality Impacts of the Fiscal System (measured at the national poverty line) (measured by the Gini coefficient) 53.7 0.361 0.359 0.359 0.356 0.349 Gini coefficient (0-100) 48.4 48.6 48.6 poverty rate (%) Market Income Net Market Disposable Consumable Market Net Market Disposable Consumable Final Income plus Pensions Income Income Income Income plus Income Income Income Pensions Source: Authors’ calculations based on IHS 2015/16 and administrative data following the CEQ Methodology. Note: (1) Market income plus pensions = pre-fiscal income; net market income = market income plus pensions - direct taxes; disposable income = net market income + direct transfers; consumable income = disposable income – indirect taxes+ indirect subsidies. (2) Poverty headcount is not measured for final income, because final income includes in-kind benefits (health and education) and we only focus on cashable interventions. Impacts on Poverty This section looks at how the poverty headcount changes across various stages of the fiscal system. For this, we assess what percentage of the total population is below the national poverty line for each income concept of the fiscal incidence analysis. The estimations are based on the 2015 national poverty line (GMD 18,039.95, annual per capita,72 real), meaning that every individual with an annual consumption level below this threshold is considered poor. The poverty headcount based on disposable income (48.6 percent) is equal to the official poverty headcount in the country since our definition of disposable income is equivalent to consumption. Figure 4 presents the estimations for different income concepts. The results show that the fiscal system in 2015 increased the poverty headcount in The Gambia by 5.3 percentage points (from 48.4 percent to 53.7 percent) when going from the market income plus pensions (pre-fiscal income) to consumable income (post-fiscal income, excluding in-kind benefits). Most of the poverty increase (5.1 percentage points) took place between disposable income (market income - direct taxes + direct transfers) and consumable income (disposable income - indirect taxes + indirect subsidies), which suggests that indirect taxes are driving the poverty increase. To understand why the poverty rate increased, Table 4 presents how the average real annual income per capita changes at various stages of the fiscal system. As it can be seen, at market income 72The national poverty line in The Gambia is based on an annual value of consumption per capita, temporally and spatially deflated. 24 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia plus pensions (before any fiscal intervention), only the poorest deciles (1–5) had an average annual income below the poverty line. However, at the stage of consumable income (after individuals pay all taxes and receive cashable taxes), the average incomes per capita had fallen in all deciles. In particular, the average income per capita in decile 6 fell from GMD 20,074.3 to GMD 18,476.2 (when going from market income to consumable income), the latter being close to the poverty line. Table 4. The Gambia: Evolution of Income per Capita, at Various Stages of the Fiscal System, 2015 Values: Per capita, annual, GMD real Market income Market income Net market Disposable Consumable Final income per plus pensions plus pensions per income plus income per income per capita per capita capita pensions capita capita (+) Indirect (+) In-kind health (-) Direct (+) Direct Pre-fiscal income subsidies (-) and education taxes transfers indirect taxes benefits Poorest 7,211.9 7,193.7 7,194.0 6,664.2 7,124.4 2 10,484.7 10,469.0 10,471.2 9,668.2 10,195.4 3 12,703.6 12,670.4 12,672.0 11,697.5 12,217.5 4 14,902.2 14,842.1 14,843.9 13,687.2 14,194.8 5 17,186.7 17,083.3 17,084.1 15,718.8 16,244.1 6 20,074.3 20,018.0 20,019.3 18,476.2 19,015.7 7 23,571.3 23,486.8 23,492.5 21,677.3 22,280.9 8 28,080.5 27,895.1 27,901.6 25,690.7 26,216.6 9 35,206.0 34,891.8 34,892.6 31,996.5 32,562.1 Richest 67,982.9 67,088.9 67,109.5 60,781.0 61,364.1 Source: Authors’ calculations based on IHS 2015/16 and GMB administrative data. Note: Cells colored in orange shows deciles where the average annual income per capita (real) is below the national poverty line of GMD 18,039.95. While the net cash position showed that all deciles were net payers of the fiscal system, the study also determined what proportion of poor and the nonpoor households became poorer as a result of the fiscal system. To ascertain this, we calculated the Fiscal Impoverishment Indicator, and the results are presented in Table 5.73 The second column shows fiscal impoverishment as a share of the total population (the share of nonpoor households that became poor) and the third column shows fiscal impoverishment as a share of post-fiscal poor (what percentage of the already poor had become poorer). The results highlight that when going from market income plus pensions (pre-fiscal income) to consumable income (post-fiscal income, excluding in-kind benefits), the proportion of the population that went from nonpoor to poor was 0.06 percent. However, the most salient result is that 100 percent of poor individuals were poorer as a result of the fiscal system. Again, the fact that most of the fiscal impoverishment happens at the last stage (consumable income) suggest that indirect taxes could be driving the results. However, it is likely that this estimation is biased upwards since the fiscal incidence analysis in The Gambia covers more taxes relative to the amount of social expenditure. 73Higgins and Lustig (2016) explain that the poverty gap ratio can be decomposed into “fiscal impoverishment� and “fiscal gains to the poor�. Fiscal impoverishment measures the percentage of the population who became impoverished by the fiscal system (e.g. income falling below the poverty line); this increases the poverty gap. Fiscal impoverishment can be decomposed to assess the pre-fiscal nonpoor who became poor post-fiscal and the pre-fiscal poor who became poorer post-fiscal. 25 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Table 5. The Gambia: Fiscal Impoverishment, 2015 With respect to the national poverty line Fiscally Fiscally impoverished as impoverished as % % of post-fiscal poor of population Market income plus pensions to net market income 0.00 5.13 Market income plus pensions to disposable income 0.00 5.13 Market income plus pensions to consumable income 0.06 100.0 Source: Authors’ calculations based on IHS 2015/16 and administrative data; following the CEQ Methodology Impacts on Inequality The analysis also looked at how inequality changed at various stages of the fiscal system. For this, it estimated the Gini coefficient for the different income concepts; the index ranges from zero (perfect equality) to 1 (maximum inequality). The Gini at disposable income (equivalent to consumption) is estimated at 0.359, similar to the official inequality estimate for The Gambia in 2015. The estimates show that the Gambian fiscal system reduced inequality slightly in 2015 (Figure 5). The Gini coefficient fell by approximately 1.2 points (from 0.361 to 0.349) when going from market income plus pensions (pre-fiscal income) to final income (post-fiscal income, including in-kind benefits). Most of the reduction in inequality (0.7 Gini points) happened when going from consumable income to final income, which suggests that in-kind health and education benefits could be driving the reduction in inequality. It is important to note that, similar to other developing countries, that inequality estimates in this study are likely to be underestimated since the household survey data do not capture well the top income of the distribution. Some studies suggest adjusting inequality estimates by merging household survey data with administrative tax records from top incomes (Lustig et al. 2019), but tax records data were not available for The Gambia to perform this adjustment. 6.3. Results by Fiscal Intervention The aggregate results show that the combined effect of the fiscal system in The Gambia is to increase poverty and slightly reduce inequality. To understand what is driving these results, this section presents the incidence, progressivity, and marginal contributions of each fiscal intervention. For direct taxes and direct transfers, the incidence and progressivity results are presented with respect to market income plus pensions (the relevant base for interventions that are defined relative to income). For indirect taxes, indirect subsidies and in-kind health and education benefits, the incidence and progressivity results are presented with respect to disposable income (the relevant base for taxes/transfers that are defined relative to consumption). Incidence The study used two indicators to measure incidence: (i) absolute incidence, which shows what share of total taxes are paid (or transfers received) by income decile; and (ii) relative incidence, which shows how much taxes households pay (or transfers they receive) with respect to their income level. Absolute incidence therefore shows the degree of targeting of taxes/transfers, whereas relative incidence shows the actual burden of taxes/transfers across households relative to income. The cumulative sum of absolute incidence by decile shows the concentration of taxes and transfers; this measure, when compared to the concentration curve of a reference income, can be used to assess if a tax or transfer is progressive or regressive (see next section). 26 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Direct taxes: Considering the total direct taxes collected (PIT from employees), most of the tax came from higher deciles. The absolute incidence shows that the richest decile alone paid 43 percent of total PIT in The Gambia in 2015 (Figure 7). Richer deciles contribute more to PIT collections since they are more likely to be formal employees (taxpayers) and also more likely to have higher taxable income (tax base). From the model used, the estimates indicate that 54 percent of total PIT taxpayers simulated (i.e., with a PIT liability of more than zero) belong to deciles 8–10. When looking at the relative incidence, PIT did not impose a large tax burden: the total PIT paid (as a share of market income plus pensions) was -0.54 percent in the poorest decile and -1.22 percent in the richest decile (Figure 6). Nevertheless, the amount is non-negligible in poorer households, which could be because PIT is a withholding tax levied on individuals and therefore does not account for household-level characteristics. The study also notes that while the relative incidence of PIT is slightly higher in decile 1 (-0.54 percent of market income plus pensions) than in deciles 2-4 (average of -0.28 percent of market income plus pensions), these differences are not significant and could be related to the assumptions used in the annualization of income variables in the survey. Such assumptions are subject to bias, which is more likely to affect the bottom of the distribution, where individuals tend to have more volatile jobs. When assessing an alternative scenario where 100 percent PIT formality is assumed, the results of absolute and relative incidence are similar.74 Direct transfers: When looking at absolute incidence, most of the social protection transfers (79 percent) were allocated among the three richest deciles (Figure 9). This is driven by the fact that social protection in the analysis means predominantly social insurance (not assistance) which is a function of formality. In effect, the social protection transfers (PSPS benefits) are associated with previous formal employment, hence it makes sense that they are more concentrated in richer households. However, in relative terms, the estimates show that the PSPS benefits were low relative to households’ market income plus pensions (Figure 8). At most, PSPS benefits represented 0.04 percent of households’ income (in deciles 8–9) and the share is similar for beneficiary households in decile 1 (0.03 percent). It is worth recalling that these results are not likely to be extrapolated to the total population since only 20 percent of total PSPS pensioners in the IHS 2015/16 were identified. 74When assuming 100 percent formality for PIT, the richest decile paid 41 percent of total PIT collections simulated. The relative incidence (as a share of market income plus pensions) was -0.61 percent in decile 1 rising to -1.33 percent in decile 10. These results are similar to those achieved when accounting for labor informality. 27 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Figure 6. Direct Taxes, Relative Incidence Figure 7. Direct Taxes, Absolute Incidence Taxes paid by decile, as a share of their market income plus Taxes paid by decile, as a share of total taxes pensions PIT on employees 50.0 PIT on employees 0.00 45.0 Equality 1 2 3 4 5 6 7 8 9 10 40.0 -0.20 35.0 -0.40 30.0 percent 25.0 percent -0.60 20.0 -0.80 15.0 10.0 -1.00 5.0 0.0 -1.20 -1.40 Deciles by market income+pensions Deciles by market income +pensions Figure 8. Direct Transfers, Relative Incidence Figure 9. Direct Transfers, Absolute Incidence Transfers received by decile, as a share of their market income Transfers received by decile, as a share of total transfers plus pensions 0.05 Public Service 40.0 0.04 Pension Scheme 35.0 0.04 30.0 Equality 0.03 25.0 0.03 20.0 percent percent 0.02 15.0 0.02 10.0 0.01 5.0 0.01 0.0 0.00 Deciles by market income+pensions Deciles by market income+ pensions Source: Authors’ calculations based on IHS 2015/16 and administrative data. Note: (1) Absolute incidence is presented as a positive number and the sum of all shares by decile adds up to 100% (total taxes or total transfers estimated). (2) For taxes, relative incidence is presented as a negative (share of income loss) while for transfers, relative incidence is presented as positive (share of income gain). Indirect taxes: Most of the total indirect tax collections (73 percent) came from the richer deciles (deciles 6–10), whereas the poorer deciles (deciles 1–5) accounted for 27 percent (Figure 11). However, when looking at the relative incidence of indirect taxes (total taxes paid relative to disposable income), the tax burden looks similar across all deciles (between -8 and -9 percent); moreover, the highest tax burden, -8.9 percent, was found in the poorest decile (Figure 10). It is worth noting that most of the indirect taxes estimated in the model stem from VAT (49 percent) and custom duties (48 percent), with only a small share being excises (2 percent). The incidence of indirect taxes could have been lower in poorer households if information on place of purchase was available to better proxy their consumption informality. It is likely that the incidence of indirect taxes is overestimated at the bottom of the distribution since informal consumption tends to be higher in poorer households (Bachas et al. 2020). Indirect subsidies: As mentioned in Section 3, in 2015 the implicit indirect subsidies on light fuel products were negative in The Gambia (similar to a tax), as a result of low international fuel prices. Considering the direct effects of indirect subsidies, the results show that these were low relative to households’ disposable 28 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia income (less than -1 percent for all deciles); however, these negative indirect subsidies (implicit taxes) were concentrated in richer deciles both in absolute and relative terms (Figure 12 and Figure 13). Figure 10. Indirect Taxes, Relative Incidence Figure 11. Indirect Taxes, Absolute Incidence Taxes paid by decile, as a share of their disposable income Taxes paid by decile, as a share of total taxes Excises Custom duties VAT VAT Custom duties Excises Equality 0.0 30.0 -1.0 25.0 -2.0 -3.0 20.0 percent -4.0 15.0 percent -5.0 10.0 -6.0 -7.0 5.0 -8.0 0.0 -9.0 -10.0 Deciles by disposable income Deciles by disposable income Figure 12. Negative Indirect Subsidies (Implicit Figure 13. Negative Indirect Subsidies (Implicit Taxes), Relative Incidence Taxes), Absolute Incidence Taxes paid by decile, as a share of their disposable income Taxes paid by decile, as a share of total taxes Kerosene Diesel Gasoline Gasoline Diesel Kerosene Equality 0.00 100.0 -0.10 90.0 80.0 -0.20 70.0 percent percent -0.30 60.0 -0.40 50.0 40.0 -0.50 30.0 -0.60 20.0 -0.70 10.0 0.0 -0.80 1 2 3 4 5 6 7 8 9 10 Deciles by disposable income Deciles by disposable income Source: Authors’ calculations based on IHS 2015/16 and administrative data. Note: (1) Absolute incidence is presented as a positive number and the sum of all shares by decile adds up to 100% (total taxes or total subsidies estimated). (2) For taxes, relative incidence is presented negative (share of income loss) and for negative subsidies relative incidence is also presented as a negative (share of income loss). In-kind health and education benefits. In 2015, the estimates on absolute incidence of the public health expenditure (outpatient and inpatient services) indicate that the five richer deciles benefitted from 61 percent of this spending whereas the five poorer deciles benefitted from the remaining 39 percent (Figure 15). Similarly, in the case of public education expenditure (ECD, primary, secondary, and post-secondary), 67 percent of the budget benefitted the five richer deciles, while the five poorer deciles benefitted from the remaining 33 percent (Figure 17). These results relate to the fact that more individuals in the five richer deciles accessed public education services (66 percent of the total public students estimated, at all levels) and public health services (61 percent of total health visits estimated). This is consistent with the Gambian context, where the public sector is the main provider of both health and education services. However, when looking at relative incidence, in-kind benefits from public health and education represented a larger share of the poorer households’ disposable income. On average, in the five poorer deciles, in-kind health benefits represented 0.9 percent of their disposable income compared to 0.5 percent in deciles 6–10 (Figure 14). 29 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Similarly with education, in-kind education benefits represented on average 2.1 percent of households’ disposable income in the five poorer deciles and most of these benefits came from primary education, where access rates are higher among poor households. In contrast, in the five richer deciles in-kind education benefits represented on average 1.7 percent of households’ disposable income (Figure 16). Figure 14. In-Kind Health Benefits, Relative Figure 15. In-Kind Health Benefits, Absolute Incidence Incidence Transfers received by decile, as a share of their disposable Transfers received by decile, as a share of total transfers income Outpatient benefits Inpatient Benefits Inpatient Benefits Outpatient benefits 1.2 18.0 16.0 1.0 14.0 0.8 12.0 percent 10.0 percent 0.6 8.0 6.0 0.4 4.0 2.0 0.2 - - Deciles by disposable income Deciles by disposable income Figure 16. In-Kind Education Benefits, Relative Figure 17. In-Kind Education Benefits, Absolute Incidence Incidence Transfers received by decile, as a share of their disposable Transfers received by decile, as a share of total transfers income ECD Primary Post-secondary Secondary Primary ECD 35.0 Secondary Post-secondary 30.0 Equality 2.5 25.0 2.0 percent 20.0 1.5 percent 15.0 1.0 10.0 0.5 5.0 - 0.0 Deciles by disposable income Deciles by disposable income Source: Authors’ calculations based on IHS 2015/16 and administrative data. Note: (1) Absolute incidence is presented as a positive number and the sum of all shares by decile adds up to 100% (total benefits estimated). (2) For benefits, relative incidence is presented as positive (share of income gain). 30 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Progressivity This section explores which taxes and transfers were overall progressive and which ones were overall regressive in 2015. To assess progressivity, the study estimated the Kakwani Index (KI) by fiscal intervention. The KI is an aggregate indicator that allows the assessment of how progressive/regressive a tax or transfer is by comparing its concentration coefficient to the Gini of the reference income.75 The KI links to the absolute incidence indicator since the cumulative sum of the latter is used to calculate the concentration shares of fiscal interventions. A positive KI means that a tax/transfer is progressive, a negative KI means that it is regressive and a KI equal to zero means that the tax/transfer is neutral. The results are considered significant when the KI is above 0.10 in absolute value; KI results between 0.05 and 0.10 are considered mildly significant. One limitation of the KI is that since it is an aggregate indicator, it does not show the heterogeneity of the distributional results across income groups. For this reason, following Lustig (2018b), it is important to complement the Kakwani Index with an analysis of concentration curves relative to the income Lorenz Curve to assess if the results of progressivity and regressivity are unambiguous (Box 3). The results for the Kakwani Indexes are presented in Figures 18–21. To maintain consistency with the incidence results, the KI is calculated with respect to market income plus pensions for the case of direct taxes and direct transfers. For the rest of fiscal interventions, the KI is calculated taking the disposable income as the benchmark. The results on progressivity as of 2015 are as follows: • Direct taxes and transfers: The PIT in The Gambia was progressive and direct transfers (PSPS benefits) were mildly regressive relative to market income plus pensions (Figure 18 and Figure 19). The progressivity of the PIT is consistent with the fact that richer households are more likely to be formal employees; these results were unambiguous when looking at the PIT concentration curve, which lies below the market income Lorenz Curve across all distributions. The regressivity of PSPS is consistent with the fact that these pension benefits are related to previous formal employment; however, these results were ambiguous since the PSPS concentration curve overlapped with the market income Lorenz Curve at the bottom and top of the distribution (Box 3). • Indirect taxes and subsidies: In the analysis of indirect taxes, the study included negative fuel subsidies (implicit taxes) along with VAT, custom duties, and excises (Figure 20). The progressivity measured by the KI relative to disposable income shows that petrol and diesel taxes were unambiguously progressive whereas the kerosene tax was neutral; this was because in the particular year of analysis these subsidies were negative due to the international drop in oil prices. However, since the direct consumption of these fuel products is concentrated in richer households, this also means that for the more usual years, in which fuel subsidies are positive (market price above subsidized price), they would be largely regressive. Second, the KI shows that the excises (largely tobacco) were mildly regressive. Recent research shows that while tobacco taxes are regressive in the short term, they could be progressive in the long term when considering the health impacts of reduced consumption among the poor (Fuchs et al. 2019). The progressivity of VAT and custom duties is close to neutral since the KI was close to zero. VAT in The Gambia was neutral despite the existence of VAT exemptions on staples in a country where poorer households spend a higher share of their budget on food.76 Hence, it is likely 75 The Kakwani Index for taxes is calculated as the difference between the concentration coefficient of the tax and the Gini of the income concept. The Kakwani Index for a transfer is calculated as the difference between the Gini of the income concept and the concentration coefficient of the transfer. This allows that in both cases, a positive KI means that a tax or transfer is progressive; a negative KI means that it is regressive, and a zero KI means neutral (Lustig 2018b). 76 The study estimated that the average household in the poorest decile spends about 57 percent of its budget on food vs 48 percent in the richest decile, which is a stylized fact from budget share studies in poor countries. 31 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia that the study underestimated the VAT progressivity in The Gambia since it was not able to model consumption informality thoroughly in the household survey; recent research shows that accounting for informality makes consumption taxes like VAT become more progressive since informality is higher in poorer deciles (Bachas et al. 2020; Inchauste et al. forthcoming). Finally, the progressivity results of all indirect taxes were unambiguous when comparing the concentration curves with the disposable income Lorenz Curve, except the kerosene negative subsidy, where the results were ambiguous since the concentration curve overlapped with the disposable income Lorenz Curve at several points (Box 3). • In-kind benefits: All in-kind health benefits (inpatients and outpatients) were progressive, as were most in-kind education benefits. The only in-kind benefit that was regressive was post-secondary education, which is consistent with the fact that poor households have less access to higher education than richer households (Figure 21). The progressivity of in-kind benefits were unambiguous in most cases, except for the secondary and post-secondary education benefits, since their concentration curves overlapped with the disposable income Lorenz Curve at the bottom of the distribution (Box 3). Figure 18. Progressivity of Direct Taxes, Based Figure 19. Progressivity of Direct Transfers, Based on the Kakwani Index on the Kakwani Index (wrt Market Income plus pensions) (wrt Market Income plus pensions) PIT (employees) 0.185 Public Service Pension -0.066 Figure 20. Progressivity of Indirect Taxes, Based Figure 21. Progressivity of In-Kind Benefits, Based on the Kakwani Index on the Kakwani Index (wrt Disposable Income) (wrt Disposable Income) Negative Kerosene subsidy Inpatient Benefits -0.006 0.197 (implicit tax) Negative Diesel subsidy Outpatient benefits, scaled 0.200 0.530 (implicit tax) Negative Gasoline subsidy Post Secondary Ed. Benefits -0.121 0.383 (implicit tax) Excises Secondary Ed. Benefits 0.097 -0.093 Primary Ed. Benefits 0.217 VAT, scaled 0.016 ECD benefits 0.256 Custom Duties, scaled -0.012 Source: Authors’ calculations based on IHS 2015/16 and administrative data following the CEQ Methodology. 32 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Box 3. Concentration Curves of Selected Fiscal Interventions Concentration Curve: Direct Transfers 100 This box contrasts the concentration curves relative 90 to the Lorenz Curve for those fiscal interventions 80 where the former overlapped with the latter; this is 70 considered as an indication of ambiguous 60 progressivity results. In the case of The Gambia 50 (2015), progressivity results could be considered as Cumulative share 40 ambiguous in the PSPS, the kerosene implicit tax 30 and the secondary/post-secondary education 20 benefits. 10 0 Deciles by market income plus pensions Equality line Market income plus pensions Direct Transfers (PSPS) Concentration Curve: Indirect Concentration Curve: Education Subsidies Benefits 100 100 90 90 80 80 70 70 60 60 Cumulative share 50 50 Cumulative share 40 40 30 30 20 20 10 10 0 0 Deciles by disposable income Deciles by disposable income Equality line Equality line Disposable Income Disposable Income ECD Gasoline subsidy Diesel subsidy Primary education Secondary education Kerosene subsidy Post-secondary education Source: Authors’ calculations based on IHS 2015/16 and administrative data following the CEQ Methodology. 33 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Marginal Contributions To assess which fiscal interventions had a greater individual impact on poverty and inequality reduction, the study estimated their marginal contributions. The marginal contribution (MC) is understood as the difference in the poverty or inequality indicator with and without the fiscal intervention, measured for a specific income concept. A positive marginal contribution means that the specific fiscal intervention reduces poverty or inequality, and a negative marginal contribution means that it increases poverty or inequality. The marginal contribution is a function of two characteristics: (i) how progressive a tax/transfer is; and (ii) how large a tax/transfer is relative to households’ income. The interaction of these two characteristics determines the final impact on poverty and inequality. Table 6 presents the main results for the marginal contributions by fiscal intervention. When assessing the total poverty increase at consumable income, the MC is -5.3 percentage points. Most of the poverty increase came from indirect taxes, particularly from custom duties and VAT with marginal contributions to poverty of -2.63 and -2.07 percentage points respectively. This is consistent with the fact that these fiscal interventions had the highest relative incidence (burden) on households’ pockets. PIT also contributed to the poverty increase (MC=-0.21 percentage points). The fiscal interventions destined to improve households’ cash position (indirect subsidies and direct transfers from the PSPS) were not able to contribute to poverty reduction: the PSPS had only small coverage and was regressive, while indirect subsidies were negative and hence implicit taxes in 2015. When looking at the total inequality reduction at final income (MC=1.2 Gini points), most of the reduction came from in-kind education benefits (MC=0.540 Gini points), mostly explained by the primary-education level (MC=0.443 Gini points), where poor households have attained more access. In-kind health benefits also contributed to inequality reduction (MC=0.195 Gini points). Table 6. Marginal Contributions of Taxes and Transfers to Poverty and Inequality Reduction Reduction in inequality Reduction in national poverty (Gini points) headcount (percentage points) Note: Positive means reduction; negative means increase With respect to market income Direct taxes: PIT on employees 0.178 -0.21 Direct transfers: Public Service Pension Scheme -0.003 0.00 (PSPS) With respect to disposable income Indirect taxes 0.202 -5.02 VAT 0.221 -2.07 Custom duties -0.038 -2.63 Excises 0.016 -0.04 Indirect subsidies 0.114 -0.01 Gasoline 0.027 -0.01 Diesel 0.087 0.00 Kerosene 0.000 0.00 With respect to consumable income In-kind benefits 0.728 n.a Health 0.195 n.a Inpatients 0.092 n.a Outpatients 0.105 n.a Education 0.540 n.a ECD 0.104 n.a Primary 0.443 n.a Secondary 0.054 n.a Post-secondary -0.053 n.a Source: Authors’ calculations based on IHS 2015/16 and administrative data, following the CEQ Methodology. Note: Due to path dependency, marginal contributions do not add up to the total effect. 34 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 7. International Comparison 7.1. Comparing Results on Poverty and Inequality In this section, we contextualize the results for poverty and inequality in The Gambia by comparing them with results from fiscal incidences studies (under the CEQ Methodology) of seven other African countries with different income levels. These comparator countries are Guinea-Conakry 2012 (World Bank 2019b); Ethiopia 2011 (Hill et al. 2017); Ghana 2012 (Younger et al. 2017); Tanzania 2011/12 (Younger et al. 2016); Uganda 2012/13 (Lustig et al. 2016); South Africa 2010 (Inchauste et al. 2017); and Namibia 2009/10s (Sulla et al. 2017). Figure 22. African Countries: Impact of Fiscal Figure 23. African Countries: Impact of Fiscal Systems on Poverty Systems on Inequality At national poverty line By Gini coefficient 70.25 0.85 60.25 0.75 50.25 0.65 40.25 0.55 30.25 0.45 20.25 10.25 0.35 0.25 0.25 Market income disposable income consumable Market Net Market Disposable Consumable Final plus pensions income Income plus Income Income Income Income pensions Ghana (2012) Tanzania (2011/12) Ethiopia (2010)* Ghana (2012) Uganda (2012/13) Gambia (2015) Tanzania (2011) Uganda (2012) Ethiopia (2011) South Africa (2010) Gambia (2015) South Africa Namibia 2009/10 Guinea (2012) Namibia (2009/10) Guinea (2012) Source: Authors’ calculations, based on own calculations for The Gambia; and secondary studies for Guinea 2012 (World Bank 2019b); Ethiopia 2011 (Hill et al. 2017); Ghana 2012 (Younger et al. 2017); Tanzania (Younger et al. 2016); Uganda 2012/13 (Lustig et al. 2016); South Africa (Inchauste et al. 2017); and Namibia (Sulla et al. 2017). Note: Results from Ethiopia, South Africa and Guinea available for market income. The findings of these international comparisons of the impact of taxes and transfers on poverty and inequality can be summarized as follows: - Impact on inequality: In all eight countries (including The Gambia), the fiscal system reduces inequality after all taxes, transfers and in-kind benefits are taken into account. The reduction (from market income plus pensions to final income) in The Gambia was -1.2 Gini points, higher than in Guinea (-0.2 Gini points), but lower than in Ethiopia (-2.0 Gini points), Ghana (-3.5 Gini points), Uganda (-3.0 Gini points), and Tanzania (-5.1 Gini points). The largest reductions in inequality have been estimated for South Africa (17.4 Gini points) and Namibia (20.6 Gini points, see Figure 23). In The Gambia, most of the inequality reduction is driven by in-kind health and education benefits and this is also the case in Ghana, Uganda, South Africa, and Namibia. - Impact on poverty: In five out of eight countries the fiscal system increases poverty after accounting for taxes and transfers that affects households’ cash position (from market incom e plus pensions to consumable income). The increase in The Gambia is only lower than in Tanzania (+6.5 percentage points), and higher than in Ethiopia (+1.2 percentage points), Ghana (+2.2 percentage points), and Guinea (+3.3 percentage points). Only South Africa and Namibia’s fiscal systems reduced the national 35 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia poverty headcount (-2.2 and -3.1 percentage points, respectively) while in Uganda the effect on poverty was neutral (Figure 22). - Fiscal impoverishment: The share of poor individuals who become poorer due to the fiscal system in The Gambia (100 percent) is higher than in Ethiopia (76.6 percent) and Uganda (83.2 percent) and similar to Tanzania (98.6 percent).77 7.2. Interpreting the Results Fiscal systems that are inequality reducing but poverty increasing are common in Africa (Beegle and Luc 2019).78 This largely reflects the fact that African fiscal systems depend heavily on indirect taxes but have limited social protection systems providing targeted cash transfers to the poor. The only two countries in the sample with fiscal systems that reduced both inequality and poverty are Namibia and South Africa, both of which have larger direct transfer programs that offset the impoverishing effects of taxes on households’ pockets. The size of social protection expenditure in Namibia (1.7 percent of GDP) and South Africa (3.8 percent of GDP) was significantly higher than in The Gambia, even after including donor-funded programs (0.9 percent of GDP). Figure 24. African Countries, Ranked by Tax Figure 25. African Countries, Ranked by Revenue Social Expenditure (share of GDP) (share of GDP) 35.0 35.0 30.0 30.0 25.0 25.0 20.0 20.0 15.0 15.0 10.0 10.0 5.0 5.0 0.0 0.0 Source: Authors’ calculations, based on own calculations for The Gambia; and secondary studies for Guinea (World Bank 2019b); Ethiopia 2011 (Hill et al. 2017); Ghana 2012 (Younger et al. 2017); Tanzania (Younger et al. 2016); Uganda 2012/13 (Lustig et al 2016); South Africa (Inchauste et al. 2017); and Namibia (Sulla et al. 2017). Nevertheless, it is worth acknowledging that Namibia and South Africa are the two larger upper-middle income countries in the sample. This explains why they have higher (more than double) tax revenue and social expenditure than the average of the rest of the countries. The Gambia had the third lowest tax revenue to GDP in the year of analysis (Figure 24), only higher than Ethiopia and Uganda, while ranking bottom for social expenditure (Figure 25). It is likely that these rankings would change if recent policy changes were taken into account, particularly, the increase in social expenditure in The Gambia after 2015. The next section discusses recent policy developments in The Gambia in both tax and social expenditure. 77 The fiscal impoverishment in The Gambia was based on the national poverty line. The fiscal impoverishment in Ethiopia, Uganda and Ghana was based on the 1.25 USD 2005 PPP extreme poverty line. Source: Jellema et al. 2018 78 Other countries with poverty-increasing fiscal systems include Mali, Togo and Zambia. 36 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 8. Recent Policy Developments and Potential Simulations 8.1. Recent Developments Since the year of analysis (2015), The Gambia has made changes in its tax policy and social expenditure. During this period, the country saw a political transition, followed by the launch of a new National Development Plan 2018–21.79 This sub-section discusses the fiscal policy changes that affect households in the fiscal incidence analysis; a more detailed description of the policy changes is presented in Annex VI. Changes in tax policy: In the area of direct taxes, there have been reductions in the main income taxes (PIT and CIT). In the case of PIT, the threshold for exempt (annual gross) income has been increased from GMD 18,000 to GMD 24,000 and the maximum PIT rate has been reduced from 30 to 25 percent. In the area of indirect taxes, one of the main changes has been the elimination of the custom duty on rice in 2019 and the increase in excises for certain tobacco/alcoholic products. No major changes have been identified in VAT, the main tax in 2015 (Table 7). Table 7. Main Tax Policy Changes, 2015–20 Category Effective in 2015 Effective in 2020 Direct Personal January 2013–January 2018: January 2018- present: taxes income tax GMD 0 to GMD18,000: 0 percent GMD 0 to GMD 24,000: 0 percent GMD 18,001 to GMD 28,000: 5 percent GMD 24,001 to GMD34,000: 5 percent GMD 28,001 to GMD 38,000: 10 GMD 34,001 to GMD 44,000: 10 percent percent GMD 44,001 to GMD 54,000: 15 percent GMD 38,001 to GMD 48,000: 15 GMD 54,001 to GMD 64,000: 20 percent percent Above GMD 64,000: 25 percent GMD 48,001 to GMD 58,000: 20 percent Above GMD 58,000: 30 percent Custom Rice 5 percent from 5 percent to 10 percent (2017); duties from 10 percent to 0 percent (2019) Used cars 15 percent 10 percent (2018) (>5 years) Other GMD 110/kg GMD 165/kg (2017) tobacco products Excises Wine 15 percent 60 percent (2019) Beer/stout 10 percent 75 percent (2019) Spirits 15 percent 60 percent (2019) Source: Authors’ elaboration, based on consultations with the GRA. Changes in social expenditure: As part of the National Development Plan, which includes health and education as part of its priorities, the GoTG has increased its budget and participation on social policies relative to 2015. Some of the main changes are described in Table 8. In health, the GoTG has almost doubled its planned budget between 2015 and 2020, from 1.2 percent to 2.1 percent of GDP; the priorities in the sector include expanding healthcare access and children’s immunization. The GoTG has drafted its National Health Insurance Policy, which is pending approval by the National Assembly. In education, the planned budget has also almost doubled between 2015 and 2020, from 2.5 to 4.7 percent of GDP; part of 79The National Development Plan 2018–21 aims to “deliver good governance and accountability, social cohesion, and national reconciliation and a revitalized and transformed economy for the wellbeing of all Gambians.� 37 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia the focus has been on improving quality of education in public schools and expanding access to public ECD. The GoTG has signaled a clear commitment to strengthen the social protection system, although no estimate is available of the budget increase. Since 2018, it has funded the expansion of the School Feeding Program in two regions where the WFP had not reached (Lower River Region and West Coast Region), with the aim of gradually taking over the funding of the program. It has partnered with the World Bank in 2020 to pilot an unconditional cash transfer (the Social Safety Net Program (SSNP) or Nafa program) targeted through a proxy means test with the aim of reaching 6,000 households in 20 districts. More recently, with the COVID-19 crisis, it implemented a package of in-kind benefits (over GMD 734 million) which provided food assistance (rice, oil, and sugar) to 84 percent of Gambian households (Gentilini et al. 2020). Lastly, another cross-sectorial change in public expenditure has been that the GoTG has increased public wages and pensions by 50 percent across the board; since most government employees belong to richer deciles,80 this change is likely to have increased income inequality. Table 8. Main Social Expenditure Changes, 2015–20 Increased in planned budget Main policy changes (as of March 2021) 2015–20 Health Planned budget (as a share of -The National Development Plan (NDP) has a focus on GDP) has almost doubled from expanding healthcare access and child immunization. 1.2 percent to 2.1 percent. -The National Health Insurance policy is at advanced stage (pending approval from National Assembly). Education Planned budget (as a share of More commitment on improving quality of education in GDP) has almost doubled from public schools and on improving access to public ECDs. 2.5 percent to 4.7 percent. Social n/a -GoTG funding the expansion of the School Feeding protection Program in 2 pending regions (LRR and WCR) -COVID-19: Over GMD 734 million allocated in in-kind benefits (including oil, sugar, rice). -GoTG partnering with World Bank (SPJ project) to pilot an unconditional cash transfer (UCT) that targeted beneficiaries using a proxy means test approach. Source: Authors’ elaboration, based on MoFEA and official sources. Data are for planned budget given limitations on executed budget data series. 8.2. Potential Simulations While the fiscal incidence analysis in The Gambia discussed above gives a picture of the distributional impacts of the fiscal system as of 2015, the model it was based on can also be used to assess the fiscal policy changes that have taken place during 2015–20 (model update) and also to assess the distributional analysis of proposed fiscal policy reforms. In addition, the hope is that the coding structure of the fiscal incidence model adopted for this analysis allows more recent household survey data to be incorporated as they become available (e.g. IHS 2020). This sub-section analyzes the potential impacts of two of the main fiscal policy changes that have taken place during 2015–20L the change in the PIT structure and the participation of the GoTG in funding of the School Feeding Program. The simulations allow us to assess: (i) What would have been the PIT incidence if the 2020 PIT structure had been in place in 2015? (Figure 26); (ii) What would have been the 80According to the IHS 2015/16, about 38 percent of total household members employed in government (public sector and parastatals) belong to the two richest deciles (9–10) compared to only 15 percent in the two poorest deciles (1–2). 38 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia incidence of direct transfers if the School Feeding Program had been fully funded by the government in 2015? (Figure 27); and (iii) Adding these two changes (PIT + SFP), what would have been the changes in poverty and inequality due to the fiscal system? (Figure 28 and Figure 29). These estimations are constructed as counterfactual scenarios, maintaining the rest of the fiscal policies as originally modelled for 2015. Nevertheless, they could shed light on how the fiscal policy changes could be shaping the distributional landscape in The Gambia in more recent years. Other ideas for policy simulations are presented in Box 4. More recent fiscal policy reforms, as well as future policy reforms under discussion, could be better evaluated once the fiscal incidence model is updated with the IHS 2020/21 household survey data, when they become available. The main results of the policy simulations are the following: - Under the new PIT scheme, the incidence of direct taxes could be lower. Using the 2020 PIT structure (increasing the PIT income exempted threshold and reducing the PIT maximum rate) on the IHS 2015/16 data, the PIT incidence would have been lower for most deciles. The estimated number of taxpayers and PIT collected would have been lower in 2015 (a reduction of 5,902 taxpayers and -30 percent in PIT collected) if the 2020 changes had been in effect. - The incorporation of the SFP in the social protection budget could improve the coverage and progressivity of direct transfers. The simulation of the SFP in the IHS 2015/16 was based on administrative data from the program and the characteristics of the target population. The results show that would have been 97,314 potential beneficiaries (i.e., students enrolled in primary/ECD in public schools or Madrassas). The model performed well in macro-validation.81 Furthermore, the simulations show that adding the SFP to the 2015 direct transfers (which only included the PSPS pensions), would have increased the social protection coverage and incidence among poor households (Figure 27). For instance, the incidence of direct transfers in the poorest decile would increase from 0.02 percent to 0.77 percent (as a share of their market income plus pensions). Moreover, since the incidence of the SFP decreases with income, adding this program would change the impact of direct transfers from regressive to progressive, with the Kakwani Index going from -0.07 to 0.36. - The combination of policy changes (PIT and SFP) improves the distributional impacts of the fiscal system, with most of the effect driven by the SFP. The results show that if the 2020 PIT structure and the SFP (fully funded under the government’s budget) had both been in place in 2015, the fiscal system would have achieved better distributional impacts. The poverty increase would have been smaller by -0.6 percentage points and the inequality reduction would have been larger by 0.2 Gini points compared to the actual effects of the fiscal system in 2015 (Figure 28 and Figure 29). While the new PIT structure, having a higher tax-exempt threshold, contributes to these results, the estimates show that most of the effect is driven by the potential full absorption of the SFP in the government’s budget.82 Therefore, changes in the PIT structure have very limited effect on redistribution and welfare. Moreover, the results also showcase that the efforts of the government to strengthen the social protection system in The Gambia, could improve the distributional impacts of the country’s fiscal system. 81 For the SFP simulations we use inference and imputation techniques based on administrative data from the School Feeding Program (WFP) and MoBSE. In the household survey (IHS 201/16), we identify the eligible population (students enrolled in ECD and primary levels, public schools and Madrassas) and also perform demographic targeting (ages 3–18) and geographical targeting (only the districts covered by WFP in 2015). The average SFP transfer was calculated (GMD 2,595 annual per student) and imputed to each eligible student. The SFP model performed well in macro-validation: We estimated 97,413 beneficiary students and a total SFP budget (on school meals) of GMD 252.8 million (covering 98 percent of the official values reported by the WFP in 2015). 82 We estimate that 83% of the reduction in the poverty (relative to the baseline) was due to the SFP. 39 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Figure 26. PIT Simulations, 2015 vs 2020 PIT Figure 27. Direct Transfers Simulations, 2015 Incidence by decile as a share of market income plus pensions (without SFP) vs 2020 (with SFP) Incidence by decile as a share of market income plus pensions 1.4% 0.9% 1.2% 0.8% 0.7% 1.0% 0.6% 0.8% 0.5% 0.6% 0.4% 0.3% 0.4% 0.2% 0.2% 0.1% 0.0% 0.0% 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Deciles by market income plus pensions Deciles by market income plus pensions Direct Transfers (without SFP) PIT 2015 PIT 2020 Direct Transfers with SFP Figure 28. Change in Poverty Rate Post-Fiscal Figure 29. Change in Inequality Post-Fiscal System, 2015 vs 2020 Simulations System, 2015 vs 2020 Simulations Effect of personal income tax + School Feeding Program Effect of personal income tax + School Feeding Program 0.00 -0.20 2015 Sim. (PIT 2020+SFP) 6.0 5.3 change in poverty rate (p.p.) -0.40 change in (Gini points) 5.0 4.7 -0.60 4.0 -0.80 -1.00 3.0 -1.20 2.0 -1.40 -1.23 -1.60 -1.38 1.0 -1.80 - -2.00 2015 Sim. (PIT 2020+SFP) Source: Authors’ calculatioIbased on IHS 2015/16 and administrative data. Note: The results presented take the 2015 model and simulate the effects of changes in the PIT and direct transfers. In Figure 26, we simulate what would have happened to direct taxes in 2015 if we had the 2020 PIT structure in place. In Figure 27, we simulate what would have happened if we included the School Feeding Program in the direct transfers in 2015 (e.g. assuming the SFP was government financed in 2015). In Figures 28 and 29 we add both changes together (using the 2020 PIT structure and including the SFP in the social protection budget) to see how these policy changes affect the impacts of the fiscal system on poverty and inequality. Box 4. Ideas for policy simulations for ex-ante distributional analysis of fiscal policies - Simulation of COVID-19 social protection policies - Targeting and distributional impacts of the new targeted unconditional cash transfers program (World Bank SPJ pilot with GoTG) - Distributional impact of VAT without exemptions - Distributional impact of expansion of excises (tobacco, alcohol, telecommunications) - Distributional analysis of the national health insurance scheme. 40 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia 9. Conclusions and Policy Implications 9.1. Main Results and Policy Implications The results show that the fiscal system in The Gambia in 2015 reduced inequality by 1.2 Gini points, but at the same time increased the national poverty rate by 5.3 percentage points. After paying taxes and receiving cashable transfers, all deciles were net payers into the fiscal system and 100 percent of the poor were fiscally impoverished. As regards the drivers of inequality and poverty impacts, the results are based on marginal contributions which depend on both the size and progressivity of fiscal interventions. Most of the inequality reduction came from the progressive benefits of primary education, an area where poor households tend to benefit disproportionally as they have more children. The poverty increase was driven by indirect taxes (mainly VAT and custom duties); although these taxes were close to neutral in terms of progressivity, they still represented a large share of households’ disposable income. One important disclaimer about the main results relate to data and methodology limitations. Specifically, because the model covers more taxes than transfers, it could overestimate the impoverishment effects of the fiscal system. In addition, since the model was not able to account for informality patterns of household consumption in great detail, it is likely that it overestimated the burden of these taxes for households at the bottom of the distribution and hence underestimated their progressivity. The main results for the Gambian fiscal system are in line with the findings from similar studies in the region. Other countries in Africa (Ethiopia, Ghana, Guinea, Tanzania, and Uganda) also have fiscal systems that reduce inequality while at the same time increasing poverty. These results show the importance of having strong and well-targeted social protection systems in place in countries that depend largely on indirect taxes (Beegle and Luc 2019). This is what the experience in Namibia and South Africa shows, where the authorities have been able to target large direct transfers that offset the impoverishment impact of indirect taxes among the poor, thus making them net receivers instead of net payers of the fiscal system. Although the results from Namibia and South Africa are reasonable for upper-middle income countries (with higher tax revenue and social expenditure than The Gambia’s peer countries), tools such as the fiscal incidence microsimulation models could improve how governments design and implement fiscal policies from an equity perspective. 9.2. Recent Policy Developments and Simulations The GoTG has implemented fiscal policy changes post 2015, most of them in the area of social expenditure. Two of the main fiscal policy changes were simulated in the core model: (i) the reduction in the PIT structure and (ii) the potential absorption of the School Feeding Program into the government’s social protection budget.83 The simulation results show that these policy changes could improve the distributional impacts of the fiscal system. If these changes had been in place in 2015, the poverty increase would have been lower, and the inequality reduction would have been higher. Considering these simulations and that most of the public expenditure increase in the NDP 2018-21 has been in health and education (in-kind benefits), it is likely that the fiscal system in The Gambia (as of 2020) could be more inequality reducing but still poverty increasing, since significant targeted cash transfers would be needed to effectively offset the large effect of indirect taxes. 83 Although the GoTG is only funding 2 regions from the SFP since 2018 (WCR and LRR), it is signaling its commitment to expanding the program. In the exercise, we simulated that the GoTG funded the implementation as it took place in 2015 (assuming the 4 regions funded by WFP as part of the government’s budget). 41 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Other fiscal policy changes that took place between 2015 and 2020 could not be included in the simulations, but the results from the baseline model could shed light on the potential direction of changes. For instance, the elimination of the custom duty on rice could improve the distributional impact of this tax, since rice consumption is higher among poorer households.84 On the other hand, the increase on tobacco/alcohol excises could affect poor households in the short term (since this tax was regressive in the analysis); however, when taking into account long-term health impacts from reduced consumption among the poor, research shows that tobacco excises could be progressive (Fuchs et al. 2019).85 On social expenditure, the planned budget on education and health expenditure (as a share of GDP) has almost doubled during 2015–20, reflecting the social priorities under the NDP 2018–21. These changes could also further enhance the distributional impacts of the fiscal system in The Gambia since health and education transfers were overall progressive in 2015. The GoTG is also taking steps towards strengthening the social protection system. One such initiative is the launch of the Social Safety Net Program (SSNP), also called the Nafa program, in partnership with the World Bank in 2020. More recently, with the COVID-19 crises, the GoTG also financed an important in-kind relief program, providing food aid to 84 percent of households. Altogether, the results from this study show that more targeted social protection transfers are needed to offset the impoverishment effects of indirect taxes. To fund such social protection programs, the GoTG will need to increase its tax collection and improve the efficiency of public expenditure components (World Bank 2020a).86 The authors of this study hope that the Fiscal Incidence Microsimulation Model that has been developed for this analysis could serve as a platform to analyze the distributional analysis of these recent and future fiscal policy changes, so that tax and expenditure policies contribute to both fiscal sustainability and equity. One first step to enable the model to become a platform for assessing the distributional impacts of more recent policy reforms would be to update it with the new household survey data (IHS 2020/21) as soon as they become available. 9.3. Data Limitations and Future Extensions One of the major limitations of the analysis was the lack of a variable to act as a proxy for informality in household purchases; although the study tried to proxy informality using custom duties liabilities, this did not allow it to capture the full heterogeneity of informality in households’ consumption patterns. Since informality is high in The Gambia, incorporating informality assumptions could have enhanced the progressivity and distributional impacts of VAT as poorer households are more likely to purchase in informal establishments (Bachas et al. 2020; Inchauste et al. forthcoming). In other words, the estimated impact of VAT on inequality reduction would likely be higher, and its impact on increasing poverty would likely be lower, if the study had been able to incorporate how informality varies across households and types of products. Another important limitation is that no input-output matrix was available for The Gambia or any similar country, which inhibited the estimation of the indirect effects of indirect taxes. To improve 84 For instance, as of 2010, the consumption shares on rice reached 5.14 percent in the poorest quintile vs. 0.8 percent in the richest quintile. Source: World Bank Global Consumption Database https://datatopics.worldbank.org/consumption/. 85 The authors highlight that when facing increases in tobacco prices, poor households are more likely to reduce their consumption. Hence, by reducing their tobacco consumption, poor households are more likely to benefit from health improvement and income-gains over the medium-long term. 86 The Gambia’s most recent Public Expenditure Review ( World Bank 2020a) shows that the country could create additional fiscal space (+4.87 percentage points of GDP) by increasing tax collection (3.3 percentage points of GDP) and reducing expenditure inefficiencies in sectors like health, education, security and public financial management (+1.57 percentage points of GDP). 42 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia future iterations of the Fiscal Microsimulation Model in The Gambia, the authors present a list of data limitations in Annex VII, along with ways to mitigate them, as well as future methodological extensions. 43 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia References Alam, S.A., G. Inchauste, and U. Serajuddin. 2017. “The Distributional Impact of Fiscal Policy in Jordan.� In The Distributional Impact of Taxes and Transfers: Evidence from Eight Developing Countries, 179–198. Aristy-Escuder, J., M. Cabrera, B. Moreno-Dodson, and M, Sánchez-Martín. 2016. “Fiscal Policy and Redistribution in the Dominican Republic.� In The Distributional Impact of Fiscal Policy: Experience from Developing Countries. Arunatilake, N., G. Inchauste, and N. Lustig. 2017. “The Incidence of Taxes and Spending in Sri Lanka.� In The Distributional Impact of Taxes and Transfers: Evidence from Eight Developing Countries, 267–293. Bachas, P., L. Gadenne, and A. Jensen. 2020. “Informality, Consumption Taxes and Redistribution.� World Bank Policy Research Working Paper 9267, World Bank, Washington, DC. Beegle, K. and C. Luc. 2019. Accelerating Poverty Reduction in Africa. World Bank, Washington, DC https://doi.org/10.1596/978-1-4648-1232-3. Beneke, M., N. Lustig, and J.A. Oliva. 2018. “El Salvador: The Impact of Taxes and Social Spending on Inequality and Poverty.� In Commitment to Equity Handbook. A Guide to Estimating the Impact of Fiscal Policy on Inequality and Poverty, 667–697. Brookings Institution Press and CEQ Institute. Benjamin, Nancy, and Ahmadou Aly Mbaye. 2012. The Informal Sector in Francophone Africa: Firm Size, Productivity, and Institutions. World Bank, Washington, DC. https://doi.org/10.1596/978-0-8213-9537-0. Cancho, C., and E. Bondarenko. 2017. “The Distributional Impact of Fiscal Policy in Georgia.� In The Distributional Impact of Taxes and Transfers: Evidence from Eight Low-and Middle-Income Countries, 113–148. Candia, B. 2018. “Taxes, Transfers, and Income Distribution in Chile.� Chisha Z., M.L. Janneh, and H. Ross. 2019. “Consumption of Legal and Illegal Cigarettes in The Gambia. Tobacco Control 2020, s254–s259 http://dx.doi.org/10.1136/tobaccocontrol-2019-055055. Daude, C., N. Lustig, A. Melguizo, and J. Perea. 2017. “On the Middle 70%. The Impact of Fiscal Policy on the Emerging Middle Class in Latin America Using Commitment to Equity.� CEQ Working Paper Series 72, Tulane University. De La Fuente, A., M. Rosales, and J. Jellema. 2017. “The Impact of Fiscal Policy on Inequality and Poverty in Zambia.� Policy Research Working Paper No. 8246. World Bank, Washington, DC. Enami, A. 2018. “Measuring the Effectiveness of Taxes and Transfers in Fighting Inequality and Poverty.� In Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty, 207–216. Brookings Institution Press and CEQ Institute. Fuchs, A., P. Marquez, S. Dutta, and F. Gonzalez Icaza. 2019. Is Tobacco Taxation Regressive? Evidence on Public Health, Domestic Resource Mobilization, and Equity Improvements. Policy Note. World Bank, Washington, DC. GBoS (Gambia Bureau of Statistics). 2019. The Gambia Labour Force Survey (GLFS 2018): Analytical Report. Gasior, K., C. Leventi, M. Noble, G. Wright, and H. Barnes. 2018. “The Distributional Impact of Tax and Benefit Systems in Six African Countries.� WIDER Working Paper 2018/155. UNU-WIDER, Helsinki. Gavrilovic, M. and Y. Dibba. 2013. Moving Towards An Integrated And Equitable Social Protection In The Gambia: Analysis of Social Protection Systems in The Gambia. Gentilini, Ugo, Mohamed Almenfi, Ian Orton, and Pamela Dale. 2020. “Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures.� COVID-19 Living Paper. World Bank, Washington, DC. Goraus, K. and G. Inchauste. 2016. “The Distributional Impact of Taxes and Transfers in Poland.� Policy Research Working Paper No. 7787. World Bank, Washington DC. GoTG (Government of The Gambia). 2012. The Gambia’s Income and Value Added Tax Act 2012. GoTG. 2017. National Development Plan 2018-21. Higgins, S. and N. Lustig. 2016. “Can a Poverty-Reducing and Progressive Tax and Transfer System Hurt the Poor?� Journal of Development Economics 122 63–75. https://doi.org/10.1016/j.jdeveco.2016.04.001. 44 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Higgins, S., N. Lustig, W. Ruble, and T.M. Smeeding. 2016. “Comparing the Incidence of Taxes and Social Spending in Brazil and the United States.� Review of Income and Wealth 62(S1) S22–S46. https://doi.org/10.1111/roiw.12201. Hill, R., G. Inchauste, N. Lustig, E. Tsehaye, and T. Woldehanna. 2017. “A Fiscal Incidence Analysis for Ethiopia�. In The Distributional Impact of Taxes and Transfers: Evidence From Eight Developing Countries . https://doi.org/10.1596/978-1-4648-1091-6_ch3. Hounsa, T., M. Coulibaly, and A. Sanoh. 2019. “The Redistributive Effects of Fiscal Policy in Mali and Niger.� Policy Research Working Paper, World Bank, Washington, DC. Inchauste, G., K. Goraus, and H. Carrasco. Forthcoming. “The Distributional Incidence of VAT in the Context of Informality.� World Bank (mimeo). Inchauste, G. and N. Lustig. 2017. “Overview: Fiscal Policy and Redistribution.� In The Distributional Impact of Taxes and Transfers: Evidence From Eight Developing Countries. https://doi.org/10.1596/978-1-4648-1091-6_ch1. Inchauste, G., N. Lustig, M. Maboshe, C. Purfield, and I. Woolard. 2017. “The Distributional Impact of Fiscal Policy in South Africa.� In The Distributional Impact of Taxes and Transfers: Evidence From Eight Developing Countries 233–266. https://doi.org/10.1596/978-1-4648-1091-6_ch8. IMF (International Monetary Fund). 2015. “2015 Article IV Consultation—Press Release; Staff Report; and Statement by the Executive Director for The Gambia.� Country Report No. 15/272. IMF, Washington, DC. IMF. 2017. Sub-Saharan Africa: Restarting the Growth Engine. Regional Economic Outlook. IMF. ISBN: 978-1-47557- 498-2 Jellema, J., M. Wai-Poi, and R. Afkar. 2017. “The Distributional Impact of Fiscal Policy in Indonesia�. In The Distributional Impact of Taxes and Transfers: Evidence From Eight Developing Countries 149–178. https://doi.org/10.1596/978-1-4648-1091-6_ch5. Jouini, N., N. Lustig, A. Moummi, and A. Shimeles. 2018. “Fiscal Policy, Income Redistribution, and Poverty Reduction: Evidence from Tunisia.� Review of Income and Wealth 64(s1) S225–S248. https://doi.org/10.1111/roiw.12372. Lara Ibarra, G., N. Sinha, R. Fayez, and J. Jellema. 2019. “Impact of Fiscal Policy on Inequality and Poverty in the Arab Republic of Egypt.� Policy Research Working Paper No. 8824, World Bank, Washington, DC. Lustig, N. 2015. “The Redistributive Impact of Government Spending on Education and Health: Evidence from 13 Developing Countries in the Commitment to Equity Project.� CEQ Working Paper Series No. 30. Lustig, N. 2016a. “Fiscal Policy, Inequality, and the Poor in the Developing World.� Working Paper No. 441. Center for Global Development. Lustig, N. 2016b. “Inequality and Fiscal Redistribution in Middle Income Countries: Brazil, Chile, Colombia, Indonesia, Mexico, Peru, and South Africa.� Journal of Globalization and Development 7(1). https://doi.org/10.1515/jgd-2016-0015. Lustig, N. 2017a. “Fiscal Policy, Income Redistribution and Poverty Reduction in Low and Middle Income Countries.� Working Paper No. 448. Center for Global Development. https://doi.org/10.2139/ssrn.2951735. Lustig, N. 2017b. “Fiscal Policy, Inequality, and the Poor in the Developing World.� CEQ Working Paper Series 23, Tulane University. https://commitmentoequity.org/wp- content/uploads/2017/08/CEQ_WP23_Lustig_July2017v2.pdf. Lustig, N. 2018a. “The Sustainable Development Goals, Domestic Resource Mobilization, and the Poor.� In The Welfare State Revisited. Columbia University Press. https://doi.org/10.7312/ocam18544-008. Lustig, N. 2018b. Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty. Brookings Institution Press. Lustig, N. 2018c. “Measuring the Distribution of Household Income, Consumption and Wealth: State of Play and Measurement Challenges.� Tulane Economics Working Paper No. 1801. Tulane University. Lustig, N. 2019. “The ‘Missing Rich’ in Household Surveys: Causes and Correction Approaches.� CEQ Working Paper Series 75, Tulane University. 45 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Lustig, N. 2020. “Measuring the Distributional Impact of Taxation and Public Spending: The Practice of Fiscal Incidence Analysis.� In Oxford Research Encyclopedia of Economics and Finance. Oxford University Press. https://doi.org/10.1093/acrefore/9780190625979.013.363. Lustig, N. and S. Higgins. 2013. Commitment to Equity Assessment (CEQ): Estimating the Incidence of Social Spending, Subsidies, and Taxes - Handbook. https://doi.org/10.2139/ssrn.2366810. Lustig, N., J. Jellema, A. Haas, and S. Wolf. 2016. “The Impact of Taxes, Transfers, and Subsidies on Inequality and Poverty in Uganda.� Working Paper No. 443. Center for Global Development. https://www.cgdev.org/publication/taxes-transfers-subsidies-uganda. Lustig, N., Jellema, J. R., & Wolf, S. (2017). The Impact of Taxes, Transfers, and Subsidies on Inequality and Poverty in Uganda. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2893765 Lustig, N. and H.E. Morán. 2019. “Fiscal Policy Incidence on Inequality and Poverty in Low-and Middle-Income Countries.� Working Paper. Intergovernmental Group of Twenty Four. Lustig, N., C. Pessino, and J. Scott. 2014. “The Impact of Taxes and Social Spending on Inequality and Poverty in Argentina, Bolivia, Brazil, Mexico, Peru, and Uruguay: Introduction to the Special Issue.� Public Finance Review 42(3). https://doi.org/10.1177/1091142113506931. Maboshe, M. and I.D. Woolard. 2018. “Revisiting the Impact of Direct Taxes and Transfers on Poverty and Inequality in South Africa.� WIDER Working Paper 79/2018. UNU-Wider. Mejia-Mantilla, C., J. Fajardo-Gonzalez, M. Goldman, J.Jellema, and H. Renda. 2019. “Impact of Fiscal Policy on Poverty and Inequality in Uganda: Fiscal Incidence Analysis Using the UNHS 2016/17. The World Bank. Minh-Le, T. (2003). Value Added Taxation: Mechanism, Design, and Policy Issues. Paper prepared for the World Bank course on Practical Issues of Tax Policy in Developing Countries (Washington D.C., April 28-May 1, 2003) MoBSE-World Bank- UNESCO (2011). The Gambia: Education Country Status Report MOTIE (2018). The Gambia Trade Strategy and Industrial Development Policy. OECD, & Lustig, N. (2015). Inequality and Fiscal Redistribution in Emerging Economies.� In In It Together: Why Less Inequality Benefits All. OECD Publishing, Paris. https://doi.org/10.1787/9789264235120-10-en. Phillips, D., R. Warwick, M. Goldman, K. Goraus, G. Inchauste, T. Harris, and J. Jellema. 2018. “Redistribution Via VAT and Cash Transfers: An Assessment in Four Low and Middle Income Countries.� CEQ Working Paper Series 78. Tulane University. PwC (PricewaterhouseCoopers). 2013. A Quick Guide to Taxation in Gambia. PwC, Accra. https://www.pwc.com/gh/en/assets/pdf/a-quick-guide-to-taxation-in-gambia-september-2013.pdf. Rawlings, L. 2015. “Overview of Social Protection.� Pensions Core Course. World Bank. https://www.worldbank.org/content/dam/Worldbank/Event/social- protection/Rawlings_Overview%20of%20Social%20Protection%20(poll)_PCC15.pdf. Sine, J., P.P. Saint-Firmin, and T. Williamson. 2019. Assessment of the Health System in The Gambia: Overview, Medical Products, Health Financing, and Governance Components. Palladium, Health Policy Plus, Washington, DC. SSHFC (Social Security and Housing Finance Corporation). 2016. Annual Report and Financial Statements for the Year Ended 31st Decemeber 2015. Sulla, V., P. Zikhali, P.M. Schuler, and J. Jellema. 2017. Does Fiscal Policy Benefit the Poor and Reduce Inequality in Namibia? World Bank, Washington, DC Ueda, J. 2017. “The Evolution of Potential VAT Revenues and C-Efficiency in Advanced Economies.� IMF Working Paper WP/17/158. International Monetary Fund, Washington, DC. WFP (World Food Programme). 2015. “Gambia: Establishing the Foundation for a Nationally-Owned Sustainable School Feeding Program (2012-2016).� Standard Project Report. World Bank 2010. “The Gambia: Improving Civil Service Performance. Volume II: Public Service Pensions Policy Reform Note.� Report No. 51655-GM. World Bank, Washington, DC. http://hdl.handle.net/10986/2849. World Bank. 2011. The Gambia: Education Country Status Report. World Bank, Washington, DC. World Bank. 2017a. “The Gambia Pension Valuation Report For Federated Pension Scheme (FPS).� Unpublished report. 46 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia World Bank. 2017b. The Gambia: Education Sector Public Expenditure Review. World Bank, Washington, DC. http://hdl.handle.net/10986/28942. World Bank. 2018. The Gambia: Social Safety Nets Diagnostics. World Bank, Washington, DC. http://hdl.handle.net/10986/30008. World Bank. 2019a. “The Gambia: Health-Public Expenditure Review.� Draftinternal document. World Bank. 2019b. Incidence of Fiscal Policy on Poverty and Inequality in Guinea. World Bank, Washington, DC. World Bank. 2020a. The Gambia Public Expenditure Review: Creating Fiscal Space. World Bank, Washington, DC. World Bank. 2020b. “The Gambia.� Poverty and Equity Brief.World Bank, Washington D.C World Bank. 2020c. Republic of The Gambia: Overcoming a No-Growth Legacy. Systematic Country Diagnostic. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/33810. Younger, S.D., F. Myamba, and K. Mdadila. 2016. “Fiscal Incidence in Tanzania: Commitment to Equity.� CEQ Working Paper Series 36. Tulane Univeristy. Younger, S.D., E. Osei-Assibey, and F. Oppong. 2017. “Fiscal Incidence in Ghana.� Review of Development Economics 21(4), e47–e66. 47 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annexes Annex I. What did Gambia Import in 2015? Source: Harvard-CID Economic Complexity Atlas. https://atlas.cid.harvard.edu/. Note: Product disaggregation level at 1 digit and 2-digits under the Standard International Trade Classification (SITC), respectively. 48 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annex II. Evolution of Price Subsidies in The Gambia, 2015 Graph 1. The Gambia: Premium Motor Spirit (gasoline/petrol) Market price vs Fix price 2015 70 60 Price: GMD/L 50 40 30 20 10 0 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Dec-15 Mkt price Fix price Graph 2. The Gambia: Automotive Gas Oil (diesel) Market price vs Fix price 2015 60 50 Price: GMD/L 40 30 20 10 0 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Dec-15 Mkt price Fix price Graph 3. Kerosene Market price vs Fix price 2015 60 50 Price: GMD/L 40 30 20 10 0 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15 Dec-15 Mkt price Fix price Source: Authors’ elaboration based on data from MoFEA. 49 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annex III. Detailed Assumptions and Parameters, by Fiscal Intervention III.1. Direct Taxes PIT by income brackets, effective 1 January 2013 Income range (per annum) Tax rate GMD 0 to GMD18,000 0% GMD 18,001 to GMD 28,000 5% GMD 28,001 to GMD 38,000 10% GMD 38,001 to GMD 48,000 15% GMD 48,001 to GMD 58,000 20% Above GMD 58,000 30% PIT by income bracket, effective 1 January 2018 Income range (per annum) Tax rate GMD 0 to GMD 24,000 0% GMD 24,001 to GMD 34,000 5% GMD 34,001 to GMD 44,000 10% GMD 44,001 to GMD 54,000 15% GMD 54,001 to GMD 64,000 20% Above GMD 64,000 25% PIT– informality assumptions First occupation: Formality for employees (100% formality for public employees, and based on SST for the rest of employees) Employer Informal Formal Total Informality Govt./public sector 0 53,257 53,257 0% Private firm 35,643 17,896 53,539 67% Public works program 461 357 818 56% State-owned 0 1980 1,980 0% NGO/humanitarian 396 780 1,176 34% Private individual 34,409 2,231 36,640 94% Total 70,909 76,501 147,410 48% Source: Authors’ calculations based on IHS 2015/16. First occupation: Formality, all employed individuals Formal Frequency Percent Cumulative No 550,996 87.74 87.74 Yes 76,997 12.26 100 Total 627,993 100 Source: Authors’ calculations based on IHS 2015/16. 50 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia First occupation: Formality for employees (100% formality for public employees; based on SST for the rest of employees) s4eq7 - which sector Informal Formal Total Informality govt./public sector 0 370 370 0% private firm 601 8 609 99% public works program 28 0 28 100% ngo/humanitarian 68 0 68 100% private individual 916 15 931 98% Total 1613 393 2,006 80% Source: Authors’ calculations based on IHS 2015/16. Second occupation: Formality, all occupied individuals Formal Freq. Percent Cum. No 38,910 98.96 98.96 Yes 409 1.04 100 Total 39,319 100 Source: Authors’ calculations based on IHS 2015/16. 51 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia III.2. Indirect Taxes Custom Duties Rate Main household products Evaporated milk; powdered milk; Vitalait; baby milk powder; bleach; electric iron; 5% purchase of animal drawing vehicles; garden tools Fish (except smoked bonga, oyster and shrimps); Fresh and Sour Milk; Cassava; Chili powder (black pepper); Kerosene; Petrol and diesel for generator; Biogas 10% (sawdust/briquette, etc.); Parts for personal transport (car, motor and bicycles parts); Fuels, Lubricants for Personal Transport; Lantern-Kerosene Pasta and pasta products; Canned meat; Smoked Bonga; Oyster; Shrimps; Cream; Cheese; Oils and fats; Fruits and nuts; Plantain; Onion; Tomato puree (paste); Sugar; Salt; Garlic; Maggi Cube; Small dry pepper; Locust beans; Vinegar; Powder Pepper; Curry Powder; Chinese Green Tea; Soft drinks; Wines (red, white, etc.); Alcoholic Beverages, Tobacco, Narcotics (except beer, palm wine and stout); Insecticides (mosquito coils, repellants, sprays etc.); Candle; Matches; Lighters; Batteries; Light bulbs; Razor Blade; Tooth Paste; Face Powder; Skin Lightening; Tooth brush; Toilet Paper; Shampoo; Torch light; Incense; Clothing Materials; Garments; Footwear; Hair drier; Electric razor; Brooms; Scrubbing brushes; Dustpans and dust brushes; Dusters; Tea towels; Floor cloths; Household sponges; Plastic Pan; Basin; Glass Ware (cup, plates and the like); Thermos flask; Bedstead; Wooden Bed; Mat (sleeping/praying); Fitted carpets; Home decorations; Bed linen/blanket; Mosquito Net; Pillows and cushions; Table cloths; Curtains & Materials 20% for Curtains; Electric Blender; Electric Toaster; Electric Kettle; Coffee Mills/Makers; Food Mixers; Deep Fryers; Negro Pot; Bucket (Galvanized); Bucket (Plastic); Frying pan and pots; Thermos flask; Lantern-Batteries; Cutlery (forks, knives, etc.); Glass and Plastic ware; Motor cars; Motor cycles; Bicycles; Wheel Barrow; Car battery; Spectacles; Spectacle frames; Contact lens; Dentures; High blood pressure apparatus; Glucometer/ Sugar monitoring apparatus; Electric appliances for personal care; Envelopes; Writing pads, envelopes, account books; Pens, pencils, fountain pens; Toner and ink cartridges; Guidebooks; Musical scores; Paper punches, paper cutters; Music or video cassette; Recording media; Records, DVD Disc, Flash Drives; Cassettes, video diskettes and CD- ROMs; Unrecorded tapes; Unexposed films, cartridges; Pets and pet products; Cement; Paints (oil and water); Windowpanes; Wallpaper Pastes; Plastering walls; Ceramic tiles; Pipes; Taps; Joints; Jewelry; Clocks; Watches Cake (pan, etc.); Chicken (imported); Duck; Eggs; Beef; Sheep (mutton); Goat meat; Pork; Yoghurt; Potatoes (Irish); Sweet potatoes; Black mint; Chewing gum; Chocolate; Ice 35% cream; Mint stick; Coffee-Nescafe; Powdered Tea; Mineral water; Washing Soap; Starch; Washing Blue; Washing Powder; Washing Powder; Soda soap; Soda powder; Toilet Soap; Nail Polish; Knitting Wool; Thread; Buttons; Zips; Gloves; Helmets Source: Authors’ elaboration. Products based on HIS 2015/16. Tax rates based on MOTIE portal and GRA- Customs consultations. 52 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Excises Rate or value Main products 5% Telephone calls (International); Wheelbarrow; Soft drinks; Mineral water 15% Motor cars 7.5 Washing Soap and Toilet Soap (GMD per kg) 175 Beer (imported) (GMD per liter) 240 Wines (red, white, etc.) (GMD per liter) 280 Spirits (GMD per liter) 495 GMD per kg Piccadilly; Bond; Rizzla; Marlboro; Prince; Benson; St. Morithz; Manise (tobacco) (include wrapped in paper; Snuff environmental tax of 165 per kg) Source: Authors’ elaboration. Products based on HIS 2015/16. Tax rates based on MOTIE portal and GRA Customs consultations. Value Added Tax Standard VAT rate=15% VAT exemptions Rice; Maize; Millet; Sorghum; Findi; Flour; Bread; Chicken; Duck; Eggs; Beef; Sheep (mutton); Goat meat; Pork; Fish (except Canned and fried fish); Fresh and sour Milk; Groundnut oil; Palm oil; Vegetable oil; Palm Kernels oil; Fruits and nuts; Potatoes (Irish); Sweet potatoes; Cassava; Vegetables; Sugar; Honey; Garlic; Small dry pepper; Locust beans (Neteetu); Cold water in plastic bag; Tea; Jambakatang; Transportation fares; Medicines; Doctor's fees; Traditional Healer, faith healer and et; Hospital fees (hospitalization); Out-patient fees (ticket paid); Dental costs; Actual and Imputed Rent; Actual rent paid; Education (except books and uniforms); Health Consultation fees; Health Medication; Health Procedures; Health Insurance; Newspapers; Magazines; Atlases; Dictionaries; Encyclopedias; Garden Tools Source: Authors’ elaboration. Products based on HIS 2015/16. Tax rates based on MOTIE portal and GRA- Customs consultations. 53 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Indirect Taxes: Informality Assumptions for Tradable Goods • Custom duties: Formal if subject to custom duties and imported in 2015 (based on COMTRADE data). • Excises: Excisable goods were formal only if they paid custom duties (as simulated in the model). • VAT: VAT-able goods were formal only if they paid custom duties (as simulated in the model). In addition, for non-tradable services, informality assumptions were applied depending on type of service, based on local technical insights and the GRA Informal Tax Brochure (see below). Indirect taxes: List of services classified as informal mobile communication (scratch cards) Laundry (Clothes) mobile communication (nopal, e-credit) Dry cleaning Cassette/DVD rental Tailoring Charges-Men Puncture repair Tailoring Charges-Women Servicing of motor vehicle (labor) Tailoring Charges-Children (girls and b Car wash Repair of furniture, furnishings, and floor Other services Repair of small electrical appliances Taxi fares Housing Repair Motorcycle fares Repair to Footwear-Men House cleaner servants-full/part time Repair to Footwear-Women Gardener- full/part time Repair to Footwear-Children Security guard Repair of small electrical appliances Drivers Dowry Cook Marriage ceremony Other domestic servants Birthdays Barber (hair cut) Funeral Beauty Saloon (hair dressing) Other ceremonies Beauticians Other hair and beauty services Source: Authors’ elaboration, based on local consultations and GRA Informal Tax Brochure. 54 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia III.3. Indirect Subsidies, 2015 Name GMD/L Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Dec. Average Market n/a n/a 43.0 48.2 51.7 53.9 55.4 54.4 49.8 49.8 50.8 price Gasoline Fix price 58.8 58.8 58.8 58.8 58.8 58.8 58.8 58.8 58.8 58.8 58.8 Subsidy -15.8 -10.6 -7.1 -4.8 -3.4 -4.4 -9.0 -9.0 -6.4 unit Market 48.0 48.0 44.8 48.2 49.4 51.8 51.1 48.2 44.1 44.1 47.8 price Diesel Fix price 56.6 56.6 56.6 56.6 56.6 56.6 56.6 56.6 56.6 56.6 56.6 Subsidy -8.6 -8.6 -11.8 -8.4 -7.2 -4.8 -5.5 -8.4 -12.5 -12.5 -8.8 unit Market 46.9 46.9 43.0 45.1 46.5 38.1 37.8 36.3 33.0 33.0 40.6 price Kerosene Fix price 53.6 53.6 53.6 53.6 53.6 53.6 53.6 53.6 53.6 53.6 53.6 Subsidy -6.7 -6.7 -10.6 -8.5 -7.1 -15.5 -15.8 -17.3 -20.6 -20.6 -12.9 unit Source: Authors’ calculations based on MoFEA data. III.4. Social Protection, 2015 Public Service Pension Scheme (non-contributory) Total pensions paid 2015 (GMD) Total pensioners Median annual pension (GMD) 80,869,846 6,612 6200 Source: Authors’ calculations based on MoFEA data. 55 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia III.5. Education Benefits, 2015 Admin data: Calculation of transfer, per student GoTG expenditure Enrollment in public Annual transfer, per Education level (exec.), in GMD schools, number of students student (in GMD) Basic (ECD+ grades 1– 966,900,000 374,131 2,584.39 6) Secondary 167,596,000 141,401 1,185.25 (grades 7–12) Tertiary or post- secondary (post 154,704,000 10,639 14,541.22 G12) Source: Authors’ calculations, based on official data and World Bank Education staff. Household Survey: Students enrolled by education level, public schools Education level, Public schools Total Classification attended last year Early childhood 48,908 48,908 Basic p-1 (grade 1) 42,191 42,191 Basic p-2 (grade 2) 36,702 36,702 Basic p-3 (grade 3) 27,667 27,667 Basic p-4 (grade 4) 24,525 24,525 Basic p-5 (grade 5) 23,805 23,805 Basic p-6 (grade 6) 23,041 23,041 Basic ls-1 (grade 7) 21,097 21,097 Secondary ls-2 (grade 8) 17,315 17,315 Secondary ls-3 (grade 9) 12,244 12,244 Secondary us-1 (grade 10) 11,451 11,451 Secondary us-2 (grade 11) 9,478 9,478 Secondary us-3 (grade 12) 23 23 Secondary us-5 7 7 n/a Teacher training 1,948 1,948 Post-secondary Nursing/public health 169 169 Post-secondary Non-tertiary 138 138 Post-secondary Tertiary diploma 3,620 3,620 Post-secondary Bachelor’s 1,545 1,545 Post-secondary Master’s 189 189 Post-secondary Doctorate 22 22 Post-secondary Total 306,085 306,085 Source: Authors’ elaboration, based on IHS 2015/16. 56 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia III.6. Health Benefits, 2015 Admin data: Calculation of transfer, per case GoTG expenditure Annual transfer per Health service (estimated) Total cases case in GMD (in GMD) Outpatient Services 155,839,119.4 23,315 168.16 (Primary and Secondary) Inpatient Services (Tertiary) 166,657,476.50 926,720 7,148.08 Source: Authors’ calculations based on MoHSW and World Bank (2020). Household survey: Identification of public outpatient services Visited a public health provider in Last 2 weeks Annualized (x26) the last 2 weeks No 1,770,282 Yes 102,161 2,656,186 Total 1,872,443 Source: Authors’ elaboration based on IHS 2015/16. Note: (1) Outpatient beneficiaries: individuals who attended health consultations in the last 2 weeks due to sickness or other reason. (2) Public health facility: individuals who attended a public hospital, public clinic or public health center. (3) Annualization by 12*26 following CEQ Handbook (2018). Household survey: Proxying public inpatient services Assumption: Setting out-of-pocket annual hospitalization expenditure threshold at GMD 10,000 (below, public) Type of hospitalizations by local Share of annual public hospitalizations government area Public inpatient Frequency Percent Cumulative Area Private Public Total % public No 1,486 8.1 8.1 Banjul 38 303 341 88.9% Yes 16,855 91.9 100 Kanifing 470 2,007 2,477 81.0% Total 18,341 100 Brikama 387 5,092 5,479 92.9% Mansakonko 112 1,762 1,874 94.0% Kerewan 271 2,393 2,664 89.8% Kuntaur 58 1,144 1,202 95.2% Janjangbureh 46 1,449 1,495 96.9% Basse 104 2,705 2,809 96.3% Total 1,486 16,855 18,341 91.9% Source: Simulations based on IHS 2015/16. Note: The threshold for out-of-pocket expenditure was calculated as follows (i) the distribution of out of pocket expenditure on hospitalization (3 scenarios87 above the median, which was GMD 5,000) were tested; (ii) for all scenarios we assessed if the share of public hospitalizations was higher for poorer households and in local government areas with less coverage of private health facilities; (iii) the scenario where GMD 10,000 was the threshold was the most stable. The three scenarios for the public hospitalization cost threshold were: (i) GMD 5,000; (ii) GMD 8,000; and (iii) 87 GMD 10,000. 57 Carrasco et al. (2022): The Effects of Fiscal Policy on Inequality and Poverty in The Gambia III.7. Non-fiscal: The School Feeding Program (WFP), 2015 Admin: Calculation of transfer, per beneficiary Food budget, Food budget Annual transfer per executed 2015 executed 2015 Total beneficiaries beneficiary (in USD) (in GMD) (in GMD) 6,080,000 258,437,745 99,603 2,595 Source: Authors’ calculations based on WFP (2015). Note: (1) Beneficiaries include children under 5 (11,104) and children between 5-18 years old (88,499). (2) Calculation in GMD based on 2015 exchange rate (GMD/USD) = 42.51, based on World Bank World Development Indicators (https://databank.worldbank.org/source/world-development-indicators). Admin: School Feeding Program: List of beneficiary districts, based on MoBSE Region MoBSE District District HH survey Code Region Banjul city council 10 Banjul City 1 BCC Banjul city council 11 Banjul City Banjul city council 12 Banjul City 1 KMC Sere Kunda West 24 Banjul City 3 Sanjally District Sabach Sanjar 56 North Bank Region 3 Upper Badibu Illiasa 55 North Bank Region 3 Upper Badibu Sabach Sanjar 56 North Bank Region 3 Central Badibu Central Badibu 54 North Bank Region 3 Lower Badibu Lower Badibu 53 North Bank Region 3 Lower Niumi Lower Niumi 50 North Bank Region 3 Upper Niumi Upper Niumi 51 North Bank Region 5 Niamina East Niamina East 72 Central River Region 5 Nianija Nianija 62 Central River Region 5 Upper Saloum Upper Saloum 61 Central River Region 5 NiaminaDankunku Niamina Dankunku 70 Central River Region 5 Niamina West Niamina West 71 Central River Region 5 Janjangbureh Janjanbureh 75 Central River Region 5 Lower Fulladou West Lower Fuladu West 73 Central River Region 5 Same Sami 64 Central River Region 5 Upper Fulladou West Upper Fuladu West 74 Central River Region 5 Niani Niani 63 Central River Region 5 Lower Saloum Lower Saloum 60 Central River Region 6 Jimmara Jimara 80 Upper River Region 6 Tumana Tumana 82 Upper River Region 6 Kantora Kantora 83 Upper River Region 6 Sandu Sandu 86 Upper River Region 6 Basse Basse 81 Upper River Region 6 Wuli East Wuli East 85 Upper River Region 6 Fulladu Jimara 80 Upper River Region Source: Authors’ elaboration based on MoBSE district list and IHS 2015/16. 58 Annex IV. Social Protection Financing The World Bank (2018) estimated that total expenditure on the main social protection programs in The Gambia could be around 0.93 percent of GDP, when considering government and donor funding together, but only 0.09 percent of GDP when considering GoTG contributions alone. Line item Source of Donor – off Most recent Most recent Remarks finance budget financing financing estimate estimate (GMD) (US$) School meals MoBSE 30,000,000 635,324 Govt estimate for 2018 (first such contribution) School meals WFP (donor – 89,740,571 1,900,000 off-budget) Social production Ministry of 3,580,000 75,815 Department of Social services Health and Welfare, excluding Social Welfare health functions, of which US$11,000 is welfare of Gambians/refugees Youth EU (donor – 175,499,520 3,000,000 Youth empowerment Empowerment off-budget) Project Songhai Initiative Ministry of 6,000,000 127,065 Horticulture and (Support to Youth Youth and Agriculture inputs Empowerment) Sports Songhai Initiative Ministry of 6,000,000 127,065 Horticulture and Youth and Agriculture inputs Sports FAO livelihoods FAO (donor 89,549,862 1,896,195 Communal gardens off-budget) and inputs Action Aid Action Aid 8,216,957 174,000 Communal gardens livelihoods/food (donor off- and inputs security budget) Antenatal CCT NaNA (donor – IDA 8,173,800 173,057 Cash transfer for (Maternal and Child off-budget) pregnant women Nutrition and Health Results Project; MCNHRP) Food and Nutrition NaNA (donor – IDA 10,800,000 228,659 Cash transfer and asset Security off-budget) for poor (MCNHRP) Targeted Food and WFP (donor – off- 2,241,247 47,452 Supplemental feeding Nutrition budget) Assistance, Protracted Relief Recovery Building Resilience EU (donor 55,733,618 1,180,000 Cash transfer for through Social off-budget) women with children Transfers (BReST) 0-2 GMD US$ 485,535,575 9,564,632 Source: World Bank (2018). 59 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annex V. Estimations of Taxpayers and Beneficiaries in The Gambia, 2015 Direct taxes and direct transfers Decile (by market Public Service Pension PIT taxpayers income plus pension) beneficiaries 1 1575 62 2 1838 11 3 2034 22 4 2290 76 5 4317 46 6 4072 62 7 5330 0 8 7748 372 9 6056 455 10 11155 233 SUM 46,415 1,339 Source: Authors’ estimations based on IHS 2015/16 and GMB administrative data. In-kind benefits from (public) health and education Decile (by Post- Inpatient Outpatient ECD Primary Secondary disposable secondary visits visits students students students income) students 1 877 87,880 1,656 5,289 1,348 251 2 1,584 152,256 3,370 10,302 2,579 314 3 1,556 164,554 4,403 14,518 4,532 303 4 1,503 194,610 4,780 17,557 6,245 102 5 1,303 182,494 4,472 17,963 5,189 197 6 1,677 228,488 6,331 20,902 7,916 496 7 1,637 222,274 6,098 21,566 8,271 382 8 1,977 257,088 6,059 26,591 9,021 1,893 9 2,517 316,108 6,179 22,404 13,484 1,347 10 2,528 211,952 5,560 20,839 13,023 2,341 SUM 17,159 2,017,704 48,908 177,931 71,608 7,626 Source: Authors’ estimations based on IHS 2015/16 and GMB administrative data. 60 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annex VI. The Gambia: Fiscal Policy Changes Between 2015 and 2020 VI.1. Summary of the Tax Policy Changes Between 2015 and 2019 Tax type Tax item changed Description of change Year Threshold (tax free) from GMD 18,000 to 24,000 2018 PIT Maximum tax rate from 30% to 25% 2018 Income tax on turnover from 2% to 1.5% 2017 (audited) from 1.5% from 1.0% 2018 Income tax on turnover from 3% to 2.5% 2017 (unaudited) from 2.5% to 2% 2018 Rental Residential income tax from 10% to 8% 2018 income tax From 15% to 10% 2018 Commercial rent tax from 10% to 20% 2019 Corporate from 31% to 30% 2016 Corporate income tax income tax from 30% to 27% 2018 Custom Used cars (>5yrs) from 15% to 10% 2018 duties from 25% to 20% 2018 New cars from 20% to 25% 2019 from 5% to 10% 2017* Rice from 10% to 0% 2019 from GMD 110/kg to 120/kg 2016 Excises Other tobacco products from GMD 120/kg to 165/kg 2017 Wine from GMD 150/L to 250/L 2019 Beer/Stout from GMD 100/L to 175/L 2019 Spirits from GMD 175/L to 280/L 2019 Excises Wine from 15% to 60% 2019 (domestic) Beer/Stout from 10% to 75% 2019 Spirits from 15% to 60% 2019 Source: Author’s elaboration based on GRA. * Motivated by implementation of the ECOWAS Common External Tariff. The table summarizes major tax policy changes in the Gambia between 2015 and 2019. As the table shows, there have been policy changes in some of the main taxes: - The burden of personal income tax (PIT) was reduced, with the threshold increased for all five bands and the top rate of tax lowered. - Corporate income tax (CIT) was reduced. - Rental income tax has been reduced for residential income tax. Commercial rent tax was initially reduced in 2018 but then increased to 20 percent. - Custom duties on used cars were initially reduced but then restored to their original rate. - Custom duties on rice were increased in line with the ECOWAS Common External Tariff but then removed in 2019, making rice currently duty free. - There has been a staggering increase in domestic excises on alcoholic beverages, from 10–15 percent to 60–75 percent; the minimum taxes per unit (L) have also increased. Excises for tobacco products have also increased but less steeply. 61 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia VI.2. Social Expenditure Changes Between 2015 and 2020 In this section we look at some policy changes in social spending between 2015 and 2020. Due to data limitations, the table presented shows data on just government budget on health and education in the period under review. Changes in the social protection sector are described qualitatively. Changes in Health and Education Budget, 2015–20 Planned budget GMD million Sector/year 2015 2016 2017 2018 2019 2020 Health 684.71 812.12 1,032.00 1,489.94 2,143.20 2,095.44 % change 19% 27% 44% 44% -2% Basic and secondary 1,354.56 1,369.07 1.600,55 2,089.05 3,063.39 3,602.14 education % change 1% 17% 31% 47% 18% Higher education 123.64 - - 873.46 1,373.59 1,097.76 % change - - - 57% -20% Source: Author’s elaboration. Data from planned budget speech available in MoFEA. Health: Except for in 2020, the health budget has been increasing since 2015 with an average growth of about 26 percent per year.88 To improve health outcomes, the National Development Plan (NDP) emphasizes the need to spend on expanding access to healthcare as well as improving access to immunization by children. Despite the expansion of budget allocations, most of the health expenditure, about 97 percent, still goes to recurrent expenditure, and the development spending goes mainly on infrastructure development (World Bank, 2020a). The introduction of a national health insurance scheme in Gambia is also at an advanced stage—just pending National Assembly approval of the bill. It is hoped that these changes will catalyze access to healthcare among the poor, thus creating more scope for fiscal policy on health to improve household welfare. Education: There have been significant changes in budget allocation to education since 2016. In the period under review, the budget allocated to basic education grew by 23 percent on average per year, with the largest increases in 2018 and 2019. Due to data limitations, we could only calculate the percentage change in budget allocation to higher education for 2019 and 2020, when the average growth was 19 percent. One significant driver of public spending on education has been spending on personnel (World Bank, 2020a). After 2015, there has also been more commitment from the government to increase access to public ECD and to improve the quality of education, particularly, in public schools (see Education Policy 2016–30). These initiatives will increase access to education among vulnerable households, and therefore can improve the impact of the fiscal system on household welfare. Social protection: While some social protection programs that promote health are still donor-funded,89 recent developments in the area of social protection have seen more participation from the GoTG, which could 88Due to data limitations, we used budget allocations rather than actual spending. 89The two main programs that are under implementation are the Maternal and Child Nutrition and Health Results Project (MCNHRP) funded via a World Bank grant and Building Resilience through Social Transfers (BReST) program funded 62 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia increase both the size and the distributional impacts of social protection expenditure in future fiscal incidence analysis: - The School Feeding Program (SFP), which provides meals to children enrolled in public schools, used to be funded and implemented solely by the World Food Program (as described above). However, starting in 2018, the GoTG has funded the expansion of the SFP into the two remaining regions that were not covered by WFP (Lower River Region and West Coast Region) meaning the SFP now covers all six regions of the country. This decision signals the willingness of the GoTG to take over the funding of this program and integrate it into the national social protection system. - In 2020, during the COVID-19 crises, the GoTG also implemented social protection relief measures in the form of in-kind benefits. Over GMD 734 million (US$14.7 million) has been allocated to support 84% of households countrywide; the in-kind benefits include rice, oil, and sugar (Gentilini 2020). - Under a World Bank Social Protection and Jobs project, The Gambia is about to pilot a flagship government social protection program. This will include an unconditional cash transfer (UCT) targeting vulnerable households. In the pilot phase, the program will be funded jointly by the government and the World Bank. The UCT will be targeted using a proxy means test indicator and the plan is to transfer GMD 3,000 bi-monthly to 6,000 households in 20 districts. After the pilot, the end goal is to make this program part of the GoTG social protection system, thus strengthening the country’s structural policies for improving welfare among the poor. via a European Union (EU) grant. These programs are targeting women and children in the North Bank Region, Central River Region, Upper River Region, and Lower River Region. 63 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annex VII. Data Limitations and Future Model Extensions Data Limitations - Direct taxes (administrative data): The lack of data on PIT taxpayers or collections by income bracket makes it harder to cross-validate the PIT model.90 Our model underestimates total PIT, something expected in standard incidence analysis with household surveys, where income tends to be under-reported and top-income taxpayers are not well-represented. To identify the potential sources of the PIT gap, ideally tax admin data would include PIT collections by income bracket and the number of taxpayers registered with the GRA. - Indirect taxes (micro/macro): The HIS 2015/16 was missing some ideal variables at the household- expenditure level which could help to better identify taxable products: product origin to identify potential imported goods and better narrow the custom duty-base; and place of purchase to proxy informality.91 Another data limitation was the lack of a recent input-output matrix, which would help estimate the indirect effects of indirect taxes, VAT exemptions and informality.92 For the excises model, we need a more complete list of price data for non-food products.93 - Social protection (micro): One limitation of the IHS 2015/16 is that the information on income from pensions was available in the Household-Income module, and not at the individual level. In future waves of the IHS, it would be ideal to include the income from pensions in the Individual-Level income questionnaire and disaggregated by type of pension scheme (PSPS, NPF, FPS). - Indirect subsidies (macro/admin): One limitation of the analysis of indirect fuel subsidies in the current model is that we were not able to estimate indirect effects, due to the lack of an input-output matrix. Another important limitation of the analysis for 2015 is that we were not able to include agricultural subsidies due to lack of administrative information about the distribution mechanisms of these subsidies. - Health (micro/admin): One data limitation on the admin side was that the MoH budget was received itemized and not aggregated by functions or health services. Other limitations relate to the IHS 2015/16 individual health module. For health consultations (outpatient services), we had information on consultations for by individuals who were sick in the last two weeks, according to type of health facility. For future waves of IHS 2015/16, it would be ideal to also ask individuals about their access to other preventive health consultations. For hospitalizations (inpatient services), we recommend that future waves of IHS 2015/16 ask individuals about “type of health facility attended� and “duration of hospitalization�; this information would allow to identify better the inpatient public beneficiaries and the size of the in-kind benefit. - Education (admin): We were not able to collect the official admin data (budget and students enrolled) for the post-secondary education level from MoHERST. For the current analysis, we complemented data collection on education with estimates from World Bank (2017b). 90 The GRA is currently developing its PIT database. 91 These variables of “product origin� and “place of purchase� by item are available in other household surveys in Africa (Burundi ECMVB 2013/14). 92 The last input-output matrix for The Gambia dates to the 1990s. 93 Goods that could not be included in the analysis, due to lack of price data to calculate quantities, were: manise (tobacco) wrapped in paper; snuff; washing soap; and toilet soap. In addition, the service of telephone calls (international). 64 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Future Model Extensions - Improving modelling of consumption informality: Total indirect taxes represented 8.3 percent of GDP in 2015. Even using the current fiscal incidence model with the existing IHS 2015/16, we could improve the modelling of indirect tax informality by having disaggregated indirect tax collection data (custom duties, VAT, excises) by type of economic activity or geographical region. Another methodological extension would entail predicting household consumption informality based on cross- country evidence available in Bachas et al. (2020). Given that recent research shows that informality could improve the progressivity of indirect taxes (Bachas et al. 2020; Inchauste et al. forthcoming), this future extension of the analysis would be truly relevant for The Gambia, given its high informality and high poverty incidence. - Improving coverage of excises: In 2015, excises collections represented 1.1 percent of GDP, but our model only estimates 0.3 percent of GDP. One potential reason for this underestimation could be if individuals are under-reporting their consumption of “sin-goods� (tobacco and alcohol). One potential methodological extension to our current model could be to combine a probabilistic model that predicts which households are likely to consume tobacco/alcohol and complement that with imputation of consumption using information from similar households in the survey. Macro data on consumption of these goods would be needed to validate this model. - Estimating an alternative model for health benefits: One alternative model to the average cost model used to allocate health transfers would be a health-insurance model. Based on this model, the allocation of health benefits is done based on eligibility94 rather than on self-reported usage of health services provided in the household survey.95 - Including households’ out-of-pocket expenditure on education: While public education is free in The Gambia, households do incur significant out-of-pocket expenditure for indirect costs (e.g. transports, books, and supplies). Subtracting households’ indirect costs from in-kind education benefits could provide a richer perspective on which households are net payers or net receivers for the public education system. 94 For countries like The Gambia, where very few households have health insurance to proxy eligibility to public health, another option could be assuming that the whole population is eligible, except those who reported attending private health facilities in the household survey. 95 One limitation of modelling health transfers based on the “average cost� imputed to individuals who reported accessing health services is that it can make households with sick members that accessed these services “better off� while there were other people who were also eligible to receive public healthcare but didn’t get sick in the period of analysis. 65 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Annex VIII. Methodological Appendix Visualization of Fiscal Impoverishment Indicator Source: Higgins and Lustig (2016), available in CEQ Handbook (Lustig 2018b). Progressivity, by the Kakwani Index Tax Progressivity Progressive Regressive Neutral 𝒑𝒂𝒚𝒎𝒆�𝒕 𝒑𝒂𝒚𝒎𝒆�𝒕 higher same higher for 𝒊�𝒄�𝒎𝒆 Taxes 𝒊�𝒄�𝒎𝒆 for poorer distribution as richer people the income people Kakwani Index (KI)= KI >0 KI <0 KI=0 CC-Market Gini 66 The Effects of Fiscal Policy on Inequality and Poverty in The Gambia Spending/ Transfer Progressivity Progressive Regressive Neutral Absolute Relative Benefit 𝒃𝒆�𝒆𝒇𝒊𝒕 𝒃𝒆�𝒆𝒇𝒊𝒕 received in 𝒊�𝒄�𝒎𝒆 𝒊�𝒄�𝒎𝒆 Spending higher for higher for same distribution as the value higher /Transfer poorer richer income for poorer people people people Kakwani KI>0, Index (KI)= KI>income KI>0 KI<0 KI=0 Market Gini Gini-CC Marginal Contribution, Example The marginal contribution of a tax is 𝑀𝐶𝑡 = 𝐺𝑥+𝐵 − 𝐺𝑥+𝐵−𝑡 Where 𝐺𝑥+𝐵−𝑡 and 𝐺𝑥+𝐵 are the Gini coefficient of incomes after the tax and transfers and after transfer only, respectively. If 𝑀𝐶𝑡 >0, the tax is equalizing. 67