EGYPT, THE IMF AND THREE SUBSIDY APPROACHES: THE UNIVERSAL, THE SELF-TARGETED, AND THE TARGETED

EGYPT, THE IMF AND THREE SUBSIDY APPROACHES: THE UNIVERSAL, THE SELF-TARGETED, AND THE TARGETED

Foreword Many countries in the Middle East and North Africa region are facing structural economic problems that have not improved, let alone stabilised, despite several rounds of IMF programmes over the last decade. Consequently, the current polycrisis has hit the region utterly unprepared with state finances, public services and social protection already threadbare. The COVID-19 pandemic and its economic fallout, the Russian war against Ukraine since February 2022 with its effects on global food and energy prices, and the global inflation that led to higher borrowing costs have all negatively impacted the region’s economies. First and foremost, the poorest households have carried the heaviest burdens, with rising prices, inflation and little economic growth. Egypt, a net importer of food and energy with a high debt exposure even before 2020, has faced especially massive challenges. Given this situation, Egypt concluded a new deal with the IMF in December 2022, the country`s fourth IMF programme in less than a decade. As always, these programmes came with a set of attached conditionalities aimed at achieving macro-economic stability. The newest programme was no exception. However, as this study shows, conditionalities on social spending were usually weak and spending for social protection programmes could not offset the negative impacts of fuel subsidy cuts for the poorest households. It is the hope of FES to contribute to the growing number of publications which critically assess the IMF and other financial institutions and their activities in the region. Together with our regional partners, we aim to shape future policies to achieve greater social justice and more equal distribution of the burdens of economic crises. I wish to thank Osama Diab very much for this insightful study and his excellent collaboration with FES on this and many other projects. Similar studies like this one have been published by FES MENA focusing on Morocco, Tunisia and Jordan in an edited volume available online(https://library.fes.de/pdf-files/bueros/tunesien/19559. pdf, also available in French and Arabic). I encourage readers to explore these studies as well to gain a better understanding of the regional dimension of the IMF-induced reforms and their repercussions on social protection and social spending. Finally, I want to thank my colleagues Imene Cherif and Salma Hussein for their contributions and tireless efforts to move this publication forward. I also would like to thank Sadri Khiari for his beautiful cover illustration and Karl Binger for proof-reading the English version. Thomas Claes Project Director Tunis, June 2023 4 Egypt, the IMF and Three Subsidy Approaches

Table of contents Introduction 7 1. Recent debates on social protection in Egypt 8 2. Review of social spending in Egypt since 2011 15 3. Review of IMF social spending recommendations and conditions 18 4. Impact of the COVID-19 lockdown and the Ukraine war on social protection 26 5. Methodology for calculating embodied and direct energy expenditure 32 6. Energy subsidies: busting the myths with evidence from Egypt 36 7. Final remarks and policy implications 41 The Universal, the Self-Targeted, and the Targeted 5

Summary of findings • Despite claims that energy subsidies disproportionately benefit the rich, its elimination in Egypt’s case disproportionately affected low-income households. • Universality of subsidies does not contradict with its progressivity. The claim that energy subsidies disproportionately benefit the rich is only true when considered in absolute terms not relative to income, and when it excludes embodied energy. Once we consider expenditure relative to income and expenditure on embodied energy, energy price subsidies can start being progressive. • Targeted subsidies have far from compensated low-income households for their loss of purchasing power; only a small fraction of the fiscal savings resulting from the elimination of universal price subsidies was used in cash transfer schemes, contradicting claims by the IMF. • Fiscal space created by the elimination of energy subsidies was mostly used to honour Egypt’s growing interest payments and to achieve a primary budget surplus. • Given their high propensity to consume, any increase in disposable income for low-income households will translate to increases in household well-being. Energy subsidies can play a significant role in achieving this without high inflation. • Due to high degrees of error in cash transfer programmes, the ability of targeted programmes to reach the poor is often rather limited, challenging its proponents’ claims of effective targeting. • Tight monetary policy(i.e., high interest rates), which is typically demanded by the IMF, causes more severe inflation for low-income households. Evidence from Egypt shows that high interest tends to control the price of non-essential products consumed by households with disposable income, while failing to control the inflation of basic commodities. It could in fact increase demand for it since households typically handle high inflation by switching to cheaper products. • Inflation in the past years was higher for low-income households. Therefore, all forms of social protection(contributory social insurance and non-contributory social assistance) need not only be indexed to inflation, but also indexed to the inflation rate of each income profile – especially the poor – based on its average consumption patterns. • The IMF’s social protection conditions are in general less binding than fiscal and monetary tightening conditions. 6 Egypt, the IMF and Three Subsidy Approaches

There is no sense in implementing policies that impoverish people, and then try to include them in social protection programmes. – Gouda Abdel-Khaliq, Egyptian economist and former minister of social solidarity. Introduction Egypt has a long history of government subsidies whose modern form dates back to at least World War I. For decades, the question of targeting has been central to social policy making and wider societal debate. Questions about the fairness of universal price subsidies, which the welloff strata of society benefit more from, were almost uninterruptedly posed. On the other hand, the high margin of error – as well as other bureaucratic constraints of targeted subsidies – has also posed questions about its feasibility in the Egyptian context of high poverty, poor data on income, and an under-resourced state apparatus. Due to such challenges, the favourite subsidy regime of the Egyptian state in the last decades was self-targeted subsidies that excludes the well-off by subsidising inferior energy and food products (e.g., bread, butane gas cylinders, leaded gasoline, etc.). This regime has become part and parcel of the government’s moral economy, and is perceived by successive governments as necessary for maintaining a delicate social contract that was severely disrupted in the 1977 Bread Riots. The riots left a great mark to this present day regarding all issues linked to price subsidies. Despite this, the International Monetary Fund(IMF), with its increasing influence on policy making in Egypt, has successfully advocated for replacing both universal and self-targeted subsidies(especially energy) with targeted subsidies. This paper explores which system is more equitable and progressive through a mix of qualitative and quantitative analysis. The paper examines which of the social assistance regimes provide the highest benefit for the most economically vulnerable households. This will be based on analysing the impact of cutting energy subsidies on different income profiles. In 2015, universal energy subsidies for consumers and households were still the dominant form of social assistance; in 2021 it had been almost-completely scrapped and"replaced" by cash subsidies. The paper will demonstrate how this transformation has differently impacted households based on their embodied and direct energy expenditure patterns, general inflation experience, and the amount of direct cash assistance they received. In this light, the paper is set out to test the IMF’s"replacement" claim, meaning whether energy subsidies were replaced by more direct and targeted social assistance by examining household income and expenditure – as well as national social spending. The Universal, the Self-Targeted, and the Targeted 7

There are a few reasons why the paper focuses on energy subsidy: first, consumer and household energy subsidies have seen the most radical transformation in recent years, from constituting most of the subsidy bill to being almost fully eliminated; 1 second, it was responsible for a large part of the inflationary waves significantly diminishing household purchasing power in recent years. Third, energy remained the closest form of subsidy to universalism on the targeteduniversal spectrum. Finally, cutting energy subsidies created the most fiscal savings that was supposed to be used for strengthening IMF-mandated social safety nets, that is, direct cash-based social assistance schemes. Before the paper gets to this point, it will present an overview on debates on social protection, as well as the history of social spending in Egypt with a larger focus on the post-2011 period. The paper will then sketch out a history of IMF recommendations and conditions related to social spending, and its enduring influence on the matter from the 1970s up to the present time. Subsequently, the paper will discuss the impact of the COVID-19 pandemic and the Ukraine war on social protection in Egypt. 1. Recent debates on social protection in Egypt Despite increasing demands for social justice since the 2011 revolution,"social justice" has not been well and concretely defined. However, one can safely argue that it targets a higher bar than social spending and protection as understood and defined by the IMF. According to the IMF narrative, social protection is mainly a mitigation technique that is composed of two components, namely social insurance and social assistance. 2 Mitigation, by definition, stops even short of the near-consensus universal objective of poverty reduction(UN SDGs 1& 2). In the 2016 IMF Egypt programme reports, there was no mention at all of poverty reduction. 3 One of the two components of social protection as defined by the IMF, social insurance is contributory and designed to protect households from shocks that can adversely impact their income and welfare. The other component is social assistance which is non-contributory, aimed at protecting households against poverty and is funded by general government revenues. According 1 There is still a limited subsidy of diesel and butane gas cylinders amounting to about EGP 18 billion in 2021/22, or about 0.25 per cent of GDP. 2 IMF.(14 June 2019). A strategy for IMF engagement on social spending. Retrieved 13 December 2022, from https://www.imf.org/en/Publications/Policy-Papers/Issues/2019/06/10/A-Strategy-for-IMF-Engagement-on-SocialSpending-46975 3 Diab, O.,& Hindy, S. I.(2022). Women and economic reform in Egypt: Impact of production changes on female waged labor force participation. Middle East Critique, 31(1), 61–79. https://doi.org/10.1080/19436149.2022.2030984 8 Egypt, the IMF and Three Subsidy Approaches

to the IMF’s 2019 Strategy for Engagement on Social Spending(hereafter referred to as"Social Spending Strategy"),"social safety net" is synonymous with"social assistance". 4 In the Strategy, social spending offers a broader framework than social protection as it includes spending on health and education in addition to spending on social protection. Figure 1. Summary of the most important social protection schemes and their financing methods SOCIAL PROTECTION Non-contributory(public spending) contributory Social assistance social care social insurance Labour Market policies and interventions Social transfers > Cash transfers > Vouchers > In-Kind transfers (including school feeding) Public works programmes Fee waivers > Cash for work > Food for work > Vouchers for work > For basic health/ education Subsidies > Fuel > Food > Family support services > Home based care > Health insurance > insurance for: - unemployment - Maternity/ Paternity - Disability - Work accidents > Old age pension > Crop/livestock insurance Active Passive > Work sharing > Training > Job-search services > Maternity benefits, ijury compensation, and sickness benefits for those in work > Changes in legislation(e.g minimum wage, safe working conditions) Notes:(1)'Non-contributory' schemes are defined by the International Labour Organisation(IL0) as those that,'normally require no direct (financial) contribution from beneficiaries or their employers as a condition of entitlement to receive benefits'(ILO, 2017). Public works programmes are usually counted as'non-contributory even though the recipient contributes labour.(2) Social transfers may be conditional or unconditional. A conditional transfer requires the recipient to meet certain behaviours[such as ensuring school attendance) to receive the benefit. Source: Adapted from O'Brien et al.(6:2018). Source: Hussein, S.(2022) Umbrellas for All,(Friedrich Ebert Stiftung). Retrieved 12 March 2023, from https://library.fes.de/pdf-files/bueros/tunesien/19560.pdf. 4 IMF.(2019). A strategy for IMF engagement on social spending. The Universal, the Self-Targeted, and the Targeted 9

Although social justice is much less concretely defined as mentioned above, it can still be considered to comprise an even broader view than social spending, including issues such as social mobility, reducing relative poverty/inequality, and adopting a more multidimensional and intersectional definition of poverty and marginalisation. Contemporary social justice demands entail a downward redistribution of wealth to counter the increasing inequality which has characterised the last few decades where most of the growth was captured by a small section of the population. 5 It could also entail equality of gender, race, religion, age, etc., as well as climate and environmental justice. 6 Nevertheless, social justice is still present in the IMF and the World Bank literature. For example, former IMF Deputy Director for Fiscal Affairs Peter S. Heller wrote in an official article about the IMF perspective on social justice that"issues of social justice and environmental sustainability are taken into account by the Fund in working with its member countries." 7 Heller agrees that"social justice" is a normative concept whose definition will differ from one culture to another. He thinks, however, that there could be some core minimum elements that almost everyone could agree on, such as inalienable human rights and the Millennium Development Goals(MDGs). Heller also defines the IMF and the World Bank’s social programmes and social impact analysis as part of their approach to take social justice seriously. Additionally, a World Bank team of researchers working on social justice in healthcare in Egypt defines it as"providing equal access to liberties, rights, and opportunities especially for the least advantaged members of society". 8 In this way we can imagine a three-tier approach for social policy: the broader tier is that of social justice, followed by social spending, before narrowing down to social protection, comprising its two elements of social insurance and social assistance. When it comes to social assistance schemes, they are typically divided into universal price subsidies and targeted social assistance. Universalism usually refers to subsidising a product to provide it at a lower price than which would have otherwise been determined by market forces. Here, the subsidy is universal because in theory anyone can buy the product and benefit from the subsidy. However, the degree of benefit from universal subsidies is rarely evenly distributed among income groups for several reasons. 5 United Nations.(21 January 2020). Rising inequality affecting more than two-thirds of the globe, but it’s not inevitable: New UN report| UN news. United Nations. Retrieved 21 December 2022, from https://news.un.org/en/ story/2020/01/1055681 6 Heller, P. S.(13 February 2003). Wealth creation and social justice: An IMF perspective. IMF. Retrieved 22 December 2022, from https://www.imf.org/en/News/Articles/2015/09/28/04/53/sp021303a 7 Heller.(2003). Wealth creation and social justice: An IMF perspective. 8 Pande, A., El Shalakani, A.,& Hamed, A.(2017). How can we measure progress on social justice in Health Care? the case of Egypt. Health Systems& Reform, 3(1), 14–25. https://doi.org/10.1080/23288604.2016.1272981 10 Egypt, the IMF and Three Subsidy Approaches

First, no product is consumed equally by different income groups. For example, subsidising gasoline only used in private cars will be of little benefit to the lowest income group who do not own a private vehicle. Conversely, subsidising basic staple food would likely benefit the poor more. Second, consumption patterns of different income groups fundamentally differ. One of the most consistent and stable differences is that the share of income spent on food increases as income declines. 9 Third, since people’s incomes vary widely, even if different income groups spend equal amounts on subsidised products, this expenditure will greatly differ relative to their income. In many cases, higher income brackets might consume more of a subsidised good, but the subsidy will still constitute a larger share of low-income group incomes. Universal subsidy schemes have come under consistent attack from both International Financial Institutions(IFIs) and successive Egyptian governments evoking its perceived inequity. This evocation is based on the argument that the better-off sections of society can still benefit from it. According to this narrative, this is wasteful of public and taxpayer money. 10 In Egypt, the government has for decades been trying to tackle this perceived problem by making subsidies, whether for commodities or services, more self-targeted, as I will discuss below. IFIs have been producing research and numerous papers arguing for the inequity of price subsidies especially in comparison with targeted cash transfers. For example, the IMF writes on fossil fuel subsidies: Subsidies are intended to protect consumers by keeping prices low, but they come at a high cost. Subsidies have sizable fiscal costs(leading to higher taxes/borrowing or lower spending), promote inefficient allocation of an economy’s resources(hindering growth), encourage pollution(contributing to climate change and premature deaths from local air pollution), and are not well targeted at the poor(mostly benefiting higher income households). Removing subsidies and using the revenue gain for better targeted social spending, reductions in inefficient taxes, and productive investments can promote sustainable and equitable outcomes, 11 (my italics). A World Bank paper similarly asserts that price"subsidies consume valuable government resources and crowd out much needed investments in human capital and infrastructure", and that"despite their high cost/low effectiveness, price subsidies can be extremely persistent and hard to remove for a number of important reasons." 12 9 Atkinson, A. B.(2019). Measuring poverty around the world. Princeton and Oxford: Princeton University Press. 10 IMF.(2012). IMF survey: Costly Mideast subsidies need better targeting. IMF. Retrieved 29 December 2022, from https://www.imf.org/en/News/Articles/2015/09/28/04/53/socar051412b 11 IMF.(2022). Fossil fuel subsidies. IMF. Retrieved December 21, 2022, from https://www.imf.org/en/Topics/climatechange/energy-subsidies 12 Bank, W.(2013). From universal price subsidies to modern social assistance: The Political Economy of Reform. DSpace. Retrieved 21 December 2022, from https://openknowledge.worldbank.org/entities/publication/de6b4852-1a185439-bd25-cd806a3c863c The Universal, the Self-Targeted, and the Targeted 11

One of the strongest and most convincing arguments against universal subsidies is on the environmental impact of subsidising fossil energy. It postulates that subsidising fossil energy is bad for the environment and causes an increase of CO2 emissions. For example, the IMF argues that"raising fuel prices to their fully efficient levels reduces projected global fossil fuel CO2 emissions 36 percent below baseline levels in 2025—or 32 percent below 2018 emissions." 13 This is because cheap, subsidised energy diverts investment to energy-intensive sectors, and promotes its wasteful consumption. While there is some truth to this reasoning, as we will discuss in the following sections, well-designed energy subsidies can have positive environmental effects, such as if it promotes public transportation; moreover, there is no strong empirical evidence that elimination of energy subsidies in Egypt discouraged investment in energy-intensive sectors, such as mining and construction. In fact, since 2016, these two sectors have expanded and have been the drivers of growth and Foreign Direct Investment(FDI). 14 Moreover, Egypt’s CO2 emissions per capita have increased since 2014 from 2.3 to 2.4 metric tons. 15 This might be partially because of the great expansion of the real estate and construction sectors, which is one of the most energy intensive economic activities; for example, real estate construction contributes up to 40 per cent of the global energy consumption annually. 16 Finally, providing subsidised electricity to the poorest segment of societies can have beneficial environmental impacts as it renders unnecessary the use of polluting and hazardous alternatives such as kerosene lamps for lighting or heating, and makeshift fire for warmth and heating water. 17 It is admittedly counter-intuitive to think that universal energy price subsidies can be progressive, but evidence this paper presents – as well as other papers in this series – shows that reality is more complex; if one casts aside thinking in absolute consumption terms and thinks instead of the benefit derived from subsidy relative to income, we can start seeing the progressive dimension and the poverty-reduction and-prevention potential of many forms of universal subsidy. Self-targeting is often regarded in mainstream literature as a form of universal subsidy. This is because self-targeting occurs when subsidies are theoretically available to everyone, but in practice only target populations that elect to participate. In other words, a universal subsidy becomes self-targeted if it subsidises inferior goods whose consumption declines with increases of income because higher-income households prefer non-subsidised equivalents. 18 Self-targeting 13 IMF.(2022). Fossil fuel subsidies. 14 Diab& Hindy.(2021). IMF’s Social Protection Between Rhetoric and Action. 15 World Bank Data.(2022). CO2 emissions(metric tons per capita)- Egypt, Arab Rep. World Bank Data. Retrieved 1 January 2023, from https://data.worldbank.org/indicator/EN.ATM.CO2E.PC?locations=EG 16 Anderson, N.; Wedawatta, G.; Rathnayake, I.; Domingo, N.; Azizi, Z.(2022). Embodied Energy Consumption in the Residential Sector: A Case Study of Affordable Housing. Sustainability 2022, 14, 5051. https://doi.org/10.3390/ su14095051 17 Mousa, S.& Mousa, D.(2018). Universal Energy Access Is a Major Challenge for Arab Countries, World Economic Forum. Retrieved 11 March 2023, from https://www.weforum.org/agenda/2018/12/even-in-the-arab-world-accessto-energy-is-a-challenge/ 18 Adams, R.(2000). Self-targeted subsidies: The Distributional Impact of the Egyptian Food Subsidy System. Retrieved 14 December 2022, from https://openknowledge.worldbank.org/handle/10986/22207 12 Egypt, the IMF and Three Subsidy Approaches

is often implemented and works best when different income groups consume significantly different products and have significantly different consumption patterns. 19 Self-targeting is often forgotten as a form of subsidy in the dichotomised debate of universal versus targeted subsidies. For example, the IMF Social Spending Strategy only speaks of universal and targeted subsidies as the only forms of social assistance. Even when it speaks about different forms of targeting, it mentions concepts such as"categorical targeting"(i.e., targeting all children, the elderly, etc.), but never of self-targeted subsidies. 20 Egypt has significant experience with self-targeted bread subsidies. Allowing free-market bread and restricting access and reducing the size and the quality of subsidised bread meant that only those who cannot afford market-priced bread would opt for the subsidised equivalent. Selftargeting in the Egyptian context usually invokes food subsidies, especially bread. However, the same could be said to apply to social services such as health and education. The quality of public health and education has deteriorated so severely that even those who can barely afford private services(e.g., private schools, private tutoring) opt for them with no hesitation. 21 As a result, Egypt has one of the highest out-of-pocket expenditures on health compared to the global average, but also compared to different groups of countries including middle-income countries, to which Egypt belongs(see Figure 2). Figure 2. Out-of-pocket expenditure as a percentage of total health expenditure in Egypt between 2000 and 2019. 65 60 55 50 45 40 35 30 25 20 15 10 2000 Egypt, Arab Rep. Low income Middle income High income 2002 2004 2006 2008 2010 2012 2014 2016 2018 19 Adams, R.(2000). Self-targeted subsidies. 20 IMF(2019). A strategy for IMF engagement on social spending. 21 Rentsch, S.(2015)."A collapse of formal schooling in Egypt": Interview with Hania Sobhy. Hypotheses. Retrieved 27 December 2022, from https://ies.hypotheses.org/815#more-815 The Universal, the Self-Targeted, and the Targeted 13

Adams shows in a research paper published in 2000 that food subsidies in Egypt, at least for the urban population, could in fact be progressive even before the deepening of self-targeting in the late 1990s. In 1998, the poorest expenditure quintile group in urban Egypt spent 20.8 per cent of its total expenditures on subsidised food, while the top quintile group spent 16.8 per cent. With the deepening of self-targeting and exclusion by proxy means testing, the progressivity is expected to be much deeper now. Also, since higher-income groups tend to save more of their incomes, relative to income the progressivity will be even starker. Finally, self-reported income and expenditure surveys tend to exclude the wealthiest households and underestimate their income, 22 hence overestimating food expenditure of the top quintile. Conversely, energy subsidies would often be perceived as less progressive since it is quite well-established that lower-income households spend more on food relative to their income, while for energy subsidies the leakage to well-off households is believed to be generally higher due to their ownership of many energy-intensive appliances(e.g., cars, airconditioning, etc.) and their higher participation in energy-intensive activities(such as travelling). This has been reflected in the IMF 2016 programme which placed a pre-condition on the gradual elimination of all energy subsidies, while advocating for an increase in food subsidies. The IMF says in its 2016 programme staff report: About 1 percent of GDP out of the achieved fiscal savings will be directed to additional food subsidies, cash transfers to the elderly and low-income households, and other targeted social programs, including more free school meals. The aim is to replace poorly targeted energy subsidies with programs that directly support poor households. 23 The paper is set to test the hypothesis of energy subsidies being"poorly-targeted"; and test the economic impact of scrapping energy subsidies on different households. As for targeted social assistance, it has been by far the IMF’s favourite means of social assistance. The IMF, as we will discuss below, has been adamant and persistent about pushing Egypt for a transition from universal and self-targeted subsidies to targeted subsidies. However, one of the weakest points of targeting in high-poverty economies is in fact poor targeting as expressed by rates of exclusion errors. A 2018 study by the International Food Policy Research Institute(IFPRI) demonstrated that 55 per cent of the poorest quintile of households did not receive Takaful(the primary conditional cash transfer programme) transfers because they either received a government pension, had a government job or simply did not know or hear about the programme; however, the study shows that having a meagre government pension or even a government job did not prevent them from being among the poorest quintile, and therefore 22 Alvaredo, F., Assouad, L.,& Piketty, T.(2016). Measuring inequality in the Middle East 1990–2016: The World’s Most Unequal Region? Wiley Online Library. Retrieved 1 January 2023, from http://piketty.pse.ens.fr/files/AAP2019RIW.pdf 23 IMF, Request, staff report. 14 Egypt, the IMF and Three Subsidy Approaches

deserving of social assistance. 24 Similarly, the World Bank’s 2022 Egypt Public Expenditure Review asserts that approximately half of the eligible poor are not being included in the programme due to exclusion errors. 25 Exclusion errors are part and parcel of cash transfer programmes everywhere, and they have ranged from 56 per cent of the intended beneficiaries in Cambodia to as high as 93 per cent in Indonesia. 26 This challenges the narrative that cash transfers are more of an effective targeting approach to subsidy. However, it must be noted that with the expansion of the programme, exclusion errors should be reduced. It remains, however, a sizable and credible concern as to how effective in reaching the poor is the form of social assistance that has been most celebrated for its effective targeting. Already two decades ago, Korpi and Palme demonstrated that"the more we target benefits at the poor(...) the less likely we are to reduce poverty and inequality" in what they described as the"paradox of redistribution" in their famous study of the same title. The reasons for this, they argue, are the likely trade-offs between the degree of low-income targeting and the size of redistributive budgets, and because the"size of redistributive budgets is not fixed but reflects the structure of welfare state institutions." 27 2. Review of social spending in Egypt since 2011 Starting in the 1980s, the Government of Egypt(GOE) decided to dismantle universal price subsidies through the slow transition to self-targeted subsidies. It allowed limited market forces to operate, which meant that most subsidised products had non-subsidised market equivalents. This included food, transportation, education, and health. Energy was the exception and remained to a large extent universally subsidised despite multiple government mandated price increases. But what remained constant is that even after price increases it was still sold lower than the international price. 24 Breisinger, C.(2018). Impact Evaluation Study for Egypt’s Takaful and Karama Cash Transfer Program. ifpri. org. Retrieved 22 December 2022, from https://ebrary.ifpri.org/utils/getfile/collection/p15738coll2/id/132719/ filename/132930.pdf 25 World Bank Group.(2022). Egypt Public Expenditure Review for Human Development Sectors. World Bank. Retrieved 17 January 2023, from https://www.worldbank.org/en/country/egypt/publication/egypt-public-expenditure-reviewfor-human-development-sectors 26 Korashi, H.(2020). Social Protection in Egypt: Who passes the means-test? Alternative Policy Solutions. Retrieved 22 December 2022, from https://aps.aucegypt.edu/en/articles/71/social-protection-in-egypt-who-passes-the-means-test 27 Korpi, W.,& Palme, J.(1998). The paradox of redistribution and strategies of Equality: Welfare State Institutions, inequality, and poverty in the Western countries. American Sociological Review, 63(5), 661. doi:10.2307/2657333 The Universal, the Self-Targeted, and the Targeted 15

As for food subsidies, market food products such as bread, oil and tea, existed alongside subsidised products; by keeping the quality at a minimum, the GOE ensured that those consuming subsidised food products were mostly those in need. The same logic applied to education and health, where – starting the 1970s – there was an explosion of private institutions(e.g., hospitals and schools) providing better-quality, more dignified and less crowded services. The quality of the public subsidised services has rapidly deteriorated to the extent that people sought the private services even if they could barely afford it. 28 After 2011, and particularly in 2014, social protection started becoming more focused on targeting by removing many people from the food subsidy system in 2014 using Proxy Means Testing(PMT) – the GOE also started the gradual elimination of universal fuel and electricity subsidies in the same year. It is worth noting that the elimination of energy subsidies in the national budget does not necessarily mean it has been generally eliminated in the economy, as state-owned energy companies can still provide cheap energy to producers as long as they do not record losses. 29 While subsidising energy for production can have a lowering effect on inflation and benefit the final household consumer, the sectors that benefit the most from such"hidden subsidies" are both energy-intensive and export-oriented, such as the steel, cement, fertilisers and ceramic industries, which means the subsidy in the end benefits foreign consumers. 30 In addition to the PMT targeting for food subsidies, households were excluded from Takaful if they met one of the six following exclusion criteria: owned a car, owned more than one feddan (4,200 m2) of land, had a government job or pension, received transfers from abroad, or had a formal private sector job. 31 Targeting has become the predominant model after 2016 with the near-full elimination of energy subsidies and an upscaling of cash transfers. As can be seen from Figure 2 below, social expenditure has increased after 2011, probably in response to the social and political demands of this period. After 2013, social spending was halved relative to GDP from about 26 per cent of GDP to about 13 per cent in 2020. 32 The COVID-19 pandemic, however, has caused social expenditure to slightly increase, but it still remains far below the pre-2014 levels. 33 This casts serious doubt on the claim that elimination of 28 Ille, S.,& Peacey, M. W.(2019). Forced private tutoring in Egypt: Moving away from a corrupt social norm. International Journal of Educational Development. Retrieved 29 December 2022, from https://www.sciencedirect.com/science/ article/pii/S0738059318303894 29 Hussein, S.(2020). Fatourat Al-Akhta’ Al-’Arba’a: Al-Khita Al-Masriya Al-Sanduqiya Li Rafi’ As’aar Al-Taqa(the Cost of the Four Flaws: the Egypt-IMF Plan to Increase Energy Prices), EIPR.org(Egyptian Initiative for Personal Right). Retrieved from https://eipr.org/sites/default/files/reports/pdf/the_four_misconception_in_egy-monetary_fund_2020.pdf 30 Ibid. 31 Breisinger. Impact Evaluation Study for Egypt’s Takaful and Karama. 32 Source: United Nations Economic and Social Commission for Western Asia(UNESCWA).(2022). Social Expenditure Monitor- Egypt. Data Dashboards| Social Expenditure Monitor. Retrieved 22 December 2022, from http://sem. unescwa.org/dashboard/social-expenditure 33 This count by the UNESCWA includes spending on arts, culture and sports, education, environmental protection, health and nutrition, housing, connectivity and community amenities, labour market and employment, and social protection, subsidies and support to farms. 16 Egypt, the IMF and Three Subsidy Approaches

universal subsidies is designed to create fiscal space for more targeted social assistance. In fact, this should be read as a simple, yet aggressive, austerity policy designed to plug the budget deficit by acutely reducing social expenditure. Figure 3. Social spending as share of GDP between 2011 and 2021. 30% 25% 20% 15% 10% 5% 0% 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Source: United Nations Economic and Social Commission for Western Asia(UNESCWA).(2022). Social Expenditure Monitor- Egypt. Data Dashboards| Social Expenditure Monitor. Retrieved 22 December 2022, from http://sem. unescwa.org/dashboard/social-expenditure The fiscal space created by elimination of energy subsidies was mostly used to honour Egypt’s growing debt service bill. Figure 3 below shows a sharp decline in energy subsidies without being compensated for with either food subsidies or social insurance pensions(including cash transfers), which remained stable although they were meant to increase as a result of the fiscal savings achieved by elimination of energy subsidies, according to the IMF narrative. On the other hand, the budget item that has expanded and consumed a significant portion of the newly found fiscal space was interest payments rising from about eight per cent of GDP in 2013 peaking at about 10 per cent of GDP in 2019, before dropping again to about nine per cent of GDP. With the recent subsequent rises of interest rates prescribed by the IMF, this figure is expected to see a renewed spike. The Universal, the Self-Targeted, and the Targeted 17

Figure 4. Interest payments versus social spending components(% of GDP) 10 8 Interest payments 6 Petroleum product subsidies Food subsidies 4 Social insurance pensions(incl. Takaful and Karama) 2 0 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 Author’s calculations based on various national budget and national account figures published by the Egyptian Ministry of Finance and the Ministry of Planning and Economic Development. 3. Review of IMF social spending recommendations and conditions 3.1 An overview: Did the IMF really change, and how binding are their social policy conditions? Since the 2008 Global Financial Crisis, the IMF has adopted a stronger social policy discourse and has abandoned, at least rhetorically, advocating for the most aggressive forms of neoliberal reform. However, there has been serious doubts about whether this new discourse has been translated to action. Assessing 55,465 individual programme conditions across 131 countries from 1985 to 2014, a 2016 study detected a fundamental shift in IMF conditionality towards more social policy. 34 The authors found that after 2008 the IMF emerged again as a central player in the post-crisis global policy response, after a period of organisational insecurity and relatively slow activity in the IMF preceding the 2007–08 financial crisis, resulting in a decline of conditionality attached to its loans; the study shows how structural conditions were revived as key components of IMF programmes after the crisis. 34 Kentikelenis, A. E., Stubbs, T. H.,& King, L. P.(2016). IMF Conditionality and Development Policy Space, 1985–2014, Review of International Political Economy, 23(4), pp. 19. 18 Egypt, the IMF and Three Subsidy Approaches

The most binding type of conditionality in IMF programmes are prior actions, which are measures that have to be implemented before any of the funding is dispatched. Quantitative Performance Criteria(QPCs), are quantifiable and measurable conditions, such as those related to primary balances, international reserves, debt levels, etc. If QPCs are not met, the IMF Executive Board will need to issue a waiver to continue dispatching loan tranches. Indicative Targets(ITs) are like QPCs but are less binding and are more akin to programme objectives than conditions, meaning that if they are not met, the IMF Executive Board would not need a waiver. 35 The study shows that despite the increase in social policy conditions, social spending was rarely part of the programmes’ binding conditions, meaning that their non-implementation is unlikely to affect countries’ access to credit. Engström also argues that social spending is commonly set out as non-binding indicative targets, which creates a structural bias in favour of binding macroeconomic goals. Structural Benchmarks(SBs) are as non-binding as ITs, but are more of qualitative objectives and hence cannot be classified as QPCs or ITs. 36 I have shown in a previous study on social protection in the Egypt 2016 programme that all the programme’s prior actions or preconditions were non-social; they were all in fact typical inflationary neoliberal policies with adverse social impact as they have directly affected poor households’ purchasing power. The three preconditions were: the devaluation of the Egyptian pound, the elimination of energy subsidies, and the enactment of a regressive 14 per cent VAT to replace a 10 per cent sales tax. 37 The programme’s two social protections measures were listed as structural benchmarks, and therefore lack the binding power as major neoliberal reforms that were listed as prior actions, that is, the most binding type of conditionality. 38 Before 1991, the IMF recommendations on social spending mostly revolved around simply reducing it. In May 1962, at the peak of Egypt’s state-led development model, it signed a standby agreement with the IMF as a result of a foreign currency crisis driven by a fall in cotton and rice export revenues due to crop failure. However, the agreement proved short-lived as Egypt found other sources of financing. This was Egypt’s only agreement with the Fund from its establishment in 1944 until the late 1970s. The shift in Egypt’s economic policy giving more space to the private sector paved the way for more intensive relations with the IMF. 39 35 IMF.(2021, February 22). IMF conditionality. IMF. Retrieved 22 December 2022, from https://www.imf.org/en/About/ Factsheets/Sheets/2016/08/02/21/28/IMF-Conditionality 36 Engström, V.(2022). Social Protection in the mandate of the IMF. The International Journal of Human Rights, 1-21. doi:10.1080/13642987.2022.2078313 37 Diab, O.,& Hindy, S. I.(2021). IMF’s social protection between rhetoric and action: The case of Egypt. Middle East Critique, 30(4), 391-409. doi:10.1080/19436149.2021.1989551 38 The two measures included expanding Egypt’s two main cash transfer programmes, Takaful and Karama, and building affordable public nurseries to encourage women to actively seek employment. 39 Al-Khãliq, J.(2001). Stabilization and adjustment in Egypt: Reform or de-industrialization. Cheltenham; Northampton, Mass: Edward Elgar Pub. The Universal, the Self-Targeted, and the Targeted 19

The 1977 Stand-by Arrangement(SBA) and the 1978 Extended Fund Facility(EFF) had no positive social protection component. Conversely, it still placed conditions on dismantling the existing social protection regime through reducing fiscal deficit by cutting food, energy subsidies and government services. 40 IMF activity in pushing Egypt to reduce its subsidy bill in the 1970s led to large-scale violent protests across the country in January 1977, popularly known as the Bread Riots in reference to the increasing cost of staple products such as bread. The food subsidy system which started in the interwar period had expanded by the 1970s to include almost 20 foods and to account for about 15 per cent of total government expenditures. By 2000, the Egyptian food subsidy system only included four items: baladi bread, wheat flour, oil and sugar. Baladi bread used to be sold to consumers without quantity restrictions by private sector bakeries, the wheat flour was sold by government warehouses, and edible oil and sugar were sold at fixed monthly quotas by private groceries. 41 As a result, the urban households in 1974/5 were spending a fifth of their total expenditure on subsidised foods. However, the real value of food subsidies and its contribution to household expenditure started declining rapidly in the 1980s. 42 In the new world economic order emerging in the 1970s, subsidy was thought of as the easiest major budget item to cut, until doing so sparked significant unrest in 1977 when protests and riots erupted across Egypt to express discontent with the new subsidy cuts plans. Hamed Latif elSayeh, Minister of Economy and Economic Cooperation at the time of the IMF 1970s agreements, said in a press interview: We had a government budget deficit of over US$2.0 billion[...] we decided that this must be reduced. There are four major items in the budget: military, investment, subsidies and debt service. Should we cut back on our military? You can’t do that. Investment? That’s growth and without growth Egypt cannot live. Debt service? We have more than$2.5 billion in medium-and long-range debt. You can’t touch that. (So the only thing left) is subsidies. A few years ago we started with bread subsidies. Then we went into all kinds of things. The subsidies went from US$175 million four years ago to US$1.7 billion. This is why we moved immediately, and without thinking politically, I guess, to reduce subsidies. The problem was it was a bad strategy. We are not trained politicians. We did not anticipate any trouble. We told the International Monetary Fund there would be no trouble. 43 After the tragic events of 1977, successive governments have become extra prudent with handling food subsidies. For example, the government resorted to reducing the relative cost of bread subsidy by limiting access to subsidised bread to holders of ration cards, revoking millions 40 Al-Khãliq,(2001). Stabilization and adjustment in Egypt. 41 Adams, R.(2000). Self-targeted subsidies. 42 Al-Khãliq,(2001). Stabilization and adjustment in Egypt. 43 Adams, R.(2000). Self-targeted subsidies. 20 Egypt, the IMF and Three Subsidy Approaches

of cards that it deemed its holders undeserving of food subsidy using proxy-mean testing, and reducing the size of the loaf from 150 grams in the 1980s to 90 grams currently. 44 As discussed above, Egyptian policymakers began implementing prudent self-targeting by means of quality differentiation allowing the market to produce and sell different and larger types of unsubsidized bread. The sale of what is perceived as higher quality breads attracted the demand of betteroff households, while leaving subsidised bread for purchase by the poor. Finally, in 1992 the government stopped subsidising the"higher quality" fino bread. Four years later, in July 1996 the government stopped subsidising the sale of white pita shami bread while keeping only baladi bread subsidised. 45 More recently, subsidised baladi bread has become rationed using a point system where each eligible household receives on average about 10 loaves a day. 3.2 The IMF’s social spending influence in the Egyptian case Table 1. History of IMF lending arrangements with Egypt from 1962 to 2020 (In Thousands of SDRs). Facilty Date of Arrangement Expiration Date Amount Agreed Amount Amount Drawn Outstanding Standy Arrangement June 26, 2020 June 25, 2021 3,763,640 2,605,600 2,605,600 Rapid Financing Instrument May 11, 2020 May 13, 2020 2,037,100 2,037,100 2,037,100 Extended Fund Facility Nov 11, 2016 July 29, 2019 8,596,570 8,596,570 8,596,570 Standy Arrangement Oct 11, 1996 Sep 30, 1998 271,400 0 0 Extended Fund Facility Sep 20, 1993 Sep 19, 1996 400,000 0 0 Standy Arrangement May 17, 1991 May 31, 1993 234,400 147,200 0 Standy Arrangement May 15, 1987 Nov 30, 1988 250,000 116,000 0 Extended Fund Facility July 28, 1978 July 27, 1981 600,000 75,000 0 Standy Arrangement April 20, 1977 April 19, 1978 125,000 105,000 0 Standy Arrangement May 23, 1964 May 22, 1965 40,000 40,000 0 Standy Arrangement May 7, 1962 May 6, 1963 42,500 42,500 0 Total 16,360,610 13,764,970 13,239,270 Source: International Monetary Fund 44 Ibid.& Awadalla, N.(2020). Egypt shrinks subsidized bread loaf by 20 grams, revises cost of flour. Reuters. https:// www.reuters.com/article/egypt-commodities-bread-idAFL8N2FJ50N 45 Adams, R.(2000). Self-targeted subsidies. The Universal, the Self-Targeted, and the Targeted 21

In terms of social spending, the Economic Reform and Structural Adjustment Programme(ERSAP) established a Social Fund for Development(SFD) with the objective to"minimise effects of reforms on the poor" and"mitigate the transient effects of the reforms especially on vulnerable groups". 46 The SFD focused on supporting small enterprises, microfinance, infrastructure expenditure and community development. 47 Former Minister of Social Solidarity Gouda Abdel-Khaliq calls it the "palliative for the bitter medicine of ERSAP", and describes the SFD as a temporary measure to provide short-term and temporary relief to groups that will be negatively affected by the programme measures. Although the 1991 programme focused on poverty alleviation, poverty rates have been persistently on the rise since the 1990s, skyrocketing from 16.7 per cent of the population in 1999 to 32.5 per cent, before modestly dropping to 29.7 per cent in 2019. No official poverty figures have been published since 2019. 48 However, after the economic hardships caused by the COVID-19 pandemic followed by the war in Ukraine, official poverty rates are expected to have increased. The IMF programme with the strongest social protection narrative was the 2016 one. The IMF recommended the increase of spending on the main conditional cash transfer programme and food subsidies by one per cent of GDP(EGP 25 billion). Additionally, another structural benchmark related to social protection was the construction of affordable nurseries to enable women to join the workforce. There was a systematic rift, however, between the 2016 IMF programme’s professed objectives and actual outcomes in the case of Egypt. 46 African Development Bank.(1999). Arab Republic of Egypt Structural Adjustment Loan 1991-1994 Project Completion Report 1999-2001, African Development Bank. Retrieved 10 May 2020, from https://www.afdb.org/ fileadmin/uploads/afdb/Documents/Project-and-Operations/ADB-BD-IF-99-130-EN-EGYPT-PCR-SAL.PDF 47 Abou-Ali, H., El-Azony, H., El-Laithy, H., Haughton, J., and R. Khandker, S.(2009). Evaluating the Impact of Egyptian Social Fund for Development Programs, World Bank. Retrieved 30 April 2020, from at: http://documents.worldbank. org/curated/en/970431468037489618/pdf/WPS4993.pdf 48 World Bank Data.(2022). Poverty headcount ratio at national poverty lines(% of population)- Egypt, Arab Rep. World Bank Data. Retrieved 29 December 2022, from https://data.worldbank.org/indicator/SI.POV.NAHC?locations=EG 22 Egypt, the IMF and Three Subsidy Approaches

Table 2: Review on IMF recommendation on social spending in Egypt(1962-2022) Type of arrangement Stand-by Arrangement Stand-by Arrangement Stand-by Arrangement Extended Fund Facility Stand-by Arrangement Stand-by Arrangement Extended Fund Facility Year of agreement 1962 1964 1977 1978 1987 1991 2016 Rapid Financing Instrument Stand-by Arrangement Extended Fund Facility 2020 2020 2022 Social spending recommendation None None None None None Establishment of the social fund for development - Expansion of cash transfers - Building nurseries and safe transportation to encourage women to join waged labour None Public Expenditure Review Increase of social spending and strengthening social safety net to protect the vulnerable Source: author’s collection of social protection components of the different IMF agreements with Egypt One major difference between the 1991 and 2016 programme when it comes to social protection is the objective of the social measures. In 1991, the establishment of the SFD was aimed at poverty alleviation. However, the objective of the 2016 social protection component was to merely mitigate the impact of the reforms on the most vulnerable groups. Poverty alleviation or reduction were never mentioned in all the 2016 programme documents. Only synonyms or near-synonyms of"mitigation" were used, namely"ease","offset","alleviate","shield", and "protect". Here is one example: To alleviate the impact of these measures on the poor, about 1 percent of GDP of the savings in 2016/17 have been set aside to be spent on social protection in addition to the amounts allocated in the budget last year. 49 49 IMF.(2017b). Arab Republic of Egypt: Request for extended arrangement under the extended fund facility press release; Staff Report; and Statement by the Executive Director for the Arab Republic of Egypt. Retrieved 8 December 2022, from https://www.imf.org/en/Publications/CR/Issues/2017/01/18/Arab-Republic-of-Egypt-Request-forExtended-Arrangement-Under-the-Extended-Fund-Facility-44534 The Universal, the Self-Targeted, and the Targeted 23

The mitigation narrative absolves the IMF and the government of their social protection responsibilities. Unlike poverty reduction, mitigation is an unmeasurable concept and one can always argue that things could have been worse if mitigation measures were not put in place. Critics have also pointed out that the IMF’s focus on mitigation as an objective of social protection ignores underlying structural issues. 50 However, the IMF has returned to using"poverty alleviation" in its 2022 EFF agreement with Egypt. 51 The IMF’s rhetoric on social spending reached its peak when it adopted the Social Spending Strategy in 2019. 52 Although the history of the IMF’s lending commitments to Egypt date back to at least 1962, it was not until 1991 that the IMF started explicitly adopting social policy. In 1991, the IMF along with other IFIs such as the World Bank and the African Development Bank, pushed Egypt to conduct its first major and completed structural adjustment programme, officially known as the ERSAP. The pillars of the ERSAP included the introduction of an indirect consumption tax (a 10 per cent sales tax), eliminating subsidies, eliminating price controls, reducing the fiscal deficit, increasing foreign reserves, increasing fuel and electricity prices, trade liberalisation and the privatisation/corporatisation of the public sector. 53 Social spending for the IMF is not an objective in and of itself; it has two objectives as set out in the IMF’s Social Spending Strategy; first, when social spending engages macro-criticality(i.e., when it has significant impact on macroeconomic stability), and second, for mitigating the adverse effects of adjustment. This prevents the setting of conditions promoting social protection if its spending cannot be justified as crucial for the achievement of the stated macroeconomic objectives. 54 This contrasts with other approaches to social protection that deal with social protection as a human rights issue. For example, the International Labour Organisation(ILO), as discussed in the book’s introduction, treats social protection as a human right that must be honoured through the pursuit of social policies and programmes in liaison with the Universal Declaration of Human Rights. This human rights-based approach is manifest in the ILO’s encouragement of countries to adopt social protection floors, which are defined as the ability to access healthcare, basic income security for the elderly and disabled, support for dependent children, and basic income security for the unemployed and the working poor, as well as government programmes providing job opportunities. 55 50 Engström, V.(2022). Social Protection in the mandate of the IMF. The International Journal of Human Rights, 1–21. Retrieved from https://doi.org/10.1080/13642987.2022.2078313 51 IMF.(2022). IMF Executive Board approves 46-month US$3 billion extended arrangement for Egypt. IMF. Retrieved 20 December 2022, from https://www.imf.org/en/News/Articles/2022/12/16/pr22441-egypt-imf-executive-boardapproves-46-month-usd3b-extended-arrangement 52 IMF.(2019). A strategy for IMF engagement on social spending. Retrieved 13 December 2022, from https://www.imf. org/en/Publications/Policy-Papers/Issues/2019/06/10/A-Strategy-for-IMF-Engagement-on-Social-Spending-46975 53 Diab, O.,& Hindy, S. I.(2021). IMF’s social protection between rhetoric and action: The case of Egypt. Middle East Critique, 30(4), 391-409. Retrieved from https://doi.org/10.1080/19436149.2021.1989551 54 Engström, V.(2022). Social Protection in the mandate of the IMF. 55 International Labour Organization.(2012). Social Protection Floors Recommendation, 2012(No. 202). Retrieved March 14, 2023, from https://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB%3A12100%3A0%3A%3ANO%3A %3AP12100_INSTRUMENT_ID%3A3065524 24 Egypt, the IMF and Three Subsidy Approaches

The UN General Assembly, in its setting out of the Sustainable Developments Goals, also treats many social protection issues from a rights-based viewpoint. In contrast, the IMF has always refrained from engaging with a human rights discourse, and insisted that its role is to ensure the stability of the international monetary system; 56 it only engages with non-monetary matters insofar as they have an impact on monetary stability. The IMF’s Guidelines on Conditionality emphasise the need for social measures to be"feasible" and "appropriate". It reads:"if feasible and appropriate, any adverse effects of program measures on the most vulnerable should be mitigated" 57 (my italics). As for the IMF’s Social Spending Strategy, "mitigating the adverse effects of adjustment on vulnerable groups and improving spending adequacy can usually be addressed by including quantitative conditionality". 58 In 2020, the IMF signed two agreements with the Egyptian government. One was a COVID-19related emergency finance, officially known as Rapid Financing Instrument(RFI), and the second was a 12-month-SBA, whose negotiations had started in 2019 long before the outbreak of the pandemic. The RFI does not require having a programme in place. As for the SBA, its only social protection measure did not entail increased social expenditure and was instead a technical one, that is, conducting a public expenditure review(PER) in order"to ensure sufficient and appropriate reallocation of resources and achieve the maximum efficiency". According to the programme’s documents, the GOE promised to undertake – with the help of the World Bank – a PER on social protection spending, health, and education. 59 In December 2022, the IMF Executive Board approved a USD 3 billion 46-month EFF arrangement. The programme remains neoliberal in essence. While the programme included only one prior action on the discontinuation of subsidised lending to the industrial, construction and agricultural sectors, most of the austerity and contractionary measures were included as QPCs. The programme’s three QPCs are the increase of net international reserves, the increase in primary fiscal balance, and no new external debt payments arrears. As for SBs and ITs, they include increases in tax revenues, increasing the maturity of local debt, enhancing public procurement disclosure, fuel price indexation to international prices, and the publishing of annual tax expenditure reports. 56 Engström, V.(2022). Social Protection in the mandate of the IMF. 57 IMF.(2014). Revised Operational Guidance to IMF Staff on the 2002 Conditionality Guidelines. IMF.org. Retrieved 23 December 2022, from https://www.elibrary.imf.org/view/journals/007/2014/047/article-A001-en.xml 58 IMF.(2019). Social Spending Strategy. 59 IMF.(2022). Arab Republic of Egypt: Request for a 12-Month Stand-By Arrangement-Press Release; Staff Report; and Statement by the Executive Director for the Arab Republic of Egypt. IMF. Retrieved 24 December 2022, from https://www.imf.org/en/Publications/CR/Issues/2020/08/10/Arab-Republic-of-Egypt-Request-for-a-12-Month-StandBy-Arrangement-Press-Release-Staff-49683 The Universal, the Self-Targeted, and the Targeted 25

As for social protection, the only measure was to expand the number of households benefiting from Takaful and Karama to five million households(SB), which is linked to the indicative target of increasing social spending by EGP 153 million in 2023(about 1.5 per cent). These social spending conditions only promote the horizontal expansion of the cash transfers by increasing the number of beneficiaries, whereas vertical expansion is also urgently needed as cash received by households has barely increased since 2015 despite several waves of very high inflation as we will discuss below. Unlike the 2016 programme, social spending was included as an indicative target, making it a degree more binding. Moreover, the 2022 EFF explicitly mentions poverty reduction and inequality, whereas the 2016 EFF only spoke of mitigation as discussed above. However, the small and horizontal scale of this single measure, the lack of measures on health and education, and other forms of social assistance is expected to make the impact very limited in the face of extreme loss of purchasing power caused by the most recent IMF-mandated pound devaluation in January 2023. 4. Impact of the COVID-19 lockdown and the Ukraine war on social protection The IMF and the GOE often rightly point to the external"force majeure" nature of the COVID-19 pandemic impact. However, it must be noted that the weakening status of social protection as Egypt entered the crisis compounded its challenges. Moreover, responses were not adequate enough to prevent millions more from falling into poverty as we discussed above. This is especially the case when we consider that Egypt was one of the few economies in the world to have a positive GDP growth due to enforcing limited lockdown measures. 60 A mix of monetary and fiscal measures were taken by the GOE during the pandemic to support vulnerable groups and the economy including: 1. Reducing policy interest rates by 400 basis points during 2020 – with the overnight deposit rate cut from 12.25 per cent to 8.25 per cent(monetary); 2. A six-month debt moratorium on existing credit(monetary); 60 Assaad, R. et al.(2022). Egypt Covid-19 country case study. International Labour Organization. Retrieved 23 December 2022, from https://www.ilo.org/africa/information-resources/publications/WCMS_838226/lang--en/index.htm 26 Egypt, the IMF and Three Subsidy Approaches

3. Reducing electricity tariffs for all industries by nine per cent(at an annual cost of EGP 4 billion); 4. EGP 500 a month to affected informal workers for a period of three months; 5. Adjusting the percentage of personal consumer loans to a maximum of 50 per cent from total monthly income; and 6. Increasing the tax exemption limit to EGP 15,000 annually. 61 However, these measures have mostly benefited those with access to formal banking credit, which is a minority of Egyptians. Only 20 per cent of the poorest 40 per cent of the population have access to a financial account, including with mobile-money service providers. 62 If we consider access to credit, this figure will be even much lower. A study conducted by the Central Authority for Public Mobilisation and Statistics(CAPMAS) issued in 2020 assessed the impact of the pandemic on the income of Egyptians, and concluded that most of those polled(73.5 per cent) have seen their incomes decrease since the start of the pandemic. Those affected have mostly depended on their informal social networks, grants and borrowing, as opposed to official government support packages. 63 The report says: About half of the household borrowed from others, and about 17% depended on the assistance of charitable people, while 5.4% of households received the informal employment grant, in the event of insufficient income. The impact was not evenly distributed among different social groups, although they were all negatively affected. The CAPMAS report shows a very clear inverse relationship between the level of education and the negative impact on income since the start of the pandemic, meaning that the lower the level of education, and hence income, the greater the negative impact. This has several reasons which government support packages did not sufficiently address. 61 KPMG.(2020). Egypt: Government and institution measures in response to COVID-19. KPMG. Retrieved 29 December 2022, from https://home.kpmg/xx/en/home/insights/2020/04/egypt-government-and-institution-measures-inresponse-to-covid.html 62 World Bank Data.(2022). Account ownership at a financial institution or with a mobile-money-service provider, poorest 40%(% of population ages 15+)- Egypt, Arab Rep. World Bank Data. Retrieved 24 December 2022, from https://data.worldbank.org/indicator/FX.OWN.TOTL.40.ZS?locations=EG 63 CAPMAS,(2022). Cairo; CAPMAS. The Universal, the Self-Targeted, and the Targeted 27

Table 3. The impact of the COVID-19 pandemic on income based on education level of earners Pandemic impact on income(% of total polled) Educational level Illiterate Can read and write Primary school certificate Preparatory school certificate Secondary school certificate Technical school certificate Two-year post-secondary degree University and post-graduate degree Source: CAPMAS, Declined 85.1 81.3 87.3 84.3 86.6 75.7 59.2 48.7 Unchanged 13.8 18.2 12.4 15.3 13.4 23.6 40.8 50.8 (2022). Cairo; CAPMAS. Increased 1.1 0.4 0.4 0.4 0.0 0.7 0.0 0.6 First is the disproportionate inability of low-income workers to work from home. Domestic work was a privilege mostly limited to higher educated"white-collar" workers. Second, low-income workers did not benefit from corporate support because employment support in the case of Egypt could at best be considered indirect support obtained by companies. Since those with lower levels of education usually do not have a stable and contractual employment in the private sector or the government, they are more likely to not have benefited from corporate support and seen their income plummet since the start of the lockdown. Since the CAPMAS study did not measure the impact according to any other social indicators, we currently only have the educational level to deduce the disproportionate impact of the crisis. What is certain is that income is correlated with levels of education. Because of this disproportionate impact of the crisis on the poorest groups, poverty and extreme poverty rates are expected to have further increased in the pandemic period after they have increased continuously and steadily in recent years and decades. 64 This is why direct individual support, not just corporate support, would have been crucial to protect vulnerable workers who have seen their incomes plummet in 2020. While the government announced financial support of EGP 500(about USD 30 according to exchange rates at the time) for three months – later extended to six months – for informal workers, it is such a meagre amount and only lasted for a very short period. Additionally, only a small share of households 64 Economy Plus.(2019, August 1). 2022, from https://economyplusme.com/11052/ . Economy Plus. Retrieved 23 December 28 Egypt, the IMF and Three Subsidy Approaches

(around 11 per cent) received such minimal support, 65 while both the poor and informal workers constitute about a third of the population/workforce. Half of the government support package estimated at EGP 100 billion and EGP 3 billion of concessional loans at low interest rates went to support the tourism sector alone, in addition to concessional loan packages worth billions of pounds for the manufacturing, agricultural and aviation sectors. 66 The study also did not show the impacts according to gender, but it is established that the representation of women among the illiterate and the poor is higher than that of men; therefore, it is also expected that women were more affected than men. A study published by the ILO and the Economic Research Forum(ERF) confirms this analysis and even shows that these disproportionate impacts on women and low-income groups could be more lasting. It demonstrates that formal wage workers were relatively protected from income declines and saw their work hours recover in June 2021. However, informal wage workers experienced a decline in their wages. The study also shows that a substantial and increasing proportion of self-employed workers saw their incomes fall relative to pre-pandemic levels. Although small and medium enterprises(SMEs) started recovering already in the second quarter of 2021, micro-enterprises saw a continued deterioration of their situation from February to June 2021, and two-thirds of micro-enterprises and SMEs were unable to access government support programs. The report finally shows that "a substantial majority of those interviewed reported low levels of wellbeing as measured by the WHO-5 scale, but even more worrying, this proportion increased from February to June 2021, especially among women." 67 In short, social protection during the pandemic had a number of gaps, namely, poor support to informal and low-income workers, as well as micro-enterprises, due to their lack of access to financial credit support. In addition to the disproportionate negative effects on the poor, the pandemic social assistance packages did not encompass a gender aspect resulting in a steeper drop in women’s well-being according to the WHO-5 scale, a measure devised by the World Health Organisation to measure mental health and well-being. The impact of the Ukraine war dealt yet another blow to low-income households. Levels of inflation have significantly increased after a relatively long period of relatively low inflation. 65 Ebeid, H.,& Hussein, S.(2022). Building Resilience Blocks: How to Improve the Quality of Work for the Egyptian Construction Precariat? euromesco.net. Retrieved 14 April 2023, from https://www.euromesco.net/wp-content/ uploads/2022/01/Paper-N%C2%BA50.pdf 66 IMF.(2022). Policy Responses to COVID-19. IMF. Retrieved 23 December 2022, from https://www.imf.org/en/Topics/ imf-and-covid19/Fiscal-Policies-Database-in-Response-to-COVID-19 67 Assaad.(2022). Egypt COVID-19 Country Case Study. The Universal, the Self-Targeted, and the Targeted 29

However, this inflationary wave is also thought to have had a disproportionate impact on lower income households based on the observation that food inflation witnessed one of the highest rates of inflation among different commodity groups. It is well-established that the lower income households spend significantly more of their income on food. According to CAPMAS 2015 HIECS data, the extreme poor households spent about 57 per cent of their income on food while high-income households spent less than 40 per cent. 68 If we consider the inflation wave caused by the Ukraine war, we find that the immediate impact hovered around 15 per cent a year. Food, as the largest expenditure item and one whose Consumer Price Index(CPI) changes has the largest impact on purchasing power, saw one of the highest CPI increases from September 2021 to September 2022. It came only third to"restaurants and hotels" and"recreation and culture". However, the weight of those two categories is quite minimal, and"restaurants and hotels" include food services, including cheap takeaway services that low-income workers heavily depend on. This means that it is very likely that low-income households have experienced higher rates of inflation as a result of the war than that of higher-income households. This is a good reminder of the calls to develop disaggregated inflation indicators for different income groups, especially for lower income. 69 This has in fact been quite typical in recent decades. For example, the global inflation rate was 2.5 per cent in 2018, while the global food inflation rate was about four per cent in the same year. 70 Similarly, in Bangladesh, the Basic Needs Price Index, which mainly consists of foodstuffs, shows a rise in basic needs prices of 85.5 per cent over the period from 2005 to 2010, compared to a 44 per cent increase in the general CPI, which includes essential and nonessential spending. 71 Also in the United Kingdom, the poorest 20 per cent of the population was found to experience a higher rate of inflation than all other income groups in all but one of the years between 2003 and 2013. 72 Reasons for this vary, but one factor is that as incomes of the consuming classes(low- and middleincome households) stagnate or squeeze, demand for non-essential goods declines. However, demand for essential goods is unlikely to be affected due to its low(and even negative) elasticity of demand. The Ukraine crisis and its consequent global energy and food shocks remind us of the need to not only index all forms of social protection(contributory social insurance and non-contributory social assistance) to inflation, but also index each income category to its own experience of inflation 68 Author’s calculation after removing imputed spending and adding expenditure on food and restaurants. 69 For advocacy efforts of creating a disaggregated inflation rates see Murphy, E.,& Garvey, E.(2004). 70 Ivanic, M.,& Martin, W.(2018). Implications of Higher Global Food Prices for Poverty in Low-Income Countries. Retrieved from http://documents.worldbank.org/curated/en/253001468150585777/pdf/wps4594.pdf 71 Atkinson.(2019). Measuring poverty around the world. 72 Monaghan, A.(2014, December 15). Britain’s poorest households face higher rates of inflation, research suggests. Retrieved from https://www.theguardian.com/business/2014/dec/15/poorest-households-inflation-spending-income 30 Egypt, the IMF and Three Subsidy Approaches

based on its consumption patterns. Needless to say, any lack of action on that front will mean that whatever social protection efforts are made, they will be very quickly eroded by high levels of inflation, especially the inflation of basic needs. Table 4. CPI of different consumer goods and services categories from September 2021 to September 2022. Goods and service category General inflation Food and non-alcoholic beverages Tobacco and alcoholic beverages Clothing and footwear Housing, water, electricity, gas, and other fuels Furniture, household equipment and routine maintenance of the house Health Transport Communication Recreation and culture Education Restaurant and hotels Miscellaneous goods and services YoY change%(Sep 2021-Sep 2022) 15.3 21.5 12.9 12.8 6.5 17.9 8.6 16.1 0.8 24 13.9 26 13.1 Source: CAPMAS, Monthly Bulletin of Consumer Price Index(CPI)- September 2022. The Universal, the Self-Targeted, and the Targeted 31

5. Methodology for calculating embodied and direct energy expenditure Many studies aim at calculating the environmental impact of households with different income profiles using a mix of national CO2/GHG data and Consumer Expenditure Surveys(CES). For example, Song et al. integrated CES data in the United States with Environmentally Extended Multi-Regional Input-Output(EEMRIO) models that count CO2 emissions on the national level. 73 Similarly, Ivanova and Wood apply a comparable method to illustrate the distribution of carbon footprints and consumption within 26 European Union countries, regions and social groups. 74 To measure the carbon intensity of different economies, they calculate kgs of CO2 for every expended Euro. The present study uses the same starting point to calculate embodied energy of household consumption as explained below. To calculate embodied energy expenditure for households, I calculated embodied CO2 in every EGP spent on consumer goods(CO2/EGP) by dividing total CO2 emissions by"GDP minus exports". Imports, whose embodied energy were used outside of Egypt, are already subtracted from the GDP. Subtracting exports and imports means that the measure is of the energy used to produce goods and services that are consumed domestically. Following this, CO2 was transformed to specific fuel types(e.g., natural gas, gasoline, etc.) in two steps: first, determining the national energy mix 75 after excluding energy consumed by the household sector to exclude direct energy consumption(e.g., consumption of house electricity, gasoline for cars, etc.), which is already included in the HIECS data(see Table 6). After determining the relative weight of each fuel type in the national energy mix, I used its CO2 content along with its relative weight in the energy mix to determine its percentage share in every spent EGP. This was then monetised by using the cost of each fuel type in 2015 to establish the energy profile of every spent EGP. Now, indirect(embodied) expenditure of households could be established by multiplying the energy profile of each EGP by households’ incomes. After this, the CPI increase of each fuel type between December 2015 and August 2019 was applied to calculate the rate of increase 73 Song, K. et al.(2019). Scale, Distribution and Variations of Global Greenhouse Gas Emissions Driven by U.S. Households, Environment International, 133, pp. 105–137. 74 Ivanova, D.,& Wood, R.(2020). The unequal distribution of household carbon footprints in Europe and its link to sustainability. Global Sustainability, 3, E18. doi:10.1017/sus.2020.12 75 Based on Gross Inland Consumption published in CAPMAS Energy Balance Report 2015-2016 where different fuel types are measured in tonnes of oil equivalent. 32 Egypt, the IMF and Three Subsidy Approaches

of embodied energy expenditure. As for direct energy expenditure, these were divided into expenditure on housing energy and on private vehicles. 76 Table 5. Method for calculating CO2/EGP Total CO2 in kgs in 2015(production-based to exclude embodied imported energy) 218,170,000,000 GDP(less exports) (EGP in 2015) 2,083,200,000,000 Kgs of CO2 for every spent EGP 0.105 Source: GDP data taken from the Egyptian Ministry of Planning website, while export to GDP data from World Bank Data website: https://data.worldbank.org/indicator/NE.EXP.GNFS.ZS?locations=EG As discussed in the paper, calculating the impact on household energy subsidy was based on creating five income profiles(extreme poor, poor, near poor, middle-income, high-income) whose consumption patterns were based on official HIECS data from 2015 – the year before the implementation of the 2016 IMF economic reform programme. For the extreme poor and poor households, national extreme poverty and poverty lines were used. The near poor was just under double the income of the extreme poor. Top income constituted the top 10 per cent of the income ladder, and the rest of the HIECS sample was considered middle income. The evolution of the general cost of living was traced using official CPI data for different goods and services categories. The contribution of the increase of energy(embodied and direct) costs to the overall increase in the cost of living for each of the respective households was then calculated by creating a consumption basket made up of 12 goods and services categories for each of the five households based on the 2015 HIECS data(see Table 7). The CPI for each category was used to calculate the overall inflation for the period from December 2015 to August 2019, and the energy contribution to overall inflation for each of our households could therefore be determined(see Table 8). 76 Here we use expenditure on operation of personal transport equipment of which is constituted mostly of expenditure on vehicle energy. The Universal, the Self-Targeted, and the Targeted 33

Table 6. The breakdown of fuel type in every expended EGP in 2015. Oil equivalent (000 tons) Weight Kgs of CO2 in 1 EGP by fuel type Kgs of oil equivalent in 1 EGP by fuel type Conversion factor Natural Gas Electric Diesel Fuel oil/ mazut Gasoline Total 9,740 7,133 14,032 2,296 7,348 40,549 24.02032109 17.59106266 34.60504575 5.662285137 18.12128536 100 0.02522133715 0.01095326642 0.0184706158 0.008021524575 0.03633529803 0.01577990086 0.005945399393 0.002582002022 0.01902734963 0.008263306123 0.105 0.0456 1 kg= 2.75 kgs CO2 1 KWH= 0.23 kg CO2 1 litre=2.7 kg CO2 1 kg=3.11 kgs of CO2 1 Litre=2.3 kgs of CO2 Fuel type in every spent EGP 0.009171395326 0.08030702521 0.01345751779 0.001911703985 0.008272760707 Unit Kg EGP average price for type of fuel in 2015 1.33 % of spent EGP by fuel type 0.01219795578 Kw/h 0.61 0.04898728538 Litre 1.88 0.02530013345 Kg 2.2 0.004205748767 Litre 2.2 0.01820007356 0.1088911969 Source: Table 53- CAPMAS Annual Electricity and Energy Bulletin 2015-2016; for electricity, the median of different consumption brackets in 2015 was used. Table 7. Consumption patterns for households with five different income profiles Extreme Poor poor Near poor Middle- High- CPI Increase(%) income income(Dec 2015 to Aug 2019) % of total expenditure Food 53.01 50.9 47.85 44.03 35.73 91.8318028033 Alcohol and tobacco 5 5.39 5.46 5.05 3.54 81.0827629123 Clothing and footwear 5.93 6.35 6.18 5.96 5.3 63.9909977494 Housing 6.52 6.57 6.58 6.68 6.59 50.3858024691 Furniture& housing equipment 4.17 4.34 4.22 4.37 4.64 64.2220699109 Health 7.03 7.11 8.46 10.11 12.91 65.2115513768 Transportation 3.81 4.36 4.94 6.03 8.22 102.6315789474 Communications 1.81 1.95 2.22 2.67 3.24 13.1416837782 Recreation and culture 1.09 1.12 1.34 1.69 3.25 99.0903922683 Education 2.71 2.76 3.36 3.4 3.75 55.5860805861 Restaurants& hotels (including cheap takeaway services) 3.85 3.79 3.78 3.69 4.41 74.6957403651 Miscellaneous 5.05 5.36 5.6 6.33 8.42 85.4184641933 Source: Author’s calculation based on the 2015 HIECS data and CAPMAS CPI reports. 34 Egypt, the IMF and Three Subsidy Approaches

These datasets and calculations show that the cumulative inflation experience was quite regressive due to two main reasons. First is the proportional difference food consumption takes up of household income, since food has experienced a higher-than-average rate of inflation as discussed above. Second is because embodied energy constitutes a larger share of poorer households’ incomes due to their propensity to spend most or all of their income compared to higher income households who progressively tend to save more of their incomes. 5.1 Limitations Surely, this approach will be a rough estimation, mostly because it treats the embodied energy in all products equally. This poses two challenges: first, different income profiles differ significantly in their consumption patterns(i.e., low-income households always spend more of their income on food); if embodied energy for different goods and services categories is different, which is the case, it will mean that households will differ in their embodied energy expenditure based on what products they consume. Second, even within the same goods and services categories, it is believed that higher income households spend more money on energy intensive goods and services(i.e., imported products, private commutes, tourism, etc.). However, while this will overestimate embodied energy for wealthier households and underestimate it for poor ones affecting the argument of the paper, this impact will be limited for a number of reasons. First, products that vary significantly in their energy intensity(e.g., private versus public commute, tourism, etc.) take up small amounts of income. Moreover, food(including from restaurants), tobacco, clothing, housing, furniture and equipment, health, communications, and education should not vary in their energy intensity between different income profiles, and they constitute about 90 per cent of total household expenditure. This is partly because even the cheap products in these categories(cheap clothing, appliances, etc.) will be fully or partly imported. Only transportation and recreation might differ significantly and those constitute 4.9 per cent of the poorest household expenditure, and 11.47 per cent of the wealthiest. This method was also used for the lack of a better alternative to calculate such a crucial measure. An alternative input-output(IO) based methodology is marred by serious problems, such as the availability and usability of EEMRIO databases for different years 77 and the"coarse-grained" (i.e., simplified) sectoral resolution of EEMRIO models. 78 However, one future area in which the data can be fine-tuned to produce more tailored results is the use of EEMRIO to establish embodied energies of different goods and services, and hence for households based on their consumption patterns. 77 Coady, D., Flamini, V.,& Sears, L.(2015). The Unequal Benefits of Fuel Subsidies Revisited: Evidence for Developing Countries. IMF. Retrieved 30 December 2022, from https://www.imf.org/external/pubs/ft/wp/2015/wp15250.pdf 78 Ivanova& Wood.(2020) The unequal distribution of household carbon footprints. The Universal, the Self-Targeted, and the Targeted 35

6. Energy subsidies: busting the myths with evidence from Egypt Of all types of universal price subsidies, energy subsidies have received the lion’s share of accusations for being inequitable and inefficient. First, subsidising energy is seen as contributing to large fiscal deficits and diverting scarce resources from growth-promoting and poverty-reducing expenditures, such as education and health. Second, it is also often accused of creating an economic bias towards capital- and energy-intensive, and hence polluting, sectors of the economy – as well as reducing incentives for the adoption of energy-efficient and innovative technologies. 79 Third, it is identified as distortionary of markets and therefore optimal allocation of scarce resources. Since the 1970s, the IMF has always pushed for energy prices to be at par with international prices. However, Abdel-Khaliq calls for the"demystification" of international prices of oil; he argues that for a domestic economy, there is no reason to apply international oil prices as the norm because the domestic price of oil is a national parameter. It should"mesh in with other parameters" of the economy like per capita income, otherwise raising energy prices to international prices – meant as a stabilising mechanism – could in fact turn out to be a destabilising force. In other words, if the IMF never advocates for incomes in Global South economies to be raised to international averages, there is no good reason why energy prices should. Perhaps most importantly from a social protection viewpoint, it is seen as disproportionately benefiting richer households, since they consume more energy in absolute terms. 80 This reasoning is often substantiated by a few questionable presuppositions. First, the fact that energy subsidies benefit the rich more is true when considered in absolute or consumption terms not relative to income, and when it excludes embodied energy. For example, an IMF study on the distribution of fuel subsidy benefits uses consumption to measure welfare loss of different income quintiles. The study finds that"the richest 20 percent of households capture[...] more than six times more in fuel subsidies than the poorest 20 percent, making universal fuel subsidies a very inefficient policy instrument for protecting poor households from fuel price increases." 81 The World Bank 2022 Egypt Public Expenditure Review argues that"poverty targeted social assistance programs in Egypt are far more progressive in terms of their incidence to the poorest households than untargeted subsidies", referring mostly to energy subsidies as an example of the 79 Breisinger, C., Mukashov, A., Raouf, M.,& Wiebelt, M.(2019). Energy subsidy reform for growth and equity in Egypt: The approach matters. Energy Policy, 129, 661–671. https://doi.org/10.1016/j.enpol.2019.02.059 80 IMF.(2015, February 11). Arab Republic of Egypt: Staff report for the 2014 Article IV Consultation. IMF. Retrieved 20 December 2022, from https://www.imf.org/en/Publications/CR/Issues/2016/12/31/Arab-Republic-of-Egypt-StaffReport-for-the-2014-Article-IV-Consultation-42692 81 Coady et al.(2015) The Unequal Benefits of Fuel Subsidies Revisited 36 Egypt, the IMF and Three Subsidy Approaches

latter. 82 However, if embodied energy is integrated into the analysis, we get a different picture of the impact on household expenditure since it is about three to four times larger than direct energy consumption. 83 Once we consider expenditure relative to income and expenditure on embodied energy, the picture becomes a lot more complex and energy price subsidies can in fact start seeming progressive as we will discuss below. As for which income groups benefited the most from energy subsidies, this is what the section examines by demonstrating how much has expenditure on energy increased for the different households, the percentage it constitutes of their income, and whether targeted subsidies have adequately compensated them for such increase. Moreover, by counting energy inflation and general inflation, we can quantify the specific contribution of energy subsidy elimination on the erosion of purchasing power vis-a-vis other inflation drivers, such as VAT hikes and currency devaluation. As per the study’s calculations, household energy expenditure contributed to about 40 per cent of the increase in the cost of living between December 2015 and August 2019. For extremely poor households, the increase in energy expenditure constituted about 35.7 per cent of their 2015 incomes, whereas for the top income group it constituted about 21.5 per cent; this can be contrasted with the increase in absolute terms which was 13.5 times more for the higher-income group. Here even the OECD and the World Bank admit that although the rich receive most of the total value of energy subsidy,"impact of energy subsidy removal can be greatest for the poor". 84 This is clear evidence of the progressive potential of universal price subsidy when embodied energy and relative expenditure are taken into account. The progressivity also becomes clearer when energy expenditure is calculated relative to income. This is partly because as people move up the income ladder, they spend less of their income and save more. For example, based on the 2015 HIECS data, the saving factor for extremely-poor households was minus-nine per cent, meaning that they spend nine per cent more than what they earn. This rate incrementally increases until it reaches 35 per cent of the top income quintile, meaning that households in this category spend on average about two-thirds of what they earn. Since income gaps are therefore much larger than expenditure gaps, the progressivity becomes clearer when calculated relative to income(see Table 8). 82 World Bank,(2022). Public Expenditure Review 83 Song.(2019). Scale, distribution and variations of global greenhouse gas. 84 OECD& World Bank(2010) Analysis of the Scope of Energy Subsidy and Suggestions for the G-20 Initiative, OECD. org(Organisation for Economic Co-Operation and Development). Retrieved 11 March 2023, from https://www.oecd. org/env/45575666.pdf The Universal, the Self-Targeted, and the Targeted 37

Table 8. Increase of household expenditure on energy as a percentage of their income Household income profile Extreme poor Poor Near poor Middle-income High-income % of increase in energy expenditure to income 35.709159 30.926799 28.935202 26.443483 21.493307 General cumulative inflation 81.02% 80.69% 80.04% 79.37% 78.45% Source: Check methodology annex for data sources and information on calculations. Cash transfers far from compensate for the increase in the cost of living driven by elimination of energy subsidies, especially that they have meagrely increased since 2015. While for the extreme poor, cumulative general inflation was 81 per cent from December 2015 to August 2018, of which a large portion of it was caused by increase in energy prices as discussed above(Table 8), maximum allowed conditional cash transfers for a four-person household under the Takaful programme were only increased by 15.2 per cent(see Table 9). This is expected when we consider that only a small fraction of fiscal savings – realised by elimination of universal price subsidies – was used in targeted cash transfer programmes. Between 2014 and 2021, Egypt has saved about 4.3 per cent of GDP in energy subsidies. However, social insurance pensions(including Takaful and Karama cash transfer programmes) have only increased by 0.07 per cent of GDP, and currently stands at 0.3 per cent of GDP. This is less than half of the global average and the lower-middle-income country average(both at 0.9 per cent of GDP) and even significantly lower than the already low MENA average of 0.42 per cent. 85 Food subsidies, which have been invoked by the IMF as a social spending item that would benefit the most from elimination of energy subsidies, have also declined relative to GDP from 1.6 per cent to 1.3 per cent. The savings were mostly used for two purposes, that is, honouring the growing debt servicing bill – mostly driven by several IMF-mandated waves of increase in policy interest rates – and achieving a primary balance surplus. 85 World Bank,(2022). Public Expenditure Review 38 Egypt, the IMF and Three Subsidy Approaches

Table 9. Maximum allowed monthly transfers for a household of four by the Takaful programme(in EGP) Base transfer Transfer for two high school students Total 2015 325 200 525 2018 325 280 605 Rate of increase 0.0% 40% 15.2% Table 10. The evolution of expenditure on social protection and interest payments from FY 2013/14 to FY 2020/2021 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 Food subsidies (% of GDP) 1.61 1.59 1.60 1.32 1.76 1.60 1.37 1.31 Petroleum product subsidies (% of GDP) 5.72 3.00 1.91 3.19 2.65 1.56 0.32 0.30 Social insurance pensions(incl. Takaful and Karama) (% of GDP) 0.22 0.27 0.26 0.36 0.38 0.33 0.31 0.30 Interest payments (% of GDP) 7.85 7.80 9.11 8.78 9.58 9.79 9.67 8.92 Source: Author’s calculations from national budget statistics published by the Ministry of Finance. In order to counter inflation largely caused by the elimination of energy subsidies, the Central Bank of Egypt implemented a few waves of interest rate increases which had severe macroeconomic consequences, most prominently among which are lower employment rates and growing debt servicing bill. High interest rates also contribute to high inflation for low-income groups because it affects demand, and hence prices, for non-essential goods, whereas essential goods that take up most of low-income households’ expenditure remain vulnerable to inflation due to their low (and even negative) elasticity of demand. The Universal, the Self-Targeted, and the Targeted 39

Numerous studies find that a rise in real interest rates increases unemployment, including longterm unemployment, and disproportionately affects young people. 86 In Egypt, increases in interest-driven financialisation of the economy had severe repercussions for job creation. 87 Unlike interest rates, energy subsidies acted as a more progressive form of inflation control benefiting lower-income households more. Neoliberal orthodoxy only considers monetary measures to control inflation, namely the increase of interest rates as discussed above. However, in practice governments around the world often resort to fiscal measures to control inflation. For example, in 2020 VAT reduction in Germany led to lower inflation rates without affecting household purchase power and without over-burdening the national budget. 88 The policy led to an increase of EUR 34 billion of household spending while only resulting in about EUR 7 billion of lost budget revenues. This demonstrates that a relatively small increase in budgetary spending or decrease in revenue can have a much larger impact on household spending, therefore helping lower-income households as well as demand-driven growth. 89 Given the law of propensity to consume, any savings for low-income households will translate to increases in household consumption since a larger share of any additional increase to low incomes is spent on consumption, whereas a larger proportion of increases to high incomes goes into savings. In the end, the fact that different income groups use different fuel types means that the government could have easily only slashed subsidies for fuels used by higher income groups. Also, unlike food, there are less ethical implications in the self-targeting mechanisms of energy subsidies. Self-targeting in food products could mean that lower income groups consume less balanced diets, or consume larger portions of harmful nutrients. However, this is usually less of a concern when it comes to energy subsidies making self-targeting a progressive social protection option. Self-targeted energy subsidies could also promote the sustainable consumption of energy, by subsidising less-polluting energy types(e.g., LNG), or energy types used in more sustainable activities(e.g., public transportation). 86 For example, see Feldmann, H.(2013). Real interest rate and labor market performance around the world. Southern Economic Journal, 79(3), 659–679. http://www.jstor.org/stable/23809689 87 Diab, O.(2022). The role of subordinate financialisation in Egypt’s employment crisis. Review of African Political Economy, 1–9. https://doi.org/10.1080/03056244.2022.2151358 88 Nienaber, M.(2020, August 31). VAT cut pushes German inflation into negative territory. Reuters. Retrieved 25 December 2022, from https://www.reuters.com/article/uk-germany-economy-inflation-idUKKBN25R1MR 89 Look, C.(2021, November 2). German vat cut unleashed$39 billion spending spree, study finds. Bloomberg. com. Retrieved 25 December 2022, from https://www.bloomberg.com/news/articles/2021-11-02/german-vat-cutunleashed-39-billion-spending-spree-study-finds 40 Egypt, the IMF and Three Subsidy Approaches

7. Final remarks and policy implications This paper demonstrated that the elimination of energy subsidies has disproportionately affected lower-income households despite wealthier households spending more on energy in absolute terms. Also, the snowball effects of eliminating energy subsidies, such as increasing interest rates to control subsequent inflation, has also disproportionately impacted lower-income households. This has a number of policy implications that I will set out in the following paragraphs. Although the environmental impact of energy subsidies is a legitimate and necessary concern, this should not come at the expense of the poorest segments of society, and should not be used to greenwash anti-poor policies. The government should pursue self- and categorical-targeting in energy subsidies. Unlike selftargeted food subsidies, there are no ethical or health concerns associated with self-targeting energy subsidies. In fact, it could have beneficial health and environmental impacts if it was used for subsidising fuel used in public transportation. This is already the case in Egypt where vehicles used for collective transportation mostly run on diesel, while private cars in Egypt very rarely use diesel fuel. Realising this, the GOE already allocates a small amount of subsidy for diesel, but such a small amount far from prevented significant increases in diesel prices, even surpassing increases in gasoline products mostly used in private cars. In short, subsidising diesel, if done with the environment in mind, could have a significant impact on CO2 emissions as well as urban pollution. This, however, requires the significant improvement of the public transportation network to encourage and absorb those who leave behind their private cars. Supporting public transportation – in addition to having many health, urban and environmental benefits – should also have the advantage of reducing the current account deficit as it will require the importation of less refined petroleum products. Other than collective transportation, diesel in Egypt is also used in the agricultural sector meaning that its increased subsidy could contribute to a downward pressure on food prices. This will also support the purchasing power of lowerincome groups since food products take up most of their expenditure. When it comes to electricity, a progressive system based on level of consumption is already in place. However, recent increases reduced the degree of progressivity since lower brackets have increased at a faster rate since 2011. Electricity bills paid by low- and middle-income groups increased since 2011 by 218 per cent and 271 per cent respectively. 90 Studies show that increasing 90 Shawkat, Y.(2018). Electricity Facts. Marsad Omran(10 Tooba for Research). Retrieved from https://marsadomran. info/en/2017/11/1266/ The Universal, the Self-Targeted, and the Targeted 41

the progressivity could also benefit the environment and reduce national energy consumption as energy-intensive wealthy households would be encouraged to reduce their electricity consumption. 91 However, categorical targeting can also be utilised to target some of the poorest areas, especially in the countryside. Of course, some wealthier households in those areas would benefit from such targeting; however, the effectiveness and low-cost nature of this measure would outweigh such small leaks. These measures can be seen as a departure from neoliberal orthodoxy that only considers monetary measures with regressive effects to fight inflation. Lower inflation caused by such subsidies would reduce the need to raise interest rates; this will save the budget large sums that can even exceed what was spent on energy subsidies, given that currently interest payments are nine per cent of GDP whereas energy subsidies are at 0.3 per cent. This means that a one per cent decrease in interest payment would allow for quadrupling the amount of energy subsidy in an equitable and environmentally responsible manner. As for targeted cash transfers to play the role they are supposed to play, they must be indexed to inflation. With price levels more than doubling since 2016, this acutely and dangerously erodes the already small amount of cash transfers. Protecting against high inflation will provide yet another argument for price subsidies vis-a-vis cash transfers. In order for cash transfers to work in a Global South high inflation context, a simple issue such as inflation indexation cannot just be brushed aside by IFIs and the GOE. Additionally, social protection must go beyond social assistance and also include its other components, that is, health and education spending which is rhetorically allowed for – and even encouraged – in the IMF Social Spending Strategy. In addition to all the well-established human development benefits accruing from increased education and health spending, public investments in these sectors would have numerous macroeconomic benefits. First, it will have a good impact on the severely affected employment rate as both sectors are among the most labour intensive. Egypt’s labour force participation rate has severely dropped from 47 per cent of the population in 2016 to 41 per cent in 2021 92 as a result of IMF-mandated policies, namely severe austerity, erosion of household purchasing power and tight monetary policy. The jobs created in these sectors will also be gender balanced as both sectors are very female intensive(see Table 11). This is of great importance given that the female labour force participation rate has collapsed even more aggressively since the start of the economic reform programme in 91 Dehmel, C.(2011). Progressive electricity tariffs in Italy and California – prospects and limitations on electricity savings of domestic customers. European Council for an Energy Efficient Economy. Retrieved from https://www.infosperber. ch/wp-content/uploads/2022/09/STROM-Progr.-Tarife-Dehmel.pdf 92 World Bank Data.(2022). Labor force participation rate, total(% of total population ages 15+)(modelled ILO estimate) - Egypt, Arab Rep. World Bank Data. Retrieved 31 December 2022, from https://data.worldbank.org/indicator/SL.TLF. CACT.ZS?locations=EG 42 Egypt, the IMF and Three Subsidy Approaches

2016, from about 23 per cent in 2016 to 15 per cent in 2021. 93 As for younger women, the percentage drops to nine per cent, which confirms the novelty of the crisis and that the future would not hold good news on that front without a serious policy intervention. In addition to being extremely labour intensive, investment in health and education would increase households’ purchasing power as Egypt has one of the highest out-of-pocket expenditures on health and education in the world due to the government’s strategy of self-targeting in these social sectors as discussed above. Table 11. Ratio of female workers in key selected economic sectors Sector Agriculture Mining Manufacturing Construction Wholesale and retail Transportation and storage Education Health and social work Male 18,83,400 36,300 24,03,500 27,04,400 13,75,700 12,69,900 9,77,300 3,04,600 Female 6,76,400 1,100 2,56,600 22,200 1,36,700 20,500 11,36,300 4,53,400 Total workers 25,59,800 37,400 26,60,100 27,26,600 15,12,400 12,90,400 21,13,600 7,58,000 Female workers share of total workers(%) 26.4 2.9 9.6 0.81 9 1.6 53.7 59.8 Source: Diab, O.,& Hindy, S. I.(2022). Women and economic reform in Egypt: Impact of production changes on female waged labor force participation. Middle East Critique, 31(1), 61–79. https://doi.org/10. 1080/19436149.2022.2030984 Finally, it needs to be noted that an effective social protection policy must abandon such a dichotomous"either-or approach" between targeted and universal social protection. Instead, the government should adopt an integrated system of social protection in which targeting is used as an additional mechanism within a broader universalist framework. This is especially the case since there is strong evidence that"universalist countries tend to have better redistributional outcomes", and that"targeting[...] seems to work best when embedded within the overall framework of universalism". 94 Gugushvili and Laenen argue that the balance between universalism and targeting will vary across countries, but it is important that selective policies are not too 93 World Bank Data.(2022). Labor force participation rate, female. 94 Gugushvili, D., and Laenen, T.(2021),‘Two Decades after Korpi and Palme’s"Paradox of Redistribution": What Have We Learned so Far and Where Do We Take It from Here?’, Journal of International and Comparative Social Policy, Vol. 37 No. 2, pp. 112–127. https://doi.org/10.1017/ics.2020.24 The Universal, the Self-Targeted, and the Targeted 43

narrowly targeted at the very poor because, as we discussed above, this seems to"paradoxically" increase levels of poverty. Egypt has also become a case in point of such dynamic with poverty rates increasing the more that social protection shifts to targeting. Integrated systems are more capable of preventing people from falling into poverty rather than only intervening once they have. Shahra Razavi of the ILO argues that comprehensive systems aim at preventing poverty in the first place and are thus better placed to respond to the dynamic nature of poverty. For Razavi, tax-financed programmes and social insurance should be fully embedded in national social protection policies or strategies to avoid the fragmented delivery of social protection. 95 Research by CAPMAS also shows that in 2017/18 if butane gas subsidy(a universal/self-targeted subsidy) was removed, poverty will increase by a staggering five per cent of the population, that is, more than five million people. This is mostly because of the large percentage of the population living just above the poverty line, comprising households that are extremely sensitive to even the slightest increases in living costs. This simply means that universal types of subsidies are crucial for poverty fighting through prevention. 95 Razavi, S.(2022). Building universal social protection systems for all: What role for targeting? Retrieved 15 March 2023, from https://socialprotection.org/discover/blog/building-universal-social-protection-systems-all-what-role-targeting-0 44 Egypt, the IMF and Three Subsidy Approaches

Publisher: Friedrich-Ebert-Stiftung Regional Office MENA Economic Policies for Social Justice 4, rue Bachar Ibn Bord 2078 La Marsa Tunisia Email: info.mena@fes.de Responsible for content: Thomas Claes| Director| Economic Policies for Social Justice https://mena.fes.de/themes/economy-and-social-justice Contact/Orders: info.mena@fes.de Design: Marc Rechdane The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung. Commercial use of media published by the Friedrich-Ebert-Stiftung(FES) is not permitted without the written consent of the FES. Publications by the Friedrich-Ebert-Stiftung may not be used for electioneering purposes. Attribution 4.0 International Licensed as: