84516 FEBRUARY 2014 • Number 134 Reforming Subsidies in Morocco Paolo Verme, Khalid El-Massnaoui, and Abdelkrim Araar The cost of the subsidy system in Morocco peaked at 6.6 percent of gross domestic product (GDP) in 2012, an amount larger than the country’s total investment budget for that year. Direct subsidies to households in 2013 (October 2013 prices) are estimated at DH34.4 billion (US$4.1 billion, or 3.9 percent of GDP). Replacing direct subsidies with a universal per capita annual cash transfer of DH749 (US$90) would leave the poverty rate un- changed, while reducing direct subsidies to DH24.6 billion (2.8 percent of GDP). It is possible to reduce direct subsidies further by targeting only a part of the population, down to DH2.4 billion (0.3 percent of GDP), if only the poor were targeted. However, to implement selective targeting, the government of Morocco would need to substantially strengthen its social protection system. Subsidies in Morocco high oil prices throughout the 1990s and was eventually sus- pended in September 2000, paving the way for the rise in sub- The subsidy system in Morocco dates back to the early 1940s. sidies that followed. The first product to be subsidized was Liquefied Petroleum Today, Moroccan subsidies are costly, inefficient, and are Gas (LPG), followed by sugar. For several decades, these two putting the medium-term sustainability of public finances at products were the only subsidized products in the country. risk. Subsidies have reached unprecedented levels since 2008, The range of subsidized products started to expand in the peaking at 6.6 percent of GDP in 2012. For the first time, 1960s, with the addition of cereals (1966), milk (1973), and subsidy outlays became higher than capital expenditures, ex- cooking oil (1974). Prices for milk and cooking oil were later plaining most of the rise in budget deficits over the last few liberalized in 1983 and 2000, respectively, while LPG contin- years. Worse, the fiscal space achieved through sound macro- ued to be subsidized by means of an equalization system, economic management during the previous decade has now whereby a special tax on liquid petroleum products (gasoline totally eroded. As a result, Morocco’s public debt has in- and diesel) helped stabilize LPG prices. Subsidies on liquid creased by more than 12 percentage points of GDP since petroleum products have been introduced more recently, in 2008, reaching close to 60 percent of GDP in 2012. In addi- conjunction with the global rise of oil prices in the early tion, subsidies are putting the sustainability of the external 1990s. The equalization system was abandoned in January accounts at risk by increasing and distorting domestic de- 1995, and a price indexation system for liquid petroleum mand for fuels, given that the country is totally dependent on products was established to allow domestic prices to be ad- imports. After generating surpluses over 2001–6, the current justed to reflect changes in world prices. This new system be- accounts have suffered increasing deficits since 2008, peaking came politically difficult to sustain given the persistence of at 9.7 percent of GDP in 2012. The deterioration of the cur- 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise rent accounts has taken its toll on international reserves, they account for 75 percent of the nonsubsidized price. Sugar which declined from an equivalent of eight months’ worth of and the low-quality flour (national) are sold at about half their imports of goods and services in 2007 to less than four free market price, while gasoline and better-quality flour (free) months in 2012. are sold at a price that is close to the free market price. There- The current price subsidy system in Morocco is universal fore, the relative importance of unit subsidies varies signifi- and includes basic food products (sugar and flour) and petro- cantly across products, and this plays an important role in un- leum products (gasoline, diesel, fuel-oil, and LPG). Its stated derstanding the distribution of subsidies. objectives are to ensure the price stability of targeted prod- The distribution of subsidies across households depends ucts, preserve the purchasing power of consumers, and help on the unit subsidies and also on the relative consumption of promote “promising” productive sectors to increase their con- subsidized products of each household. This analysis consid- tribution to the economy. In addition, given its welfare fea- ers only the direct effects of subsidies on households. Subsi- ture that helps control the prices of basic commodities (main- dies also affect consumers indirectly—via the increase in pric- ly food and LPG), the subsidy system constitutes a de facto es of nonsubsidized products due to the increase in prices of social protection mechanism for low-income households. subsidized products—and producers. Subsidies also affect the The government of Morocco has publicly stated on sev- functioning of various markets, including the labor market. eral occasions its intention to radically reform subsidies by These effects can be very large of course, and should be ana- reducing them significantly or eliminating them altogether. lyzed within a general equilibrium framework, but in this The analysis that follows looks at the distribution of subsidies note, analysis is limited to direct effects using the latest house- and the potential impact that total elimination may have on hold consumption survey available in Morocco—the 2007 household welfare and on government revenues. This exami- Living Standards Survey (LSS).1 The survey information was nation will provide some basic information that policy mak- updated to 2013 using macroeconomic, population and price ers can use to better understand the implications of subsidy statistics, so that all estimates presented here are for 2013, us- reforms. Finally, this note discusses recent reforms and some ing October 2013 prices. of the key issues that complicate subsidy reforms. Table 2 shows that household expenditures on subsi- dized products are large, but also that subsidies associated Who Benefits from Direct Subsidies? with these products are almost as large. Moroccan households Table 1 reports the official subsidized prices and estimates of spend about DH38.9 million on subsidized products, about unit subsidies and the unsubsidized prices published by the 7.7 percent of their total annual expenditures. Flour is the government of Morocco for all subsidized products as of Octo- most consumed item in terms of expenditure (DH11.7 mil- ber 2013. The list of products includes LPG, gasoline, diesel, lion) followed by LPG (DH10.9 million). Subsidies on the sugar, and flour. Subsidies are particularly large for LPG, where products considered totaled DH34.4 million, an amount al- most as large as total household expenditures on subsidized products. In a nutshell, it is as if households are currently only Table 1. Prices of Subsidized Products (October 2013, Moroccan DH) paying—on average—half of what they should pay if Subsidized subsidized products were freely exchanged on the price as % market. Subsidized Unit Unsubsidized of unsubsidized Subsidized products are relatively more im- Product Unit price subsidy price price portant for the poor, as shown in figure 1 (left LPG Liter 3.50 6.50 10.00 35.0 panel), where it is evident that the expenditure Gasoline Liter 12.77 0.80 13.57 94.1 share of subsidized products is much higher for the very poor compared to the very rich. In the Diesel Liter 8.84 2.60 11.44 77.3 poorest households, more than 15 percent of their Sugar, in total expenditure is on flour, about 10 percent on Kg 6.00 5.00 11.00 54.5 pieces LPG, and about 8 percent on sugar. Using the aver- Sugar, in age for the poorest 20 percent of the population, Kg 6.00 5.00 11.00 54.5 cubes results show that this group spends, on all subsi- Sugar, in dized products combined, 13 percent of its total Kg 5.00 5.00 10.00 50.0 granules expenditures compared to an average of only 4.3 Flour, free Kg 5.00 0.70 5.70 87.7 percent for the richest 20 percent of the popula- Flour, national Kg 2.00 1.90 3.90 51.3 tion. The only exceptions are gasoline and diesel, which are irrelevant items for the very poor. Inter- Source: Authors’ compilation. estingly, the rich spend between 2 and 3 percent of 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise Table 2. Household Expenditures on Subsidized Products and Total Subsidies (DH millions) Total Total Quintile LPG Gasoline & diesel Sugar Flour expenditures subsidies 1 1,416.4 76.4 1,108.8 1,721.4 4,323.1 4,952.3 2 1,702.8 312.9 1,343.7 2,249.0 5,608.3 5,789.3 3 2,126.3 558.1 1,489.2 2,402.0 6,575.5 6,622.7 4 2,353.4 1,479.8 1,648.4 2,822.1 8,303.7 7,323.6 5 3,273.4 6,446.5 1,790.4 2,545.1 14,055.4 9,693.6 Total 10,872.3 8,873.6 7,380.5 11,739.5 38,866.0 34,381.6 Source: Authors’ compilation from 2007 LSS (updated to 2013). their total expenditures on subsidized items across all subsi- ents a simulation that removes all subsidies and then shows dized products. prospective impacts of this removal on household welfare, However, subsidies are clearly pro-rich in per capita abso- poverty, and government budget. Potential impacts of com- lute terms. Figure 1 (right panel) shows that LPG, flour, gaso- pensatory cash transfers on poverty and government revenues line, and diesel are all pro-rich, with subsidies per capita high- are also considered. er for the rich than for the poor. For example, the richest The total loss in welfare and the total gain for the govern- households receive more than four times what the poorest ment can be estimated according to the total amount of subsi- households receive in per capita benefits, and the poorest dies, which amount to DH34.4 million or 6.9 percent of total quintile of the population receive on average only a little more household expenditure. than half of what the richest quintile receive. The only excep- The impacts of the elimination of susbsidies on poverty tion is flour, which benefits mostly the middle class. It is clear and inequality is also very significant. Estimations show that— that the greater share of public expenditures on subsidies do with a poverty line of DH4,318 per person per year (US$526 not accrue to the poor, but to the rich. or US$1.44/day)—the poverty rate would increase from 5.5 to 9.9 percent, while the Gini inequality measure would in- Impacts of Subsidy Reforms crease from 40.7 to 42.8 percent. These are very significant The incidence analysis presented in the previous section changes. Figure 2 (left panel) shows the impact of price in- shows that subsidized products are important for the poor, creases of subsidized products on poverty at between 0 and but mostly benefit the rich. Hence, a sensible reform would 100 percent. Price rises in LPG, bread, and flour would in- remove the subsidies from the rich and keep subsidies for the crease poverty significantly, with flour being the most pover- poor. Unfortunately, this is difficult to do because subsidies ty-increasing product. Only increases in the price of gasoline are not amenable to selective targeting. A more viable alterna- and diesel have very small direct effects on poverty, for the tive would be to eliminate subsidies altogether and compen- simple reason that poor people do not typically own vehicles sate the poor with a targeted cash transfer. This section pres- in Morocco. Figure 1. Expenditure Shares on Subsidized Products (left) and Subsidies per Capita (right) .2 1,000 LPG LPG gasoline and diesel gasoline and diesel sugar sugar 800 flour total benefits per capita .15 flour expenditure shares 600 .1 400 .05 200 0 0 0 .18 .36 .54 .72 .9 0 .18 .36 .54 .72 .9 percentiles (p) percentiles (p) Source: Authors’ illustration. 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise The impacts on government revenues are also significant, The x-axis plots the level of individual transfers, while the y- but quite different across products. For example, figure 2 axis plots the headcount poverty index (left panel) and gov- (right panel) shows that LPG has the best potential to increase ernment revenues (right panel). The left panel shows that re- government revenues, while gasoline and diesel have less po- placing direct subsidies with a universal cash transfer of tential to increase revenues because prices are increased in per- DH749 (US$90.40) would maintain the poverty rate, while centage terms. Of course, equal increases in percentage terms reducing direct subsidies to DH24.6 billion (US$3 billion or correspond to different price increases across products, and 2.8 percent of GDP). If only the poor were targeted, this cost certain products are closer to the market price than others, so would decrease further to DH2.4 billion (US$0.3 billion or not all products can be increased by the amounts shown (0– 0.3 percent of GDP). That is to say, providing a flat transfer to 100 percent). However, figure 2 shows how impacts on pover- all would bring the poverty level back to its pre-reform level ty and government revenues are product specific, as well as and be more efficient and less costly than providing a univer- possible trade-offs when it comes to reforming subsidies: for sal subsidy, and targeting only the poor and vulnerable would example, flour is the most poverty-increasing product, yet con- significantly reduce public cost even more. tributes relatively little to government revenues compared to Reforming Subsidies under Budget LPG for the same percentage increases in prices. Constraints Finally, figure 3 shows the impacts on poverty and gov- ernment revenues of a cash transfer designed to compensate Reforming the subsidy system has become a high priority the poor for the negative consequences of subsidy removal. for the Moroccan government so that it can insulate public Figure 2. The Impact of Price Changes on Poverty (left) and Government Revenues (right) 1.5 LPG LPG 1.00e+10 impact on governement revenue gasoline and diesel gasoline and diesel sugar sugar impact on poverty headcount flour 8.00e+09 flour 1 6.00e+09 4.00e+09 .5 2.00e+09 0 0 0 20 40 60 80 100 0 20 40 60 80 100 increase in prices (%) increase in prices (%) Source: Authors’ illustration. Figure 3. The Impact of Cash Transfers on Poverty and Government Revenues .1 initial level of poverty headcount ratio 3.50e+10 (.05592) change in government revenue .09 required transfer to offset the 3.00e+10 change in poverty (748.909) .08 2.50e+10 headcount 2.00e+10 .07 1.50e+10 .06 1.00e+10 .05 0 200 400 600 800 0 200 400 600 800 level of individual transfer level of individual transfer Source: Authors’ illustration. 4 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise finances from the negative effects of subsidized products’ Morocco has such a system in place. Targeting the poor and world price volatility while also generating fiscal space and vulnerable with direct cash transfers implies an accurate reinforcing social programs that target the poor and vul- knowledge of who is poor and where the poor are located, and nerable. collecting this information requires a complex nationwide ex- In 2013, the Moroccan government started to reform the ercise that Morocco has not yet implemented. subsidy system in line with the commitment to decrease the Moreover, food and commodity prices have been on the budget deficit to the sustainable target of 3 percent of GDP rise for over a decade and the expectation is that this trend over the medium term. In addition to measures aimed at con- will continue for the foreseeable future. This is a fact that trolling the wage bill and other recurrent expenditures, and to puts the government budget and households simultaneously rationalize capital spending, the government adopted a Head under pressure, leaving the government little space for ma- of Government Order (HGO) in the summer of 2013 allow- neuvering, especially during this period of fragile political re- ing for a partial price indexation of selected domestic fuel forms. These price increases already affect the poor and vul- products to changes in world fuel prices. Therefore, the gov- nerable via nonsubsidized products and highlight the need ernment reintroduced the price indexation mechanism that for a stronger social protection system. Hence, one way for- operated between 1995 and 2000, although the new system ward would be to focus on increasing the efficiency of public differs from the previous one in that it is limited to a few spending by removing subsidies and use the savings to rein- products and uses only a partial indexation mechanism. The force the social protection system to protect the poor and vul- budget support for these subsidies is defined by the budget nerable not just from the effects of subsidy reforms, but also law. This is expected to be the first phase of a wider subsidy from the wider effects of price increases. This requires a sub- reform that could eventually extend the indexation system to stantial reform of the social protection system. other subsidized products, including food and LPG, while es- About the Authors tablishing a compensatory transfer system targeting poor and vulnerable families. Paolo Verme and Khalid El-Massnaoui are Senior Economists in This first reform is expected to deliver some positive re- the Poverty Reduction and Economic Management (PREM) Net- sults on the budget side, but remains short of a full subsidy work, Middle East and North Africa Region of the World Bank. system reform. The estimations provided in this note show Abdelkrim Araar is a Researcher at the University of Laval and that the current subsidy system is vital for the poor, but clear- Senior Consultant for the World Bank. ly pro-rich and wasteful. A more efficient instrument to pro- tect the poor and the vulnerable would be to completely elim- Note inate subsidies and compensate the poor and vulnerable with 1. www.hcp.ma. a direct cash transfer. Estimates show that this approach would cost only a small fraction of current expenditures on Reference subsidies. However, such reform remains politically difficult Araar, A., and P. Verme. 2012. “Reforming Subsidies: A Toolkit for to implement and ultimately relies on an efficient system of Policy Simulations.” World Bank Policy Research Working Paper targeting the poor and vulnerable, and it is unclear whether No. 6148, Washington, DC. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. They are produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at: www.worldbank.org/economicpremise. 5 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise