Impact of the Covid-19 Pandemic on Monetary Child Poverty in Morocco

As we stated, the pandemic and measures taken by the government inside and outside the country generated a strong contraction of GDP (-6.7%) (Ministre de l’économie et des Finances-MEF, 2021). In terms of inflation, for the month of March, the HCP released data with the CPI rising by 0.4% during the month of March and noting an increase in the CPI of 1.5% over 12 months. Breaking down the index, the HCP noted that the food index rose 1.3 percent while the non-food index fell 0.2 percent. The largest price increases were for vegetables (6.3%), fish and seafood (1.9%), and fruit (1.7%). (CDG, 2020). By the end of the year, a 0.7% increase over 2019 was observed. As a result of the increase in the index of food products by 0.9% and that of non-food products by 0.5%. The variations recorded for non-food products range from a decrease of 1.6% for "Transport" to an increase of 2.7% for "Education" (HCP, 2021). In the first months of the pandemic (April and May 2020), the biggest price drop was in fuel prices, which fell by 6.8%.

For public finances, according to (Ministre de l’économie et des Finances-MEF, 2021) the state budget deficit increased from 3.6% of GDP in 2019 to 7.7% and the treasury debt increased from 64.9% of GDP in 2019 to 77.6% and the unemployment rate rose from 9,2% to 11,9% for 2020 following the pandemic.

At the household level in Morocco, with the confinement of the population and the total or partial cessation of several economic activities, both in the formal and informal sectors, many workers suddenly lost their jobs, whether permanent or seasonal, and have therefore temporarily or permanently lost their main source of income. On April 22, after one month of confinement, the HCP published the main results of its qualitative survey on the effect of the economic situation and COVID-19 on business activity (Haut-Commissariat au Plan, 2020a).⁴

The survey showed that 3% of the companies in the sectors surveyed had stopped their activities permanently and 54% have stopped their activities temporarily. The most affected sectors are accommodation and food services, with 89% of companies in this sector having to stop their activities either permanently or temporarily. The industrial sector has also had a marked impact with permanent or temporary closures ranging from 55% in the chemical and parachemical industries to 76% in the textile and leather industries (Table 1).

On the other hand, 43% of the companies surveyed remained active since the beginning of the crisis. Of these, 81% halved their activity and nearly 50% are very small, small, and medium-sized enterprises. The construction and service sectors are the ones that have slowed down the most (62% and 54% respectively).

The Confédération générale des entreprises du Maroc (Confédération générale des entreprises du Maroc-CGEM, 2020) in its preliminary report published on April 24^th based on a survey found that nearly two-thirds of respondents have decreased their activity by at least 30% and that only 7.5% have maintained a stable activity following the appearance of the virus in Morocco.

This slowdown in economic activity has had a de facto impact on employment. The International Labour Office (ILO) estimates that nearly 230 million workers have lost their jobs worldwide, including 30 million in sub-Saharan Africa and the Arab States. Moreover, the ILO estimates that 26.4% of the jobs lost are in high-risk sectors in Africa and 33.2% in the Arab States. Finally, the analysis shows that almost three-quarters of non-agricultural jobs are in the informal sector on the African continent. For the Arab States, this figure is 64%. (International Labour, 2020).

In Morocco, in April 2020, the HCP estimated that 20% or nearly 726,000 jobs (excluding agriculture and the financial sector) would be lost, the majority (57%) in very small, small, and medium-sized enterprises (Haut-Commissariat au Plan, 2020a). For the CGEM, more than 55% of the jobs in companies that participated in the survey are at risk and for nearly 50% of them, it was anticipated that the pandemic could have an impact on more than 75% of their jobs (Confédération générale des entreprises du Maroc-CGEM, 2020). In addition, among the sectors that experienced the greatest slowdown in March 2020, the tourism sector suffered the greatest drop in the number of jobs (78.24% job loss in the sector) followed by the textile (70.39%) and real estate (70.31%) sectors. The primary sector (agriculture, poultry farming, fishing), despite a slowdown in its activity (between 30 and 40%), seemed to be less affected in terms of job losses.

In the summary note dated May 19, 2020, the HCP publishes the preliminary results of its household survey, conducted between April 14 and 23, 2020 on some 2,350 households representative of the socio-economic classes of the Moroccan population and living in urban and rural areas (Haut-Commissariat au Plan, 2020c). The survey found that 34% of households reported having lost all sources of income during the initial stages of the lockdown period. Almost half of the households have at least one member who has been forced to stop working, and 40% of them have received assistance from either their employer or the government.

Moreover, in many countries, including those in the MENA region, remittances from abroad constitute a significant source of household income. These remittances have a positive impact on household welfare, both in monetary (poverty reduction) and non-monetary (health, nutrition, education) terms. However, the pandemic also had an impact on the host countries of many Moroccans living abroad (European countries, Gulf countries, North America, etc.). In the early stages of the pandemic and lockdowns, the World Bank anticipated a decline in remittances of around 20% globally, with a similar percentage in MENA countries (Bank, 2020). In the case of Morocco, Iñigo Moré estimates that remittances from the European Union fell by 30% between February and March 2020 and that this drop could be greater than 50% in April (Azougah, 2020).⁵ The World Bank estimates that in 2019, Moroccan households received 66.75 billion dirhams (US$6.7 billion) and 64.78 billion according to the Moroccan Exchange Office (Figure 1). In addition, it is estimated that 20 to 30% of Moroccan households receive remittances from migrants living abroad.

Figure 1

Download asset Open asset

In the early stages of the pandemic, many organizations presented reports on the impacts that COVID-19 could have on the poorest populations. Unanimously, these reports pointed to a significant setback in the progress that had been made on poverty reduction objectives. As many households are dependent on a single member for their income. The lockdown impacted directly the economic activity in certain sectors likely pushing many of these households into poverty.

For example, in March 2020, (Vos et al., 2020) use a global CGE model in combination with a micro-simulation model to estimate that extreme poverty could increase by 2 to 3 percentage points in developing countries. Thus, according to their first estimates, between 14 and 22 million people could fall into extreme poverty and nearly half of them would be from sub-Saharan Africa. One month later, the same authors conclude that 140 million people are at risk of falling into extreme poverty, including 80 million on the African continent (Laborde et al., 2020). This change in the authors' estimates is due in large part to the lockdown enacted in many countries that have resulted in a sharp slowdown in economic activity in some sectors, and certain cases a complete halt in activities. From the detailed early data,⁶ it was estimated by Haut-Commissariat au Plan (2020a) in the early stages of the pandemic that 30,000 people could fall into extreme poverty in Morocco, an increase of 8.5%.⁷

Gerzon Mahler et al. (2020), on their part, apply a nowcasting approach using household dataset and growth forecasts and estimate that between 40 and 60 million people, including 2.8 million in MENA countries, could be pushed into extreme poverty, depending on the growth assumptions used, resulting in a reversal of the gains of at least three years (Figure 2). In addition, beyond the loss of the main income in poor and vulnerable households, the sanitary measures put in place to limit the spread of the virus could also increase inequalities, especially in the urban area (United Nations Development, 2020).

Figure 2

Download asset Open asset

Gerzon Mahler et al. (2021) using the same approach (nowcasting) provide an estimate of the impact of COVID-19 on global poverty. They perform a nowcasting approach for their analysis using high-frequency phone surveys. They first estimated, in January 2021, that between 119 and 124 million people were pushed into extreme poverty globally. By using more recent data, they revised this number down to 97 million by June 2021. Albeit, this number is reduced, according to the authors, this is an unprecedented increase in extreme poverty.

Summner et al. (2020) also use nowcasting using the PovcalNet database and growth estimates. They simulate some scenarios to measure the short-term impact of Covid. They estimate that with a 20 percent income or consumption contraction in the world, the number of people living in poverty could increase by 420-580 million, relative to the latest official recorded figures for 2018.

In this paper, for the micro-simulation model part, we use data from the 2019 wave of the Enquête panel des ménages (EPM) of the Observatoire national du développement humain (ONDH). This survey was first implemented in Morocco in 2012. It consists of annually following the same households and members of the sampled households visited in the reference year (2012), the first wave of the panel.

In 2017 and 2019, ONDH sought to ensure regional representativeness (the 12 regions of the country in addition to national representativeness and by area of residence). In 2019, 16,879 households were surveyed.

In 2019, the questionnaire for this survey has 17 components and provides information on the main dimensions of human development such as demography, education, health and access to basic services, income, consumption expenditure, detailed food expenditure, detailed non-food expenditure of each household (in volume and value) as well as all transfers paid and/or received by them. It also covers all relevant information on household members (socio-demographic characteristics, literacy, education, employment, unemployment, economic activity, health insurance coverage, morbidity and access to health care, under-five mortality, borrowing, sources of income, housing and living conditions, etc.).

The wealth of information contained in this survey, and to which we had access, allows us to conduct the analysis for this paper according to the methodology chosen.⁸ Table 2 presents the distribution of the EPM sample of the ONDH (wave 2019) at the household and individual levels. Table 3 shows the distribution of the Moroccan population in 2019.⁹

In this paper, the model is broken down into three interdependent modules. The first module is a macro-sectoral model that allows to simulate variations in the prices of goods following different shocks (price block). The other two modules are directly linked to the EPM survey used. One deals with household expenditure on different products (expenditure block). The last block is decomposed into two parts related to the characteristics of individuals in the labor market and their sources of income. From the results generated by these three modules, we proceed to apply a detailed distributional analysis. In the following subsections, we briefly present the functioning of these three modules.

3.2.4. Distributional analysis

To begin our distributional analysis, we determined the average annual expenditure per capita for each household for the reference year, 2019. Then, to perform our analysis at the individual level on the one hand and at the level of children on the other, we retained a partition of the population into five classes (poor, vulnerable, low average (middle class), high average (middle class), wealthy) based on the average annual expenditure per capita in 2019. Thus, to ensure comparability with the other approaches by official national institutions, concerning the middle class (HCP approach using ONDH data), we have retained the following definitions:

• The poor class is composed of households (individuals) that have a per capita expenditure below the national poverty line differentiated according to the area of residence (identified with Z_poverty) ;

• The vulnerable class is composed of households (individuals) with a per capita expenditure above the poverty line (Z_poverty) but below the vulnerability threshold (Z_{vulnerability}) associated and differentiated according to the area of residence. In our case, we use the recommended threshold of the World Bank, and hence, we use a vulnerability threshold equal to the poverty line plus 50% (Z_{vulnerability} = 1,5x Z_poverty);¹²

• The other three classes are lower middle, upper middle, and wealthy.¹³

This partition was repeated on the new vectors of average annual expenditure per capita adjusted after each scenario and for each area, with poverty lines (Z_poverty) and (Z_{vulnerability}) constant and variable medians.¹⁴

To derive the vectors of average annual household expenditure under each of the three scenarios implemented, we applied the growth rates calculated on incomes at the end of the income module considering the following assumptions:

• The rates of income change associated with the different simulations described in Figure 3 are fully reflected in the expenditures of poor, vulnerable, and other households in the database (in 2019). The assumption made here is therefore simply that the marginal savings of households in these three classes are zero;

• We transmitted 50% of income variation in simulations (1 to 3) on the spending of the non-poor, vulnerable, and lower middle class. Our assumption implies that the expenditures of these households are not fully transferred given that they drew from their accumulated savings;

• The expenditures of wealthy households are not affected at all, regardless of the change in their income. The hypothesis adopted here is that these households have sufficient savings or other sources of financing to absorb the shock and keep their consumption expenditure constant, at least in the short term.¹⁵

Figure 3

Download asset Open asset

Based on these assumptions, vectors of average annual (nominal) expenditure per household and per capita are derived. However, they cannot be used in this form to conduct poverty or distributional analyses, since they do not take into account price effects.

In our analysis, the adjustments were considered and applied according to the approach that relies primarily on the concept of equivalent income of (King, 1983).¹⁶ The resulting per capita expenditure is used to conduct poverty and vulnerability analyses with constant thresholds and different distributional analyses.

Indeed, based on these variables, all the usual measures of income poverty of the FGT (Foster et al., 1984) and inequality (Gini) are calculated, at the national level. Decompositions of these same measures, according to the place of residence in particular, or to bring out child poverty more specifically, have been produced. Incidence curves have also been constructed for cases taken in pairs and by area. These curves make it possible to highlight the winners and losers from one scenario to another.

In this section, we present in detail the approach used to construct and implement the three simulations, which are broken down into a base case that we will call COVID, in which we simulate the impact of the crisis. Scenario 1 consists of applying the measures adopted by the government to mitigate the impact of the pandemic on households. Finally, scenario 2 consists of simulating complementary measures to further reduce the impact of the shock in addition to the measures put in place by the government. The three simulations are interrelated and complement each other sequentially (Cf. Figure 4).

Figure 4

Download asset Open asset

To determine the impact of COVID-19 on Moroccan households (i.e., the base scenario applied to the micro-simulation model), we decomposed the impact into two components:

• The impact on the prices of the household consumption basket and

• The impact on revenues.

At the individual level, the average annual per capita expenditure of Moroccans is 18,769.68 MAD. As in the case of households, individuals living in urban areas have a higher average annual expenditure per capita (22,724.15 MAD) while individuals in rural areas have an average annual expenditure per capita more than 36% lower than the national average (12,069.16 MAD) (Table 4).

To compute monetary measures of poverty, as is standard practice at the official level in Morocco, we used separate absolute poverty lines for urban and rural areas. We thus adjusted the official HCP thresholds used in the Enquête nationale sur la consommation et les dépenses des ménages (ENCDM) of 2013-2014 (Haut Commissariat au Plan, et Banque Mondiale, 2017) by applying the annual change in the consumer price index produced by the HCP between 2014 and 2019. Thus, the adjusted poverty lines used in this report are respectively 4,939.10 MAD per capita per year in the urban area and 4,563.40 MAD per capita per year in the rural area.

The Moroccan population is predominantly urban, with almost 2/3 of the population living in cities (62.89%). In 2019, according to data from the ONDH survey and the HCP’s official basic thresholds, Morocco has 1.19% of poor individuals, divided between 13.06% in the urban area and 86.94% in rural area, where the incidence of poverty is estimated at 2.79%. In terms of depth (FGT1) and severity (FGT2), rural areas are also more affected than urban areas Table 5.

Since our objective is to approximate the impact of the pandemic on households and their children, we examined child poverty by distinguishing three groups: children under 5 years of age, those between 5 and 17 years of age, and individuals 18 years and older. From the data, 8% of the Moroccan population are children under the age of 5 and 23.2% are individuals between 5 and 17 years old. Thus, nearly one-third of the Kingdom’s population is under 18 years old.

In 2019, more than one million children under the age of 18 were poor in Morocco (1.03 million) and 24% of these poor children were under 5 years old (250,000). At the national level, 1.8% of children under the age of 5 are poor, contributing to the 12.5% national poverty rate. It is the group of children aged between 5 and 18 years old that has the highest poverty rate with nearly 2% of this population considered poor and contributing 36.1% to national poverty. Thus, nearly 50% of poverty in Morocco affects Moroccans under the age of 18 (Table 6).

When we break down this distribution of poverty by age group, considering urban and rural areas, we find that the poverty incidence among young people under 18 is higher in rural areas than in urban areas (Figure 5). Thus, the poverty rate for youth in the rural area is higher than the poverty rate observed at the national level for this environment. This is because there are more poor people in the rural area and rural households are also larger on average.

Figure 5

Download asset Open asset

Then Figure 6 presents the Gini inequality indices for Morocco and rural and urban areas in 2019. Not surprisingly, we observe higher inequality in urban areas compared to rural areas (0.347 versus 0.280). Urban inequality is also lower than that of the country (0.361).

Figure 6

Download asset Open asset

Table 7 reports the incidence of poverty and vulnerability in Morocco decomposed for urban and rural areas, following the simulated COVID-19 shock and the implementation of mitigation measures (scenario 1 and scenario 2).

In terms of poverty incidence, our results confirm the negative impact of COVID-19. Indeed, poverty among Moroccans increases by 6 percentage points. The finding is that individuals in urban areas are the most affected by this shock in terms of percentage change in poverty incidence. This is explained by the cessation of economic activities and, as explained above, by the partial or total loss of income for some Moroccans.

Although the shock appears to be smaller for rural dwellers, the poverty rate increased by 7.1 percentage points. Thus, the pandemic has affected urban Moroccans to a greater extent, but the number of poor people in rural areas, where we observed a higher incidence of poverty in 2019, has also increased significantly. This corroborates the findings of the UNDP report on the impact of the pandemic (United Nations Development, 2020).

Our results also highlight that the aid quickly put in place by the government (scenario 1) had a positive effect in attenuating the shock, since the incidence of poverty in urban areas fell and was 2.2 percentage points higher than in 2019. The effect in the rural area is also positive, with the incidence going from 9.9 to 5.1%.

Scenario 2, which complements the government’s income support measures, targets workers in the informal sector as well as those working in the tourism sector. This scenario has a positive effect on urban individuals as the incidence of poverty decreases to around 1.4 percent. Another measure introduced in scenario 2 targeted rural households by providing an allowance for households with children under certain conditions related to schooling. The impact of this measure is to practically absorb the negative effects of COVID-19 shock on poverty.

The effects on vulnerability are in the same direction, with an increase of 2.4 percentage points following the COVID-19 shock at the national level. This confirms the fact that the pandemic has pushed part of the Moroccan population into poverty and vulnerability. The government’s measures partially offset this negative effect, but we highlight that part of the population that was vulnerable before COVID remained in poverty despite the measures.

In terms of inequality, Figure 7 shows that the pandemic has increased inequality in both urban and rural areas. This is coherent with the findings of (Gerzon Mahler et al., 2020) on the distributive impact of the pandemic.

Figure 7

Download asset Open asset

As expected, the general shapes of the density function curves, based on the area of residence (urban and rural), are generally the same. However, the curve for average per capita expenditure during the pandemic shifted sharply to the left, regardless of their location of residence, confirming the shift of a portion of the population towards the vulnerable and poor classes (Figure 8).

Figure 8

Download asset Open asset

To show the combined effect of the negative shock of the pandemic and the different measures simulated, we analyze the shift of the density curves of the baseline situation with those obtained under scenario 2 (Figure 9).²¹

Figure 9

Download asset Open asset

For the rural population, the policies put in place, in addition to those we have identified and simulated, are likely to neutralize most of the negative effects of the shock on the poor and vulnerable classes. For the urban group, some individuals persist in poverty, which can be explained by the fact that the shock was more profound and since the additional mitigation policies proposed are likely to benefit poor and vulnerable rural households by their nature or design. This finding is the same as that obtained with the poverty and Gini indices.

To illustrate the effects of the shock and the mitigation measures, we plot the incidence curves (CI).²² These curves allow us to verify the "pro-poor" character of shocks and measures. Figure 10 depicts the incidence curves on per capita spending between the baseline situation and the three scenarios for rural and urban. This analysis shows that the poorest percentiles are the most affected by the pandemic and remain so compared to the pre-crisis situation, despite the government’s measures and our proposed measures.

Figure 10

Download asset Open asset

For percentiles below 50% (p < 50%), the values of the incidence curve (noted g(p)), are negative while these values are always negative but less for the other percentiles. It is therefore clear that households in the poor and vulnerable classes of the population were the most affected, regardless of the place of residence. In the urban area, the workers are generally active in the private sector and especially in the informal sector. They obtain their income from activities that were negatively impacted by the cessation of economic activities or by the lockdown. Therefore, we observe a general deterioration in the population’s standard of living as measured by average household expenditure per capita, although the decline is very slight for the top 5% of the distribution. Moreover, despite the aid put in place or proposed in favor of the most disadvantaged classes, the overall impact of the pandemic and the interventions was not "pro-poor" in the sense (Ravallion and Chen, 2003), even with very high rates of growth in the median and average per capita expenditure (14.41% and 10.93% respectively).

Now, as our objective is to study the impact of the pandemic on children, based on the age group decomposition presented above, we look at how the poverty and vulnerability indices vary (Table 8). It appears that it is those in urban areas who are most affected by the pandemic. While the poverty rate for children under 5 years of age was 0.6% in 2019, it rose to 8.9%. Thus, nearly 150,000 children under the age of 5 would have been pushed into poverty because of the pandemic. For the group aged 5 to 17, it is 108,500 additional individuals living in poverty in the urban area. These children belong to families where some members have seen their income drop in whole or in part, children are directly affected by these drops in their parents' resources.

In the rural area, where the poverty rate is estimated at 2.79% in 2019, the incidence of poverty among children under 5 has increased from 3.70% (approximately 43,000 children) to 11.80%, representing nearly 140,000 children. Thus, the pandemic would have driven nearly additional 96,000 children under the age of 5 into poverty in the rural area. For group 5 to 18, nearly 200,000 individuals living in the rural area are poor, whereas they were not poor in 2019. Thus, more than half a million Moroccan children under the age of 18 would have fallen into poverty because of the pandemic.

As expected, we find that the transfers offered by the government to compensate for the loss of household income due to confinement partially offset the effect of COVID-19 for Moroccan children under 5 years of age and those aged 5 to 18 years. These offsetting effects are observed in both urban and rural areas. However, our complementary measures (scenario 2) targeting rural households with a cash transfer to compensate parents for the loss of education-related assistance and the additional costs of children not attending school since mid-March results in a larger decline in poverty among rural children. As a result, the increase in poverty among children under 18 is reduced by almost 60% because of complementary measures.

When we look at the incidence of vulnerability by age group, we find that young Moroccans under 18 are almost twice as vulnerable as their elders. In terms of the effect of the pandemic, we find that the shock was stronger in the urban area. The compensatory measures simulated in scenarios 1 and 2 are still playing their role here, with the incidence rates of vulnerability decreasing to their initial rate for children under 5 years of age and decreasing very slightly for young people aged 5 to 17 years at the level of Morocco. When we distinguish the effects of the measures by area of residence and age group, we find positive effects in both areas and for all age groups. However, while in the rural area the proportion of vulnerable youth under 18 decreases relative to the baseline situation, in urban areas this proportion remains much higher despite the compensatory measures.

We also evaluated the contribution of the components associated with our scenarios to the variation in the poverty and Gini indices between the baseline situation and the results obtained with scenario 2. By construction, our scenarios combine five components that are added to the baseline situation in terms of per capita expenditure. Our five components are 1- the nominal effect associated with the pandemic (covid); 2- the nominal effect of simulation 1 in which some individuals received transfers (scenario 1); 3- the nominal effect of the extension of transfer payments that we propose for workers in the tourism sector and workers in the informal sector (scenario 2 – no school aid); 4- the nominal effect of the payment of school aid to each child who is already a beneficiary of an aid program (school aid); and finally 5- the price effect (deflator). We computed these contributions by place of residence(Figure 11) but also for children under 18 years of age (still by place – Figure 12).

Figure 11

Download asset Open asset

Figure 12

Download asset Open asset

As we have already seen, the contribution of the pandemic is negative for all groups and indices. The measures implemented by the government partially offset this negative effect. However, this decomposition allows us to conclude that the variation in prices observed in Morocco had no significant effect on the variation in the poverty and inequality indices. Also, the negative effect of the pandemic is due mainly to the partial or total loss of income for some people and to the decrease in transfers received from households living abroad.

By breaking down scenario 2 into two (government transfer effect and school aid effect), we see that their contribution differs according to the area of residence. Indeed, while scenario 2 without school, transfers has a positive contribution to the reduction of poverty and inequality in the urban area, we find that this effect is negative in the rural area. This could be explained by the fact that Moroccans living in rural areas do not benefit from the complementary transfers that we designed. Finally, the contribution of school aid is positive for both groups (urban and rural), but with a greater contribution to poverty reduction in rural areas since this measure essentially targets this poor population.

When we perform the same analysis restricted to children under 18, we obtain similar results. Thus, the price effect has no significant impact on the variation in poverty and inequality (Figure 12). The COVID effect helps explain a large part of the variation in the indices among children. This negative contribution is stronger for children in the rural area. A final difference that we observe when we target children is that the contribution of school aid to reducing poverty and inequality is stronger in rural areas than in urban areas, but it remains below the contribution we obtained when we considered the entire Moroccan population. Thus, even though the measure targets children between the ages of 5 and 17, the effects of this measure are felt by the households in which these children live.

Finally, we analyze the cost of scenario 2 in its entirety and isolate the school aid provided in this scenario. Scenario 2 complementary measures would cost the government nearly 173 million MAD, representing a per capita cost of 4.86 MAD, or 6.33 MAD per beneficiary (Table 9). Finally, this scenario has a cost equal to 789.86 MAD per child under the age of 18 who is lifted out of poverty thanks to these measures.

When we compare rural and urban areas, we find that 55% of the cost of the full scenario 2 falls on rural areas. However, even though the per capita cost is more than double in the rural area (7.17 MAD versus 3.50 MAD in urban areas), the cost per child out of poverty with this measure is only 5% higher in the rural area. We have done the same analysis by isolating the effect of the school aid provided in scenario 2. The total cost of this measure amounts to 20.62 million Dirhams, or 0.58 MAD per Moroccan and 2.12 MAD per beneficiary of the measure. This measure, which targets children, would cost almost 100 MAD (94.18 MAD) per child lifted out of poverty by scenario 2. As these programs are mainly implemented in rural areas, more than 71% of the cost of this measure is allocated to them. This school aid program would thus cost 125.18 MAD per child lifted out of poverty with scenario 2.

We find that the school aid measure has a low marginal cost when spread over the entire Moroccan population and that its contribution to poverty reduction, particularly among rural children, is significant. Thus, a measure of this nature could easily be considered, especially since we believe that it could have long-term effects by encouraging households receiving this allowance to send their children back to school.

https://www.imf.org/en/News/Articles/2019/07/16/pr19279-morocco-imf-executive-board-concludes-2019-article-iv-consultation-with-morocco.

https://www.tresor.economie.gouv.fr/Pays/MA/indicateurs-et-conjoncture, https://www.afdb.org/en/countries/north-africa/morocco/morocco-economic-outlook, https://www.imf.org/en/News/Articles/2022/02/09/pr2227-imf-executive-board-concludes-2021-article-iv-consultation-with-morocco.

Ahmed Lahlimi (High Commissioner for Planning) to EFE: Morocco should record its worst economic year of the century, by Rahim Sefrioui on 14/04/2020 at 12:11 (updated on 14/04/2020 at 12:19 and https://www.hcp.ma/file/216987/.

This survey was conducted by telephone between April 1 and 3, 2020 among 4,000 companies operating in the manufacturing, construction, energy, mining, fishing, trade and non-financial market services sectors.

In fact, this negative impact did not materialize. In fact, transfers increased to 68,2 billions dirhams in 2020 reprsenting a 4,9% increase (Office des changes 2021: Indicateur Mensuels des échanges extérieurs, OC-Maroc décembre 2021).

Available at: https://public.tableau.com/profile/laborde6680#!/vizhome/EconomicImpactEconomicslowdownduetoCovid-19IFPRIblogApril16th2020/Blog04162020_IFPRI?publish=yes.

No other estimate of poverty impact of COVID has been produced in Morocco to our knowledge.

In this sense, the ONDH has sent us 11 raw sub-files of this survey for the 2019 wave (without any processing carried out on the base). All our processing was carried out on the basis of these raw files.

In some of the treatments carried out in this work, the numbers appear slightly different from what is expected. Indeed, we did not impute any missing data in the different files.

For more details on this type of model, see (Miller and Blair, 2009), (Kamionkowski et al., 1995) and (Chowdhury and Kirkpatrick, 1994).

See the Haut-commissariat au plan’s information note on the consumer price index (CPI) for March 2020.

The vulnerability analysis was conducted by considering all individuals below the vulnerability threshold. Thus, the figures obtained are the cumulative number of poverty and people in the vulnerable class.

A lower middle class composed of households (individuals) that have a per capita expenditure above the vulnerability threshold (Z_{vulnarability}) but below the median of the distribution; An upper middle class composed of households (individuals) with a per capita expenditure above the median of the distribution but below 2.5 times this median. The wealthy class is composed of households (individuals) with a per capita expenditure greater than 2.5 times the median of the distribution.

This implies that for each scenario, we have fixed thresholds (poverty and vulnerability) but variable ones when they are calculated from the median of distributions. See Figure A1 in the Appendix.

These assumptions were postulated based on a combination of information coming from different sources namely Haut-Commissariat au Plan (2020a, CDG (2020) and Confédération générale des entreprises du Maroc-CGEM (2020). Information was not available directly but we used indirect information for these sources to construct these hypothesis.

This concept implies that when prices faced by different households are the same, the original level of equivalent income would be equal to the level of original income and the post reform level of equivalent income would equal original income plus the equivalent gain (King, 1983).

Cf. Information note from the Haut-Commissariat au Plan on the consumer price index (CPI) for March 2020.

For validation purposes, we compared the price variations (our 39 prices) with the 12 prices published by the HCP to ensure that our variations were comparable to those of the HCP.

Indeed, measures aimed at firms can only be properly captured using another modeling tool (a computable general equilibrium model (CGEM), for example) in order to trace the indirect effects of these measures on households.

Tayssir is a cash transfer program implemented to fight against school dropout.

All other density curves are available upon request.

They were introduced by (Ravallion and Chen, 2003) to analyze the dynamics of poverty in relation to economic growth and changes in inequality.

By definition of this decomposition, the sum of the components contributing to the variation of the indices considered is equal to 100. It can therefore be interpreted in relative terms.

No specific funding for this article is reported.

The authors wish to thank ONDH for giving us early acces to survey data, John Cockburn and Luca Tiberti for proprosing this research topic to us and finally Medhi Halmi of l’UNICEF Rabat for valuable input on this paper.

This research work was carried out with financial support from the Partnership for Economic Policy (PEP) and the Oxford Policy Management (OPM) with funding from UNICEF MENA. We thank ONDH for giving us early access to the 2019 household panel survey, Mehdi Halmi, John Cockburn and Luca Tiberti for valuable comments to improve this paper. Regular caveats apply.

1. Version of Record published: December 31, 2022 (version 1)

© 2022, Abdelkhalek et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.