Document de travail

The Morocco Policy Analysis Model:

Theoretical Framework and Policy Scenarios

by Aya Achour, Aleš Bulíř, Omar Chafik, and Adam Remo

Avril 2021
Les opinions exprimées dans ce Document de Travail sont celles des auteurs et
ne reflètent pas nécessairement la position de Bank Al-Maghrib. Afin de
garantir une meilleure qualité et rigueur scientifique, les documents de travail
publiés sont évalués par des arbitres externes, universitaires et chercheurs de
Banques Centrales modernes.

Aucune reproduction ou traduction de la présente publication ne peut être faite

sans l’autorisation des auteurs.

L’objet de la publication du présent Document de Travail est de susciter les

débats et d’appeler commentaires et critiques.

Si vous avez des commentaires sur ce Document de Travail, veuillez les faire
parvenir par e-mail : [email protected]

Ou par courrier à l’adresse suivante :

Bank Al-Maghrib, Département de la Recherche

277, Avenue Mohammed V - B.P 445 Rabat

Ce document peut être téléchargé sans frais par voie électronique sur :
www.bkam.ma
 The Morocco Policy Analysis Model: Theoretical Framework and Policy

Scenarios1

Prepared by Aya Achour (Bank Al-Maghrib), Aleš Bulíř (International Monetary Fund),
Omar Chafik (Bank Al-Maghrib), and Adam Remo (International Monetary Fund)

April 2021

Abstract

The Morocco Policy Analysis model (MOPAM) was created in the Bank Al-Maghrib
to simulate the impact of external developments, domestic macroeconomic policies,
and structural reforms on key macroeconomic aggregates. We describe its structure
and demonstrate its operation on two medium-term scenarios: (1) fiscal consolidation to
stabilize the debt-to-GDP ratio and (2) the effects of the COVID-19 shock, including the
endogenous fiscal and monetary policy response.

JEL Classification Numbers: E32, E52, E58, E62, F47

Keywords: Macroeconomic modeling, Morocco, fiscal stabilization, COVID-19
pandemic

1
This paper greatly benefited from seminars and presentations at the Bank Al-Maghrib and an IMF seminar.
We are grateful for comments and support from Abdessamad Saidi, Andy Berg, and Roberto Cardarelli.
Authors’ E-Mail Addresses: [email protected]; [email protected]; [email protected]; [email protected]
 3

Table of contents
Abstract............................................................................................................................. 2
I. Introduction.................................................................................................................... 4
II. Morocco: Economic Growth Since the Global Financial Crisis ....................................... 4
III. The Modeling Framework ............................................................................................ 7
A. The MOPAM Building Blocks................................................................................... 7
B. Policy Scenarios: How Should We Read Them? ......................................................... 8
IV. The Fiscal Consolidation Scenario ................................................................................ 9
A. Lessons from the 2019 Conference and Other Communications .................................. 9
B. Scenario Assumptions.............................................................................................. 11
C. Simulation Results................................................................................................... 13
V. The COVID-19 Pandemic and MOPAM Policy Options............................................... 15
A. Economic Impact of the Pandemic: Demand, Supply, Both, or Something More?...... 16
B. Scenario Assumptions.............................................................................................. 18
C. Simulation Results................................................................................................... 18
VI. Conclusions............................................................................................................... 23
ANNEX I ........................................................................................................................ 27
ANNEX II....................................................................................................................... 29
 4

I. INTRODUCTION

The paper describes the Morocco Policy Analysis Model (MOPAM), a large-scale
macroeconomic model developed at the Bank Al-Maghrib (BAM), and demonstrates how it
can be used to simulate complex structural and long-term policy scenarios. We present two
scenarios, one analyzing the long-term effects of fiscal consolidation and the other providing
an early assessment of the impact of the COVID-19 pandemic and the results of the fiscal
policy measures adopted.
The MOPAM is based on the Flexible System of Global Models (FSGM) modeling
framework (Andrle and others 2015) as adjusted to single-country setup to capture the main
features of the Moroccan economy and its policies. The MOPAM is able to capture a range
of monetary and fiscal regimes—a welcome feature in an economy that is undergoing a
series of gradual reforms as it moves toward greater exchange rate flexibility, a more open
capital account, and a wider social safety net. The model strikes a balance between economic
theory and the empirical and policy idiosyncrasies of the Moroccan economy. The model
itself is a mix of traditional dynamic stochastic general equilibrium (DSGE) segments and ad
hoc, data-driven segments, with the choice of segments driven by their ability to support a
clear economic interpretation.
Repeated application of the model, which the BAM has been using since the mid-2010s, has
demonstrated how useful it is in scenarios requested by the Moroccan policymakers. For
example, the MOPAM has been used to assess the economic impact of pension reform of
2016, the transition to a more flexible exchange rate regime in 2017, and the minimum wage
increase of 2019. We hasten to say that the MOPAM is not a forecasting framework but a
simulation tool designed to stimulate policy discussion. The BAM uses a different modeling
framework for forecasting, and official BAM and IMF forecasts for Morocco should not be
compared with the MOPAM simulation results. Instead, the MOPAM simulations explore
counterfactual scenarios based on defined assumptions and expert judgment with respect to a
baseline that either reflects the official macroeconomic forecast or assumes that policies are
unchanged.
The paper has various readers in mind. Those who want to understand the broad structure of
MOPAM may find Section III useful. Modelers will find detailed model description in the
Annex II. Economists and policymakers looking for answers to questions related to the long-
term effects of fiscal policy reforms and potential medium-term impacts of the COVID-19
pandemic should turn their attention to Section IV and Section V, respectively.
In the paper, we (1) outline recent economic developments and the policy scenarios being
considered; (2) sketch the MOPAM structure (see the detailed description in Annex II); (3)
lay out the policy scenarios; (4) discuss the simulation results and summarize their policy
implications; and (5) draw conclusions.

II. MOROCCO: ECONOMIC GROWTH SINCE THE GLOBAL FINANCIAL CRISIS

Although a textbook countercyclical response to the global financial crisis (GFC) helped
Morocco to escape recession, it failed to address the ensuing growth slowdown (Figure 1).
 5

For 2001– 08, real GDP growth averaged 5.25 percent, reducing unemployment from more
than 13 percent to less than 10 percent in 2008, and lowering the gross debt-to-GDP ratio
from 65 to 45 percent. After a massive fiscal expansion as the 2008 structural balance of
about 0 percent of GDP turned to –8 percent in 2012 and –6 percent in 2013, post-GFC real
GDP growth has averaged only 3.5 percent and unemployment is stubbornly high at about 10
percent.

Figure 1. Morocco: Real GDP Growth, Unemployment, and Fiscal Indicators, 2000–19

Sources: The National Statistical Agency of Morocco (HCP) calculates real GDP using the 1993 System of National
Accounts (SNA) and publishes the unemployment rate, using the Harmonized ILO definition. Morocco’s Ministry of
Finance (MoF) publishes the central government deficit, which includes grants, privatization receipts, and monopoly
revenues. The MoF also publishes central government gross debt, which includes all creditors, maturities, and currencies;
the 2019 observation is preliminary. The structural balance, potential GDP, and the fiscal impulse (the year-on-year change
in the structural balance) are calculated using the IMF’s WEO estimates in the October 2019 World Economic Outlook.

The recent period of loose fiscal policy—in 2010–16 the overall fiscal deficit exceeded
4 percent of GDP every year—pushed up public debt considerably. The gross debt-to-GDP
ratio rose from its trough of 45 percent in 2008–09 to more than 65 percent in 2019. While
continuously assessed as sustainable,2 as early as 2012 rising public debt was recognized as a
long-term risk, leading Morocco to request three successive IMF-supported, precautionary
arrangements and launch important fiscal reforms which initially gave priority to removing
costly budget subsidies and restructuring the pension system. 3 As oil prices fell in 2017–19,
the budget deficits narrowed to less than 4 percent of GDP.

2
For debt sustainability assessment, see IMF 2019a.
3
The first two Precautionary and Liquidity Line (PLL) arrangements were designed as purely precautionary.
The authorities made a purchase in April 2020, however, at the start of the COVID-19 pandemic.
 6

The main budget ratios were stable (Figure 2). Most notable was the drop in subsidies, from
12.5 percent of total spending in 2014 to 4.5 percent in 2019. The share of the wage bill in
spending was high—at almost 40 percent of total spending or about 10 percent of GDP. With
the debt-to-GDP ratio up by about one-half from its pre-GFC level, the cost of servicing
stabilized at about 10 percent of total spending, helped by long maturities.

Figure 2. Morocco: Central Government Revenues and Expenditures, 2014–20,

Percent

Sources: The General Treasury Department of the MoF publishes data on central government revenues and spending; the
2020 figures are the projected budget. Annual data are calculated from monthly published data.

For the third IMF-supported arrangement, approved in December 2018, Morocco made two
fiscal commitments: (1) to stabilize the overall deficit at about 3 percent of GDP by 2020 and
(2) to bring public debt down to 60 percent of GDP over the medium term. These ambitious
objectives required higher revenues and lower spending. Revenue measures considered
included boosting tax collections, reducing tax exemptions, better enforcing tax payment
discipline by the self-employed and the liberal professions, and simplifying and re-aligning
value-added tax (VAT) rates. Spending measures included stabilizing the government wage
bill, removing subsidies, and switching to targeted social transfers.
The impact of these planned reforms would be felt outside of public finance as most of them
will affect the supply side, the demand side, or both sides of the economy. Assessing the total
 7

impact of these reforms on the economy has been challenging, given the offsetting effects of
some measures. Using a rich modeling framework, such as the one described here, should
help in this task.

III. THE MODELING FRAMEWORK

The Morocco Policy Analysis Model (MOPAM) is an annual model based on the FSGM, a
semi-structural, multi-country modeling framework originally designed by the IMF Research
Department for macroeconomic analysis of structural problems or issues that involve a large
number of countries.4 The BAM has used the framework, adapted to fit the needs of a single
small-economy analysis, to study counterfactual macroeconomic and structural scenarios that
are of interest to policymakers (see, for example, Achour and Chafik 2019). Significant
adjustments have been made to the original FSGM single-country block to adapt the model to
the structure of the Moroccan economy, such as adding sector-specific dynamics, including
the traditional agricultural and phosphate sectors; extending the monetary block for capital
account regulation; and adapting the policy reaction functions to capture monetary and fiscal
regimes as they evolve. The monetary regime when this paper was written combined a fixed
exchange rate with a partially open capital account, thus giving the central bank a degree of
monetary autonomy. On the fiscal side, the MOPAM incorporates energy subsidies and a
flexible system of fiscal rules.
The MOPAM is designed to strike a balance between economic theory and empirical
insights. It incorporates both traditional theory-driven DSGE segments and ad hoc, data-
driven segments when these perform better than the DSGE ones— examples include the
supply side, especially the labor market, and the monetary and capital account blocks. For a
detailed description of the model see Annex II.

A. The MOPAM Building Blocks

The model is subdivided into four blocks: core, supply, government, and monetary. The core
block, which is micro-founded describes private consumption and private investment of
overlapping-generation (OLG) households. Households optimize utility with respect to their
budget constraint while accounting for human and financial wealth. The use of OLG
households allows us to break down Ricardian equivalence and government debt neutrality.
We also introduce in this block financially constrained (“hand-to-mouth”) households whose
consumption equals their wage and transfer income and that can neither save nor borrow.
Private investment follows the Tobin’s Q model augmented with real adjustment costs.
The supply block is semi-structural: Output follows the Cobb-Douglas production function
with steady-state labor and capital. Total factor productivity (TFP) incorporates second-
round effects of commodity prices and public investment in aggregate supply. The labor
force and the participation rate are treated as exogenous; steady-state labor is calculated with
respect to the nonaccelerating inflation rate of employment (NAIRU). Core inflation is
modeled with an open-economy hybrid Phillips curve that reflects price stickiness, inflation

4
The MOPAM was developed between 2015 and 2019 by BAM staff, with support from the IMF Research
Department and Institute for Capacity Development as part of a project financed by the government of Canada.
 8

expectations, and changes in marginal costs as captured by the output gap and the real
exchange rate gap. Nominal wage inflation follows a similarly defined Phillips curve and
Okun’s law. Noncore inflation, namely food and energy, reflects world prices and the
nominal exchange rate dynamics.
The MOPAM government block is rich, as necessitated by the policy questions the model is
tasked to answer. We capture detailed revenue and expenditure breakdown as well as
alternative fiscal rules that can target either a long-term debt-to-GDP ratio or a long-term
deficit. The government chooses among alternative fiscal instruments, which affect the
demand side (the value added tax, transfers to households), the supply side (tax on capital),
or both (public investment). The MOPAM includes distortionary taxes on both labor and
capital) in addition to lump-sum taxes to further relax Ricardian equivalence.
Finally, we strive to capture the current monetary regime in Morocco, which deviates from
the two regimes commonly studied in small open economies: either a float or a peg, typically
with an open capital account. Morocco has maintained as an exchange rate anchor a
composite comprising the euro and the US dollar. In January 2018 Morocco changed its
official arrangement from “conventional peg” to “pegged exchange rate within horizontal
bands” when the fluctuation band for the dirham was widened to ±2.5 percent (IMF 2019b);
in March 2020 the band was again widened to ±5 percent.5 As the capital account is only
partly open, such a regime creates some policy space to actively stabilize inflation and
growth (Obstfeld et al. 2005). In addition, in the MOPAM we can easily introduce a range of
intermediate exchange rate regimes and policy reaction functions that reflect different stages
of the process of making the exchange rate more flexible.6

B. Policy Scenarios: How Should We Read Them?

We demonstrate the richness and flexibility of the MOPAM framework using two
counterfactual policy scenarios where all variables are expressed as deviations from a control
simulation. (1) Building on the government commitments listed in the IMF-supported
Precautionary and Liquidity Line arrangement and the 2019 National Conference on
Taxation, we draft a scenario that encompasses all the main elements of the medium-term,
multipronged fiscal consolidation. (2). We draft a second scenario that captures the impact of
the COVID-19 pandemic on Morocco and some of the government’s policy choices. Unless
stated otherwise, all variables are expressed as a percentage point deviation from a control
simulation that would have been expected without either the fiscal consolidation or the
COVID pandemic The steady state corresponds to an “unchanged policy” scenario, where
the long-term ratios reflect past behavior. Both scenarios then incorporate a new set of initial
conditions and a path for the long-term ratios leading to the new steady states that reflect the
post-consolidation or the post-COVID situation.

5
Decision N°49/W/20 available at:
http://www.bkam.ma/content/download/700120/8116644/D%C3%A9cision%20Bank%20Al%20Maghrib_49W
20.pdf.
6
In the mid-2010s the Moroccan authorities launched a program of gradually making the exchange rate more
flexible, with an eventual transition to a floating rate and inflation targeting (IMF 2018, Benlamine et al. 2018).
 9

Both simulations need to be understood as counterfactual policy scenarios rather than

forecasts. The difference? A forecast is a statement of what is judged likely to happen in the
future. In macroeconomics, it is typically our best guess about future GDP growth, inflation,
and the corresponding policies. A counterfactual scenario is based on a hypothesis that is
contrary to some facts in the control scenario. In policy modeling we call such facts “the
control” and our counterfactual scenarios are expressed as deviations from controls. In
practice, the controls can be previous forecasts, even with a different model; projections
based on unchanged policies; and so on. The scenario then asks “what if” some components
of the scenario evolve differently from the control—for example, “What if the Moroccan
authorities introduce a fiscal consolidation program? How would macroeconomic
developments compare to the (control) scenario without consolidation?”
The question and the design of the scenario matter, of course. The fiscal consolidation
scenario enumerates potential long-term benefits from gradual, long-lasting fiscal efforts.
Most of the measures are structural, such as changes in the tax rates or redirection of public
spending. These measures are firmly guided by announced policy decisions. In contrast, the
COVID-19 scenario enumerates the possible costs of the recent pandemic, effectively
focusing on downside risks. The largely exogenous effects are driven by experience in other
countries, economic theory, and authors’ judgment.

IV. THE FISCAL CONSOLIDATION SCENARIO

Morocco’s fiscal outlook has been anchored to stabilization of the overall deficit at about 3
percent of GDP and a gradual decline in public debt to 60 percent of GDP. How to achieve
these objectives inclusively and equitably? What would be the macroeconomic impact of the
related policy actions? Regarding the former, the 2019 National Conference on Taxation
provided some guidance for policymakers. Regarding the latter, we build a scenario
consistent with guidance from the Conference and simulate its macroeconomic effects. Our
simulations suggest that a well-executed fiscal reform, accompanied by appropriate monetary
and structural policies, could very well deliver the expected results of stabilizing debt and
permanently increasing real GDP: we estimate the cumulative gain in present-value terms at
about 7 percent of GDP.

A. Lessons from the 2019 Conference and Other Communications

Let us reiterate that the scenario being considered is the basis for analyzing long-run effects
of a comprehensive fiscal consolidation plan on the assumption that the COVID-19 pandemic
has not happened and relative to a no-consolidation scenario. Naturally, any fiscal
consolidation plans formulated before the pandemic would need to be reassessed once the
pandemic is over.
The Government of Morocco convened the National Conference on Taxation with
stakeholders on May 3–4, 2019 to debate how fiscal consolidation could proceed. Although
most of the recommendations were general, the conference did provide specific guidance on
several aspects of this process:
1. The fiscal reforms will be anchored in a Framework Law that would describe the tax
measures to be implemented during the first 5 years.
 10

2. To balance distributional and efficiency concerns, revenue gains, from both better
enforcement and a wider tax base, will be split between funding the social safety net
and reducing income tax rates.
3. The complicated system of multiple VAT rates will be simplified, with a default rate
of 20 percent.
4. The conference recommended a more progressive tax system in order to increase the
disposable incomes of middle-class and poor households.
5. Liberal professions, which have been paying very little personal income tax, will be
brought into the system.
6. Digitizing the tax system will address tax exemptions and leakages.
Changes to the VAT are expected to have pronounced effects. The current system of four
nonzero VAT rates plus excluded and zero-rated items will be simplified to one zero-rate and
two nonzero rates, with a default rate of 20 percent for most goods and services. In the
current system, with rates of 0, 7, 10, 14, and 20 percent, about 80 percent of total VAT
revenue comes from 20 percent-rated items and 12 percent from 10-percent rated items
(Figure 3), giving us the reference VAT rate of almost 17 percent (0.1685).

Figure 3. Morocco: Contributions to VAT Collection, 2017, Percent

5.1%

12.3%

2.3%

80.3%

VAT rate 20% VAT rate 14% VAT rate 10% VAT rate 7%

Source: Morocco Tax Administration, https://www.lavieeco.com/economie/rendement-de-la-tva-le-taux-de-20-

rapporte-80-des-recettes/.

The current VAT system leaks: the effective VAT tax rate—VAT collection divided by the
VAT base—is only about 11.5 percent, implying the so-called C-efficiency ratio well
below 1, with the remainder accounted for by policy and compliance gaps (Keen 2013;
Hutton 2017). 7 The Ministry of Finance stated that in 2017 the “policy gap” in terms of tax
expenditures—effectively exemptions from tax obligations,—reached DH 28.5 billion,

7
The C-efficiency ratio is defined as the share of the VAT in consumption divided by the standard VAT rate
and it is a widely used as a measurement of authorities’ ability to collect indirect taxes.
 11

almost 2.75 percent of GDP. 8 A part of the policy gap will be closed in the proposed new
system of zero-rated items (necessities); 10-percent items (mass consumption products); and
20-percent items (luxury goods) that is expected to raise the effective VAT tax rate by about
half a percentage point, to almost 12 percent, based on 2017 final consumption data.
Furthermore, the authorities want to gradually address some of the compliance gaps in the
tax system. The former head of the General Tax Administration said in a press interview that
total VAT tax leakages could be as high as 3.5 percent of GDP, DH 40 billion. 9 Taking this
estimate at face value and using the VAT tax expenditure of DH 14 billion, the VAT
compliance gap (tax noncompliance, avoidance, etc.) can be estimated at almost DH
26 billion, 2.5 percent of GDP. In our simulations we do not expected that the policy and
compliance gaps will be closed fully or immediately.
The key expenditure measure in the scenario is keeping the public wage bill below
10.5 percent of GDP over the medium term, a measure carried over from previous policy
statements. Convergence to the wage bill target is expected to be slow, however, because in
2019 long-delayed wage increases were granted.

B. Scenario Assumptions

We now translate these announcements and policy guidance into steady-state model
variables. Because the scenario assumes less crowding-out and a higher share of
consumption (non-distortionary) taxes, the consolidation scenario should expand output. We
first summarize the main scenario simulation assumptions, as listed in Table 1.
• Over the medium term the debt-to-GDP ratio drops from 65 to 60 percent.
• Over the medium term the public wage bill stabilizes at 10.5percent of GDP.
• Realignment of the VAT rates and the partial recovery of leakages pushes the
effective VAT rate up from 13 to 14.5 percent, and total collections up by almost 1
percent of GDP. 10 These collection gains result from: (1) merging the 7 percent rate
into the 10 percent rate and the 14 percent rate into the 20 percent rate yield about
0.5 percentage point increase in the VAT effective tax rate; and (2) transferring one-
half of the VAT exemptions and zero-rated items to the 10 percent rate yields an
additional 1 percentage point.

8
About half of all tax expenditures appear to be in VAT (DH14 billion) and one-sixth each in corporate income
tax (CIT) and personal income tax (PIT). The main beneficiaries have been the construction and energy sectors
and these exemptions have proven difficult to repeal. The Tax Expenditures Report is a part of the 2019 budget
law draft: https://tax.gov.ma/wps/wcm/connect/d9b1e5b0-a320-479a-8367-
181dd4c027e3/Rapport+sur+les+d%C3%A9penses+fiscales+-
+Ann%C3%A9e+2019+_fr.pdf?MOD=AJPERES&CACHEID=d9b1e5b0-a320-479a-8367-181dd4c027e3.
9
The interview is available at https://lnt.ma/m-omar-faraj-dg-impots-question-de-lequite-fiscale-place-
publique/. Given the 2017 VAT collection of 81 billion dirhams, we see his figure as being on the high side of
plausible estimates. He also estimated CIT evasion at DH 12 billion and PIT evasion by liberal professions at
about DH 5 billion, or about 1.5 percent of GDP in total.
10
In these calculations we assume that the elasticity of consumer demand to the tax rate is zero, i.e.,
consumption of an item does not change when its VAT rate changes.
 12

Table 1. Fiscal Reform: Scenario Assumptions, Deviation from the Control

Time to
Change
Converge to
Variable over the Comments
the New
Long Term
Level

Simulation 1
The judgment captures
Government debt-to-GDP
–6 p.p.* 6 years stabilization of the government
ratio
debt at 60 percent of GDP.
The judgment captures
Government current stabilization of the government
–1.1 p.p. 5 years
spending-to-GDP ratio wage bill at 10.5 percent of
GDP.
Simulation 2 (judgments applied jointly with the judgments above)
The judgment captures the VAT
Consumption tax rate +1.5 p.p. 1 year
tax reform.
Simulation 3 (judgments applied jointly with the judgments above)
The judgment captures the
Capital revenue tax rate –1 p.p. 1 year
corporate income tax reform.
Simulation 4 (judgments applied jointly with the judgments above)
The judgment captures
Fuel subsidy-to-GDP ratio –0.75 p.p. 5 years
elimination of the fuel subsidy.
50% of the savings from fuel
subsidy removal used for public
Government investment-
+0.38 p.p. 5 years investment. The remaining part
to-GDP ratio
is devoted to poor households in
the form of cash transfers.
Note: p.p. = percentage point.

• The corporate income tax change from a proportional to a progressive rate

calculation leaves CIT collection slightly lower. Although the tax rates of both
systems are almost identical (Table 2), our calculations using the 2017 Moroccan firm
balance sheet data suggest that the effective tax rate would fall from 29.5 to
28.5 percent for nonfinancial firms. The reason is that the underlying tax bases are
calculated differently: in the proportional system the same rate is applied to the entire
profit; in the progressive system a different rate is applied for each of the three profit
brackets.
 13

Table 2. Morocco: The Corporate Tax Rates, Percent

Profit (DH
Proportional Tax Rates 2016 Progressive Tax Rates 2019
Thousands)
10 percent for profits up to
[0 to 300] 10 percent all profits
DH 300,000
17½ percent for profits of
]300 to 1,000] 20 percent all profits
DH 300,000 to 1,000,000
]1,000 to 5,000] 30 percent all profits 31 percent for profits above
> 5,000 31 percent all profits DH1,000,000
Source: Ministry of Finance, 2016 and 2019 budget laws.

C. Simulation Results

The impact of fiscal consolidation on output and its composition is positive and the welfare
gains are substantial; however, the consolidation program as outlined in Section IV.A has the
usual features of a Keynesian contractionary stabilization because we discount the possibility
of a non-Keynesian expansionary stabilization (see, e.g., Giavazzi and Pagano 1996 and
Alesina and Ardagna 2010). 11 All simulations are expressed as deviations from the control.
We start with the short-term results: Real GDP falls below the control through 2025 and
unemployment stays high (see Figure 4 and Annex Figure A.1). However, government
deficits fall, financed by permanently higher revenue and temporary cuts in spending—real
government absorption falls permanently relative to the control. A lower wage bill pushes
the gross debt-to-GDP ratio from almost 65 percent in 2019 to the new steady state of
60 percent (Simulation 1; blue line), helped by higher indirect taxes (Simulation 2; red line);
the income tax changes have negligible impact (Simulation 3; yellow line). Finally, gradual
removal of the remaining fuel subsidies brings down public debt, but the drop is offset by
higher public investment and targeted social transfers (Simulation 4; purple line).
The removal of fuel subsidies has a notable negative impact on real GDP in the short and
medium term, through both consumption and investment channels. While efficiency gains
might be expected from the subsidy removal because fuel subsidies tend to benefit the
wealthiest households, those effects are largely offset by a temporary fall in private
consumption and investment because the OLG households have less disposable income.
Unlike liquidity-constrained households (LIQ), which are fully compensated through cash
transfers and whose consumption remains unchanged, OLG households are only partly
compensated. Moreover, higher retail energy prices depress private investment to the extent
that demand for fuel is inelastic: especially in agriculture, higher prices force reallocation of
resources to energy at the expense of other production factors. Furthermore, the cut in energy
subsidies is gradual and the efficiency gains and increases in public investment are small, not
enough to offset the immediate negative income effect on OLG households.

11
It has been argued that a large fiscal adjustment today removes expectation of both future harsher adjustments
and future tax hikes, thus stabilizing expectations, increasing expected consumer disposable income, boosting
the confidence of investors, and thus stimulating private demand on impact.
 14

In summary, the mix of fiscal measures in the scenario is expected to have an offsetting long-
run impact on output (see Figure 4). (1) the reduction in public indebtedness is expected to
lower the country risk premium and the cost of funds, and (2) the switch from direct to
indirect taxes can be expected to generate efficiency gains, with both channels generating a
substantial long-run increase in private investment. These two positive effects are partly
offset by removing the fuel subsidy that leads to higher energy prices, which at first have a
negative impact on private investment.

Figure 4. Morocco: Developments Resulting from Fiscal Consolidation

(Deviation from the control unless noted otherwise)

Notes: The control simulation is a “no-policy-change” scenario in which steady-state parameters remain constant at either
their historic averages or the latest expert-judgment values. The baseline simulation encompasses all fiscal consolidation
elements decomposed into four layers: the debt target and wage bill reductions (Simulation 1); plus the VAT reform
(Simulation 2); plus the income tax reform (Simulation 3); plus removal of the fuel subsidy.

Source: MOPAM; authors’ simulations.

The general equilibrium, long-run benefits of fiscal consolidation are considerable: GDP
increases by about 1.5 pp in Simulation 4 (Figure A1) relative to the control and the economy
eventually stabilizes at a permanently higher level of output, equivalent to about 7 percent of
GDP in present value terms. Private consumption—after the initial decline brought about by
the negative fiscal impulse—is expected to recover after about three years and stabilize some
2 pp above the control, which underscores the positive welfare effect of fiscal consolidation.
 15

The fiscal measures in our simulation are broadly supported by accommodative monetary
policy (Figure A.1; bottom panel). However, in the case of the fuel subsidy removal, the
slight increase in the policy interest rate follows an upward pressure on the neutral interest
rate which results from higher public investment and improvement in potential GDP. Of
course, the need to keep the exchange rate within the ±5 percent band limits monetary policy
options. The BAM is expected to adjust its stance only marginally with inflation in line with
the control. Given the energy share of the headline consumer price index, removing the fuel
subsidy causes a short-lived spike in headline inflation.
The nominal and real exchange rate both depreciate. Because the exchange rate is managed,
the nominal exchange rate depreciates only slightly, leaving the real exchange rate to
depreciate through internal devaluation—in the baseline scenario headline inflation stays
below the control and by 2040 the price is expected to be more than 3 pp below the control.
In the long run, the trade and current account balances are expected to be close to the
controls.
The fiscal consolidation simulations suggest that the long-term gains offset the short-term
costs. The initial drop in GDP peaks in 2022 at –0.7 percent. However, after the initial
decline in output, output stabilizes at a permanently higher level, equivalent to about 7
percent of GDP in present value terms. A permanently lower debt-to-GDP ratio lowers the
country risk premium and creates space for eventual expansion of targeted social transfers
(the transfers-to-GDP ratio goes up by more than 2.5 pp). Higher public investment
eventually boosts potential output, observed GDP, and private consumption. While
consolidating debt by 1 pp annually may seem ambitious, note that in 2000–08 Morocco’s
public debt declined twice as fast as in this simulation.

V. THE COVID-19 PANDEMIC AND MOPAM POLICY OPTIONS

The ambitious fiscal consolidation scenario naturally did not anticipate the COVID-19
pandemic. The planned fiscal reforms were temporarily mothballed, and attention turned to
policies to ease the impact of the pandemic. We therefore built a “harsh reality” scenario that
encompassed the authorities’ actions through the first half of 2020 and some assumptions
about the short and medium term both abroad and at home that seemed realistic at the time of
writing. Our simulations—predicated on these assumptions—suggest that even with a well-
coordinated fiscal-monetary mix, returning to the pre-COVID level of real GDP will take
several years and that pandemic-related debt will be difficult to extinguish. With hindsight,
some of these assumptions may have been too pessimistic.
Besides the obvious external demand assumptions, we also make a judgment call about
medium-term supply-side disruptions due to the pandemic. Tourism travel restrictions
extending into 2021, or even 2022, and a sustained decline in demand for Morocco’s exports
are likely to slow investment in both physical and human capital. The corresponding
slowdown in productivity growth and disruptions in supply chains will depress potential
GDP growth for some time. The judgment call was guided by the estimated declines in the
natural rate of interest due during past pandemics (Jordà and others 2020); Penn Wharton
model estimates of lasting macroeconomic impacts of the coronavirus on the U.S. (Dinerstein
and Huntley 2020); and corresponding calculations for emerging market economies (S&P
Global Ratings 2020).
 16

Note that the fiscal scenario was designed by the authors of this paper, not by the Moroccan
authorities, the IMF, or its Executive Board. However, we did draw on all available
information and attempted to limit judgment calls. Early on, to combat the COVID-19
pandemic the Moroccan authorities imposed a lockdown whose stringency was in the top
decile of the Oxford database, and the negative impact on economic activity was
correspondingly strong. 12 And because the scenario is designed as a counterfactual COVID-
19 scenario, we excluded from the initial conditions and near-term forecasts some events that
manifested themselves in 2020 but were not related to the pandemic, such as a below-average
harvest. 13 These were included in the control. Finally, we decided to limit the content of the
initial conditions to information available in June 2020 and did not update the simulations for
new information that became available later.

A. Economic Impact of the Pandemic: Demand, Supply, Both, or Something More?

We divide our discussion of the COVID-19 scenario into four parts: domestic demand and
supply factors; external demand and supply factors; developments in commodity markets;
and budgetary factors. Because the scenario assumes massive joint supply and demand
shocks (Guerrieri and others 2020), it is expected to lead to a pronounced, though short-lived,
recession, with fiscal and monetary policies only partly offsetting the first-round shock.
Domestic factors: demand and supply
The strictly enforced domestic lockdown had the dual supply and demand effects seen in
other countries. As this paper was being written, nonagricultural total factor productivity was
estimated to be down by some 5 percent in and labor participation rate by about 4 percent, on
the base of firms’ surveys (HCP). As for demand, households and firms reacted to the
general heightened uncertainty by cutting back on consumption (by about 3 percent) and
investment (by about 10 percent).

During the lockdown, according to the July 2020 national survey, 84 percent of Moroccan
firms interrupted their activities. 14 Hotels, restaurants, and textile industries were totally shut
down and nearly 50 percent of all respondents said they will need one year or more to again
reach pre-pandemic production levels. Mid-2020 macroeconomic data released by HCP
supported these statements: 2020 nonagricultural GDP was expected to decline by 5.25
percent, year on year; unemployment rate jumped to 12.25 percent; and average hours
worked per week plunged from 45 to 22 hours. Of the firms surveyed, 44 percent did not
expect to meet their 2020 investment plans. For publicly owned firms, the 2020 budget
amendment anticipates a drop of 28 percent in investment relative to pre-pandemic plans. 15

12
Morocco scored more than 90 points on the 0-to-100 scale. See the Coronavirus Government Response
Tracker: https://www.bsg.ox.ac.uk/research/research-projects/coronavirus-government-response-tracker. The
medical emergency was declared on March 19, 2020 and extended twice until June 9, 2020.
Such forward-looking judgmental adjustments are done through so-called “tunes,” where we override the
13

model simulations with an explicit value or impose a shock in future periods of the simulations.
14
See “2 ème enquête sur l’impact de la Covid-19 sur l’activité des entreprises“, HCP, July 2020;
https://www.hcp.ma/Reprise-d-activite-des-entreprises-suite-a-la-levee-du-confinement_a2578.html.
15
See https://www.finances.gov.ma/Publication/db/2020/np-plfr2020-fr.pdf.
 17

Despite government financial transfers that boosted household incomes, household

consumption is expected to decline by more than 5 percent year on year as projected by HCP.

External factors
The slowdown in trading partner countries interrupted supply channels for Moroccan
manufacturers, especially in the automotive sector. Lack of international travel hit tourism
hard; for 2020 revenues were expected to fall by 60–70 percent. Recessions in host countries
displaced Moroccan workers; as a result, compared to 2019 remittances in 2020 were
expected to be down 25 percent. 16 The COVID shock was mostly disinflationary in
developed countries, thus pushing down the contribution of imported inflation to headline
CPI. Figure 5 shows projections for the main foreign sector and commodity market variables

Figure 5. Projected External Developments, 2020–30

(Deviation from the control)

Notes: All projections were extracted from the databases in Spring 2020.
Sources: Global Projection Model Network; World Bank; IMF World Economic Outlook.

Commodity markets
The impact on Morocco’s terms of trade was balanced. For exports, the international price of
phosphates, its main export commodity, declined by about 15 percent in 2020. As for
imports, in 2020 oil prices were lower by about 33 percent and food prices by about 7 percent
(Figure 3).

16
The assumption of 25 percent decline in remittances is hypothetical and linked to preliminary information at
the time of writing. Recent data show a greater resilience in Moroccan remittances.
 18

Budget developments
Evolution of the fiscal balances reflects four major developments: (1) As economic activity
slowed, so did revenue collection, especially nontax revenue and consumption taxes. (2)
Spending on health went up, mostly for COVID prevention, testing, and treatment. (3)
During the lockdown, cash transfers to vulnerable households jumped by almost 1.5 percent
of GDP. (4) The government cut most nonurgent expenditures. Overall, we estimate that the
fiscal impulse only partly offset the shortfall in private demand—at the cost of pushing the
projected debt-to-GDP ratio to about 75 percent in 2020.

B. Scenario Assumptions

This scenario incorporates assumptions that affect initial conditions and near-term
developments, as shown in Table 3. Particularly significant is our assumption about the fiscal
stimulus in 2020 and 2021: In 2020, the Moroccan government chose to mitigate the
economic and social impact of the COVID crisis primarily through cash transfers and
guarantees. 17 At the date of this writing, there is considerable uncertainty as to the stance of
fiscal policy for 2021, we thus assume a generally neutral fiscal stance that would stabilize
public debt at about 75 percent of GDP, with transfers being the residual item in our
simulations. Please note that this frequently used modeling choice to balance the fiscal block
is only one of many possible spending strategies, it is not a policy recommendation.

C. Simulation Results

The impact of the pandemic on Morocco crucially depends on the assumptions about its
duration in the industrial countries (as outlined in Section IV.B and summarized in Figure 5).
Because EU and US demand for imports is expected to stay depressed until the late 2020s
and the phosphate price is not expected to recover until 2025, the impact on the Moroccan
economy is likely to be long-lasting (see Figure 6 and Annex Figure A.2). The instantaneous
drop in real GDP—without the fiscal package adopted by the Moroccan government—is
projected to be equivalent to some 10 pp below the control, 18 and the fiscal response would
limit the decline to about 6 pp below the control. In 2020–21 the fiscal stimulus would push
public debt some 10 pp above the control, to about 75 percent of GDP, delaying fiscal
consolidation. Note that while real GDP is expected to be below its potential over the
simulation horizon, in 2021 economic growth is expected to again turn positive. Given the
magnitude of the shock, however, the rate of growth would not be sufficient to bring real
GDP back to its pre-COVID-19 level by 2030.

17
See the cover note of the amended Finance Law for 2020;
https://www.finances.gov.ma/Publication/db/2020/np-plfr2020-fr.pdf.
18
The steady-state rate of growth of real GDP was calibrated at 3.5 percent in the control.
 19

Table 3. The COVID-19 Pandemic: Scenario Assumptions and Judgment Calls

(Deviation from the control)

Magnitude
Variable Comments
in 2020
External environment 19
Gradual recovery with demand; in 2030 still below
EU demand for imports –9.7%
the control by 2 percent.
Gradual recovery with demand; in 2030 still below
US demand for imports –8.3%
the control by 2 percent.
Inflation is expected to pick up starting in 2022,
US CPI inflation –0.9 pp
with the inflation gap largely closed by 2024.
Deviation from the control peaks in 2021 (–1.9 pp)
US Fed funds rate –1.2 pp
and dissipates by 2027.
World oil price –33% Oil prices are expected to recover by 2022.
World food prices –7% Food prices are expected to fully recover by 2025.
World phosphate prices –15% Phosphate prices are expected to recover by 2025.
Perception of risk
Sovereign premium +100 bp A temporary shock, largely dissipating in 2021.
The corporate risk premium goes gradually back to
Corporate risk premium +100 bp
zero over the next five years.
Retail risk premium +100 bp A temporary shock, dissipating by 2022.
Supply-side judgments
Unemployment rate +8.5 pp Model-determined after 2020.
Labor participation rate 20 –4 pp Model-determined after 2020.
Total factor productivity in the
–5.2% Expected to mostly recover by 2022.
nonagricultural sector

19
External environment judgments are based on forecasts from the Global Projection Model Network, the
source of global economy projections for the BAM. The judgments are derived from revision of these forecasts
between January and June 2020. The only exception is the forecast for phosphate prices, which is based on
revision of World Bank projections for October 2019 through June 2020.
20
Although not considered in the presented scenarios, the model could also be used to simulate a hypothetical
scarring effect of the pandemic in which case the decline in labor force would be more persistent (even
permanent in the extreme case).
 20

Magnitude
Variable Comments
in 2020
Demand-side judgments
Households’ marginal
–3% Model-determined after 2020.
propensity to consume
Model-determined partial recovery starting in
Remittances –25%
2022.
Model-determined after 2020. The shock is in
addition to weaker export demand from abroad
Idiosyncratic drop in export
–10% because lockdowns and travel restrictions have
demand
had a more pronounced impact on the export
sectors (e.g., tourism).
Real private investment –10% Model-determined after 2020.
Fiscal judgments
Public deficit-to-GDP ratio +4 pp Model-determined after 2020.
Fiscal consolidation delayed beyond 2021; debt
Delayed
Public debt-to-GDP ratio kept at 75% of GDP (debt ceiling) in 2021.
consolidation
Model-determined after 2021.
Public consumption-to-GDP
+0.2 p.p. Model-determined after 2020.
ratio
Public investment-to-GDP
+0.8 p.p. Model-determined after 2020.
ratio

Source: MOPAM; authors’ calculations.

The simulation results for output are predicated on a fiscal multiplier of about 0.4 because the
2020 fiscal stimulus package was implemented mostly through transfers, which typically
have a higher short-term multiplier than the spending multipliers of about 0.1–0.3
documented in Batini et al. (2014). The multiplier assumption and the nature of the stimulus
package explain the substantial difference in GDP between the scenarios with and without
fiscal measures. In addition, in the scenario without fiscal measures, the government would
cut transfers to offset the revenue loss and stabilize public debt, effectively introducing an
additional negative fiscal impulse. Furthermore, one would expect that a lack of government
response to the pandemic would lead to high uncertainty that is not explicitly modeled in the
MOPAM: a wave of bankruptcies, financial instability, and permanent job losses.
 21

Figure 6. Morocco: The Impact of the COVID-19 Pandemic, 2020–30

(Deviation from the control unless noted otherwise)

Notes: The control simulation is a “no-policy-change” and “no-COVID” scenario in which the steady-state
parameters hold constant at either their historic averages or the latest expert-judgment values. The baseline
simulation (blue line labeled “with fiscal measures”) encompasses all the elements of the COVID-19 pandemic
scenario: (1) the external shock; (2) domestic effects of the lockdown; (3) monetary easing; and (4) the fiscal
stimulus as of mid-2020. The simulation labeled “without fiscal measures” (red dashed line) encompasses only
elements (1), (2), and (3).
Source: MOPAM; authors’ simulations.

The simulation paints a picture of massive simultaneous negative demand and supply shocks.
In the no-fiscal-package simulation, all demand components decline precipitously, especially
remittances from Moroccans abroad, which are assumed to drop by a staggering 25 percent
relative to the control. Labor supply, which declined as a result of the lockdown, recovers
only slowly. Private investment dives; it is expected to stay below the control until 2024.
What is the appropriate response of monetary policy and how should we model it? Should we
tune the policy response so that it is identical with and without fiscal policy options? Or
should we leave the policy response to be fully state-dependent? And should we apply
judgments and impose, for example, a limit on the lower bound of the policy response? Each
of these options could be justified by alternative objectives for the simulations. Since the
objective of the paper is to demonstrate MOPAM capabilities, we chose to let the policy
response to be fully endogenous, to reflect the state of the economy. Evolution of the policy
 22

interest rate and the exchange rate therefore reflects only the current policy reaction function
and its calibration.
Consistent with the scenario of a massive demand shock and a sluggish recovery is a decline
in the policy interest rate toward the zero lower bound and some depreciation of the dirham.
Given the persistence of the policy rate in the reaction function and the observed increase in
the country risk premium in 2020, in the scenario with fiscal measures the decline in the
policy rate is about 50 basis points relative to the control, with an additional 100 basis points
in 2021, when the risk premium shock is expected to dissipate.
Predictably, endogenous monetary easing is larger in the scenario without fiscal measures
because the much larger negative output gap at the beginning of the simulation necessitates
further rate cuts to stabilize the economy. Without fiscal measures, the policy rate would be
expected to decline to zero—a somewhat unrealistic prediction given the policymaker’s
preference for keeping a healthy positive interest rate differential vis-à-vis the euro and the
US dollar. Morocco has little the room for monetary policy maneuver because of its
exchange rate regime and limited international reserves.
Furthermore, the endogenous monetary policy stance should be interpreted with caution as
the BAM has executed monetary policy through channels in addition to the policy rate in its
multi-dimensional framework. For example, movements in the required reserve ratio have
been used more often than interest rate adjustments (Benlamin and others 2018). The
MOPAM, however, does not consider a nexus between required reserves, lending conditions,
and the credit risk premium. This is a problem common to most DSGE models: the simulated
endogenous policy rate path does not map the headline rate exactly and it should be
interpreted as a general direction of loosening or tightening. In other words, we interpret our
simulations as monetary loosening that can be effected through a combination of policy rate
cuts and other instruments at the disposal of the BAM.
The nominal exchange rate is projected to depreciate by about 1 pp against the control—still
well within the current ±5 percent fluctuation band of the dirham. Such a mild depreciation is
supported by two policy decisions. 21 First, a higher level of international reserves, following
the PLL purchase, that created a credible buffer to support the current exchange rate
framework. It is worth mentioning that no FOREX interventions were carried out by the
BAM during 2020, however. Second, a 10-pp increase in import tariffs, primarily designed to
discourage imports of consumption goods, improve the trade balance, and thus limit short-
term depreciation pressures on the dirham. Although the trade balance is expected to improve
initially relative to the control as imports in local currency terms decline by more than
exports in local currency terms, it is projected to worsen in 2021 and stay below the control
until 2025.
The projected fiscal response assumes accumulation of additional debt in response to the
pandemic, which will be gradually brought back to 66 percent of GDP over the next 10 years

21
On April 7, 2020 the Moroccan authorities used all available resources (equivalent to about US$ 3 billion)
under the Precautionary and Liquidity Line (PLL) arrangement with the IMF, thus creating a buffer to support
the current exchange rate framework. The March 2020 budget amendment proposed increasing import duties
applicable to a range of finished consumer products from 30 percent to 40 percent, with an explicit objective of
protecting COVID-affected industries and alleviating the pressure on international reserves:
https://www.finances.gov.ma/Publication/db/2020/np-plfr2020-fr.pdf.
 23

or so. This results in public debt that in 2021–24 is higher on average by 3–5 percentage
points than in the no-fiscal-response simulation. The fiscal response is a combination of
announced fiscal measures and expert judgment about gradual removal of the stimulus.
Regarding the former, we incorporated measures announced and applied by mid-2020.
Regarding the latter, we assumed that the implicit debt target of 65 percent of GDP will be
temporarily disactivated in 2020–21 and reactivated only in 2022.
The MOPAM simulations suggest a difficult tradeoff. On one hand, a gradual withdrawal of
the stimulus would limit the short-term negative impact on growth. Of course, higher public
debt will crowd out private investment and will come at the cost of a somewhat higher
country risk premium and a higher neutral real interest rate that could further depress private
investment and long-term growth. 22 We can see this channel operating in Figure A2: in the
no-fiscal-response simulation real private investment is 2–3 pp higher than in the fiscal-
response simulation. On the other hand, a faster fiscal consolidation that would stabilize
debt-to-GDP ratio at 66 percent in 2–3 years would limit debt sustainability risks and bring
down the neutral real interest rate, thus supporting private investment. However, we need to
remind ourselves of the central lesson from the GFC: excessive austerity at the time of crisis
is often counterproductive—in our simulations output would stay below potential for longer.

VI. CONCLUSIONS

Here we describe an approach to simulating macroeconomic variables using MOPAM, a

large-scale macroeconomic, semi-structural model developed at the Bank Al Maghrib.
MOPAM is based on the IMF’s Flexible System of Global Models, which has been adjusted
to single-country setup to capture the main features of the Moroccan economy and policy,
striking a balance between economic theory, empirics, and policy idiosyncrasies. The model
is a mix of traditional DSGE segments and ad hoc, data-driven segments, with the choice of
segments driven by their ability to provide a clear economic interpretation. We present the
model and demonstrate its policy simulation capabilities using two scenarios, one analyzing
the long-term effects of fiscal consolidation and the other providing an early assessment of
the impact of the COVID-19 pandemic and the attenuating role of the fiscal policy measures
adopted. The simulations in the paper describe theoretical scenarios based on a set of
assumptions and expert judgments and should not be interpreted as either Moroccan
authorities or IMF forecasts.
The fiscal consolidation scenario summarizes the assumptions and then simulates the impact
of the framework proposed at the 2019 National Conference on Taxation, finding that
consolidation has a generally positive impact. The starting assumptions include (1) adopting
a credible Framework Law; (2) splitting gains from additional revenue and removal of the
fuel subsidy between funding the social safety net and reducing income tax rates;
(3) simplifying the complicated system of multiple VAT rates; (4) increasing the
progressivity of the tax system; (5) widening the tax base; and (6) addressing tax exemptions

See Engen and Hubbard 2004, Jaramillo and Weber 2012, and Schumacher and Żochowski 2017 for
22

empirical estimates of this channel.

 24

and leakages. The scenario assumes that there are no changes to the current monetary and
exchange rate system.
The outcome of this scenario has the usual features of a short-lived Keynesian contractionary
stabilization: government deficits fall, financed by permanently higher revenue and
temporary cuts in spending, and the gross debt-to-GDP ratio falls to the target of 60 percent.
On one hand, the reduction in public indebtedness and a switch from direct to indirect taxes
generate efficiency gains and lead to a substantial increase in private investment. On the
other, removal of the fuel subsidy lowers disposable income, thus limiting investment. The
total medium-term impact on output is positive and large. The permanently lower debt-to-
GDP ratio lowers the country risk premium and creates space for eventual expansion of
targeted social transfers.
The COVID-19 pandemic scenario encompasses the authorities’ actions during the first half
of 2020 and makes assumptions about short- and medium-term developments, finding that
even with a well-coordinated fiscal-monetary mix, the return to the pre-COVID level of real
GDP will take several years, with a lasting increase in public debt. The starting assumptions
include the initial demand shortfall and medium-term supply-side disruptions. Productivity
slowdown and disruptions in the supply chains would depress potential GDP growth for
some time.
External developments drive the impact of the COVID pandemic on the Moroccan economy;
demand for Moroccan imports and tourism is expected to remain depressed. The demand
developments are amplified by a combination of supply shocks. A two-year stimulus in 2020
and 2021 would push public debt 10 pp above the control, to about 75 percent of GDP. The
fiscal measures introduced by mid-2020 softened the blow from the negative demand shock,
shielded liquidity-constrained households from deprivation, and prevented financial sector
scarring. Still, real GDP growth in 2020 is estimated at about 6 pp below the pre-COVID
control.
The pandemic scenario is consistent with a decline in the monetary policy rate and a small
depreciation, well within the current fluctuation band. The key policy dilemma is the how
quickly to withdraw the fiscal stimulus after 2021: a gradual withdrawal will limit the short-
term negative impact on growth but at the cost of a higher country risk premium, smaller
private investment, and lower long-term growth. In contrast, faster fiscal consolidation would
limit debt sustainability risks but keep output below its potential longer. The dilemma cannot
be decided in a macroeconomic model; it needs to be addressed by the public authorities,
though credible modeling and well-specified scenarios may help guide the discussion.
MOPAM has proved its worth as an analytical tool for a variety of policy and structural
simulations. The BAM has at its disposal a rich and flexible model that can be used for a
wide range of theoretical scenarios, two of which are demonstrated here. The main advantage
of MOPAM is that it can be readily adapted to changes in the fiscal, monetary, and exchange
rate regimes. Such adaptability will be crucial in assessing the impact of the numerous
reform challenges that lie ahead.
 25

REFERENCES
Alesina, Alberto, and Silvia Ardagna, 2010, “Large Changes in Fiscal Policy: Taxes versus
Spending,” Tax Policy and the Economy, Vol. 24, pp. 35–68.
Andrle, Michal, Patrick Blagrave, Pedro Espaillat, Keiko Honjo, Benjamin Hunt, Mika
Kortelainen, René Lalonde, Douglas Laxton, Eleonora Mavroeidi, Dirk Muir, Susanna
Mursula, and Stephen Snudden, 2015, “The Flexible System of Global Models –
FSGM,” IMF Working Paper 15/64. (Washington: International Monetary Fund).
Available at: https://www.imf.org/~/media/Websites/IMF/imported-full-text-
pdf/external/pubs/ft/wp/2015/_wp1564.ashx.
Achour, Aya, and Omar Chafik, 2019, “Salaire minimum au Maroc: faits stylisés et impacts
économiques,“ Bank Al-Maghrib Working Papers Series, No 2019-3. (Rabat: Bank Al-
Maghrib:). Available at:
http://www.bkam.ma/content/download/683186/7841547/SMIG_Maroc_dt_final.pdf.
Benlamine, Mokhtar, Aleš Bulíř, Meryem Farouki, Ágnes Horváth, Faical Hossaini, Hasnae
El Idrissi, Zineb Iraoui, Mihály Kovács, Douglas Laxton, Anass Maaroufi, Katalin
Szilágyi, Mohamed Taamouti, and David Vávra, 2018, “Morocco: A Practical Approach
to Monetary Policy Analysis in a Country with Capital Controls,” IMF Working Paper
18/27. (Washington: International Monetary Fund). Available at:
http://www.imf.org/~/media/Files/Publications/WP/2018/wp1827.ashx.
Dinerstein, Marcos, and Jon Huntley, 2020, “Lasting Macroeconomic Impacts of the
Coronavirus Crisis, Absent Fiscal Policy Response,” Penn Wharton Budget Model Blog.
Available at: https://budgetmodel.wharton.upenn.edu/issues/2020/3/27/macroeconomic-
impacts-coronavirus.
Engen, Eric, and R. Glenn Hubbard, 2004, “Federal Government Debt and Interest Rates,”
NBER Working Paper 10681 (Cambridge, MA: National Bureau of Economic
Research). Available at: http://www.nber.org/papers/w1068.
Giavazzi, Francesco, and Marco Pagano, 1996, “Non-Keynesian Effects of Fiscal Policy
Changes: International Evidence and the Swedish Experience,” Swedish Economic
Policy Review, Vol. 3, pp. 67–112.
Guerrieri, Veronica, Guido Lorenzoni, Ludwig Straub, Iván Werning, 2020,
“Macroeconomic Implications of COVID-19: Can Negative Supply Shocks Cause
Demand Shortages?” NBER Working Paper 26918 (Cambridge, MA: National Bureau
of Economic Research). Available at: https://www.nber.org/papers/w26918.pdf.
Hutton, Eric, 2017, “The Revenue Administration – Gap Analysis Program: Model and
Methodology for Value-Added Tax Gap Estimation,” FAD Technical Notes and
Manuals 17/04. (Washington: International Monetary Fund). Available at:
https://www.elibrary.imf.org/doc/IMF005/24113-9781475583618/24113-
9781475583618/Other_formats/Source_PDF/24113-9781475592856.pdf.
IMF (International Monetary Fund), 2018, “Morocco: 2017 Article IV Consultation – Staff
Report,” (Washington: International Monetary Fund). Available at:
https://www.imf.org/~/media/Files/Publications/CR/2018/cr1875.ashx.
 26

——, 2019a, Morocco: 2019 Article IV Consultation –Staff Report,” (Washington:

International Monetary Fund) Available at:
https://www.imf.org/~/media/Files/Publications/CR/2019/1MAREA2019003.ashx.
——, 2019b, Annual Report on Exchange Arrangements and Exchange Restrictions 2019,
(Washington: International Monetary Fund). Available at: https://www.elibrary-
areaer.imf.org/Documents/YearlyReport/AREAER_2019.pdf.
——, 2020, “Morocco Draws on Funds Available under the Precautionary and Liquidity
Line to Address COVID-19 Pandemic,” (Washington: International Monetary Fund).
Available at: https://www.imf.org/en/News/Articles/2020/04/08/pr20138-morocco-
draws-funds-available-under-precautionary-liquidity-line-covid19-pandemic.
Jaramillo, Laura, and Anke Weber, 2012, “Bond Yields in Emerging Economies: It Matters
What State You Are In.” IMF Working Paper 12/198 (Washington: International
Monetary Fund). Available at: https://www.imf.org/~/media/Websites/IMF/imported-
full-text-pdf/external/pubs/ft/wp/2012/_wp12198.ashx.
Jordà, Òscar, Sanjay R. Singh, and Alan M. Taylor, 2020, “Longer-run Economic
Consequences of Pandemics,” NBER Working Paper No. 26934 (Cambridge, MA:
National Bureau of Economic Research). Available at:
https://www.nber.org/papers/w26934.pdf.
Keen, Michael, 2013, “The Anatomy of the VAT,” IMF Working Paper 13/111,
(Washington: International Monetary Fund). Available at:
https://www.imf.org/external/pubs/ft/wp/2013/wp13111.pdf.
Obstfeld, Maurice, Jay C. Shambaugh, and Alan M. Taylor, 2005, “The Trilemma in History:
Tradeoffs Among Exchange Rates, Monetary Policies, and Capital Mobility.” Review of
Economics and Statistics, Vol. 87, pp. 423–438.
S&P Global Ratings, 2020, “Economic Research: COVID-19 Deals A Larger, Longer Hit To
Global GDP,” Comments published on April 16, 2020. Available at:
https://www.spglobal.com/ratings/en/research/articles/200416-economic-research-covid-
19-deals-a-larger-longer-hit-to-global-gdp-11440500.
Schumacher, Malte, and Dawid Żochowski, 2017, “The Risk Premium Channel and Long-
term Growth,” ECB Working Paper No 2114 (Frankfurt: European Central Bank).
Available at: https://www.ecb.europa.eu/pub/pdf/scpwps/ecb.wp2114.en.pdf.
 27

ANNEX I

Simulation Results
1. The Fiscal Consolidation Simulation Results

Figure A1. Morocco: Selected Variables, 2020–30

(Deviation from the control unless noted otherwise)

Notes: The control simulation is a “no-policy-change” scenario in which the steady-state parameters remain
constant at either historic averages or the latest expert-judgment values. The baseline simulation encompasses
all the fiscal consolidation elements decomposed into four layers: debt target and wage bill reduction
(Simulation 1); plus VAT reform (Simulation 2); plus income tax reform (Simulation 3); plus fuel subsidy
removal.

Source: MOPAM; authors’ simulations.

 28

2. The COVID-19 Simulation Results

Figure A2. Morocco: Selected Variables, 2020–30

(Deviation from the control unless noted otherwise)

Notes: The control simulation is a “no-policy-change” scenario in which the steady-state parameters remain
constant at either historic averages or the latest expert-judgment values. The baseline simulation (blue line
labeled “with fiscal measures”) encompasses all the elements of the COVID-19 pandemic scenario: (1) the
external shock; (2) domestic effects of the lockdown; (3) monetary easing; and (4) the fiscal stimulus as of mid-
2020. The simulation labeled “with fiscal measures” (red dashed line) encompasses only elements (1), (2), and
(3).

Source: MOPAM; authors’ simulations.

 29

ANNEX II

A T S M P A M

A.1 Model Overview

The Morocco Policy Analysis Model (MOPAM) is derived from the Flexible System of
Global Models (FSGM) which the IMF Research Department developed to facilitate macroe-
conomic analysis of a wide range of issues in open economies that experience major cross-
country spillovers. Working with the FSGM, we customized the original model to the Mo-
roccan economy. Several aspects were altered or added to take into account specific sec-
toral dynamics, capital account regulation, current Moroccan policy, and the importance of
traditional agriculture, and the adaptation paid special attention to phosphate exports. The
model also addresses monetary policy, which in Morocco combines a fixed exchange rate
with a partly open capital account. Tailored to fiscal policy, the MOPAM incorporates en-
ergy subsidies and proactive use of import tariffs to alleviate pressures on the currency.
The eclectic MOPAM approach strikes a balance between theory and empirical insights. It
has four blocks: The core block, fully micro-founded, describes private consumption and
the private investment of overlapping generations (OLG) households. Households optimize
their utility with respect to their budget constraints while accounting for different sources
of wealth, human and financial. The use of OLG households rather than infinitely-lived
ones, breaks down the Ricardian equivalence and removes government debt neutrality. The
model also comprises households that because of their financial situation can neither save
nor borrow; their consumption equals their income from wages and government transfers.
Private investment follows the Tobin’s Q model augmented by real adjustment costs, and
capital stock can be inferred from the law of capital accumulation.
The supply block is semistructural. The production function is Cobb-Douglas, with labor
and private capital as production factors. Total factor productivity (TFP) includes the ef-
fects of accumulation of public capital and the second-round effects of commodity prices to
capture their impacts on aggregate supply. Total labor force is determined by an exogenous
participation rate, and the labor supply is then determined by the unemployment rate, which
is modeled by a reduced-form version of Okun’s law. In the MOPAM core consumer price
inflation is described by a reduced-form, hybrid, open-economy Phillips curve, to reflect the
price stickiness and changes in marginal costs captured by the output gap and developments
in the real exchange rate and oil prices. Non-core prices of food and energy are determined
by world prices and the exchange rate, and for food also by the domestic food supply. Wage
inflation is modeled by a wage Phillips curve.
As for the government block, both revenue and expenditure sides of the government bal-
ance are captured, and different fiscal rules can be explored depending on the government’s
objectives for long-term debt or the long-term deficit. To meet its objectives, the govern-
 30

ment chooses fiscal instruments that affect either the demand side (consumption tax, trans-
fers to households); the supply side (tax on capital); or both (public investment). Distor-
tionary taxes weaken the Ricardian equivalence.
Finally, we model carefully Morocco’s current monetary regime, which is quite different
from the floating or fixed exchange rate regimes that are commonly studied, and which are
accompanied by an open capital account. As this paper was being written, Morocco fixed
its exchange rate against a basket of currencies, with weights equal to 60% for the euro and
40% for the US dollar. With the partly open capital account and active use of import duties,
such a regime opens up maneuvering space to achieve such internal objectives as stabiliz-
ing inflation and growth. The MOPAM can also capture a wide range of intermediate and
hybrid exchange rate regimes and monetary policy reaction functions that correspond to
different stages in making the dirham more flexible.
Before we describe the main equations of the model we must make a crucial comment about
notation: All nonstationary real variables in the equations below are detrended by an exoge-
nous trend in technology (G), all nominal prices by the price of the core consumption bas-
ket (P core) and all other nonstationary nominal variables by the trends in technology and the
core consumption price (P coreG). Also, variables with periods after the current period refer
to rational expectations for the variables in the future period (i.e., for brevity, we omit the
expectations operator). Further, an increase in the exchange rate, whether nominal or real,
always means appreciation of the currency. The MOPAM is an annual model and, hence, all
growth rates, inflation rates and interest rates are annualized rates.

A.2 Private Consumption

To introduce non-Ricardian features into the MOPAM, the model includes two types of
households: those with overlapping generations (OLG) features and those that are liquidity-
constrained (LIQ). The share of the LIQ households in the population is λ LIQ . OLG house-
holds can accumulate wealth and draw it down, but LIQ households consume only out of
current labor income, remittances, and net transfers from the government. Total household
consumption is
Ct = CtOLG + CtLIQ (1)

A.2.1 OLG Households

OLG consumption is a time-varying proportion of wealth with respect to the marginal propen-
sity to consume:
PtC CtOLG = MPC t Wt (2)
The marginal propensity to consume is given by expected future nominal interest rates (rC ),
∗
the tax rate on consumption (τ C ), the tax rate (τ B ) on changes in household net foreign
assets positions (B ∗ ), the probability of dying (ρ) and the stochastic pricing kernel (j):
 31

∗ ( )2
τB ∗
Bt∗ − Bt−1 ρjt πt+1
MPC −1
t =1+ τtC + + C
MPC −1
t+1 (3)
2 PtC CtOLG Zt 1 + rt
( ) σ1 ( )γ σ−1
CtOLG
σ
1 + rtC 1 + τtC
jt = β C OLG
∆G (4)
πt+1 1 + τt+1 Ct−1

The wealth of OLG households has four components: financial wealth (WF ), capital wealth
(WK), human wealth (WH) and other wealth (WO),

Wt = WF t + WK t + WH t + WOt (5)

Financial wealth is the current value of domestic government bonds and net foreign assets
converted to the domestic currency:
( B
) Bt−1 ( B∗
) Bt−1∗
WF t = 1 + rt−1 + 1 + r t−1 (6)
πtcore∆G Zt πt∗ ∆G

Capital wealth is the value of accumulated capital at the beginning of the current period; the
price of the capital (QR) is the Tobin’s Q:
Kt−1
WK t = QRt (7)
∆G

Human wealth is the present value of expected future labor income after tax, dis-counted
by both the OLG household’s probability of dying (ρ) and the decline in labor productivity
over the life of the household (χ),
( ) ρ χ ∆G
WH t = 1 − τtL WN t LtOLG + WH t+1 (8)
(1 + rtC )

where τ L is the labor income tax rate, WN is the nominal wage, and L OLG is the supply of
labor by the OLG household.
Other wealth is the present value of current and future lump-sum transfers from the gov-
ernment, both general (TF ) and targeted (TF OLG )), less lump-sum taxes (TAX ls ), royalties
from phosphate production paid to the government (RY LT ) and remittances received from
abroad (REMIT OLG )
( )
WOt = (1 − λ LIQ ) TF t − TAX tls + TF tOLG − RY LT t + REMIT tOLG
(9)
ρ ∆G
+ WOt+1 .
(1 + rtC )
 32

A.2.2 Liquidity-constrained Households (LIQ)

Liquidity-constrained households consume up to their net labor income, net government

transfers, and remittances from abroad:
( ) ( ) ( )
1 + τtC PtC CtLIQ = 1 − τtL WN t LtLIQ + WN tAgr LtAgr + TF tLIQ + λ LIQ TF t − TAX tls
(10)
The MOPAM assumes that some LIQ households work in the traditional agriculture sec-
tor (TAS), which is not subject to income taxes. LIQ household income therefore includes
income from work in TAS (W Agr L Agr ).

A.3 Aggregate Supply

A.3.1 The Traditional Agricultural Sector (TAS)

The TAS produces unprocessed food, uses only labor as an input, is not subject to business
cycles, and is not the target of any special transfers,

YtAgr = AtAgr LtAgr . (11)

All TAS labor is provided by LIQ and the share of such households employed in TAS is
λ Agr . Thus, the shares of OLG and LIQ households in non-TAS labor are

1 − λ LIQ λ LIQ (1 − λ Agr )

and . (12)
1 − λ LIQ λ Agr 1 − λ LIQ λ Agr

A.3.2 Nonagriculture Sectors

Nonagricultural production follows the Cobb-Douglas production function

YtExAgr = At Kt−1
αt
L1−α
t
t
, (13)

The labor share αt in equation 13 is time-varying in order to capture the observation that the
wage share increases during cyclical upswings.

A.3.3 Private Capital and Investment

Private capital evolves according to the capital accumulation equation:

Kt−1
Kt = (1 − δ) + It . (14)
∆G
 33

The shadow price of investment (Tobin’s Q) is based on the after-tax return on the invest-
ment (PRK ) adjusted by tax deductions for depreciation,
( )
1 − τt+1
K
t+1 + τt+1 δ QRt+1 + (1 − δ) QRt+1
PRK K
QRt = , (15)
1 + rtcorp

where δ is the depreciation rate of capital and τ K is the capital tax rate, and r corp is the cor-
porate interest rate.
The supply of investment is then given by the following equation:
( ( )2 ( ))
I (I
t t − I t−1 ) ρ∆Gπ t+1 I It+1 It+1
QRt = PtI + q1 PtI − P
corp t+1
−1 (16)
It−1 1 + rt It It

The equation assumes investment adjustment costs of

( )2
q1 It
−1 (17)
2 It−1

A.3.4 Labor Supply

For use in the MOPAM, potential labor supply is defined with respect to the nonaccelerat-
ing inflation rate of unemployment (NAIRU),
( )
Lt = 1 − U t LF t . (18)

where U is the NAIRU and LF is labor force.

Because the TAS labor supply is exogenous and acyclical,

LtAgr = λ LIQ λ Agr LF t (19)

,
the labor supply in non-TAS sectors is given by:
( )
Lt = (1 − Ut ) LF t − LtAgr , (20)

1 − λ LIQ
LtOLG = Lt , (21)
1 − λ LIQ λ Agr
λ LIQ (1 − λ Agr )
LtLIQ = Lt . (22)
1 − λ LIQ λ Agr
 34

A.4 Business Cycle Features

A.4.1 The Phillips Curve for Core Inflation

The MOPAM assumes a sticky-price Phillips curve for core CPI inflation with forward-
looking inflation expectations; the output gap as a measure of domestic producer marginal
costs; and depreciation of the real exchange rate as a measure of importer costs and a spe-
cific term for pass-through from oil price inflation (after oil subsidies):
( )
( core ) ( core ) Yt
log (πt ) =c1 log πt−1 + (1 − c1 ) log πt+1 + c2 log
core
Yt
( ) (23)
πtoil π
+ c3 ∆ log (Zt ) + c4 log core
+ εt
πtarget

A.4.2 Domestic Oil Prices

Domestic oil prices are determined by international oil prices and imposed oil subsidies.
The oil subsidy costs are:
( )
∗,oil
P
GSU Btoil = t
− Ptoil Ctoil. (24)
Zt

In the MOPAM domestic oil prices after subsidy (P oil) can be determined in two ways:
They are either directly set at a government-defined price or endogenously adjusted, de-
pending on the total costs of the subsidy the government is willing to absorb (GSU B oil)

A.4.3 Domestic Food Prices

Domestic food prices depend on(domestically produced

) food, YtAgr , which is inelastic to
prices, and net imports of food, Mtfood − Xtfood :

Pt∗,f ood ( food )

PtfoodCtfood = PtAgr YtAgr + Mt − Xtfood , (25)
Zt

where PtAgr is the price of domestically produced food and Pt∗,f ood is the price of traded
food.
 35

A.4.4 The Wage Phillips Curve

Wage inflation is modeled by a wage Phillips curve linking wages to the output gap and the
time-varying share of labor in production (1 − α):
( ) ( W ) ( ) ( W )
log πtW = cW 1 log πt−1 + 1 − c1
W
log πt+1
( ) ( ) (26)
W Yt W 1−α W
+ c2 log + c3 log + επt
Ȳt 1 − αt

The output gap in the equation captures the procyclical nature of wages, and the labor share
term ensures the long-term stability of the labor share.

A.4.5 Cyclical Unemployment and Okun’s Law

Because the NAIRU is exogenous, cyclical unemployment is based on Okun’s law,

( )
( ) Yt
Ut = U t + c1 Ut−1 − U t−1 − c2
U U
, (27)
Ȳt

and cyclical unemployment is negatively linked to the output gap.

A.5 Monetary Policy, the Exchange Rate Regime and Capital Controls

The MOPAM is flexible in order to capture alternative monetary and exchange rate regimes,
the degree of capital controls, and residents’ access to foreign assets.
The MOPAM assumes first that the monetary policy interest rate, rtM P , follows the forward-
looking reaction function, which takes into account rate-setting inertia, the neutral interest
rate, the inflation differential, and the output gap. When the monetary policy pursues also
exchange rate stabilization objective and capital account is not fully open (more details be-
low), the interest rate rule also takes into account an interest rate differential stemming from
capital account restrictions (Cprem).
( ( )cr2 ( )cr3 )(1−cr1 )
( MP
) ( )r
M P c1
( ) πtcore Yt
1 + rt = 1 + rt−1 1 + rtN eutral core
πtarget Ȳt
(28)
∑
6
MP
(1 + Cpremt+i ) ∗ exp(εrt )
i=1

If the central bank also chooses to control exchange rate movements while preserving (some
degree of) monetary policy autonomy, it needs to adopt capital account restrictions. The
 36

restrictions cause a discrepancy in the uncovered interest rate parity (UIP) that is captured
by Cprem. The UIP conditions with capital restrictions are
∗ −B ∗ π ∗ ∆G
(Bt+1 t) t
St B ∗ C C OLG Z
Pt+1
1 + rtM P = (1 + Cpremt ) (1 + rt∗ ) τ t+1 t+1
∗ exp(εSt ) (29)
St+1 Bt∗ (Bt∗ −Bt−1
∗
)
1+ PtC CtOLG Zt

where the last term results from restrictions on changes in households net foreign asset
(NFA) positions. The parameter capopen determines the extent of tolerated deviations of
exchange rate (S) from the targeted value achieved by imposing capital account restric-
tions. The effect of restrictions is an endogenous “premium” (Cprem) which weakens the
relationship between the domestic and foreign returns on capital. The trade-off between the
exchange rate stabilisation objective and resulting premium is captured by the equation:
( )(1−capopen) ( )capopen
St 1
1= . (30)
S target 1 + Cpremt

When the capopen is set to zero, the model operates as a closed capital account regime with
a currency peg, effectively decoupling domestic and foreign interest rates (as if the UIP
condition was not part of the model). When the capopen is equal to one, the capital account
is fully open, the UIP condition holds, and thus Cprem equals zero at all times.
It is worth mentioning that capopen does not represent any specific measures (e.g. taxes or
administrative restrictions). It rather represents a weight the monetary policy puts on ex-
change rate stabilization achieved by imposing unspecified capital control measures. Stable
capopen does not imply stable capital flow restrictions either. The implicit capital control
measures can be changing with stable capopen depending on exchange rate pressures.

A.6 Fiscal policy

A.6.1 Debt Accumulation

The government budget constraint is

Bt−1 Bt−1
Bt = + DEF t = + EXP t − REV t . (31)
∆Gπtcore ∆Gπtcore

A.6.2 Government Expenditures

Expenditures (EXP ) are broken into seven categories: government consumption (GC);
public investment (GI); general subsidies (GSU B); transfers targeted to OLG and LIQ
 37

households (ST F OLG and ST F LIQ ); general (not targeted) transfers (GT F ); and interest
payments (IN T P ):
EXPt = GCt + GIt + GSU Bt + ST FtOLG + ST FtLIQ + GT Ft + IN T Pt . (32)

Nominal interest payments are given by

B Bt−1
INTP t = rt−1 . (33)
∆Gπtcore

The model takes into account the government energy subsidy system, GSU B, which is de-
fined in equation 24.
The other expenditure items—government consumption, public investment, specific trans-
fers to OLG and LIQ households, and general targeted transfers—are treated as a fixed pro-
portion of GDP unless otherwise specified in scenario simulations.

A.6.3 Government Revenues

Similarly, revenues are broken into seven components: taxes on consumption (T AX c );

labor (T AX l ) and capital (T AX k ) income; tariffs on imports (T AX m ); lump sum taxes
paid by OLG and LIQ households (T AX ls, OLG and T AX ls, LIQ ); and phosphate royalties
(RY LT phos ).
ls, OLG
REV t = TAX C L K M
t + TAX t + TAX t + TAX t + TAX t
(34)
+ TAX tls, LIQ + RY LT phos
t .

The individual categories are treated as a product of the relevant tax rate and tax base:
TAX C C
t = τt Ct (35)
TAX Lt = τtL Wt Lt (36)
( K ) Kt−1
TAX K K
t = τt PRt − δ Qt (37)
∆G
TAX M M M
t = τt Pt Mt (38)

Except for the tax on imports, the tax rates are exogenous; the import tax rate adjusts en-
dogenously in order to stabilize the real exchange rate:
M M
τtM = τ̄ M − cτ ∆ log (Zt ) + ετt . (39)

Lump sum taxes are divided proportionally between the OLG and LIQ households:
TAX ls, OLG = (1 − λ LIQ ) TAX tls (40)

TAX ls, LIQ = λ LIQ TAX tls . (41)

 38

A.6.4 Fiscal Rules

The government targets a constant debt-to-GDP ratio (Btrat,tar). Overall and primary deficit-
to-GDP ratios consistent with the debt target are thus:

Btrat,tar rat,tar
DEF rat,tar
t = Btrat,tar − core
+ εDEF
t , (42)
∆Gπtarget

Btrat,tar rat,tar
PDEF rat,tar
t = DEF rat,tar
t − rtb core
+ εPDEF
t (43)
∆Gπtarget

The MOPAM has two main alternative fiscal policy rules. Policy can follow the rule for the
overall deficit:
( )
Yt ExAgr ( )
rat rat,tar
DEF t = DEF t − 100 ∗ c1 log
DEF
− c2
DEF
Bt − Bt
rat rat,tar
(44)
ȲtExAgr

or the rule for the primary deficit:

( )
YtExAgr
PDEF rat
t = PDEF rat,tar
t − 100 ∗ cPDEF
1 log
ȲtExAgr (45)
( )
− cPDEF
2 Btrat − Btrat,tar .

Both rules aim at keeping government debt as targeted while allowing for countercyclical
fiscal policy.
Because the fiscal rule on the overall or the primary deficit determines the total budget
envelope, the MOPAM includes simple policy rules for revenue and expenditure compo-
nents. Most of the tax rules assume constant rates, and expenditure rules assume constant
spending-to-GDP ratios. For the fiscal block to work with either of the deficit rules, one
of the budget components is always endogenously determined by the deficit rule and the
rules on all other budget components. The choice of which budget component is treated as
endogenous depends on the simulation.

A.7 External Sector

A.7.1 Current Account Balance and Net Foreign Assets

The current account balance, CBt , is a sum of the trade balance, T Bt , net remittances in-
flow (REMIT t ), and the current value of the previous-period interest on net foreign assets
(B ∗ ):
 39

( ∗
)
B∗ St−1 Bt−1
CBt = T Bt + REMIT t + rt−1 ∗ , (46)
St Zt−1 πtcore∆G

The export and import components of the trade balance will be explained later.
Remittances are modeled as
(( )cREMIT )1−cREMIT
cREMIT Zt 2
1
( )
REMIT t = REMIT t−1 1
Act∗t cREMIT
3 exp εREMIT
t , (47)
Z̃t

where Z̃t is a five-year moving average of the real exchange rate, Zt , centered at the current
year t. Act∗t measures the extent of foreign economic activity and is defined as the weighted
average of the outputs of foreign countries. The formulation of equation (47) makes remit-
tances increase in foreign currency terms when the domestic currency appreciates (relative
to the moving average exchange rate, Z̃) or foreign activity goes up (which implies that the
incomes of the Moroccan diaspora also go up).
The NFA accumulation process assumes that the assets depend on the current account bal-
ance and the last-period value of NFA:

( ∗
)
Bt∗ St−1 Bt−1
= CBt + . (48)
Zt St Zt−1 πtcore∆G

A.7.2 Exports

The MOPAM covers three export products: manufactured goods (X man ), food items (X f ood ),
and phosphate (X phos ):

man food phos

PtX Xt = PtX Xtman + PtX Xtfood + PtX Xtphos. (49)

Manufactured Exports

∆ log (Act∗t ) − cX
man man
∆ log (Xtman ) = cX
1 2 ∆RCIt
( ( ) )
∗
man man Act man
+ cX
3 −cX
4 RCIt−1 + log t−1
man
+ cX
5 (50)
Xt−1
man
+ ∆εX
t ,
where Xtman denotes manufactured exports, Act∗t is the foreign activity index, and RCI
denotes the manufactured goods competitiveness index (see equation 56).
 40

Food Exports
( ) ( )
f ood ∗,f ood
∆ log Xtf ood = cX1 ∆ log Actt
( )
X f ood ∗,f ood
− c2 ∆ log RPt
(( ) X f ood Act∗,f ood )
f ood ∗,f ood −c7
+ cX
3 log RPt−1 ∗ t−1
f ood
Xt−1
( ( ) ) (51)
∗,f ood
X f ood Pt−1 X f ood
+ c5 log Agr
+ c8
Pt−1
( ( ) ( ))
∗,f ood ∗,f ood
f ood P P
+ cX
6 log t
− log t−1
PtAgr Pt−1Agr

f ood
+ ∆εX
t ,

where Xtf ood denotes food exports, Act∗,f

t
ood
is the foreign food activity index, RPt∗,f ood
denotes world food prices relative to the world overall price level, Pt∗,f ood denotes world
food prices, and PtAgr denotes prices of domestic food production.

Phosphate Exports
( ) ( )
∗,phos
∆ log Xtphos = cphos
1 ∆ log Actt
( )
∗,phos
− cphos
2 ∆ log RP t
( ( ) ) (52)
( )−cphos Act ∗,phos
∗,phos 4
+ cphos
3 log RPt−1 t−1
phos
+ cphos
5
Xt−1
phos
+ ∆εX
t ,
where Xtphos denotes phosphates exports, Act∗,phos
t is the foreign phosphates activity index,
∗,phos
and RPt denotes world phosphate prices relative to world overall price level.

Foreign Activity Indexes and World Prices

Foreign activity is defined in terms of the trade weights of US and EU imports:

( )
Act∗t = M U S ∗ cU1 S + M EU ∗ cEU
2 . (53)
 41

Similarly, foreign food and foreign phosphate activity indexes are based on food and phos-
phate imports, taking into account the applicable trade weights.
The export prices of manufactured goods are given as:

( ) X man
( )
∆ log PtX,man = cP1 ∆ log PtGDP,ExAgr
( X man
)
+ 1 − cP1 ∆ log (Pt∗,man )
( ) (54)
GDP,ExAgr
X man P t−1
+ cP2 log X,man
Pt−1
X man X man
+ cP3 + εPt ,

where P X,man represents the domestic price of manufactured exports and P ∗,man the world
price of manufactured goods which is defined as:

( ) ( )
log (Pt∗,man )
US EU
= wUshare
S log PtX ∗ ZtU S + share
wEU log PtX ∗ ZtEU
(55)
− log (Zt ) .

The difference relates to the manufactured goods competitiveness index:

( )
PtX,man
RCIt = log . (56)
Pt∗,man

For relative world food prices (RP ∗,f ood ) we differentiate between trend and cyclical com-
ponents: ( ) ( ) ( )
log RPt∗,f ood = log RP ∗,f t
ood
+ log d
RP ∗,f ood
t , (57)
d ∗,f ood is the cyclical component, where:
RP ∗,f ood is the trend component and RP
( ) ∗,f ood ∗,f ood
log RP ∗,f
t
ood
= cRP
1 + εRP
t (58)
and ( ) ( ) ( ∗)
d ∗,f ood d ∗,f ood
RP d ∗,f ood d ∗,f ood
RP Yt
log RP t = c1 log RP t−1 + c2 log
Y ∗t (59)
d ∗,f ood
+ εRP
t .

World phosphate and oil prices relative to the world price level (RP ∗,phos and RP ∗,oil ) are
defined the same way as food prices.
 42

A.7.3 Imports

Imports consist of manufactured good (M man ), food (M f ood ), and oil (M oil ):

manadj f ood oil

PtM Mt = PtM Mtman + P Mt Mtf ood + P Mt Mtoil . (60)

Manufactured Imports

Demand for manufactured goods imports, Mtman, is a function of the domestic activity index
man
Madj
(Actt ), applicable import tariffs, (τtM ), their price after import tariffs, (Pt ), and the non-
agriculture real GDP gap, Yb tExAgr :

( ) ( )
man man M man
∆ log Mtman = cM
1 ∆ log (Actt ) − cM
2 ∆ log Pt adj
( )
man 1 + τtM
− cM
3 log M
1 + τt−1
(( ) (61)
)−m5
man M man
Act t−1 man
+ cM
4 log Pt−1adj + cM
6
Mt−1

∆Yb tExAgr + ∆εM

man man
+ cM
7 t .
man
The price of imported manufactured goods before import tariffs (PtM ) is determined by
the relative import shares of the US and the EU, their real exchange rates, export prices, and
the domestic exchange rate:
( man
) ( US ) ( EU )
log PtM = mshare
US log P t
X
∗ Z US
t + mshare
EU log P t
X
∗ Zt
EU

(62)
M man M man
− log (Zt ) + cP + εPt .

Oil Imports

Demand for oil imports, Mtoil, is modeled similarly:

( ) oil
( )
M oil
∆ log Mtoil = cM 1 ∆ log (Act t ) − c 2 ∆ log Ptoil
( ( ) )
( ) Actt−1
M oil M oil M oil
+ c3 −c4 log Pt−1 + logoil
oil
+ c5 (63)
Mt−1

∆Yb tExAgr + ∆εM

oil oil
+ cM
6 t .
 43

Food Imports

The food import function also incorporates the effects of prices of domestic food production
(PtAgr ):
( )
( ) food food P ∗,food
∆ log Mtfood
= cM1 ∆ log (Actt ) − cM 2 ∆ log t
Zt
( ( ) ( )
∗,food ∗,food
food food P food P
+ cM3 −cM4 log t−1
− cM 5 log t−1
Agr
Zt−1 Zt−1 Pt−1
( ) ) (64)
Actt−1
∆Yb tExAgr
food food
+ log food
+ cM 6 + cM 7
Mt−1
( )
∗,food
food P food
− cM
8 ∆ log t
Agr
+ ∆εM t .
Zt Pt

A.7.4 Market Clearing Conditions

The MOPAM has three consumption goods: core, food, and oil. Therefore, total consump-
tion is

PtC Ct = Ptcore Ctcore + Ptfood Ctfood + Ptoil Ctoil. (65)

The total production consists of TAS and non-TAS production

PtY Yt = PtGDP,ExAgr YtExAgr + PtAgr YtAgr , (66)

and must equal aggregate demand:

( ( )2 )
q1 It
PtY Yt = PtC Ct + PtI It 1 + −1 + PtG Gt + PtX Xt − PtM Mt (67)
2 It−1

Consumption of food and oil must satisfy

Ctcore = Mtoil (68)

Ctfood + Xtfood = YtAgr + Mtfood, (69)

and total consumption is the sum of consumption of both types of household:

Ct = CtOLG + CtLIQ . (70)

 44

A.8 Calibration

Estimating the model size of a MOPAM for a country like Morocco is not feasible. In-
stead, we calibrate the model parameters to correspond with Moroccan data. We also use
extensive predefined simulations to assess the behavior of the calibrated model and fur-
ther improve the calibration. The tables below summarize the steady-state values for main
MOPAM variables, the monetary policy rule coefficients and the assumptions about capital
account openness.

Table 4. Real and Nominal Growth Rates, Percent

Variable Value
Real GDP growth 3.500
Nominal GDP growth 5.570

Table 5. Nominal GDP by Expenditures, Percent

Variable Value
Private consumption to GDP 58.60
Private investment to GDP 26.00
Government absorption to GDP 23.50
Trade balance to GDP −8.10

Table 6. Nominal GDP by Source, Percent

Variable Value
Export commodity production to GDP 82.90
Food production to GDP 13.00
Phosphates production to GDP 4.10
 45

Table 7. Consumption Shares, Percent

Variable Value
LIQ households share of consumption 47.82
Oil share of consumption 7.00
Food share of consumption 37.00

Table 8. Government Overview, Percent of Nominal GDP

Variable Value
Government deficit 3.48
Primary government balance 2.29
Public debt 66.00

Table 9. Government Revenues, Percent of Nominal GDP

Variable Value
Tariff revenues 0.80
Tax revenue 21.00
Consumption tax revenue 10.50
Capital tax revenue 5.24
Labor tax revenue 2.96
Phosphate royalties 0.52
Lumpsum tax revenue 0.98

Table 10. Government Expenditures, Percent of Nominal GDP

Variable Value
Government expenditures 24.48
Government consumption 17.00
Government investment 6.50
Government subsidies 0.75
Interest cost 5.77
 46

Table 11. Balance of Payments, Percent of Nominal GDP

Variable Value
Current account balance −4.33
Trade balance −8.10
Remittances 6.10
Net foreign assets −82.16

Table 12. Exports, Percent of Nominal GDP

Variable Value
Total exports 35.00
Manufactured goods 26.00
Food 4.50
Phosphates 4.50

Table 13. Imports, Percent of Nominal GDP

Variable Value
Total imports 43.10
Manufactured goods 32.10
Oil 7.00
Food 4.00

Table 14. Capital

Variable Unit Value

Capital stock % of nominal GDP 179.310
Capital share of income fraction 0.500
Depreciation rate of private capital fraction 0.110
Return on private capital Local currency yield (LCY) per unit of capital 0.243
 47

Table 15. Labor, Percent

Variable Unit Value

Labor force % of total population 40.95
Employment in agriculture % of labor force 40.00
Unemployment rate % of labor force 9.00
Labor age population % of total population 65.00
Participation rate % of labor age population 63.00
Labor income % of nominal GDP 50.00

Table 16. Inflation Measures, Percent

Variable Value
Headline CPI inflation 2.0
Core CPI inflation 2.0
Wage inflation 5.7

Table 17. Monetary policy and capital account strategy

Coefficient Description Value

capopen Capital account openness 0.1
∗
τB NFA adjustment costs 0.6
cr1 MP rule inertia 0.3
cr2 Inflation gap weight 0.2
cr3 Output gap weight 0.0