Economic Modelling 63 (2017) 128–142

Contents lists available at ScienceDirect

Economic Modelling
journal homepage: www.elsevier.com/locate/econmod

Potential socio-economic implications of future climate change MARK

and variability for Nigerien agriculture: A countrywide dynamic
CGE-Microsimulation analysis
⁎
Jean-Marc Montauda, , Nicolas Pecastaingb, Mahamadou Tankaric
a
CATT, Université de Pau et des Pays de l’Adour (UPPA), 8 allées des platanes, 64100 Bayonne, France
b
Universidad del Pacífico, Av. Salaverry 2020, Jesus María, Lima, Perú
c
International Food Policy Research Institute (IFPRI), Dakar, BP 24063, Senegal

A R T I C L E I N F O A BS T RAC T

JEL classification: This study explores the potential economic and social effects induced by a possible future deterioration of
C68 weather conditions on Niger's agriculture. Our scenarios over a 25 year period rely on both long term climate
O13 change deterministic features and climate variability stochastic features which take into account a higher than
Q54 normal frequency of severe droughts. Using a microsimulated Dynamic Recursive Computable General
Keywords: Equilibrium model, we show how a long run mean decrease in main crop yields could adversely affect growth,
Climate change and variability poverty and food security in the country and how more severe drought would worsen these negative effects.
General equilibrium However, we also show that there would appear to be some room for manœuvre for coping strategies for Niger
Agriculture
such as improving the rural road network, adopting modern crop varieties or, to a lesser extent, extending
Poverty
irrigation capacity.
Food security
Niger

1. Introduction
The Nigerien economy depends strongly on its agricultural sector, a
sector that occupies more than 80 percent of the workforce and
contributes to more than 40 percent of GDP. However, this agriculture
is handicapped by its low productivity and its exposition to a number of
risks. Among the latter, the adverse climatic conditions are often
considered to be the most severe (World Bank, 2013). Over the last
few decades, the Sahel region has indeed been hit by rising temperatures,
reduction in average rainfall and higher climate variability reflected, in
particular, in more frequent major droughts (Druyan, 2011; IPCC,
2013). Within this context, the living conditions of the fast-growing
population are particularly precarious, especially in rural areas where
agriculture is the main source of income but also the main source of
food. In 2010, 48,2% of Nigerien households were poors and rural areas
contributed 90% to total poverty with a poverty rate of 54.6%; 48.1%
suffered chronic malnutrition and more than 20% were affected by lack
of food security and identified as being in a severely vulnerable situation
(Herderschee et al., 2014). Noting the apparent consensus in agro-
climate literature that Niger's future weather conditions could deterio-
rate and be characterized by both a drier climate and a higher frequency
of severe droughts (Sylla et al., 2010; Mohino et al., 2011; Roudier et al.,
2011; IPCC, 2013; Sultan 2013; Niang, 2014), the present study seeks to
assess the risk that these possible deteriorations might entail for the
Niger's economy and the welfare of its population. We draw here on
several ways on the studies attempting to assess developing countries’
vulnerability to possible changes in weather conditions.
First, regarding the modelling framework, we chose to build a
Dynamic Recursive Computable General Equilibrium (DRCGE) model
for the Nigerien economy. This framework seems indeed better suited
to an economy-wide analysis of the potential effects of future agro-
climatic shocks than others approaches in the literature (relying on
Ricardian or partial equilibrium analysis). On the one hand, because
the general equilibrium logic does not restrict the effects of the shocks
to agriculture alone but also captures explicitly all the linkages between
prices, income, supply and demand in the whole economy. On the other
hand, because the use of a dynamic specification enables to generate
time paths of the evolutions of the country's economic variables under
different hypotheses for potential shocks over a given period. In the
recent economic literature CGE studies on developing counties clima-
te's vulnerability have known a rapid growth, either in a global or
national framework (see for instance Gebreegziabher et al., 2016, for a

⁎
Corresponding author.
E-mail addresses: jean-marc.montaud@univ-pau.fr (J.-M. Montaud), n.pecastaing@up.edu.pe (N. Pecastaing), M.Tankari@cgiar.org (M. Tankari).

http://dx.doi.org/10.1016/j.econmod.2017.02.005
Received 9 November 2016; Received in revised form 31 January 2017; Accepted 10 February 2017
Available online 20 February 2017
0264-9993/ © 2017 Elsevier B.V. All rights reserved.
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

review). However, to our knowledge, such analysis has never been rainfed subsistence farming, is characterised by structural fragility. The
conducted for Niger. Our DRCGE model involves fairly standard decline in natural resources and in soil fertility, the traditional nature
general equilibrium assumptions for low income countries (Decaluwé of systems of production, the lack of rural infrastructures, the weakness
et al., 2013). But, unlike the majority of other CGE studies, we also of state institutions or rural organizations, the low level of incomes
chose to consider some specific features of the Nigerien economy, which limit farmers’ ability to invest, etc., are all handicaps which have
particularly its dual nature, both in rural areas and urban areas (Stifel a powerful impact on agricultural production and its ability to meet the
and Thorbecke, 2003). Moreover, using a top-down approach, we chose needs of an ever-increasing population (3.5% growth per annum). But,
to link this model to a dynamic microsimulation (DMS) model in order beyond these constraints, which are relatively common in the least
to also assess the potential social outcomes of the future adverse developed African countries, Nigerien agriculture is also constrained by
climatic conditions at the micro level. particularly unfavourable climate conditions. The first reason has to do
Second, regarding the nature of the climatic shocks simulated with with the limit in the amount of arable land: only 12,5% of land is
the models, we chose here to use a three-pronged approach in order to cultivated, mainly in the Sudan-Sahel and Sahel regions, the only areas
isolate the respective potential effects of climate change and climate to have rainfall in excess of 300 mm per year (FAO, 2016a). The second
variability. In a first step, as with a majority of CGE studies (see for has to do with the detoriation of historical climate conditions over the
instance, Juana et al., 2008; Reid et al., 2008; Bezabih et al., 2011; last few decades. On the one hand, the continuous long term effects of
Hertel et al., 2010; Bosello et al., 2013; Calzadilla et al., 2013; climate change have contributed to an increase in average tempera-
Gebreegziabher et al., 2016), we adopt a deterministic approach of tures and a reduction in rainfall even in the more agriculturally
climate change by considering solely the potential effects of long-run productive southern regions on the country (see for instance Mohino
mean changes in agriculture (mainly crop yields reductions). We thus et al., 2011; Druyan, 2011). On the other hand, climate conditions have
define a first group of agro-climatic scenarios over the next 25 years for also been characterised by a high inter-decadal variability as well as
Niger, on the basis of the main forecasts of average yield evolutions significant interannual variability, particularly reflected in a higher
found in the agro-climate literature. However, limiting the agro-climatic frequency and intensity of drought events (Epule et al., 2014; Masih
shocks to average changes clearly underestimates their potential effects. et al., 2014). According to the World Bank (2013), there have already
In the literature, whenever possible, other CGE studies thus try to been three such major episodes leading to food crises since 2000
include some climate variability features using stochastic or probabilistic (2004/2005, 2009/2010 and 2011/2012), while only three episodes
scenarios (see for instance, Thurlow et al., 2009; Yu et al., 2010; Arndt were observed between 1970 and 2000 (1973/1974, 1984/1985 and
et al., 2011; Arndt et al., 2012; Sassi and Cardaci, 2013; Arndt et al., 1997/1998).
2015; Arndt and Thurlow, 2015). Others, though in limited number, Under these adverse, changing climatic conditions, Nigerien farm-
focus even more on future changes in extreme events and the disaster ers have preferred crops able to thrive despite low rainfall such as
risks that they could represent for a country (see for instance World millet, sorghum or cowpea which together represent 92.7% of the
Bank, 2008; Pauw et al., 2011; Al-Riffai et al., 2012 or Zhong et al., harvested area and 61% of the crops’ gross production value (Table 1).
2014). In a second step, we therefore chose to focus on the risk of a However their yields remain particularly low and unstable. Over the
potential future greater climate variability by defining a second group of period 1980–2011, millet and sorghum yields fell 11.7% and 44.5%
agro-climatic scenarios including more frequent and more intense respectively even though an average increase (23.6% and 50.4%
drought events for the next 25 years in Niger. These scenarios rely here respectively) was observed over the period 2000–2011 (World Bank,
on stochastic annual weather sequences including drought events with 2013). This decade-by-decade instability has moreover been reinforced
sizeable one-off impacts on yields, rainfed land depreciation or livestock
capital. Finally, in a last step, we combine previous analyses in a range of Table 1
scenarios including both climate change and climate variability features. Share of Harvested area, Production and Gross Production Value of major
Third, some CGE studies (see for instance, Eboli et al., 2010; World agricultural crops in Niger (in %).
Source: Own calculations based on FAO (2016a)
Bank, 2010; Robinson et al., 2012; Bosello et al., 2013; Bandara and
Cai, 2014; Calzadilla et al., 2014; Chalise and Naranpanawa, 2016) CROPS Harvested Production(1) Gross
explore also different options for coping with future climate conditions area(1) Production
focusing on either autonomous price-driven decisions of agents Value(2)
(mainly regarding primary factors-use change) or adaptation strategies
Food crops 62.47 54.59 41.11
that governments may promote. We thus chose to also asses the Rainfed 62.38 53 39.74
opportunity of such coping strategies for Niger. For that purpose, we Millet 42.21 37.65 29.90
selected some planned strategies specifically designed to boost agri- Sorghum 20.17 15.35 9.84
cultural productivity with a focus on improvements in the rural road Irrigated 0.09 1.59 1.37
Cassava 0.04 1.47 0.77
network, expansions of irrigation capacity or changes in farm manage- Rice, paddy 0.05 0.12 0.60
ment practices. Cash crops 35.20 38.33 47.82
The rest of the paper is structured as follows. Section 2 presents the Rainfed 34.34 20.77 27.42
current vulnerability of Nigerien agriculture to climate conditions and Cowpeas, dry 29.82 16.95 21.22
Groundnuts, with 4.52 3.82 6.20
the hypotheses we selected for defining the different future agro-
shell
climatic scenarios. Section 3 describes the main features of our Irrigated 0.86 17.56 20.40
economic models and Section 4 presents the results of the different Fruit, fresh nes 0.31 3.43 5.15
simulations. Finally, Section 5 explores the net impact of the coping Mangoes, 0.22 2.06 5.32
strategies we selected in order to determine if they are capable of mangosteens,
guavas
counterbalancing the effects of the changing weather conditions. Onions, dry 0.14 4.29 3.88
Cabbages and other 0.06 2.44 1.57
2. Background brassicas
Tomatoes 0.06 2.06 3.28
Vegetables, fresh 0.04 1.11 0.90
2.1. Niger's current vulnerability to adverse climatic conditions
nes
Sugar cane 0.03 2.17 0.30
Despite the major efforts made since the 1970s by the Nigerien
(1) (2)
government to develop agriculture, this sector, which is essentially Notes: Mean of 2012, 2013 et 2014; Mean of 2012 and 2013.

129
J.-M. Montaud et al.
Mean variation of temperatures (°C)
Minimum/maximum of GCMs results
Fig. 1. Level of uncertainty of GCM climate change forecasts for Niger under three SRES scenarios(1) Mean variation of temperatures (°C) Mean variation of precipitations (mm)
Minimum/maximum of GCMs results + Mean of GCMs results — Median of GCMs results Notes: (1) Based on 9 GCM (CGCM3.1-T47, CNRM-CM3, GFDL-CM2.0, GFDL-CM2.1,
ipsl_cm4, MIROC3.2, ECHO-G; ECHAM5/ MPI-OM and MRI-CGCM2.3.2) and 3 Special Report on Emissions Scenarios (B1, A1B and A2); Sources: Own calculations with the Climate
analysis tool powered by Climate Wizard (http://climatewizard.ciat.cgiar.org/index1.html).
by exceptional periods of drought. For instance, each major drought
observed since 2000 led to a dramatic fall in yields of between 21% and
24% for millet and between 19% and 28% for sorghum.
2.2. Defining future agro-climatic scenarios for Niger: the need to
deal with uncertainty
Given the already difficult socio-economic situation which weakens
the ability of the populations concerned to adapt, the potential risks
posed by a further deterioration of climate conditions in Niger are thus
far from negligible. In the agro-climate literature, it seems commonly
accepted that, on the one hand, long-run climate change could lead to
greater variations in average temperatures and rainfall over the Sahel,
which could exacerbate disturbances in crops’ biological cycles and
thus further reduce the country's agricultural potential (see for instance
Roudier et al., 2011). On the other hand, future weather conditions
could also be characterized by increased variability, including higher
frequency of severe droughts which could have serious one-off impacts
on agricultural production (see for instance Sylla et al., 2010).
However, producing reliable agro-climatic scenarios of such impacts
remains challenging because of large uncertainties grounded in both
climate and agricultural sciences.
2.2.1. Long run climate change perspective
In order to characterize the possible future evolutions of weather
conditions in Niger, we have no choice but to use the forecasts provided
by the General Circulation Models (GCMs) which simulate the effect of
variables that might affect the climate. However, these predictions are
by their very nature uncertain (see for instance Burke et al., 2015).
First, the complexity of climate is still not completely understood,
specifically in the Sahelian zone. Secondly, these forecasts depend on
scenarios (of greenhouse gas emissions for instance) which are
themselves uncertain. And thirdly, even though the
Intergovernmental Panel on Climate Change stresses the improvement
in the ability of climate models to simulate surface temperature on a
regional scale (IPCC, 2013), these predictions must be used with care
in a national framework. Fig. 1 reflects these uncertainties in the
specific case of Niger. If the different GCMs agree on an average
increase of temperature (between 1.8 °C and 2.5 °C) from the historic
to the future time period whatever the scenario, there is much more
uncertainty regarding rainfall patterns. Although the average results
show relative stability, some models indicate an increase while others
point to a reduction.
Economic Modelling 63 (2017) 128–142
Mean variation of precipitations (mm)
+ Mean of GCMs results Median of GCMs results
A second source of uncertainty concerns the potential impact of
these climatic forecasts on Nigerien agriculture (see for instance Lobell
and Burke, 2008). In agronomic literature, even if studies attempting to
assess such impacts have increased significantly in recent years, there
are unfortunately very few studies focussing exclusively on Niger, or if
they do, they only take into account a limited number of crops (see for
instance, Ben Mohamed et al., 2002 or Salack et al., 2008). It is
therefore necessary to draw on studies with a more extensive scope,
focussing on the African continent, Western Africa or the Sahel.
However, results vary as a function of the methodologies used
(mechanical or empirical)1, the geographical areas on which they
concentrate, the crops they include and the climatic scenarios they
select (see for instance Roudier et al., 2011 or Niang et al., 2014 for a
literature review). Even though some analyses do not rule out increased
yields in some cases (for instance Liu et al., 2008 for millet and rice),
most forecast a long term downturn in average yields for main crops.
For example, Knox (2012) estimates an average fall of 8% between
2010 and 2050. Thornton et al. (2011) forecasts a fall in maize yields of
19% as a result of a 5 °C increase in temperature in sub-Saharan Africa.
Ben Mohamed (2011) calculates a fall in millet yields of 20% and 40%
respectively for 2 °C and 3 °C temperature increases in the Sahel
region. Van Duivenbooden et al. (2002) estimate falls in cowpea and
groundnut yields of between 12% and 30% depending on the climate
scenarios selected (temperatures of +10% et +20% and precipitations
−10% and −20%). Schlenker and Lobell (2010) estimate a reduction in
sub-Saharan Africa for 2050 of −22%, −17%, −17%, −18%, and −8%
for maize, sorghum, millet, groundnut, and cassava respectively. Sultan
et al. (2013) forecast reductions in millet and sorghum yields in the
Sahel of up to 41% for a 6° temperature increase and a 20% drop in
rainfall. Thomas and Rosegrant (2015) calculate changes in Western
African yields of −5.8% to +0.3% for groundnuts, −9.5% to −15.9% for
sorghum and +0.5 to −4.4% for rice.
Given the state of the art, our long run agro-climatic scenarios rely
in the end on the major above-mentioned reference studies that
quantify future crop yield changes in Niger, the Sahel region or
Western Africa, for Millet, Sorghum, Cowpeas, Goundnuts, Cassava
1
The mechanical approach relies on bio-physical agronomic crop model (DSSAT or
SARRA-H for instance for Africa). The empirical approach establishes statistical links
between agronomic and climate observations from the past to forecast future yields
without necessarily explaining the nature of the causes involved (cf. Lobell and Burke,
2010 or Porter et al., 2014 for a literature review). Empirical projections exhibit greater
dispersion than mechanical projections given the variability of yields over the last few
decades.
130
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

Table 2 subjective prior distribution and to adopt a stochastic approach. We

Main hypotheses for long-run scenarios of climate change for Niger (Annual
crop yield reduction over 2040).
Crops Optimistic Medium Pessimistic Based on
Millet −0.311% −0.386% −0.833% Knox et al. (2012) and
Ben Mohamed et al.
(2002)
Sorghum −0.277% −0.386% −0.479% Thomas and Rosegrant
(2015)
Cowpeas −0.354% −0.796% −0.986% Van Duivenbooden et al.
(2002)
Groundnuts −0.01% −0.166% −0.268% Thomas and Rosegrant
(2015)
Cassava −0.232% −0.232% −0.232% Schlenker and Lobell
(2010)
Rice +0.12% +0.03% +0.01% Thomas and Rosegrant
(2015)
and Rice. In the absence of reliable projections for other crops, which
represent a minority in Niger, it was decided not to take them into
account. On this basis, three deterministic scenarios (pessimistic,
medium and optimistic) were defined for the next 25 years. (see
Table 2). They include 97% of harvested areas and 70.5% of Nigerien
crop production for which we assume a linear continuous variation of
yields that we will be able to simulate with our economic models.
2.2.2. Interannual climate variability perspective
In order to define future climate variability scenarios for Nigerien
agriculture, we chose to focus on the occurrence of severe drought.
However, here again, we have to deal with uncertainty. Regarding the
GCM predictions, it should be stressed that they tend to underestimate
climate variability. This is particularly true with regard to extreme
droughts in the Sahel because their factors are still insufficiently
understood and their rarity makes evaluation of model performance
less robust than for average climate features (Burke and Brown, 2008;
Zhao and Dai, 2016). In this context, we can however find some
indications based on GCM forecasts which may suggest that extremely
hot seasons will probably become more frequent in the future. For
instance, Fig. 2 shows that Dry periods indicate a recurrence interval
value between 2.2 and 3.4 for scenario B1 (with a median of 2.8),
between 2.9 and 3.8 for scenario A1B (with a median of 3.2) and
between 2.3 and 3.1 for scenario A2 (with a median of 3.1). For
Consecutive dry days, the recurrence interval value is between 2.4 and
3.1 for scenario B1 (with a median of 2.6), between 2.2 and 3.4 for
scenario A1B (with a median of 2.4) and between 2.1 and 4.4 for
scenario A2 (with a median of 2.2).
These uncertaincies are even greater regarding the potential
impacts of future droughts on the agricultural sector. Althougth
drought has been widely considered as a major problem in Sahelian
agriculture, there is still very little modelling and observation evidence
of their ex-post impact (see for instance Gautier et al., 2016 for a
literature review for West-african countries). Such assessment is
indeed rather difficult given that drought is not a straightforward
concept raising problems as to its definitions (agricultural, meteorolo-
gical or hydrological), its spatial extent (areas affected in the country),
its temporal dimension (date of appearance, duration) or its associated
risks (such as locust invasions, water scarcity, crop or animal diseases,
etc.). Moreover, an ex-ante modelling of cropping systems’ response to
droughts is difficult as, by definition, these events occur rarely and
process-based crop models cannot be adequately calibrated and tested
(Porter et al. 2014).
With such an uncertain background, we chose to use a conservative
sensitivity analysis procedure to define agro-climatic scenarios of
potential greater variability over the next 25 years period. First,
regarding the frequency of severe drought, given that we have no idea
of the distribution of such events, we have no choice but to develop a
then assume that the appearance of an extreme drought follows a
Bernoulli law of probability p every year. According to historical data,
the frequency observed in Niger was close to one event every ten years
from 1970 to 2000 and one event every five years from 2000 (World
Bank, 2013). Therefore our variability scenarios envisage three hy-
potheses for annual probability of drought occurrence: normal (p=0,1),
as observed before 2000, high (p=0,2) as observed between 2000 and
2015, and very high (p=0,3) with potentially greater climate variability.
Second, regarding the intensity of drought events, it is difficult to
isolate and assess the mean impact of past droughts on agricultural
yields, rainfed land depreciation or livestock in Niger. Each one is
indeed unique depending on its location within the country, its period
of occurrence and duration, or the co-occurrence of certain associated
risks which increase its impacts. In this context, we decided to adopt a
conservative approach with different hypotheses of impact levels,
severe (is=−10%), high (ih=−12.5%) and very high (ivh=−15%), which
will affect rainfed crop activities and livestock as described in the next
section.
Finally, combining the three hypotheses for drought intensity and
frequency, we thus define nine scenarios of potential greater climate
variability for the next 25 years. For each one, given that it relies on the
probability of occurrence p, we use a random process to generate
different (thirty) stochastic annual weather sequences including
drought events. Each sequence should then be simulated with our
economic models in order to average the representative impact of each
scenario.
3. Key model features
Our DRCGE model for Niger is adapted from the PEP-1-t Standard
Model (Decaluwé et al., 2013). It features four groups of households
(Subsistence farmers, Other agricultural workers in export agriculture,
Urban formal and Urban informal), one government agent, one firm
agent, twelve agricultural activities (irrigated or rainfed), seven in-
dustrial activities and five service activities (the relevant equations and
variables are presented in Supplementary Material).
Its within-period specification relies on fairly standard general
equilibrium assumptions for low income countries. On the supply side,
each activity maximises profit by combining capital factors with the
labour factor. For agricultural activities, the capital is assumed to be
composite between capital and land. The latter may be rainfed or
irrigated. On the income side, the different agents receive revenues on
the basis of their factor endowments. In addition, the government
receives various taxes. On the demand side, households’ consumption
follows a linear expenditure system function, and demand from the
government is exogenous. Nominal investments are savings driven in
the capital market. The numeraire is the nominal exchange rate (for
further details, see Decaluwé et al., 2013). Moreover, in order to
capture the dualist characteristic of this under-developed economy, we
chose to modify the PEP standard model by including a dual-dual
framework (Stifel and Thorbecke, 2003). First, urban areas reflect a
formal–informal dualism. As the urban informal sector is assumed to
use only labour, the income of informal workers is just their average
revenue product. In the formal urban sector, owners of capital are
profit maximizers and pay their workers their marginal revenue
product (urban wage rate). Informal sector workers devote all of their
labour to the sector until marginal revenue product falls below the
alternative wage in the market. Second, in rural areas, two other sectors
are distinguished: a subsistence sector (mainly food crop activities) and
an export sector (mainly cash crop activities). The wage earned in the
subsistence sector (mainly by family workers) is assumed to be lower
than in the export sector and the subsistence farmers household group
is supposed to earn the entire value added generated by this sector.
Within this framework, Nigerien workers can flow freely between all
four sectors of the economy. We assume moreover that disguised

131
J.-M. Montaud et al.
Recurrence interval of dry periods (2)
Minimum/maximum of GCMs results
Fig. 2. Level of uncertainty of GCM climate extreme event forecasts for Niger under three SRES scenarios(1) Recurrence interval of dry periods(2) Recurrence interval of consecutive dry
days(3)• Minimum/maximum of GCMs results + Mean of GCMs results — Median of GCMs results Notes: (1) Same GCMs as in Figure 1 (2) Number of consecutive dry day periods of
length > 5 days, per year; (3) Largest number of consecutive dry days (with daily precipitation < 1mm) per year. Sources: Own calculations with the Climate analysis tool powered by
Climate Wizard (http://climatewizard.ciat.cgiar.org/index1.html).
unemployment exists in the rural sector where the wage rate is lower
than the formal urban alternative wage and that rural workers migrate
to urban areas in line with Harris and Todaro's hypothesis until the
rural wage reaches equivalence with the expected wage in the urban
sector.
The between-period specification of the model is recursive, which
means that the behaviour of agents is based on adaptive expectations,
rather than on forward-looking expectations. The first assignment
concerns the accumulation of capital between two periods (see
Decaluwé et al., 2013 for further details). The second dynamic
specifications mainly concern the different climate impacts on the
agricultural sector that we chose to include in the model. We assume
(Eq. 1) that, in any one period, the total factor productivity parameter
( βjVA
,t
) of each agricultural crop activity j (rainfed or irrigated) exposed to
changing climatic conditions is firstly a function of an average
exogenous annual depreciation rate ( Climatej, t ). It is secondly a
function of the impacts of drought events ( Droughtj, t ) that occur at
random over the 25 year period on the basis of the probability p
characterizing each variability scenarios. It should be noted that this
second effect is only considered for rainfed crop activities and not for
irrigated crop activities. We moreover consider (Eqs. 2a, 2b and 2c)
that, when a drought occurs, there is a higher than normal annual
depreciation rate (δk , j, t ) for rainfed land stock (reflecting deterioration
or greater underutilization of this land) and livestock capital (reflecting
reduced pasture, increased diseases and animal losses) with a two-year
post-drought recovery period.
BjVA VA
, t +1 = Bj, t . (1 − Drought j, t +1) − Climatej, t +1 ∀ and exogenous variables from one year to the next. For this purpose,
j ∈ Agriculturalactivitiesexposedtochangingclimaticconditions (1)
δk , j, t +1 = δk , j, t − Drought j, t +1 For k = RainfedLand and Livestockcapital
(2a)
δk , j, t +2 = δk , j, t +1 + Drought j, t +1/2 (2b)
δk , j, t +3 = δk , j, t +2 + Drought j, t +1/2 (2c)
W ith Droughtj, t = iS or iH or iVH for a year of occurrence and 0
for other years
Some specific indicators are built in order to capture these climate
impacts on household welfare. The first ones are calculated directly
with the DRCGE model: a Food Availability Index measures the
volume of food production per capita in the country; a Food Access
Index measures households’ real consumption per capita of food
Economic Modelling 63 (2017) 128–142
Recurrence interval of consecutive dry days (3)
+ Mean of GCMs results Median of GCMs results
products. Moreover, as the DRCGE model's results are only relevant
at the household group level (in accordance with the representative
agent hypothesis of CGE modelling), a second group of welfare
indicators are calculated with a Dynamic Microsimulation model
(DMS). The two models are linked with a top-down approach. Their
consistency is insured by a reweighting procedure in order to account
for changes between the initial weightings in the base sample of the
DMS and those of the simulated situations over the period
(Buddelmeyer et al., 2012). The DMS model is therefore used to assess
the impacts of climate change at the individual level such as inequality
indices, poverty indicators (with two distinct poverty lines for rural and
urban areas) and, finally, another aspect of Food Security with a Food
insecurity ratio which represents the proportion of individuals with
insufficient dietary diversity (further details of the DMS model are in
the supplementary material).
Finally, the initial equilibrium of the DRCGE model is given by the
last Social Accounting Matrix for Niger which has the advantages of
focussing on the agricultural sector and distinguishing between irri-
gated and rainfed agriculture2. Most of the CGE model parameters can
be calibrated directly from this matrix. When such calibrations were
not possible, we obtained the parameters from extant literature,
including those used in CGE models previously established for Niger
(Go et al., 2013 and Devarajan et al., 2015). The microeconomic data
used for the DMS model are from the 2008 Household Food
Vulnerability Survey of Nigerien National Statistical Institute.
The models run for the period of 25 years. In a first step, a Business
As Usual (BAU) scenario is simulated by updating various constants
we chose the annual medium population growth rate projected for
Niger (United Nations, 2015), except for land stock for which a 1.39%
annual rate was selected given the observations of the last decade
(FAO, 2016a). In a second step, the different scenarios are simulated
and presented in comparison with those of BAU. Given the internal
complexity of our DRCGE framework and the multiple circular forces
between demand, incomes, prices or supply it considers, we chose here
to isolate the respective potential effects of each agro-climatic change.
In the first group of simulations we only consider the long run socio-
economic incremental impacts of climate change: the exogenous
variables Climatej, t in Eq. 1are defined as in Table 2 and variables
Droughtj, t =0 for all years. In a second group of simulations, we only
2
The informal sector was integrated into the Social Accounting Matrix on the basis of
the National Accounts and employment survey data (Institut National de la Statistique,
2012) completed by data from Herrera et al. (2012).
132
 J.-M. Montaud et al.

Table 3
Selected impacts of various long–run climate change scenarios for selected periods (Average annual deviation from BAU value in %).
Source: Own calculations with GAMS software for DRCGE results and STATA software for DMS results

Optimistic scenario Medium scenario Pessimistic scenario

2015-20 2020-25 2025-30 2030-35 2035-40 2015-40 2015-20 2020-25 2025-30 2030-35 2035-40 2015-40 2015-20 2020-25 2025-30 2030-35 2035-40 2015-40

Economic indicators
National real GDP −0.33 −0.93 −1.59 −2.29 −3.02 −1.63 −0.49 −1.39 −2.37 −3.44 −4.58 −2.45 −0.86 −2.44 −4.22 −6.21 −8.39 −4.42
National Price Index 0.90 2.06 3.14 4.28 5.49 3.17 1.13 2.60 4.07 5.67 7.41 4.18 2.28 5.43 8.75 12.66 17.32 9.29
Agricultural GDP −0.34 −0.98 −1.68 −2.43 −3.23 −1.73 −0.55 −1.59 −2.73 −3.95 −5.25 −2.81 −0.89 −2.57 −4.46 −6.53 −8.78 −4.65
Agricultural Price Index 0.96 2.39 3.86 5.45 7.16 3.96 1.33 3.38 5.59 8.08 10.85 5.85 2.53 6.51 11.03 16.48 23.05 11.92

Welfare indicators
Real income per capita
Subsistence farmers −0.02 −0.29 −0.77 −1.32 −1.92 −0.85 −0.05 −0.59 −1.36 −2.23 −3.18 −1.48 −0.01 −0.86 −2.14 −3.61 −5.25 −2.37
Other agricultural workers −0.97 −2.12 −3.12 −4.14 −5.19 −3.11 −1.30 −2.90 −4.36 −5.88 −7.48 −4.38 −2.44 −5.46 −8.25 −11.23 −14.47 −8.37
Formal workers −0.63 −1.39 −2.05 −2.73 −3.44 −2.05 −0.81 −1.80 −2.70 −3.64 −4.62 −2.72 −1.56 −3.53 −5.41 −7.47 −9.76 −5.55
Infomal workers −0.37 −0.95 −1.56 −2.20 −2.87 −1.59 −0.51 −1.34 −2.22 −3.16 −4.16 −2.28 −0.92 −2.44 −4.08 −5.89 −7.88 −4.24
Poverty

133
Headcount Poverty Rate
National 0.14 0.42 0.75 1.11 1.48 0.78 0.19 0.61 1.10 1.64 2.23 1.15 0.34 1.08 1.98 3.00 4.14 2.11
Rural areas 0.16 0.49 0.86 1.26 1.69 0.89 0.22 0.69 1.24 1.83 2.48 1.29 0.40 1.24 2.24 3.37 4.61 2.37
Urban areas 0.10 0.33 0.61 0.92 1.25 0.64 0.16 0.51 0.96 1.46 2.00 1.02 0.27 0.90 1.70 2.65 3.73 1.85
Headcount Food Poverty Rate
National 0.18 0.55 0.98 1.44 1.94 1.02 0.25 0.79 1.43 2.14 2.91 1.50 0.44 1.42 2.59 3.94 5.45 2.77
Rural areas 0.21 0.64 1.13 1.66 2.22 1.17 0.29 0.90 1.62 2.41 3.26 1.70 0.52 1.63 2.95 4.46 6.13 3.14
Urban areas 0.12 0.38 0.71 1.07 1.45 0.74 0.18 0.59 1.11 1.68 2.32 1.18 0.31 1.04 1.97 3.07 4.34 2.15
Food Security
Food Insecurity Ratio
National 0.05 0.17 0.30 0.45 0.60 0.31 0.08 0.24 0.44 0.66 0.90 0.46 0.14 0.43 0.80 1.21 1.67 0.85
Rural areas 0.06 0.19 0.33 0.48 0.64 0.34 0.08 0.26 0.47 0.70 0.95 0.49 0.15 0.47 0.86 1.29 1.77 0.91
Urban areas 0.04 0.14 0.26 0.39 0.54 0.28 0.07 0.22 0.41 0.62 0.86 0.43 0.12 0.38 0.73 1.13 1.59 0.79
Food Access Index
National −0.21 −0.61 −1.06 −1.55 −2.07 −1.10 −0.30 −0.90 −1.59 −2.33 −3.14 −1.65 −0.53 −1.56 −2.74 −4.05 −5.49 −2.87
Rural areas −0.50 −1.33 −2.21 −3.11 −4.06 −2.24 −0.67 −1.85 −3.10 −4.41 −5.78 −3.16 −1.22 −3.32 −5.53 −7.87 −10.33 −5.65
Urban areas 0.13 0.26 0.34 0.40 0.43 0.31 0.14 0.25 0.29 0.30 0.27 0.25 0.31 0.64 0.88 1.10 1.30 0.84
Food Availability Index −0.28 −0.80 −1.37 −1.99 −2.65 −1.42 −0.41 −1.18 −2.05 −2.99 −4.00 −2.13 −0.71 −2.05 −3.58 −5.27 −7.11 −3.75
Economic Modelling 63 (2017) 128–142
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

Table 4
Selected average impacts of various variability scenarios (Average annual deviation from BAU value in % for 2015–2040(1)).
Source: Own calculations with GAMS software for DRCGE results and STATA software for DMS results

Level of drought intensity

Severe (is=−10%) High (ih=−12.5%) Very high (ivh=−15%)

Probability of drought p=0.1 p=0.2 p=0.3 p=0.1 p=0.2 p=0.3 p=0.1 p=0.2 p=0.3

Economic indicators
National real GDP −0.75 −1.19 −1.90 −0.92 −1.79 −2.82 −1.29 −2.21 −3.30
National Price Index 1.05 2.03 3.17 1.47 2.94 4.70 2.06 3.92 6.07
Agricultural GDP −0.94 −1.53 −2.19 −1.11 −2.14 −3.27 −1.59 −2.60 −3.81
Agricultural Price Index 1.89 3.28 4.81 2.47 4.83 7.50 3.59 6.27 9.28
Welfare indicators
Real income per capita
Subsistence farmers −1.03 −1.63 −2.05 −1.37 −2.78 −2.95 −2.15 −3.35 −4.39
Other agricultural workers −0.46 −1.13 −1.48 −0.55 −1.32 −1.76 −0.44 −1.43 −2.78
Formal workers −0.40 −0.89 −1.10 −0.53 −1.15 −1.42 −0.62 −1.41 −2.43
Infomal workers −0.62 −1.10 −1.40 −0.77 −1.61 −1.85 −1.06 −1.87 −2.80
Poverty
Headcount Poverty Rate
National 0.44 0.68 1.14 0.55 1.05 1.62 0.73 1.11 1.93
Rural areas 0.47 0.74 1.27 0.58 1.11 1.72 0.78 1.22 2.15
Urban areas 0.44 0.64 1.04 0.56 1.06 1.60 0.71 1.01 1.75
Headcount Food Poverty Rate
National 0.57 0.88 1.49 0.71 1.37 2.10 0.95 1.44 2.46
Rural areas 0.62 0.98 1.67 0.77 1.47 2.28 1.04 1.61 2.75
Urban areas 0.51 0.75 1.22 0.66 1.24 1.89 0.84 1.19 2.01
Food Security
Food Insecurity Ratio
National 0.18 0.27 0.46 0.22 0.42 0.65 0.30 0.45 0.69
Rural areas 0.18 0.27 0.42 0.22 0.43 0.66 0.30 0.43 0.68
Urban areas 0.18 0.28 0.44 0.24 0.45 0.68 0.31 0.47 0.71
Food Access Index
National −0.55 −0.90 −1.29 −0.63 −1.24 −1.88 −0.88 −1.45 −2.14
Rural areas −0.84 −1.46 −2.19 −0.97 −1.94 −3.00 −1.30 −2.24 −3.43
Urban areas −0.16 −0.15 −0.09 −0.15 −0.26 −0.30 −0.26 −0.28 −0.25
Food Availability Index −0.68 −1.11 −1.61 −0.75 −1.51 −2.45 −1.04 −1.74 −2.63

(1)
Note: Results computed from 30 randomized simulations for each scenario.

consider the impacts of a greater climate variability ( Climatej, t = 0 and
Droughtj, t = iSoriH oriVH depending on scenarios assumptions). Finally, a
last group of simulations combine both scenarios in a unique perspec-
tive.
4. Results
4.1. Potential socio-economic impacts of climate change scenarios
Selected results for the three long-term climate change scenarios
are given in Table 3 (more detailed results are given in the supple-
mentary material). They are consistent with the findings of the other
CGE studies that assess (even in different frameworks) the impacts of a
long-run crop yields reductions for other developing countries (see for
instance Bosello et al., 2013; Calzadilla et al., 2013; Gebreegziabher
et al., 2016 for recent studies in Africa). As expected, the economic and
social impacts are proportional to the extent of climate change
envisaged. They are also progressive with effects which are amplified
over the period. For the period 2015–2040, falls in yields for the main
crops would cause a mean annual reduction in agricultural GDP against
BAU of 1.73%, 2.81% and 4.65%, respectively for the optimistic,
medium and pessimistic scenarios, and an increase in agricultural
prices of 3.96%, 5.85% and as much as 11.92% in the pessimistic
scenario. These effects are relatively evenly spread over subsistence and
export cash crops even though, given the production structure of the
food crop sector, price effects would be relatively greater (as high as
+25.56% in the pessimistic scenario). Given the importance of the
agricultural sector and its direct or indirect links with the rest of the
economy, these effects would cascade down through the entire national
economy with a fall in GDP of 1.63%, 2.45% and 4.42% and an increase
in the price index of 3.17%, 4.18% and 9.29% for the three scenarios
respectively.
In this new context of a general fall in income and price increases,
especially as regard agricultural prices, social indicators would be
sharply degraded with respect to BAU. First, climate change would
have a strong impact on real income per capita for all households. It
should be stressed, however, that subsistence farmers, whose incomes
depend mainly on the sale of their products, benefit from the increase
of the prices in this sector and experience a less of a decrease than the
households who work in the other sectors. Second, the effects in terms
of poverty would also be particularly striking. The proportion of the
national population living below the poverty line would significantly
increase with respect to BAU (+0.78%, +1.15% and +2.11% depending
on the scenario). Although this increase is logically greater in rural
areas than in urban areas, the latter are also affected given the fall in
the income of both formal and informal workers and the increase in the
urban poverty line. At the end of the period, given the demographic
growth and the cumulative effects of climate change, this increased
poverty would represent nearly 470,000, 550,000 and 1,400,000 more
poor people compared to BAU. Moreover, as indicated by the poverty
gap index, the Nigerien poor would not only be more numerous but
also poorer. On average, during the period, the national percentage
shortfall in consumption with respect to the poverty line for the
population would be annually higher than in the BAU by about
1.02%, 1.50% and 2.77% for the optimistic, medium and pessimistic
scenarios respectively. This increasing depth of poverty is particularly
true for rural residents given their exposure to climate risks. Third,
climate change would also have a strong impact on the inequality of
distribution among the poor as indicated by the Poverty severity index

134
J.-M. Montaud et al.
that would be annually higher than in the BAU scenario by about
1.17%, 1.73% and 3.20% depending on the scenario. Finally, the effects
on food security would also be very strong. The national headcount
food poverty rate would be 1.02%, 1.50% and 2.77% higher depending
on the scenario, indicating that more people would be below the food
poverty line compared to BAU. At the end of the period, it would
represent a relative increase of nearly 104,000, 157,000 and 280,000
people without sufficient income to buy food. On average, the Food
Insecurity ratio would increase annually by 0.31%, 0.46% and 0.85%
depending on the scenario. The Food Availability Index indicates that
per capita supply of foodstuffs would fall with respect to the reference
situation by 1.42%, 2.13% and 3.75% as a result of the drop in
agricultural production. Similarly, the Food Access Index shows that,
in this context of falling incomes and increasing prices, per capita food
consumption in households would fall (−1.10%, −1.65% and −2.87%)
particularly in rural areas.
4.2. Potential socio-economic impacts of variability scenarios
Table 4 presents some selected results of simulations for the nine
stochastic scenarios of climate variability (more detailed results are
given in the Supplementary Material). Because of the random process
used to generate climatic conditions over the period, each result is the
mean of the thirty sequences which define each scenario. However, as
droughts occur at different dates in each sequence of each scenario, it is
therefore no longer possible to detail the results year by year as for the
previous deterministic scenarios. We just present here the average
annual deviation from BAU for 2015–2040. Although obtained within
a different framework and a different methodology, these results
prolong the results of CGE studies that focus on extreme events (see
for instance Pauw et al., 2011 for Malawi or Al-Riffai et al., 2012 for
Syria) and confirm the usefulness to better take such events into
account in the analysis of the outcome of future agro-climatic changes.
They show that, for a given level of intensity, the increase in drought
frequency is reflected in a degradation of macroeconomic and welfare
indicators. Similarly, for a given level of frequency, the increase in
intensity of droughts also degrades these indicators. For instance, it is
observed that the introduction of droughts with a frequency of 10% and
severe intensity causes a mean annual deviation of GDP of −0.75% with
regard to the BAU growth trend, an increase of +0.44% and +0.57%
respectively in absolute poverty and food poverty rates, and a +0.18%,
−0.68% and −0.55% change in food insecurity indicators (Food
Insecurity Ratio, Food Availability Index and Food Access Index). In
the most unfavourable hypotheses, a 30% increased drought frequency
and an increase in intensity to the very high level would cause a
potential mean deviation of −3.30% of GDP, +1.93% and +2.46%
respectively for absolute and food poverty rates and +0.69%, −2.63%
and −2.14% for the food security indicators. In this context, as the
drought's effects are concentrated in the subsistence sector, subsistence
farmers are always more affected in terms of real income than other
households.
In order to illustrate the crisis effects generated by a drought, we
chose to simulate the impacts of a fictitious scenario corresponding to
very high intensity events regularly distributed every five years during
the period (Fig. 3). The results are comparable to what has been
observed in Niger in the last decades when the country suffered a
severe drought event. They show that each event causes an average fall
with regard to the BAU GDP growth trend of almost 7% (with
progressive return to normal in the two years following the event)
and effects on prices in excess of +30%. In this context, periods of
drought are characterized by a very high increase in poverty rates and
food poverty. At the end of the period, given the cumulative dynamics
involved in the model, the effects of these droughts grow ever-stronger
with an increase of more than 1,200,000 extra poor people with
880,000 in a situation of food poverty caused by the previous drought
(as against 550,000 and 430,000 in food poverty for the first drought).
135
Economic Modelling 63 (2017) 128–142
4.3. Combining climate change and climate variability perspectives
In that last group of simualtions, we combine the two previous
approaches in a single one. To keep the results to a manageable size,
only the medium long-run climate change scenario is simulated along
with the nine variability scenarios (see Table 5 and the Supplementary
Material). The results confirm the potentially strong impacts of future
weather deterioration seen in its dual perspective. As expected, a
greater climate variability leads to a degradation of macroeconomic
and welfare indicators with regard to the long-term scenario without
variability (column 1). For example, in the most unfavourable hypoth-
esis, the negative shocks caused by drought multiplies on average the
effects on GDP by 2.43 (2.41 for agricultural GDP), on the price index
by 2.83 (2.97 for agricultural prices), on the poverty rate by 2.90 and
on the food insecurity indicators by 2.93, 2.33 and 2.40 respectively. In
this context, it is rural households which will see the greatest degrada-
tion of their situation.
5. Investigating some strategies to cope with new climate
conditions
The final step of our analysis is to investigate whether Niger migth
have some room for manœuvre for coping strategies. Over the years, a
number of adaptation options have been identified for developing
countries (see for instance IPCC, 2014, chap.14) and some CGE studies
have tried to include them in their analyse (for instance, Bandara and
Cai, 2014; Calzadilla et al., 2014 or Chalise and Naranpanawa, 2016).
In Niger, many planned adaptation options have already been defined
by Nigerien authorities (CNED, 2010). The latter are numerous and it
is impossible to consider all of them in the present study. We therefore
only select three, specifically designed to boost agricultural productiv-
ity: an improvement of rural road networks, an expansion in the
capacity of irrigated agriculture and a change in farm management
practices, reflected here in the use of crop varieties better adapted to
the new climate conditions. For each one, we simulate different
scenarios (reflecting different underlying assumptions) in order to
determine if it is capable of counterbalancing the effects of the
changing weather conditions. To save space, these net impacts are
only calculated for the scenario which combines the medium long-run
agro-climatic impacts hypothesis with the medium values for drought
frequency (p=0,2) and drought intensity (ih=−12.5%). Even if they
should be considered cautiously3, some cost-benefit analysis elements
are moreover introduced by comparing the potential benefits of each
strategy with the potential financial efforts that they could require.
5.1. Improving the rural road network
The rural road network is still very limited in Niger representing
only 7371 km of classified road with a density of 0.6 km for 1000 km2
of land area, one of the lowest in Africa. Mainly used by buses that
carry farm families and trucks that transport products, this network
connects land areas to local input or output markets and strongly
determines the opportunities to trade or the costs of production for the
agricultural sector. Many studies point out that an improvement in
African rural road network could help agriculture to achieve better
performance (see for instance Calderon and Serven 2010; Gwilliam,
2011 or Domínguez-Torres and Foster, 2011). The Government of
3
It should be stressed that a better cost-benefit analysis would be quite difficult to do
here for several reasons. The first is that our estimated costs should be considered
cautiously. Second, given the absence of accurate indications, we did not introduce in the
simulations the demand effect that each strategy generates on the activities that provide
the capital goods for these investments. Third, as such strategies may involve different
types of stakeholders (government, NGOs, private firms or farmers themselves, bi and
multilateral cooperation agencies) and therefore cannot be considered only as public
policy, fiscal space effects are not included.
J.-M. Montaud et al.
Fig. 3. Selected impacts of a fictitious scenario of regular occurrence of drought events every five years a. Effect on real national GDP (base 100 for 2015) b. Effect on prices index (base
100 for 2015) c. Effect on poverty rate (base 100 for 2015) and absolute variation from BAU of poor people (thousand).
Source: Own calculations with GAMS software for DRCGE results and STATA for DMS results
Niger, which is aware of this challenge, has even made access to remote
rural areas one of the pillars of Niger's programme of commitment for
the Comprehensive Africa Agriculture Development Programme and
defined an annual target of 500 km of roads (Road-tracks) repaired,
upgraded or built.
In this context, we thus chose to assess if a better access to the road
network could be an effective strategy for adapting to climate change.
Our methodology relies on several key points. First, we chose to use the
Rural Accessibility Index (RAI) as an indicator of connectivity level of
the rural network. This RAI measures the proportion of the rural
population within a two-kilometer walking distance of an all-season
road and it is particularly low for Niger at 15% (Carruthers et al.,
2009). Second, using the Road Network Evaluation Tools (RONET)4,
we calculated how many kilometers of additional roads would need to
be built to reach a percent target for this RAI. Results show, for
instance, that the tertiary classified rural road network would need to
grow in length by around 175 percent to attain 100 percent rural
accessibility. Third, in the DRCGE model, we chose to consider the
rural road network as a pure public good with a Hicks's neutral
productivity effect on agricultural input (for a similar approach see,
for instance, Estache et al., 2012 or Chitiga et al., 2016). On this basis,
we change previous Eq. 1 into equation 1’ where agricultural produc-
tivity is also now a function, with a lagged effect, of the road network
growth rate in the preceding period given a spillover effect ( Road ) of
road infrastructure.
4
RONET is developed by the Sub-Saharan Africa Transport Policy Program of the
World Bank (see ssatpdev.worldbank.org).
136
Economic Modelling 63 (2017) 128–142
BjVA VA
, t +1 = ROADt +1. Bj, t . (1 − Drought j, t +1) − Climatej, t +1 ∀
j ∈ Agriculturalactivitiesexposedtochangingclimaticconditions
(1’)
( Roadnetwork
t +1 − Roadnetwork )
t
With ROADt +1 = ΨRoad . Roadnetworkt
+1
In this framework, we simulate different hypotheses of road
network improvement under different alternatives of RAI targets for
2040. Moreover, given that empirical estimates of the value of the
spillover effect display variation across studies economic literature, we
follow, in the absence of better information, the sensitivity analysis
procedure of Thurlow et al. (2007) using alternatively three values:
lower (ѰRoad = 0.113), medium (ѰRoad = 0.150) and upper (ѰRoad =
0.188). Finally, the overall spending needs of each alternative of RAI
targets has been evaluated on the basis of the estimates of Carruthers
et al. (2009). These costs involve upgrading the quality of the tertiary
road network (to a level adequate to the demands made upon it),
improving the network of unclassified roads (so as to minimize ongoing
maintenance costs), adding additional links to the unclassified network
to reach the most isolated villages, and ultimately, maintaining the final
network in its improved, upgraded, or expanded form. The relation
between the RAI target and the network length is not linear. Each
successive increment in the RAI becomes increasingly costly, as more
kilometers of road are needed to reach increasingly isolated segments
of the rural population. On this basis, reaching a RAI target of 40% or
100% would involve a respective spending about USD 536 million and
USD 1336 million.
Selected results of the simulations are presented in Table 6 (see also
the Supplementary Material) and appear consistent with other findings
within CGE literature (for instance World Bank (2010) or Arndt et al.,
(2015) for Ghana). They confirm for instance that a progressive
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

Table 5
Selected average impacts of agro-climatic scenarios combining long run yield reduction(1) and increasing drought intensity and frequency (Average annual
deviation from BAU value in % for 2015–2040(2)).
Source: Own calculations with GAMS software for DRCGE results and STATA software for DMS results

Level of drought intensity

Severe (is=−10%) High (ih=−12.5%) Very high (ivh=−15%)

Probability of drought p=0 p=0.1 p=0.2 p=0.3 p=0.1 p=0.2 p=0.3 p=0.1 p=0.2 p=0.3

Economic indicators
National real GDP −2.45 −3.18 −3.89 −4.74 −3.44 −4.38 −5.49 −3.61 −4.65 −5.97
National Price Index 4.18 5.52 6.90 8.55 6.16 8.15 10.45 6.66 8.76 11.82
Agricultural GDP −2.81 −3.66 −4.47 −5.42 −3.96 −5.03 −6.26 −4.15 −5.21 −6.77
Agricultural Price Index 5.85 8.02 10.16 12.66 9.00 12.06 15.52 9.73 12.72 17.53
Welfare indicators
Real income per capita
Subsistence farmers −1.48 −2.47 −3.31 −4.16 −2.97 −4.18 −4.65 −3.35 −4.67 −6.17
Other agricultural workers −4.38 −4.92 −5.62 −6.62 −4.96 −5.82 −6.18 −4.93 −5.89 −7.35
Formal workers −2.72 −3.19 −3.73 −4.46 −3.34 −4.05 −4.28 −3.43 −4.20 −5.35
Infomal workers −2.28 −2.89 −3.50 −4.24 −3.10 −3.90 −4.23 −3.24 −4.12 −5.27
Poverty
Headcount Poverty Rate
National 1.15 1.58 2.01 2.46 1.76 2.21 2.80 2.02 2.66 3.34
Rural areas 1.29 1.74 2.19 2.67 1.92 2.38 3.00 2.22 2.84 3.53
Urban areas 1.02 1.46 1.90 2.34 1.66 2.15 2.75 1.85 2.62 3.36
Headcount Food Poverty Rate
National 1.50 2.06 2.63 3.21 2.30 2.89 3.66 2.67 3.48 4.36
Rural areas 1.70 2.29 2.90 3.53 2.53 3.15 3.98 2.95 3.77 4.68
Urban areas 1.18 1.69 2.20 2.71 1.92 2.49 3.19 2.12 3.05 3.91
Food Security
Food Insecurity Ratio
National 0.46 0.64 0.81 0.99 0.67 0.89 1.13 0.71 1.08 1.35
Rural areas 0.49 0.67 0.84 1.02 0.70 0.91 1.15 0.74 1.09 1.35
Urban areas 0.43 0.62 0.81 1.00 0.66 0.91 1.17 0.71 1.12 1.43
Food Access Index
National −1.65 −2.15 −2.64 −3.22 −2.29 −2.93 −3.68 −2.39 −2.90 −3.96
Rural areas −3.16 −3.91 −4.68 −5.63 −4.11 −5.07 −6.26 −4.23 −4.95 −6.63
Urban areas 0.25 0.15 0.07 0.04 0.10 −0.01 −0.07 0.07 −0.06 −0.14
Food Availability Index −2.13 −2.73 −3.34 −4.08 −2.89 −3.68 −4.63 −2.99 −3.98 −4.96

(1) (2)
Notes: Medium long run climate change scenario Results computed from 30 randomized simulations for each scenario.

improvement in access to rural zones would be enough to counter the
negative effects of climate change and climate variability. Logically, this
neutralization effect takes place all the more rapidly as the assumed
spillover effect is high. For a high level, the impacts of climate are
neutralized from an RAI threshold close to 60%. For an intermediate
level, this neutralization is reached around an RAI value close to 80% in
2040. For a low level, a value of 100% would be required to achieve
quasi neutralization of the effects of climate. In regards to the costs-
benefits, results show that such a policy would be a potentially
profitable investment. For the low level of spillover effect, reaching
the RAI threshold of neutralization (100%) could generate a national
real GDP net gain of about USD 437 million annually (USD 251 million
if we only consider the agricultural real GDP). For a high level of
spillover effect, these net gains would be respectively of about USD 466
million and USD 269 million for the RAI threshold of neutralization of
60%.
5.2. Increasing irrigation capacities
The Niger's irrigation potential is currently under-exploited, parti-
cularly in the Niger River valley in the South-West of the country. Of
the 15 million hectares of arable land, only 0.67% is irrigated (FAO,
2016b). You et al. (2011), combining biophysical and socioeconomic
factors, indicate for instance that Niger could potentially expand the
irrigated area about 162% over existing area with a good level of
profitability, with large-scale irrigation programs (reflecting only
specific infrastructures) or small-scale irrigation projects, for an overall
investments spending about USD 788 million. Current investments are
insufficient to realize this potential even though the government
announced an objective of increasing irrigated land area by 125,000
ha by 2015 (Domínguez-Torres and Foster, 2011). In this context, our
second adaptation strategy assumes an expansion in the capacity of
irrigated agriculture (see for instance Beyene and Engida, 2013 for a
similar approach), that will take place at the expense of unirrigated
land without necessarily achieving the total potential in 2040.
Selected results of the net impact of such strategy are presented in
Table 7 (see also the Supplementary Material). They indicate that
increasing the area of irrigated land could be a profitable investment.
Achieving the full irrigable potential of the Niger could, for instance,
generate a national real GDP net gain of about USD 117 million
annually (USD 72 million for agricultural real GDP). They also show
that even if it could improve economic and social indicators in the
country (except for the export agricultural workers whose income suffer
from the mechanical decrease of the irrigated land rental rate), it would
not be enough to compensate entirely for the effects of new climatic
conditions. This outcome can be compared with those of Bosello et al.
(2015) which, for instance, find a same result for Mozambique, even in
a different framework. For Niger, one reason is that not all crops,
particularly cereals which form the basic staple of the Nigerien diet,
lend themselves necessarily to irrigation. Another reason is that as a
result of the country's physical conditions and climate, the potential of
land able to be irrigated is ultimately fairly low. However, this type of
strategy should not be neglected, because it concerns primarily cash
crops that form a major part of agricultural GDP, and it remains a
potentially cost-effective means of mitigating the damages from future
adverse climatic changes.

137
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

Table 6
Selected average net impacts(1) of a rural road network improvement policy under various hypotheses for spillover effects and RAI level reached in 2040
(Average annual deviation from BAU value in % for 2015–2040(2)).
Source: Own calculations with GAMS software for DRCGE results and STATA software for DMS results

Net impacts with adaptation strategy

Level of spillover effect of road network on agricultural productivity

Low (ѰRoad = 0113) Medium (ѰRoad = 0,15) High (ѰRoad = 0188)

Without RAI level reached in 2040 RAI level reached in 2040 RAI level reached in 2040
adaptation
strategy 40% 60% 80% 100% 40% 60% 80% 100% 40% 60% 80% 100%

Economic indicators
National real GDP −4.38 −2.73 −1.75 −0.96 −0.30 −2.18 −0.89 0.15 1.03 −1.63 −0.01 1.29 2.46
National Price Index 8.15 4.80 3.16 1.90 0.88 3.87 1.79 0.22 −1.02 2.95 0.46 −1.38 −2.89
Agricultural GDP −5.03 −3.02 −1.79 −0.77 0.10 −2.33 −0.68 0.70 1.89 −1.62 0.48 2.24 3.88
Agricultural Price Index 12.06 6.54 3.71 1.53 −0.22 4.94 1.34 −1.37 −3.55 3.35 −0.96 −4.17 −6.89
Welfare indicators
Real income per capita
Subsistence farmers −4.18 −2.65 −1.70 −0.94 −0.29 −2.12 −0.87 0.15 1.01 −1.58 −0.01 1.26 2.46
Other agricultural workers −5.82 −3.23 −1.75 −0.54 0.48 −2.40 −0.43 1.17 2.54 −1.55 0.92 2.94 4.66
Formal workers −4.05 −2.84 −2.21 −1.70 −1.29 −2.49 −1.66 −1.01 −0.47 −2.13 −1.11 −0.32 0.35
Infomal workers −3.90 −2.50 −1.69 −1.04 −0.49 −2.04 −0.98 −0.12 0.59 −1.58 −0.26 0.80 1.74
Poverty
Headcount Poverty Rate
National 2.21 1.34 0.79 0.35 −0.01 1.03 0.31 −0.25 −0.70 0.72 −0.16 −0.83 −1.42
Rural areas 2.38 1.46 0.87 0.40 0.01 1.13 0.35 −0.25 −0.75 0.79 −0.21 −0.90 −1.55
Urban areas 2.15 1.26 0.72 0.31 −0.03 0.96 0.27 −0.25 −0.67 0.66 −0.19 −0.79 −1.31
Headcount Food Poverty Rate
National 2.89 1.75 1.03 0.46 0.00 1.34 0.41 −0.31 −0.89 0.93 −0.30 −1.06 −1.81
Rural areas 3.15 1.92 1.14 0.52 0.02 1.48 0.47 −0.32 −0.97 1.04 −0.32 −1.15 −1.98
Urban areas 2.49 1.46 0.84 0.36 −0.03 1.11 0.32 −0.28 −0.76 0.76 −0.28 −0.90 −1.50
Food Security
Food Insecurity Ratio
National 0.89 0.54 0.32 0.14 0.00 0.41 0.13 −0.10 −0.28 0.29 −0.06 −0.33 −0.57
Rural areas 0.91 0.56 0.33 0.15 0.00 0.43 0.14 −0.10 −0.29 0.30 −0.09 −0.34 −0.59
Urban areas 0.91 0.54 0.31 0.13 −0.01 0.41 0.12 −0.11 −0.29 0.28 −0.05 −0.34 −0.57
Food Access Index
National −2.93 −1.53 −0.66 0.05 0.67 −1.05 0.12 1.08 1.92 −0.55 0.93 2.17 3.32
Rural areas −5.07 −2.91 −1.59 −0.50 0.43 −2.18 −0.40 1.07 2.35 −1.42 0.84 2.73 4.45
Urban areas −0.01 0.32 0.59 0.84 1.06 0.47 0.86 1.21 1.54 0.63 1.15 1.64 2.14
Food Availability Index −3.68 −1.98 −0.95 −0.09 0.64 −1.41 −0.02 1.13 2.12 −0.81 0.95 2.42 3.77
Costs-benefits indicators of strategy
Average annual cost (3) – 21.4 32.2 42.9 53.4 21.4 32.2 42.9 53.4 21.4 32.2 42.9 53.4
Average annual national net gain (4) in real GDP – 183.6 285.1 367.3 436.6 240.2 374.7 483.4 575.0 298.3 466.4 602.2 724.1
Annual real GDP gains/costs – 8.56 8.85 8.56 8.17 11.21 11.64 11.27 10.76 13.91 14.48 14.04 13.55
(4)
Average annual net gain in agricultural real GDP – 102.4 161.0 209.4 251.0 134.9 213.9 279.4 336.1 168.7 269.1 353.0 431.1
Annual real agricultural GDP gains/costs – 4.77 5.00 4.88 4.70 6.29 6.64 6.52 6.29 7.87 8.36 8.23 8.07

Notes: (1) Medium long run climate change scenario combined with variability scenario with p = 0.2 and ih = −12.5% (2) Results computed from 30 randomized simulations for each
scenario (3) In USD million, based on Carruthers et al. (2009) (4) In USD million, compared to the scenario without adaptation strategy.

5.3. Substituting traditional cultivars with modern varieties of Millet
and Sorghum
In this last a strategy for adaptation, we consider an improvement
of farm management practices. As it has already been emphasized that
cultivation of millet and sorghum occupies a central place in Niger, in
terms of the total area of land cultivated and as a source of food for the
population, we envisage that farmers could replace traditional varieties
with modern ones. In an early CGE study, Breisinger et al. (2011) had
already point out the potential benefits of such a green revolution in
Africa. According to Bezançon et al., (2009), millet and sorghum are
heavily dominated by the traditional varieties in Niger (89.1% and
95.2% respectively), whose yields are among the lowest in the world.
Yet, Sultan et al. (2013) indicate that modern varieties can have much
better yields (from +61% for millet and +31% for sorghum5 [5]).
5
The traditional varieties include HKP and Somno for millet and Timari and Kénikéba
for sorghum; for the modern varieties, Souna III for millet and CSM 63 for sorghum were
selected (See Sultan et al., 2013).
Given the difficulties involved in setting up such policy and the
possibility of resistance on the part of the farmers, we tested several
hypotheses for the substitution to achieve a 25%, 50% 75% and 100%
target for the surface area cultivated with modern varieties at the end of
the period. Moreover, as Sultan et al. (2013) stress that yields of
modern varieties are potentially more sensitive to climate risks (in
65.2% of their GCM models’ forecasts, they could suffer a greater loss
of yield by −10% than traditional varieties) we also tested two types of
hypotheses. First, we consider that the two types of crops react in the
same way to climate conditions. Second, we consider that modern
varieties are more fragile than traditional varieties. However, estimat-
ing the cost of such strategy is not as straightforward as in the previous
cases. That is due, on the one hand, to the fact that there are different
types of modern or traditional varieties, according to the crops and
types of agro-ecological zones in Niger. On the other hand, there is the
fact that the success of such policy also depends on the setting up of
additional programs dedicated to improve management and farming
practices, that would include technical assistance (training, access
to improved seeds or appropriate phytosanitary products, etc.),

138
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

Table 7
Selected average net impacts(1) of an irrigation capacity improvement strategy under different hypotheses for potential irrigation achieved in 2040 (Average
annual deviation from BAU value in % for 2015–2040(2)).
Source: Own calculations with GAMS software for DRCGE results and STATA software for DMS results

Net impacts with adaptation strategy

Without Potential of irrigation achieved in 2040

Adaptation
strategy 50% 60% 70% 80% 90% 100%

Economic indicators
National real GDP −4.38 −3.96 −3.80 −3.65 −3.51 −3.38 −3.35
National Price Index 8.15 7.21 6.96 6.73 6.52 6.33 6.18
Agricultural GDP −5.03 −4.52 −4.31 −4.12 −3.94 −3.77 −3.63
Agricultural Price Index 12.06 10.31 9.78 9.28 8.82 8.40 8.29
Welfare indicators
Real income per capita
Subsistence farmers −4.18 −3.84 −3.68 −3.53 −3.40 −3.27 −3.03
Other agricultural workers −5.82 −5.76 −5.77 −5.79 −5.81 −5.83 −6.40
Formal workers −4.05 −3.72 −3.64 −3.56 −3.49 −3.42 −3.37
Infomal workers −3.90 −3.61 −3.50 −3.40 −3.31 −3.23 −3.28
Poverty
Headcount Poverty Rate
National 2.21 2.35 2.30 2.24 2.20 2.15 2.15
Rural areas 2.38 1.99 1.89 1.81 1.73 1.66 1.60
Urban areas 2.15 2.81 2.72 2.64 2.57 2.50 2.48
Headcount Food Poverty Rate
National 2.89 3.11 3.03 2.96 2.90 2.84 2.83
Rural areas 3.15 2.30 2.20 2.10 2.00 1.92 1.86
Urban areas 2.49 −1.24 −1.19 −1.14 −1.10 −1.07 −1.04
Food Security
Food Insecurity Ratio
National 0.89 0.86 0.84 0.81 0.79 0.77 0.76
Rural areas 0.91 0.90 0.88 0.86 0.84 0.82 0.82
Urban areas 0.91 0.85 0.81 0.77 0.74 0.71 0.68
Food Access Index
National −2.93 −2.60 −2.46 −2.32 −2.20 −2.08 −1.99
Rural areas −5.07 −4.80 −4.69 −4.59 −4.49 −4.40 −4.48
Urban areas −0.01 0.33 0.49 0.64 0.78 0.92 1.21
Food Availability Index −3.68 −3.29 −3.14 −2.99 −2.85 −2.73 −2.66
Costs-benefits indicators of strategy
Average annual cost (3) – 15.8 18.9 22.1 25.2 28.4 31.5
Average annual national net gain (4) in real GDP – 54.6 71.0 86.5 100.7 113.9 117.2
Annual real GDP gains/costs – 3.46 3.75 3.92 4.00 4.02 3.72
(4)
Average annual net gain in agricultural real GDP – 30.7 40.2 49.3 57.7 65.5 71.9
Annual real agricultural GDP gains/costs – 1.95 2.13 2.23 2.29 2.31 2.28

Notes: (1) Medium long run climate change scenario combined with variability scenario with p = 0.2 and ih = −12.5% (2)Results computed from 30 randomized simulations for each
scenario (3) In USD million, based on You et al. (2011) (4) In USD million, compared to the scenario without adaptation strategy.

agro-meteorological assistance (reinforcement of early warning sys-
tems to manage meteorological risks or pests attacks, etc.), as well as
research programs to develop the varieties better adapted to the
different agro-ecological zones. However, we can ascertain an indirect
estimate of this cost by relying on the previsions made by the Nigerian
government in the context of its adaptation program for 2005–2030
(CNEDD, 2010). It was estimated at USD 540 million (including USD
491 million investment and USD 49 million maintenance operations).
The results of the simulations are presented in Table 8 (see also the
supplementary material). They show that replacing traditional varieties
with modern ones can be an effective strategy for neutralizing the
effects of climate change. In the case in which modern varieties react in
the same way as traditional varieties (34.8% of GCM forecasts), the
threshold of neutralization is reached when modern varieties represent
between 50% and 75% of traditional varieties at the end of the period.
When these modern varieties have greater sensitivity (in 65.2% of GCM
forecasts), this threshold is logically less high (between 75% and 100%)
because the positive effects on yields are in part counterbalanced by the
negative sensitivity to climatic risks. Regarding the costs-benefits
criteria, this strategy also appears to be the most effective compared
to previous ones as its potential benefits greatly exceed its costs.
Reaching the thresholds of neutralization could for instance generate a
national real GDP net gain about USD 483 million annually (USD 200
million for agricultural real GDP) for the first sensibility hypothesis and
about USD 495 million and USD 207 million, respectively, for the
second sensibility hypothesis.
6. Conclusion
This study uses a microsimulated DRCGE model to explore the
potential economic and social impacts over a 25 year period of various
new climate conditions scenarios for Niger. First, it shows how a long
term fall in the main crop yields could reduce the country's economic
growth and considerably degrade its poverty and food insecurity
indicators. Secondly, it also shows how these negative impacts could
be aggravated by a higher frequency of extreme drought events. Third,
it shows that there is some room for manœuvre to anticipate these new
conditions. Improving the rural road network or adopting new crop
varieties could, for example, turn out to be worthwhile coping strategie
whose benefits greatly exceed their costs. Similarly, extending irrigated
areas could usefully contribute to broader policies to combat the effects
of the climate.
Even if these results offer an indication of Niger's vulnerability to
future climate conditions, some caution should however be exercised
when interpreting the absolute magnitudes of estimated economic or
social losses. There are several reasons for this. The first is that our

139
J.-M. Montaud et al. Economic Modelling 63 (2017) 128–142

Table 8
Selected average net impacts(1) of a variety substitution strategy for Millet and Sorghum under different hypotheses for climate sensitivity for varieties and for
substitution achieved in 2040 (Average annual deviation from BAU value in % for 2015–2040(2)).
Source: Own calculations with GAMS software for DRCGE results and STATA software for DMS results

Net impacts with adaptation strategy

Scenario with a same sensibility to climate change for Scenario with a different sensibility to climate change
both variety for variety

Share of modern variety on total cultivars adopted in 2040

25% 50% 75% 100% 25% 50% 75% 100%

Economic indicators
National real GDP −4.38 −2.76 −1.28 0.21 1.46 −3.33 −2.05 −0.73 0.41
National Price Index 8.15 3.33 −0.27 −3.50 −5.98 4.92 1.67 −1.37 −3.80
Agricultural GDP −5.03 −3.65 −2.32 −0.94 0.25 −4.14 −3.00 −1.80 −0.73
Agricultural Price Index 12.06 6.61 2.49 −1.20 −4.08 8.39 4.66 1.16 −1.63
Welfare indicators
Real income per capita
Subsistence farmers −4.18 −3.52 −2.84 −2.18 −1.64 −3.73 −3.10 −2.45 −1.91
Other agricultural workers −5.82 −2.48 0.44 3.34 5.79 −3.69 −1.23 1.31 3.52
Formal workers −4.05 −1.53 0.57 2.58 4.22 −2.41 −0.60 1.20 2.73
Infomal workers −3.90 −2.23 −0.72 0.77 2.02 −2.82 −1.53 −0.20 0.95
Poverty
Headcount Poverty Rate
National 2.21 1.50 0.74 −0.03 −0.67 1.78 1.13 0.44 −0.15
Rural areas 2.38 1.56 0.74 −0.22 −0.98 1.89 1.13 0.34 −0.36
Urban areas 2.15 1.54 0.67 0.36 −0.10 1.78 1.24 0.70 0.26
Headcount Food Poverty Rate
National 2.89 1.94 1.20 −0.06 −0.89 2.31 1.45 0.55 −0.22
Rural areas 3.15 2.05 0.94 −0.28 −1.26 2.49 1.49 0.45 −0.46
Urban areas 2.49 1.79 0.89 0.42 −0.12 2.06 1.43 0.82 0.30
Food Security
Food Insecurity Ratio
National 0.89 0.60 0.30 −0.01 −0.26 0.72 0.45 0.18 −0.06
Rural areas 0.91 0.60 0.26 −0.08 −0.37 0.72 0.43 0.13 −0.14
Urban areas 0.91 0.66 0.40 0.15 −0.05 0.76 0.53 0.30 0.11
Food Access Index
National −2.93 −1.92 −0.89 0.23 1.25 −2.29 −1.44 −0.49 0.38
Rural areas −5.07 −2.91 −0.71 1.71 3.93 −3.70 −1.91 0.12 2.01
Urban areas −0.01 −0.57 −0.98 −1.32 −1.54 −0.39 −0.74 −1.08 −1.33
Food Availability Index −3.68 −2.34 −1.03 0.36 1.60 −2.83 −1.72 −0.53 0.55
Costs-benefits indicators of
strategy
Average annual cost (3) – 5.4 10.8 16.2 21.6 5.4 10.8 16.2 21.6
Average annual national real GDP – 179.5 332.1 482.8 609.2 118.1 247.8 381.0 495.7
net gain (4)
Annual real GDP gains/costs – 33.25 30.75 29.80 28.2 21.86 22.95 23.52 22.95
Average annual agricultural real – 73.0 136.0 200.3 255.7 48.6 101.7 157.7 206.9
GDP net gain (4)
Annual real agricultural GDP – 13.51 12.59 12.36 11.84 9.00 9.42 9.73 9.58
gains/costs

Notes: (1) Medium long run climate change scenario combined with variability scenario with p = 0.2 and ih = −12.5% (2) Results computed from 30 randomized simulations for each
scenario (3) In USD million, based on CNEDD-PNUD (2010) 4) In USD million, compared to the scenario without adaptation strategy.

methodology does not fully address the issue of uncertainty about the
nature and amplitude of climate change or climate variability or their
respective effects on agriculture. Second, our modelling framework
does not allow us to take into account all the effects associated with
new climatic conditions by including certain collateral effects (health
risks, conflict over water, external migrations, etc.). Third, although
some autonomous adaptation responses to climate change are already
endogenously determined in the DRCGE model (accumulation dy-
namics, intersectoral-labour migration, changes of relative price or
patterns of consumption, etc), the reality would probably be more
flexible over the long time period simulated. Fourth, our methodology
does not really fully examine here the question of costs-benefits of the
planned adaptation strategies which are here only taken into account
through their exogenous supply effects on the agricultural sector
without any considerations about their demand effect on the others
activities or even their fiscal space effects. Moreover, we do not address
here the critical question of funding for coping strategies. Contrary to
most low income countries where agricultural production is weather
sensitive and adaptive capacities are low, Niger now occupies a
particular position as it will benefit from additional “windfall revenues”
expected from the new Imouraren uranium deposit as well as new
petroleum production (see for instance Go et al., 2013 or Devarajan
et al., 2015). In this context, an important policy question that could be
also investigated is how this country could harness the increased
income from natural resources to cope with new climate conditions
and achieve growth and poverty reduction over a long period of time.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.econmod.2017.02.005.

140
J.-M. Montaud et al.
References
Al-Riffai, P., Breisinger, C., Verner, D., Zhu, T., 2012. Droughts in Syria: an assessment of
impacts and options for improving the resilience of the Poor. Q. J. Int. Agric. 51 (1),
21–49.
Arndt, C., Thurlow, J., 2014. Climate uncertainty and economic development: evaluating
the case of Mozambique to 2050. Clim. Change 130 (1), 63–75.
Arndt, C., Schlosser, A., Strzepek, K., Thurlow, J., 2015. Climate change and economic
growth prospects for malawi: an uncertainty approach. J. Afr. Econ. 23 (suppl. 2),
ii83–ii107.
Bandara, J.S., Cai, Y., 2014. The impact of climate change on food crop productivity, food
prices and food security in South Asia. Econ. Anal. Policy 44 (4), 451–465.
Ben Mohamed, A., van Duivenbooden, N., Abdoussallam, S., 2002. Impact of climate
change on agricultural production in the sahel - Part 1. Methodological approach and
case study for millet in Niger. Clim. Change 54 (3), 327–348.
Ben Mohamed, A., 2011. Climate change risksin Sahelian Africa. Reg. Environ. Change
11 (S1), 109–117.
Beyene, L.M., Engida, E., 2013. Public Investment in Irrigation and Training for an
Agriculture - Led Development: A CGE Approach for Ethiopia, Partnership for
Economic Policy Working Paper 2013-02. Available at SSRN: http://dx.doi.org/10.
2139/ssrn.2352907.
Bezabih, M., Chambwera, M., Stage, J., 2011. Climate change and total factor
productivity in the Tanzanian economy. Clim. Policy 11 (6), 1289–1302.
Bezançon, G., Pham, J.-L., Deu, M., Vigouroux, Y., Sagnard, F., Mariac, C., Kapran, I.,
Mamadou, A., Gérard, B., Ndjeunga, J., Chantereau, J., 2009. Changes in the
diversity and geographic distribution of cultivated millet (Pennisetum glaucum (L.)
R. Br.) and Sorghum (Sorghum bicolor (L.) Moench) Varieties in Niger between 1976
and 2003. Genet. Resour. Crop Evol. 56 (2), 223–236.
Bosello, F., Campagnolo, L., Eboli F., 2013. Climate Change and Adaptation: The case of
Nigerian Agriculture, Working Papers No.2013.35, Fondazione Eni Enrico Mattei.
Bosello, F., Campagnolo, L., Eboli, F., Parrado, R., Ponce, R.D., 2015. ‘Agricultural
impacts and adaptation through irrigation: a focus on sub-saharan African countries.
Paper presented at In: Proceedings of the 21st Annual Conference European
Association of Environmental and Resource Economists, 24–27 June 2015,
Helsinki, Finland, Retrieved from 〈http://www.webmeets.com/eaere/2015/m/
viewpaper.asp?Pid=884〉.
Breisinger, C., Diao, X., Thurlow, J., Hassan, R.M.A., 2011. ‘Potential impacts of a green
revolution in Africa-the case of Ghana’. J. Int. Dev. 23 (1), 82–102.
Buddelmeyer, H., Hérault, N., Kalb, G., de Jong, M.V.Z., 2012. Linking a
microsimulation model to a dynamic CGE model: climate change mitigation policies
and income distribution in Australia. Int. J. Microsimulation 5 (2), 40–58.
Burke, E.J., Brown, S.J., 2008. Evaluating uncertainties in the projection of future
drought. J. Hydrometeorol. 9 (2), 292–299.
Burke, M., Dykema, J., Lobell, D.B., Miguel, E., Satyanath, S., 2015. Incorporating
climate uncertainty into estimates of climate change impacts. Rev. Econ. Stat. 97 (2),
461–471.
Calderon, C., Serven, L., 2010. Infrastructure and economic development in sub-saharan
Africa. J. Afr. Econ. 19 (Supplement 1), Si13–Si87.
Calzadilla, A., Zhu, T., Rehdanz, K., Tol, R.S.J., Ringler, C., 2013. Economywide impacts
of climate change on agriculture in sub-saharan Africa. Ecol. Econ. 93, 150–165.
Calzadilla, A., Zhu, T., Rehdanz, K., Tol, R.S.J., Ringler, C., 2014. Climate change and
agriculture: impacts and adaptation options in South Africa. Water Resour. Econ. 5,
24–48.
Carruthers, R., Krishnamani, R.R., Murray, S., 2009. Improving Connectivity: Investing
in Transport Infrastructure in Sub-Saharan Africa, AICD Background Paper 7, Africa
Region, World Bank, Washington, DC.
Chalise, S., Naranpanawa, A., 2016. Climate change adaptation in agriculture: a
computable general equilibrium analysis of land-use change in Nepal. Land Use
Policy 59, 241–250.
Chitiga, M., Mabugu, R., Maisonnave, H., 2016. Analysing job creation effects of scaling
up infrastructure spending in South Africa. Dev. South. Afr. 33 (2), 186–202.
CNEDD, 2010. Rapport sur l’Evaluation des Investissements et des Flux Financiers pour
l’Adaptation du Secteur Agriculture/Elevage aux Changements Climatiques,
CNEDD-PNUD.
Decaluwé, B., Lemelin, A., Robichaud, V., Maisonnave, H., 2013. PEP-1-t Standard
Model: single-country, Recursive Dynamic Version, Politique économique et
Pauvreté/Poverty and Economic Policy Network. Université Laval, Québec,
(Retrieved from)〈www.pep-net.org〉.
Devarajan, S., Dissou, Y., Go, D.S., Robinson, S., 2015. Budget Rules and Resource
Booms and Busts: A Dynamic Stochastic General Equilibrium Analysis, World Bank
Econ Rev, lhv059.
Domínguez-Torres, C., Foster, V., 2011. Niger’s Infrastructure: A Continental
Perspective, World Bank Policy Research Working Paper No. 5698, World Bank,
Washington, DC.
Druyan, L., 2011. Studies of 21st- century precipitation trends over West Africa. Int. J.
Climatol. 31 (10), 1415–1424.
Eboli, F., Parrado, R., Roson, R., 2010. Climate-change feedback on economic growth:
explorations with a dynamic general equilibrium model. Environ. Dev. Econ. 15 (5),
515–533.
Epule, E.T., Peng, C., Lepage, L., Chen, Z., 2014. The causes, effects and challenges of
sahelian droughts: a critical review. Reg. Environ. Change 14 (1), 145–156.
Estache, A., Perrault, J.-F., Savard, L., 2012. The impact of infrastructure spending in
sub-saharan Africa: a CGE modeling approach. Econ. Res. Int. 2012, 1–18.
FAO, 2016a. FAOSTAT Statistics Database. Food and Agriculture Organization of the
United Nations, Rome, (Website accessed on [2016/06/28)(. Retrieved from)
141
Economic Modelling 63 (2017) 128–142
〈http://faostat3.fao.org〉.
Food and Agriculture Organization of the United Nations, 2016b. AQUASTAT Website.
Food and Agriculture Organization of the United Nations (FAO). Website accessed
on [2016/06/20]. Retrieved from http://www.fao.org/nr/water/aquastat/.
Gautier, D., Denis, D., Locatelli, B., 2016. Impacts of Drought and Responses of Rural
Populations in West Africa: A Systematic Review 7. Wiley Interdisciplinary Reviews:
Climate Change, 666–681.
Gebreegziabher, Z., Stage, J., Mekonnen, A., Alemu, A., 2016. Climate change and the
ethiopian economy: a CGE analysis. Environ. Dev. Econ. 21 (02), 205–225.
Go, D. S., Robinson, S., Thierfelder, K., Utz, R., 2013. Dutch Disease and Spending
Strategies in a Resource-Rich Low-Income Country — The Case of Niger, Policy
Research working paper; No WPS 6691. Washington D.C; World Bank.
Gwilliam, K., 2011. Africa's Transport Infrastructure: Mainstreaming Maintenance and
Management, Directions in development; infrastructure. Washington, D.C: World
Bank.
Herderschee, J., Backiny-Yetna, P.R., Steele, D.E., 2014. Niger - Trends in Poverty,
Inequality and Growth. D.C.: World Bank Grou, Washington, 2005–2011 ,
(Retrieved from)〈http://documents.worldbank.org〉.
Hertel, T.W., Burke, M.B., Lobell, D.B., 2010. The poverty implications of climate-
induced crop yield changes by 2030. Glob. Environ. Change 20 (4), 577–585.
Herrera, J., Kuépié, M., Nordman, C.J., Oudin, X., Roubaud, F., 2012. Informal Sector &
Informal Employment: Overview of Data for Eleven Cities in Ten Developing
Countries, WIEGO Working Paper no. 9, January. Retrieved from http://wiego.org.
Institut National de la Statistique, 2012. Enquête Nationale sur l’Emploi et le Secteur
Informel au Niger, 2012. Ministère des finances - Institut National de la Statistique.
IPCC, 2013. ‘Summary for Policymakers'. In: Stocker, T.F., Qin, D., Plattner, G.-K.,
Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M.
(Eds.), Climate Change 2013: The Physical Science Basis. Contribution of Working
Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge University Press, Cambridge, United Kingdom and New York,
NY, USA.
Juana, J.S., Strzepek, K.M., Kirsten, J.F., 2008. Households welfare analyses of the
impact of global change on water resources in South Africa. Agrekon 47, 309–326.
Knox, J., Hess, T., Daccache, A., Wheeler, T., 2012. Climate change impacts on crop
productivity in Africa and South Asia. Environ. Res. Lett. 7 (3), 034032.
Liu, J., Fritz, S., van Wesenbeeck, C.F.A., Fuchs, M., You, L., Obersteiner, M., Yang, H.,
2008. A spatially explicit assessment of current and future hotspots of hunger in Sub-
Saharan Africa in the context of global change. Glob. Planet. Change 64 (3–4),
222–235.
Lobell, D.B., Burke, M.B., 2008. Why are agricultural impacts of climate change so
uncertain? The importance of temperature relative to precipitation. Environ. Res
Lett. 3, 034007.
Lobell, D.B., Burke, M.B., 2010. On the use of statistical models to predict crop yield
responses to climate change. Agric. For. Meteorol. 150 (11), 1443–1452.
Masih, I., Maskey, S., Mussa, F.E.F., Trambauer, P., 2014. A review of droughts on the
African Continent: a geospatial and long-term perspective. Hydrol. Earth Syst. Sci.
18 (9), 3635–3649.
Mohino, E., Janicot, S., Bader, J., 2011. Sahel rainfall and decadal to multi-decadal sea
surface temperature variability. Clim. Dyn. 37 (3–4), 419–440.
Niang, I., Ruppel, O.C., Abdrabo, M.A., Essel, A., Lennard, C., Padgham, J., Urquhart, P.,
2014. Africa. In: Barros, V.R., Field, C.B., Dokken, D.J., Mastrandrea, M.D., Mach,
K.J., Bilir, T.E., Chatterjee, M., Ebi, K.L., Estrada, Y.O., Genova, R.C., Girma, B.,
Kissel, E.S., Levy, A.N., MacCracken, S., Mastrandrea, P.R., White, L.L. (Eds.),
Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional
Aspects. Contribution of Working Group II to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change(eds.). Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 1199–1265.
Pauw, K., Thurlow, J., Bachu, M., Van Seventer, D.E., 2011. The economic costs of
extreme weather events: a hydrometeorological CGE analysis fr malawi. Environ.
Dev. Econ. 16 (02), 177–198.
Porter, J.R., Xie, L., Challinor, A.J., Cochrane, K., Howden, S.M., Iqbal, M.M., Lobell,
D.B., Travasso, M.I., 2014. Food security and food production systems. In: Field,
C.B., Barros, V.R., Dokken, D.J., Mach, K.J., Mastrandrea, M.D., Bilir, T.E.,
Chatterjee, M., Ebi, K.L., Estrada, Y.O., Genova, R.C., Girma, B., Kissel, E.S., Levy,
A.N., MacCracken, S., Mastrandrea, P.R., White, L.L. (Eds.), Climate Change 2014:
Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects.
Contribution of Working Group II to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change(eds.). Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 485–533.
Reid, H., Sahlén, L., Stage, J., MacGregor, J., 2008. Climate change impacts on namibia's
natural resources and economy. Clim. Policy 8, 452–466.
Robinson, S., Willenbockel, D., Strzepek, K., 2012. A dynamic general equilibrium
analysis of adaptation to climate change in Ethiopia. Rev. Dev. Econ. 16 (3),
489–502.
Roudier, P., Sultan, B., Quirion, P., Berg, A., 2011. The impact of future climate change
on west african crop yields: what does the recent literature say? Glob. Environ.
Change 21 (3), 1073–1083.
Salack, S., Sarr, B., S., Traore, B., Baron, C., Dinghun, M., 2008. Evaluation of Climate
Change Impacts on Pearl Millet (Pennisetum Glaucum (L.) R. Br.): development and
Production in Niger Using a Crop Model. Centre Régional Agrhymet, Niamey, Niger.
Sassi, M., Cardaci, A., 2013. Impact of rainfall pattern on cereal market and food security
in Sudan: Stochastic approach and CGE model’. Food Policy 43 (C), 321–331.
Schlenker, W., Lobell, D.B., 2010. Robust negative impacts of climate change on African
agriculture. Environ. Res. Lett. 5 (1), 014010.
Stifel, D.C., Thorbecke, E., 2003. A dual-dual CGE model of an archetype African
economy: trade reform, migration and poverty. J. Policy Model. 25 (3), 207–235.
J.-M. Montaud et al.
Sultan, B., Roudier, P., Quirion, P., Alhassane, A., Muller, B., Dingkuhn, M., Ciais, P.,
Guimberteau, M., Traore, S., Baron, C., 2013. Assessing climate change impacts on
sorghum and millet yields in the Sudanian and Sahelian Savannas of West Africa.
Environ. Res. Lett. 8 (1), 014040.
Sylla, M.B., Gaye, A.T., Jenkins, G.S., Pal, J.S., Giorgi, F., 2010. Consistency of projected
drought over the sahel with changes in the monsoon circulation and extremes in a
regional climate model projections. J. Geophys. Res. 115 (D16).
Thomas, T., Rosegrant, M., 2015. Climate change impact on key crops in Africa: using
crop models and general equilibrium models to bound the prediction. In: Aziz
Elbehri (Ed.), Climate Change and Food Systems: Global Assessments and
Implications for Food Security and Trade. Food Agriculture Organization of the
United Nations (FAO), Rome.
Thornton, P.K., Jones, P.G., Ericksen, P.J., Challinor, A.J., 2011. Agriculture and food
systems in Sub-Saharan Africa in a 4 °C+ World. Philos. Trans. R. Soc. A: Math.
Phys. Eng. Sci. 369 (1934), 117–136.
Thurlow, J., Kiringai, J., Gautam, M., 2007. Rural Investments to Accelerate Growth and
Poverty Reduction in Kenya, Discussion paper no. 723. Washington, DC:
International Food Policy Research Institute. Retrieved from http://www.ifpri.org.
Thurlow, J., Zhu, T., Diao, X., 2009. The Impact of Climate Variability and Change on
Economic Growth and Poverty in Zambia, IFPRI Discussion Paper 00890.
Washington DC: International Food Policy Research Institute, Development Strategy
and Governance Division, and Environment and Production Technology Division.
Retrieved from http://www.ifpri.org.
142
Economic Modelling 63 (2017) 128–142
United Nations, 2015. Probabilistic Population Projections based on the World
Population Prospects: The 2015 Revision, Population Division, DESA.
Van Duivenbooden, N., Abdoussalam, S., Ben Mohamed, A., 2002. Impact of climate
change on agricultural production in the sahel - Part 2. Case study for groundnut and
cowpea in Niger. Clim. Change 54, 349–368.
World Bank, 2008. Ethiopia: A Country Study on the Economic Impacts of Climate
Change Environment and Natural Resource Management., Report No. 46946-ET.
Sustainable Development Department, Africa Region. Washington, DC: World Bank.
World Bank, 2010. The Economics of Adaptation to Climate Change (EACC): synthesis
Report. The World Bank Group, Washington, DC.
World Bank, 2013. Agricultural Sector Risk Assessment in Niger: Moving from Crisis
Response to Long-Term Risk Management, Technical Assistance Report Number:
74322-NE. Washington D.C.
You, L., Ringler, C., Wood-Sichra, U., Robertson, R., Wood, S., Zhu, T., Nelson, G., Guo,
Z., Sun, Y., 2011. What is the irrigation potential for Africa? A combined biophysical
and socioeconomic approach. Food Policy 36 (6), 770–782.
Zhao, T., Dai, A., 2016. Uncertainties in historical changes and future projections of
drought. Part II: model-simulated historical and future drought changes. Clim.
Change, 1–14.
Zhong, S., Okiyama, M., Tokunaga, S., 2014. Impact of natural hazards on agricultural
economy and food production in china: based on a general equilibrium analysis. J.
Sustain. Dev. 7 (2), 45.