cahier de recherche

Institut de Recherche en Management et en
Pratiques d’Entreprise
The Groupe ESC PAU Institute for Research in
Management and Best Practices
CAHIER DE
RECHERCHE
N°21
Décembre 2014
1
Cahier de Recherche 21
Sommaire
TRANSFERTS DE FONDS, CAPACITE D’ABSORPTION ET SYNDROME
HOLLANDAIS : CAS DU MAROC
PAR FARID MAKHLOUF
P.3
THE
IMPACT OF
EXCHANGE RATE POLICY
ON
REMITTANCES
IN
MAROCCO : A THRESHOLD VAR ANALYSIS
P.25
PAR FARID MAKHLOUF
DOES
EDUCATION
INFORMATION
AND
MATTER
FOR
THE
ADOPTION
COMMUNICATION TECHNOLOGIES (ICT)
OG
IN
DEVELOPING COUNTRIES? EVIDENCE FROM SENEGAL
PAR MAZHAR MUGHAL, BARASSOU DIAWARA
P.45
2
Cahier de Recherche 21
Transferts de fonds, capacité
d’absorption et syndrome
hollandais : cas du Maroc
Farid MAKHLOUF
Professeur Groupe ESC Pau
IRMAPE
3
Cahier de Recherche 21
RESUME
Pour le Maroc, les transferts de fonds des migrants augmentent de manière continue
et constituent une source non négligeable de financement. Ce papier diagnostique
la présence du syndrome hollandais au Maroc. Pour ce faire, il examine la relation
entre les transferts de fonds et le taux de change réel effectif. En utilisant la
technique bayésienne, nous avons trouvé que les transferts de fonds n’engendrent
pas une appréciation du taux de change effectif.
Mots clés : Maroc, Transferts de fonds, Syndrome hollandais, Analyses bayésiennes
ABSTRACT
Migrant remittances are a steadily rising external source of capital for Morocco, and
constitute a large source of income. This paper studies the empirical relationship
between remittances and Dutch disease in Morocco. To do this, it examines the
relationship between remittances and the real effective exchange rate. Using the
Bayesian technique, we found that remittances do not cause the appreciation of
Morocco’s real exchange rate.
Keywords: Morocco, Remittances, Dutch Disease, Bayesian analysis
4
Cahier de Recherche 21
INTRODUCTION
Les transferts de fonds représentent un phénomène complexe. Ce qui a suscité un foisonnement
d’études et de recherches ces dernières années. De plus, les transferts de fonds effectués par les
migrants vers leur pays d’origine constituent une source de financement importante pour un bon
nombre de pays en développement, le Maroc en fait partie (Makhlouf, 2013). Cette manne financière
peut être utilisée dans les pays en développement comme un substitut à d’autres flux financiers afin
de promouvoir leurs institutions économiques et financières. A cet égard, beaucoup de pays en
développement utilisent cette source pour financer leur développement local (Grabel, 2008).
Cependant, les effets des transferts de fonds sur les économies des pays d’origine restent ambigus.
Pour certains économistes, ces transferts ont un impact positif sur la balance des paiements (Chami
et al., 2005). Pour d’autres, ils peuvent avoir des effets inflationnistes et apprécient le taux de change
1
en causant ce que l’on appelle le « syndrome hollandais » (Bourdet et Falck, 2006). Les effets macroéconomiques des transferts de fonds des migrants sont donc complexes (Grabel ,2008) et diffèrent
d’un pays à l’autre. Ceci est dû principalement aux politiques économiques mises en place dans les
pays d’origine des migrants, mais aussi à la manière selon laquelle ces transferts sont utilisés. De
plus, la plupart des gouvernements des pays en développement interviennent de manière fréquente
2
sur le marché des changes (Krugman et Obstfeld, 2011, p.492).
La littérature économique souligne les risques d’appréciation du taux de change suite aux transferts
de fonds, puisque cela peut provoquer des pertes de compétitivité prix pour les pays bénéficiaires. Par
exemple, Amuedo-Dorantes et Pozo (2004) montrent en utilisant les données de panel pour 13 pays
d’Amérique Latine et des Caraïbes, qu’une augmentation de 100% des envois de fonds engendre une
appréciation de 22% du taux de change réel. Ce risque peut être plus important dans les petits pays
(Kapur, 2004).
Dans ce papier, nous allons étudier, nous vérifierons l’hypothèse selon laquelle les transferts de fonds
provoquent le syndrome hollandais, en étudiant leur impact sur le taux de change effectif et sur la
réallocation des ressources.
La suite de travail sera organisée comme suit : la deuxième section traite la revue de la littérature
selon deux approches; dans la troisième section, nous testerons l’hypothèse selon laquelle les
transferts de fonds engendrent le syndrome hollandais dans le cas du Maroc en utilisant deux types
d’estimations (fréquentiste et baysésienne) ; enfin dans la dernière section nous conclurons ce travail.
1
Le syndrome hollandais se réfère à l’appréciation de la monnaie suite à une entrée massive de capitaux.
Il est à signaler que dans un système de change flexible, le taux de change corrige le déséquilibre de la balance courante.
Dans un régime de taux de change fixe, la variation de la balance courante engendre une variation de la masse monétaire.
2
5
Cahier de Recherche 21
1. REVUE DE LA LITTERATURE : IMPACT DES TRANSFERTS DE FONDS SUR
LE TAUX DE CHANGE
Les ressources financières, engendrées par la migration internationale, peuvent servir à alléger
certaines contraintes budgétaires des ménages dans les pays d’origine. Elles peuvent, également, les
aider à améliorer leur niveau de vie. Par ailleurs, elles peuvent aussi engendrer certains effets non
souhaités sur les économies bénéficiaires. En effet, la littérature économique s’intéresse, également,
aux effets néfastes des transferts de fonds, sur le plan macro et micro-économique. En outre, ces
effets négatifs peuvent se manifester sous forme d’un phénomène connu sous le nom de syndrome
hollandais. Ce phénomène a été analysé initialement par Corden et Neary (1982) et Corden (1984)
dans le cas d’une rentrée massive de devises. Les transferts de fonds peuvent apporter de nouveaux
3
ajustements sur l’économie marocaine, en particulier par l’effet dépense .
La théorie du syndrome hollandais est associée à une appréciation du taux de change suite à une
entrée massive des capitaux étrangers. Cette théorie a été formulée au départ pour les pays
développés, notamment à partir de la découverte du gaz en Hollande dans les années 50. La
pertinence de cette théorie a poussé les économistes à l’appliquer pour les pays en développement.
Cependant, l’application de cette théorie pour les pays en développement peut donner des résultats
différents comparés aux pays développés. Cette différence est due en partie aux différences du
régime de change. En effet, Krugman et Obstfled (2011, p. 664) soulignent certaines rigidités des
régimes de change des pays en développement. Dans le cas des transferts de fonds, Vargas-Silva
(2009) note que leurs effets sur le taux de change effectif réel restent encore ambigus. En effet,
d’après Barajas et al. (2010b), l’impact des transferts de fonds sur le taux de change réel dépend de
la part des transferts de fonds dépensée dans les biens échangeables du degré de l’ouverture
économique, de la mobilité des facteurs entre les secteurs et du comportement cyclique des transferts
de fonds. In hoc sensu, Grabel (2008) considère que les effets à court terme des transferts de fonds
sont similaires à d’autres flux financiers. Par ailleurs, leurs impacts à long terme sont différents selon
les politiques économiques engagées dans les pays bénéficiaires.
Comme nous l’avons déjà précisé, les transferts de fonds représentent une source importante de
capitaux étrangers dans les pays en développement. De plus, ils sont globalement contra-cycliques et
moins volatiles que d’autres flux financiers (Makhlouf, 2014). À première vue, les résultats des études
antérieures concernant l’impact des transferts de fonds des migrants sur le taux de change restent
encore ambigus (Vargas-Silva, 2009). En effet, certaines études montrent l’occurrence du syndrome
hollandais alors que d’autres trouvent des résultats opposés (effets bénéfiques des transferts de
fonds).
Bourdet et Falck (2006) qui ont analysé les effets des transferts sur le Cap-Vert, ont montré que les
aides ainsi que les transferts de fonds ont un effet négatif sur la compétitivité. En utilisant des
données de panel, Lartey et al. (2012) soulignent que les transferts de fonds engendrent un effet de
dépense et de réallocation des ressources. Dans une étude récente (Makhlouf et Mughal, 2013), nous
avons montré, en utilisant les techniques bayésiennes, que les transferts de fonds provoquent le
phénomène du syndrome hollandais au Pakistan. Dans une autre étude (Makhlouf et Chnaina, 2012),
nous avons noté, en utilisant un modèle à correction d’erreur (VCEM), qu’une augmentation de 1% du
ratio des transferts de fonds sur le PIB provoque une appréciation du taux de change réel d’équilibre
de 0,38 %, en Tunisie. Barajas et al. (2010b), en utilisant la technique de cointégration pour des
3
La corrélation entre les transferts de fonds et la consommation finale des ménages marocains est positive est proche de
1(Makhlouf, 2013, p. 47).
6
Cahier de Recherche 21
données de panel, ont abouti aux résultats selon lesquels une appréciation du taux de change effectif
réel suite à un choc des transferts de fonds. A contrario, les résultats de Mongardini et Rayner (2009),
issus d’une étude sur des pays d’Afrique Subsaharienne, indiquent que les transferts de fonds ne
causent pas l’appréciation du taux de change réel d’équilibre. Nous pouvons également indiquer
l’étude de Rajan et Subramanian (2005) qui n’observent aucune preuve de la présence d’une telle
relation.
Il est à noter que les transferts de fonds ont un impact différent selon le régime de change.
Cependant, les autorités peuvent adopter plusieurs régimes de change (Lahrèche-Revil, 2000). En
effet, Singer (2008) est convaincu que les transferts de fonds des migrants exercent des pressions sur
le choix du régime du taux de change.
Nous venons de voir que les transferts affectent le taux de change. Cependant, ils sont eux-mêmes
déterminés, en partie, par la variation du taux de change. La manière d’utiliser les transferts de fonds
détermine en partie les effets de ces derniers sur le taux de change. Selon Glytsos (1997), l’impact
des transferts de fonds est largement lié à la manière dont ils sont utilisés. En d’autres termes, leur
effet est lié à leurs utilisations pour la consommation des biens importés, ou des biens fabriqués
localement, ou des biens échangeables ou non échangeables. Vargas-Silva (2009) explique que suite
à un choc des transferts des migrants mexicains, la réaction de la demande de monnaie est positive,
ce qui peut provoquer un accroissement de la masse monétaire et qui impliquerait une hausse de
l’inflation.
Selon Amuedo-Dorante et Pozo (2006), les envois de fonds peuvent alléger la contrainte budgétaire
des ménages. De ce fait, les transferts peuvent diminuer la demande de crédit. Ainsi, les transferts de
fonds ont un effet plus au moins nuancé sur le développement financier du Maroc (Bouoiyour et
Makhlouf, 2011). Cependant, selon la Banque Mondiale, les transferts de fonds contribuent au
développement financier des pays d’origine. Une étude réalisée sur 99 pays pour la période 19752003 concernant l’impact des transferts sur les dépôts et les crédits, montre que les transferts
contribuent à l’accroissement des crédits et des dépôts par l’intermédiaire du secteur bancaire
(Banque Mondiale, 2006). Selon la même référence, les transferts pourraient ne pas augmenter les
dépôts bancaires s’ils sont utilisés dans la consommation immédiate. Ainsi, le débat sur l’impact des
transferts de fonds sur le taux de change et la demande de monnaie est loin d’être clos. Les
principaux résultats sur le syndrome hollandais sont regroupés dans le tableau 1 ci-après.
Auteurs
4
Période
Méthode
Résultats
Amuedo-Dorantes et Pozo (2004)
1979-1998
Panel 13 pays LAC effet fixe (OLS)
+
Petri et Saadi Sedik (2006)
1964-2005
Jordanie (VCEM)
+
Bourdet et Falck (2006)
1980-2000
Cape Vert (OLS)
+
4
OLS (Ordinary Least Square), VCEM (Vector Correction Errors Model), SVAR(Structural Vector Auto regression), DSGE
(Dynamic Stochastic General Equilibrium), IV (Instrumental Variable) GMM Generalized Moments Method) ASS (Sub-Saharan
Africa countries). PMG: Pooled Mean Group, 2OLS : 2 Ordinary Least Square.
7
Cahier de Recherche 21
Izquierdo et Montiel (2006)
1960-2004
VAR 6 pays
**
Vargas-Silva (2009)
1996-2006
Mexique (SVAR)
+
Lopez, Molina et Bussolo (2007)
1990-2003
Panel effet fixe dynamique modèle IV 20 pays LAC (OLS)
+
1991
El Salvador
+
Acosta et al. (2009)
(DSGE)
Fajnzylber et López (2007)
1990-2003
Singh et al. (2009)
Corrélation, transferts / Taux de change 8 pays LAC
+
36 pays SSA
-
(panel)
Lartey et al., (2012)
1990-2003
109 Pays en développement Panel dynamique (GMM)
+
Mongardini et Rayner (2009)
1980-2006
Panel 15 pays ASS (dynamic fixed-effect)
-
Sy et Tabarraei (2009)
1970-2004
39 Pays
+
PMG
Makhlouf et Mughal (2013)
1980-2008
Pakistan
+
IV- Bayesien
Barajas et al. (2010b)
1980-2007
Panel (138 pays)
+
Cointégration
Makhlouf et Chanaina (2011)
1980-2008
Tunisie
+
VCEM
Beja (2011)
1984-2008
Fayad (2011)
1980 et
1990
Panel (20 pays)
Coupe transversale
+
(*)
IV (2OLS)
8
Cahier de Recherche 21
Kemegue et al. (2011)
1980-2008
SSA
+
Panel dynamique
Hassan et Holmes (2012)
1987-2010
Panel 24 pays
+
VECM
(+) Appréciation de taux de change. (-) dépréciation de taux de change
(**) résultats mitigés
(*) les transferts de fonds n’entravent pas la croissance économiques et les exportations
Comme nous l’avons signalé précédemment, la littérature économique est encore loin de répondre
d’une manière précise à la question de l’impact des transferts des migrants sur le taux de change. Les
causes de la divergence des résultats peuvent être expliquées par la manière dont les transferts de
fonds sont utilisés, mais également par la capacité des économies en développement à « absorber »
les chocs engendrés par ces flux financiers, ainsi que le degré d’ouverture économique. De plus, les
politiques de change adoptées par les pays bénéficiaires peuvent atténuer les effets néfastes des
transferts de fonds sur le taux de change, et bien évidemment sur la compétitivité prix. Dans cet ordre
d’idée, Singer (2008) a montré que les transferts des migrants poussent les autorités monétaires à
opter pour un régime de change fixe. Le choix du régime de change a un impact important sur
l’inflation, les investissements et les relations commerciales (Singer, 2008). Selon le même auteur, les
transferts de fonds peuvent compenser les imperfections de la politique monétaire.
La stérilisation des transferts de fonds peuvent constituer un remède contre le syndrome hollandais,
mais Lopèz et al. (2007) ont déconseillé la politique de stérilisation des transferts de fonds, qui selon
eux s’avère coûteuse en matière budgétaire. La question qui se pose alors est de savoir dans quelle
mesure les interventions de la Banque Al Maghreb (BAM) sur le marché monétaire en général, et sur
le marché de change en particulier, permettent de stabiliser la variabilité du taux de change. Pour
essayer de répondre à cette interrogation, il nous faut commencer par étudier le régime de change
5
marocain, même si les travaux de Bouoiyour et al. (2004) nous ont fourni certains éléments clés de
réponses sur la parité de change du dirham. L’ancrage du dirham à deux principales devises que sont
l’euro et le dollar américain lui assure une certaine stabilité. Mais il nous reste à savoir si cette
politique est préjudiciable pour la compétitivité des exportations marocaines ou non. La BAM intervient
sur le marché de change afin de le maintenir dans une fourchette bien définie. Pour éponger la
surliquidité engendrée en partie par les transferts de fonds, la BAM utilise la politique des réserves
obligatoires. Le marché monétaire est caractérisé par une surliquidité, notamment pour la période
1999 -2007, ce qui s’explique par la politique de libéralisation au Maroc. En ce qui concerne l’évolution
5
Leurs travaux portent sur le taux de change réel d’équilibre et la politique de change au Maroc : une approche non
paramétrique.
9
Cahier de Recherche 21
du taux de change effectif, Makhlouf (2013) montre qu’il y a une dépréciation jusqu’à l’année 1990,
puis une appréciation dans les années 1990, période où l’inflation est relativement forte (par exemple
le taux d’inflation est proche de 8% en 1991).
2. ANALYSE EMPIRIQUE
2.1.
APPROCHE FREQUENTISTE
Dans cette section, nous allons estimer l’impact des transferts de fonds sur le taux change effectif au
Maroc en utilisant la méthode GMM. Le modèle estimé est inspiré des travaux de Lartey et al. (2012).
Les variables explicatives sont principalement les fondamentaux du taux de change utilisé par
Makhlouf, (2013). Les résultats sont donnés dans le tableau2. Ils montrent que les transferts de fonds
n’ont pas un impact significatif sur le taux de change effectif. Les transferts de fonds n’engendrent pas
une appréciation du taux de change. Ce résultat peut s’expliquer par : la stérilisation des transferts de
fonds ; le manque de lien avec le cycle des affaires au Maroc ; les données sur les transferts de fonds
ne représentent pas la réalité ; une capacité d’absorption adéquate des transferts de fonds.
6
Les interventions de la BAM s’avèrent efficaces puisqu’elles maintiennent une certaine stabilité des
taux de change. Cependant, il existe un paradoxe concernant l’impact des transferts de fonds sur le
TNT. Cet impact est négatif et significatif, ce qui revient à dire que les transferts de fonds engendrent
une réallocation des ressources du secteur échangeable vers le secteur non échangeable. Ainsi, le
secteur échangeable perd en compétitivité sans que le taux de change ne s’apprécie. Les estimations
montrent que les ressources se déplacent du secteur de l’agriculture (signe négatif et significatif des
transferts) vers les services (signe positif et significatif).
Le cas marocain semble être intéressant de par le fait que le syndrome hollandais touche d’une
manière partielle l’économie marocaine. De plus, le syndrome hollandais est la réaction optimale
d’une économie suite à une entrée non anticipée et massive de capitaux étrangers. Cependant, dans
le cas marocain, où son économie est en plein développement, non seulement ces ressources
étrangères ne provoquent pas le syndrome hollandais, mais elles aident le secteur manufacturier à se
développer. Les autorités marocaines semblent avoir compris l’enjeu et l’impact des transferts de
fonds. Cela se traduit par ses interventions sur des indicateurs monétaires nominaux comme la masse
monétaire (via les réserves obligatoires) et le taux de change, par le biais du marché monétaire et du
marché des changes. En revanche, il s’avère difficile d’orienter l’utilisation des transferts de fonds.
Nous savons que la manière d’utiliser ces transferts de fonds a des conséquences non négligeables
sur l’économie.
6
Banque Al-Maghrib c’est la banque centrale du Maroc.
10
Cahier de Recherche 21
Tableau 2 : Résultats (GMM)
Table 2 MCO variable dépendante taux de change effectif réel
Constante
Transferts de fonds
PIB par tête
Masse Monétaire/PIB
Croissance (-1)
Ouverture
Termes de l'échange
R2 Ajusté
Taux
de
Change
effectif
TNT
V,Agriculture
V,Industrie
V,Service
6,470
-0.055
-7.833
1.645
-0.320
(1.738)*
(-0.210)
(-2.942)**
(0.830)
(-0.254)
0.059
-0.279
-0.351
-0.03
0.132
(0.397)
(-2.64)***
(-3.293)***
(-0.487)
(2.620)*
-1.322
(-2.934)***
0.488
0.930
-0.09
-0.251
(1.528)
(2.885)***
(-0.396)
(-1.645)
0.344
-0.0762
-1.105
-0.006
-0.04
(1.431)
(-0.447)
(-0.615)
(-0.047)
(0.463)
0.412
0.100
0.327
-0.097
-0.048
(1.474)*
(0.601)
(1.934)*
(-0.774)
(-0.601)
0.076
-0.545
-0.562
-0.145
0.260
(0.281)
(-2.83)***
(-2.895)**
(-1.007)
(2.831)***
0.912
-1.035
-0.635
-0.503
0.501
(2.095)**
(-3.35)***
(-2.041)**
(-2.174)*
(3.402)**
0.22
0.72
0.64
0.663
0.718
TNT
V,Agriculture
V,Industrie
V,Service
22.434
-8.331
-21.739
-6.672
12.709
(3.493)***
(-2.39)***
(-5.880)
(-3.704)***
(4.453)***
MMG
Taux
change
effectif
Constante
de
11
Cahier de Recherche 21
Transferts de fonds
PIB par tête
Masse Monétaire
Croissance (-1)
Ouverture
Termes de l'échange
-0.154
(-0.88)
-0.210
-0.513
0.157
0.143
(-2.036)**
(-4.312)***
(1.839)***
(2.443)**
1.310
2.436
0.369
-1.479
(3.499)***
(5.674)***
(1.468)
(-4.437)***
1.100
-0.331
-0.848
-0.314
0.577
(3.117)***
(-1.942)**
(-4.618)***
(-2.893)***
(3.484)***
1.200
-0.335
-0.157
-0.296
0.268
(3.984)***
(-1.429)
(-0.781)
(-3.265)***
(2.031)**
0.511
-1.024
-0.677
0.286
-0.104
(2.087)**
(-4.88)***
(-4.312)***
(2.717)****
(-0.680)
0.216
-0.850
-0.447
0.775
-0.303
(0.434)
(-6.03)***
(-2.945)***
(2.983)***
(0.102)
-2.889
(-4.052)***
*** significativité à 1%,**à 5%, * à 10%
( ) : t test
12
Cahier de Recherche 21
Dans le cas du Maroc, les transferts de fonds sont davantage orientés vers la construction et les
services, ce qui nuit au secteur agricole. En d’autres termes, la main-d’œuvre quitte l’agriculture pour
aller vers les secteurs industriels par le fait des transferts de fonds. Cela peut provoquer une hausse
des salaires dans ce secteur, donc une hausse des prix des produits agricoles, ce qui peut engendrer
à l’horizon une perte de compétitivité du secteur agricole. Cependant, l’abondance de la main d’œuvre
permet de garder les salaires stables malgré l’augmentation de la demande. Dans le but de se
prémunir contre le syndrome hollandais, le gouvernement marocain ne doit pas se focaliser
uniquement sur la variation du taux de change, mais également sur la manière d’utiliser ces transferts.
Le degré de liberté dans cette estimation est égal à 22, cela peut nuire à la qualité de l’ajustement.
Pour remédier à ce problème de données, nous proposons dans la section ci-dessous une estimation
bayésienne.
2.2.
APPROCHE BAYESIENNE
Le cadre fréquentiste utilisé précédemment emploie un raisonnement déterministe, afin d’estimer les
paramètres décrivant les effets des facteurs explicatifs sur le taux de change, notamment les
transferts de fonds. Dans cette sous-section, nous utilisons deux niveaux d’incertitudes sur les
paramètres (a priori et a posteriori). Dans ce cas, les paramètres sont considérés comme des
variables.
Comme nous l’avons rappelé précédemment, notre période d’estimation est relativement courte
(1980-2009). Nous ne disposons pas d’assez d’observations pour pouvoir utiliser les méthodes
économétriques usuelles sans risque concernant la fiabilité des résultats. L’utilisation de l’analyse
bayésienne nous permet de surmonter ce problème. La loi conditionnelle f(θ/X) (distribution a
posteriori) s’obtient par la formule de Bayes. La théorie de Bayse stipule la relation suivante :
f ( | X ) 
f ( ) f ( X |  )

 f ( ) f ( X |  )d

Où : θ paramètres et X les données.
Ou encore le paradigme bayésien peut être synthétisé dans le schéma 1 ci-après
13
Cahier de Recherche 21
Schéma 1 : Paradigme bayésian
hypothèses
𝜋(𝜃)
informations a priori
données / hypothèses
hypothèses /données
Où : π() la distribution a priori de paramètre
paramètre
.

𝜋(𝑋/𝜃)
𝜋(𝜃/𝑋)
et π(x/) la distribution de la densité x sachant le
π(/x) est appelé distribution a posteriori.
L’approche bayésienne permet d’intégrer l’information a priori et de l’actualiser avec les données
observées. En effet, la distribution a priori peut être interprétée comme des croyances ou des
informations subjectives sur les paramètres. Cette information peut venir des études antérieures ou
d’un expert sur le sujet à étudier. Et la distribution a posteriori peut être considérée comme
l’actualisation des informations a priori sur les paramètres avec les données observées. Par ailleurs,
dans certain cas, il est difficile de trouver des informations a priori sur les paramètres. Dans ce cas, la
meilleure méthode est d’utiliser un a priori non informatif (Box et Tiao, 1973). De plus, dans l’inférence
bayésienne, la tache la plus difficile est de trouver l’information a priori (Parent et Bernier, 2007).
L’analyse bayésienne admet que les distributions de probabilité soient connues. Cependant, la
connaissance des distributions a priori et a posteriori ne permettent pas souvent de calculer les
distributions marginales a posteriori. En effet, dans de nombreux cas il n'y a pas de solution
analytique ; d’où le recours aux techniques de (Monte Carlo, l'échantillonnage de Gibbs) pour calculer
les distributions marginales a posteriori. L’essor de l’informatique a rendu ces techniques
d’approximations abordables pour les chercheurs et les scientifiques. Parmi ces techniques, nous
trouvons l'échantillonnage de Gibbs. Ce dernier est très adapté à notre cas.
Nous utilisons ici la spécification de l’échantillonnage de Gibbs avec un modèle à variable
7
instrumentale (Gibbs Sampler for Linear 'IV' Model ). Le recours aux variables instrumentales se
7
L’estimation est réalisée à l’aide du package ‘bayesm’ sur le R
http://cran.r-project.org/web/packages/bayesm/bayesm.pdf
14
Cahier de Recherche 21
justifie par le fait que les transferts de fonds sont endogènes. L’instrument adéquat dans ce genre de
situation est le PIB des pays d’accueil.
8
De manière plus précise, selon Rossi (2012) nous considérons la modélisation suivante:
 X  Z '   1

Y   X  W '    2
( ,  ) ~ N (0, )
 1 2
Où:
X : correspond aux transferts de fond.
Z: est le PIB du pays d’accueil.
Y: est le taux de change réel effectif.
W: est l’ensemble des variables explicatives à savoir:
Remit: Transferts de fonds ;
TOT: Termes de l’échange ;
OPEN: Ouverture ;
M2: Masse monétaire ;
GDPpercapita: PIB par tête ;
Growth: Croissance.
8
Pour plus de détails voir http://cran.r-project.org/web/packages/bayesm/bayesm.pdf
15
Cahier de Recherche 21
Les expressions des « a priori » sont données par suivant (Rossi, 2012) :
 ~ N(m , A-1 ) , ( ,  ) ~ N(m , A1 ) and  ~ IW( , V)
Où:
m : la moyenne a priori de  .
A  : la matrice de variance-covariance de l’a priori  .
m  : la moyenne a priori du vecteur des paramètres  ,  .
A : la matrice de variance-covariance du vecteur des paramètres  ,  .
 : d.f. parm for IW prior on 
(5)
V : pds location matrix for IW prior on

Nous allons considérer les mêmes valeurs a priori que celles de McCulloch et Rossi. En effet, nous ne
9
disposons pas d’a priori informatif .
Les résultats du tableau 3 donnent la moyenne a posteriori de chaque paramètre et son écart type.
La distribution a posteriori est donnée en annexe (figure 1)
La variable endogène est le taux de change réel effectif. Les résultats présentés concernent le
modèle bayésien avec comme variable instrumentale le PIB du pays d’accueil, pondéré par le poids
des transferts. Un signe positif (négatif) dans le tableau 3 correspond à une appréciation
(dépréciation) du taux de change.
Les résultats montrent que le Maroc ne souffre pas du syndrome hollandais. En d’autres termes, les
transferts de fonds ne causent pas une baisse de compétitivité du pays. Les résultats de la technique
bayésienne corroborent les résultats du GMM et le modèle VAR structurel. Donc, nous pouvons
infirmer, dans le cas marocain, l’hypothèse selon laquelle les transferts de fonds engendrent le
syndrome hollandais.
9
m =0; A  =0.01; m  = 0; A = 0.01;  =5; V =0.
16
Cahier de Recherche 21
Tableau 3 : Impact des transferts de fonds sur le taux de change – Variable instrumentale (PIB
du pays hôte)
Mean/Median
SD
-0.715069
2.640014
Remit
-0.057
0.38
TOT
0.7141
0.6773
Open
0.1898
0.4312
M2
-0.4825
0.2100
GDPpcapita
0.3288
0.5390
Growth
-0.0057
0.0093
Intercept
Le taux de change est au certain. Un signe positif (négatif) est
équivalent à une appréciation (dépréciation). SD : standard deviation
17
Cahier de Recherche 21
CONCLUSION
L’objectif principal de ce travail est d’étudier l’impact des transferts de fonds sur l’économie
marocaine. Il s’agit de vérifier l’hypothèse selon laquelle les transferts de fonds engendrent le
syndrome hollandais. L’apport de ce papier ne se limite pas à cette idée car il examine aussi cette
hypothèse d’une manière approfondie en utilisant deux techniques d’estimations. Dans le cas du
Maroc, la question que nous nous somme posée, dans l’introduction, est de savoir si les transferts des
migrants peuvent engendrer une appréciation du taux de change et donc s’ils freinent la compétitivité
prix du secteur exposé à la concurrence internationale. Nos résultats montrent clairement que les
transferts de fonds n’engendrent pas une appréciation de taux change et n’engendrent pas une perte
de compétitivité. Ces résultats sont robustes aux modèles utilisés et aux différentes spécifications.
La gestion des transferts des migrants par les autorités monétaires marocaines semble être efficace
(du moins par rapport à cette problématique des envois de fonds). La gestion des liquidités bancaires
dans ce pays permet d’éviter une augmentation de la demande de monnaie suite à un choc des
transferts. Le régime de change marocain et les interventions sur le marché monétaire, et notamment
sur le marché des changes de la BAM ont permis d’écarter le risque du syndrome hollandais. Ce
résultat marque une rupture avec l’idée selon laquelle les transferts de fonds affectent généralement
d’une manière négative la compétitivité.
18
Cahier de Recherche 21
ANNEXES
Figure1 : Distributions a posteriori
1 : TOT ; 2 : OPEN, 3 : M2, 4 : GDPpercapita, 5 : Growth.
Distribution a posteriori (Transferts de fonds)
19
Cahier de Recherche 21
BIBLIOGRAPHIE
Acosta, P. A., Lartey, E. K.K., Mandelman, F. S., (2009) “Remittances and the Dutch disease,”
Journal of International Economics, Elsevier, vol. 79(1), pages 102-116, September.
Amuedo-Dorantes, C., Pozo, S., (2004) “Workers'
remittances
and the real exchange rate:
a paradox of gifts”, World Development 32, pp. 1407–1417.
Amuedo-Dorantes, C., Pozo, S., (2006) “Remittances as Insurance: Evidence from Mexican
immigrants.” Journal of Population Economics, 19:227-254.
Banque Mondiale (2006) Global Economic Prospcets: Economic Implications of Remittances
and Migration, Washington, D C., World Bank.
Barajas, A., Chami, R., Hakura, D., Montiel. P., (2010b) “Workers' Remittances and the
Equilibrium Real Exchange Rate: Theory and Evidence”. Department of Economics Working
Papers 2010-14, Department of Economics, Williams College.
Beja Jr., E. L., (2011) “Do international remittances cause Dutch disease?,” Migration Letters,
Migration Letters and The London Publishers, London, UK, vol. 8(2), pages 132-140, October.
Bouoiyour, J., Marimoutou V., Rey S. (2004) Taux De Change Réel D’équilibre Et Politique De
Change Au Maroc : Une Approche Non Paramétrique. Économie internationale 97, p. 81-104.
Bouoiyour, J., Makhlouf. F., (2011) “Est-ce que les Transferts de fonds promeuvent le
développement financier? Cas du Maroc ”. CATT 2011- UPPA.
Bourdet, Y., Falck, H., (2006) “Emigrants’ Remittances and Dutch Disease in Cape Verde”.
International Economic Journal, 20: Volume 20, Issue 3, 2006 267-284.
Box, G.E.P., Tiao, G.T., (1973) “ Bayesian inference in statistical analysis.” Reading:AddisonWesley.
20
Cahier de Recherche 21
Chami, R., Fullenkamp, C., Jahjah. S., (2003) “Are Immigrant Remittance Flows a Source of
Capital for Development?” IMF Working Papers 03/189 (Washington: International Monetary
Fund).
Corden, W. M., (1984) “Booming Sector and Dutch Disease Economics: Survey and
Consolidation.” Oxford Economic Papers, Vol. 36, p. 359-380.
Corden, W. M., Neary, J.P., (1982) “Booming Sector and Deindustrialization in a Small Open
Economy.” Economic Journal, Vol. 92, n°368, p. 825-48.
Corden, W. M., Neary. P., (1982) “Booming sector and de-industrialisation in a small open
economy.” The Economic Journal, vol. 92, n° 368, 1982, p. 825-849.
Fajnzylber, P., López, H., (2007) “Close to Home: The Development Impact of Remittances to
Latin America”. World Bank, Washington DC.
Fayad, G., (2011) “Remittances: Dutch disease or export-led growth?,” OxCarre Working
Papers 057, Oxford Centre for the Analysis of Resource Rich Economies, University of Oxford.
Glytsos, N.P., (1997) “Remitting Behavior of ‘Temporary’ and ‘Permanent’ Migrants: The Case
of Greeks in Germany and Australia.” Labour CEIS 11(3): 409–35.
Grabel, I., (2008) “The Political Economy of Remittances: What Do We Know? What Do We
Need to Know?” Political Economy Research Institute Working Paper series N° 184.
Hassan, G., Holmes, M., (2012) “Remittances and the real effective exchange rate,” MPRA
Paper 40084, University Library of Munich, Germany.
Izquierdo, I., Montiel, P., (2006) “Remittances and Equilibrium Real Exchange Rates in Six
Central American Countries.” Department of Economics, Williams College.
Kapur, D., (2004) “Remittances: The New Development Mantra?” UNITED NATIONS. No. 29,
April 2004.
21
Cahier de Recherche 21
Kemegue, F. M., Eyden, R. V. , Owusu-Sekyere, E., (2011) “Remittances and the Dutch
disease in Sub-Saharan Africa. A Dynamic Panel Approach,” Working Papers 259, Economic
Research Southern Africa.
Krugman, P., Obstfeld, M., (2011) “Economie Internationl” Edition Pearson Education ISBN:
978-2-7440-7327-4.
Lahrèche-Revil, A., (2000) “Les régimes de change” CEPPI, l’Economie Mondiale 2000.p9310. http://www.cepii.fr/%5C/francgraph/publications/ecomond/lepointsur/2000ch8.pdf
Lartey, E.K.K., Mandelman, F., Acosta, P. A., (2012) “Remittances, exchange rate regimes,
and the Dutch disease: a panel data analysis.” Review of International Economics, Wiley
Blackwell, vol. 20(2), pages 377-395, 05
Lopez, H., Molina, L., Bussolo, M., (2007) “Remittances and the real exchange rate,” Policy
Research Working Paper Series 4213, The World Bank.
Makhlouf, F., (2013) “Transferts de fonds vers le Maroc, enjeux, comportement et impacts”
Thèse soutenue à l’Université de Pau et des Pays de l’Adour.
Makhlouf, F. (2014) "Propriétés cycliques des transferts de fonds des migrants marocains,"
Working Papers 2013-2014_9, CATT - UPPA - Université de Pau et des Pays de l'Adour,
revised Feb 2014
Makhlouf, F., Chnaina, K., (2011) “Impact des Transferts de Fonds sur le Taux de Change
Effectif Réel en Tunisie.” Première rencontre internationale méditerranéenne : nouveaux
enjeux, perspectives, économiques et sociales des pays de la rive sud. Les 29, 30 Septembre
et 1er Octobre 2011 à Hammamet (Tunisie).
Makhlouf, F., Deisting. F., Naamane. A., (2012) “Développement Financier, Flux Financiers Et
Croissance Economique.” Ecole de Commerce Pau, IRMAPE (Institut de Recherche en
Management et en Pratiques d’Entreprises).
22
Cahier de Recherche 21
Makhlouf, F., Mughal, M., (2013) “Remittances, Dutch Disease, and Competitiveness - A
Bayesian Analysis”. Journal of Economic Development à paraître.
Mongardini, J., Rayner,B., (2009) “Grants, Remittances, and the Equilibrium Real Exchange
Rate in Sub-Saharan African Countries.” IMF Working Paper No. 09/75.
Parent, E., Bernier, J., (2007) “Le raisonnement bayésien modélisation et inférence” ISBN :
978-2-287-33906-6 Springer-Verlag Paris.
Petri, M., Saadi-Sedik, T., (2006) “To Smooth or Not to Smooth - The Impact of Grants and
Remittances on the Equilibrium Real Exchange Rate in Jordan,” IMF Working Papers 06/257,
International Monetary Fund.
Rajan, R. G., Subramanian, A. n., (2005) “Aid and Growth: What Does the Cross-Country
Evidence Really Show?” NBER Working Papers 11513, National Bureau of Economic
Research, Inc.
Rossi, P., (2012) “bayesm: Bayesian Inference for Marketing/Micro-econometrics” http://cran.rproject.org/web/packages/bayesm/index.html
Singer, D. A., (2008) “Migrant Remittances and Exchange Rate Regimes in the Developing
World.” Prepared for presentation at the Midwest Political Science Association annual meeting,
April 3-6, 2008.
Singh, R. J., Haacker, M., Lee, K.-W., (2009) “Determinants and Macroeconomic Impact of
Remittances to Sub-Saharan Africa.” IMF Working Paper 09/216. International Monetary Fund,
Washington, DC.
Sy, M., Tabarraei, H., (2010) "Capital inflows and exchange rate in LDCs: The Dutch disease
problem revisited," PSE Working Papers halshs-00574955, HAL.
23
Cahier de Recherche 21
Vargas-Silva, C., (2009) “The Tale of Three Amigos, Remittances, Exchange Rate and Money
Demand in Mexico” Review of Development Economics, Wiley Blackwell, vol. 13(1), pages 114, 02.
24
Cahier de Recherche 21
The Impact of Exchange Rate
Policy on Remittances in Morocco:
A Threshold VAR Analysis
Farid MAKHLOUF
Professeur Groupe ESC Pau
IRMAPE / CATT
25
Cahier de Recherche 21
ABSTRACT
The aim of this paper is to study the effect of nominal exchange rate movements (MAD to EUR) on
remittances in the case of Morocco. It analyses monthly data from 2005 to 2014 using a Threshold
Vector Auto Regression (TVAR) model to document the impact of exchange rate policy on
remittances to Morocco. The results indicate that there is one best unique threshold at
Euro/Moroccan Dirham= 11.2048: under the threshold, the effect of nominal exchange rate
appreciation on remittances is positive, and above the threshold the effect is negative. These
empirical results provide significant implications for the Central Bank of Morocco.
Keywords: Switch regime, TVAR, Remittances, Exchange rate, Morocco
26
Cahier de Recherche 21
INTRODUCTION
Remittances have become an important source of financing for developing countries due to their
volume as well as their impact on the economies of developing countries. They have increased
significantly since the 2000s and are resilient in times of economic crisis in recipient countries
(Ratha, 2007). The average annual growth of remittances sent by migrants in the world is 10.32%
for the period 1990-2008. The inflow of remittances to Morocco increased at an annual rate of
13.75 % from 2000 to 2008 according to the World Bank data. Morocco is among the major
recipients of remittances. Further, Morocco is among the top 15 largest foreign remittance
receiving developing countries in the world (Makhlouf and Naamane, 2013). Furthermore, over
19% of remittances in the Arab world are destined for Morocco. This is due to the strategies
implemented by the Moroccan authorities. Indeed, Moroccan authorities consider that remittances
can be used as one of many tools for development.
This paper focuses on the impact of exchange rate policy on remittances to Morocco. The
exchange rate policy remains the important macroeconomic policy in developing countries
(Cooper, 1999). In fact, the exchange rate influences the price of goods and services. Exchange
rate also exerts a strong influence on remittances (Makhlouf, 2013). Remittances are one of the
most visible consequences of the international migration process. They are considered as countercyclical with respect to the income in recipient countries (Frankel, 2009). Bettin et al. (2014) find
that remittances are negatively correlated with the business cycle in country of origin. Sayan
(2006) highlights that remittances can be pro-cyclical or a-cyclical. Sending money is a complex
decision involving different variables such as exchange rate and interest rate. Remittances are the
result of a mixture of pure altruism and self-interest (Lucas and Stark, 1985). Remittances are
generated by individual decisions which are influenced by a macroeconomic environment in the
host countries. However, the macroeconomic environment in the home countries also affect the
decisions to send money by migrants. Furthermore, several factors explain the behavior of
remittances over time and among host countries. Vargas -Silva and Huang (2006) highlight that
remittances are more sensitive to shocks operating in the host country than in home countries.
27
Cahier de Recherche 21
Most studies on remittance behavior are microeconomic (Köksal, 2006). Very little empirical
research is interested on the relationship between exchange rate policy and remittances behavior.
Further, few studies focus on the behavior of remittance in the
Arab Maghreb Union countries
(Miotti et al., 2010). In the case of Morocco, there is no research on remittance behavior. It should
be noted that the behavior in terms of migrant remittances may vary depending on the sensitivity
of remittances on the exchange rate. Similarly, the uncertainties related to the business cycle
affect the behavior of remittances (Mughal and Makhlouf, 2011). For example, in times of natural
disasters remittances may increase, because they are motivated by altruistic behavior. They are
considered counter-cyclical and stable (Ratha, 2007). In addition, remittances play a significant
role in reducing the amplitude of business cycles in the country of origin (Mughal and Makhlouf,
2011).
The Moroccan government has implemented policies that include mobilizing and channeling
savings of its migrants to the local economy to promote the development of the country
(Bouoiyour, 2006). The Moroccan government also aims to simplify the procedures for remitting
money. These initiatives are very interesting because they can help Moroccan migrants to stay
connected with their country of origin and participating in its development. Remittances occupy a
prominent place in the economic policies of most developing countries (Agunias, 2006). In this
sense, since the 1960s, the Moroccan government encourages emigration policies (MPI, 2005).
Indeed, the Moroccan migration policies have been designed to strengthen the links between
Moroccans living abroad and Morocco (Bouoiyour, 2006). The reminder of this paper is organized
as follow: section two provides a brief discussion on the relationship between remittances and
exchange rate; furthermore, section three talks about econometric methodology and provides the
empirical results; finally, section four concludes this paper by providing some policy implications of
this study.
28
Cahier de Recherche 21
1. REMITTANCES AND EXCHANGE RATE: AN OVERVIEW
The relationship between remittances and exchange rate is bi-causal. This section is divided into
two parts. The first part addresses the impact of remittances on exchange rate. The second part
highlights the impact of exchange rate on remittances.
1.1.
THE IMPACT OF REMITTANCES ON EXCHANGE RATE
Bourdet and Falck (2006) study the impact of remittances on exchange rate in Cape Verde, their
results show that remittances cause the real effective exchange rate to appreciate. Using panel
data, Lartey et al. (2012) note that remittances prompt an appreciation of exchange rate. Makhlouf
and Mughal (2013) using Bayesian techniques, show that remittances appreciate the real
exchange rate in Pakistan. In another study Makhlouf and Chnaina (2011) by using a vector
correction error model (VCEM), find that a 1% increase in the ratio of remittances to GDP causes
an appreciation of the real equilibrium exchange rate by 0.38% in Tunisia. Barajas et al. (2010),
using the technique of cointegration in panel data, find that a shock of remittances causes an
appreciation of the real effective exchange rate. In contrast, the results of Mongardini and Rayner
(2009), from a study of Sub-Saharan Africa countries, indicate that remittances do not cause the
appreciation of the equilibrium real exchange rate. Rajan and Subramanian (2005) observed no
relationship between remittances and exchange rate. It should be noted that remittances have a
different impact depending on the exchange rate regime. Indeed, Singer (2008) is convinced that
remittances from migrants are putting pressure on the choice of exchange rate regime. Table 1
summarizes various effects of remittances on the exchange rate.
TABLE 1: THE IMPACT OF REMITTANCES ON THE EXCHANGE RATE
Authors
Amuedo-Dorantes and Pozo (2004)
Authors
+
Acosta et al. (2009)
+
29
Cahier de Recherche 21
Petri and Saadi-Sedik (2006)
+
Lartey et al., (2012)
+
Bourdet and Falck (2006)
+
Mongardini et Rayner (2009)
-
Izquierdo and Montiel (2006)
*
Sy and Tabarraei (2009)
+
Vargas-Silva (2009)
+
Beja (2011)
+
Lopez, Molina and Bussolo (2007)
+
Fayad (2011)
(*)
(-): negative effect , (+):positive effect , (*):no evidence
1.2.
THE IMPACT OF EXCHANGE RATE ON REMITTANCES
Some macroeconomic factors such as inflation and exchange rate may affect the flow of
remittances. This part is specifically interested in how previous studies investigated the impact of
exchange rate on remittances. El-Sakka and McNabb (1999) show that both exchange rate and
interest rate are an important determinant of remittances in the case of Egypt. The volatility of the
exchange rate may also influence the decision of migrants to remit money (Barro et al., 2007).
According to Faini (2007), changes in the real exchange rate causes two main effects: income and
substitution effects. Conversely, Straubhaar (1986) notes that remittances are not affected by
changes in exchange rates in the case of Turkey. A depreciation of the Indian currency leads to an
increase in remittances in short term, but in the long term, it leads to a decrease of the remittances
(Sirkeci et al., 2012). Yang (2008) notes that in the case of the Philippines, remittances increase
as a result of depreciation of the Peso. Table 2 summarizes various effects of exchange rate on
remittances
30
Cahier de Recherche 21
TABLE 2: THE IMPACT OF EXCHANGE RATE ON REMITTANCES
Authors
Authors
Russell (1986)
*
IMF(2005)
-
Elbadawi and Rocha (1992)
-
Aydas et al. (2006)
-
Faini (1994)
+
Lueth et Ruiz-Arranz (2007)
-
Lianos (1997)
-
Freund et Spatafora (2008)
-
Buch et al. (2002)
*
Singh et al. (2009)
-
Gupta (2005)
*
Bouoiyour, (2013)
+
(-): negative effect , (+):positive effect , (*):no evidence
2. EMPIRICAL ANALYSIS
This section assesses the impact of exchange rate on remittances. A threshold VAR model was
established in this section. As mentioned in section 2, the relationship between exchange rate and
remittances is bi-directional. Since both variables are endogenous, to study the interrelationships
between those variables, the VAR model is considered as an optimal model (Joiner, 2001).
Indeed, remittances are influenced by the variation of exchange rate and the exchange rate is also
impacted by remittances (Singer, 2008). The idea here is to use a nonlinear estimations approach
to explain the impact of devaluation and revaluation of the Moroccan exchange rate (MAD/ Euro)
on remittances. Most of empirical studies modeling the impact of exchange rate on remittances
31
Cahier de Recherche 21
assume that the relationship between those variables is linear. However, the genuine effect can be
nonlinear. In this study, a high exchange rate means depreciation of the Moroccan Dirham. And a
low exchange rate means appreciation of the Moroccan Dirham.
2.1.
DATA
Monthly data used in this paper span from January 2005 to June 2014. Monthly average exchange
Rate between Eurozone and Morocco is considered. Most Moroccan migrants are in the Europe.
More than 80% of remittances are from Europe (Makhlouf, 2013). Remittances are measured in
Euro. A shock of the exchange rate is defined as devaluation of the exchange rate by the Central
Bank of Morocco. It represents a positive variation of exchange rate. Note that the Bank of
Morocco can use the exchange rate as a tool to boost remittances.
Table 3 gives a summary statistics. For example, table 3 shows that exchange rate varies from 1€
=10.93 MAD to 1€ =11.49MAD.
32
Cahier de Recherche 21
TABLE 3 : DESCRIPTIVE STATISTICS
Exchange rate MAD/EUR (Q) Remittances in Millions (R)
Min.
10.93
240.747
1st Quintile 11.09
353.249
Median
11.18
389.796
Mean
11.18
397.181
3 rd Quintile 11.27
435.305
Max
11.49
586.297
Source
Banque de France (2014)
World Bank WDI (2014)
To explore whether the relationship between Q and R is linear or not figure 1 illustrates a scatter
plot of remittances and exchange rate. Figure 1 clearly shows that the relationship between
exchange rate and remittances is not monotone. The blue line represents the linear relationship
between exchange rate and remittances, and the black line represents the kernel regression. The
possibnility of a non-linear propagation of remittances according to changes in exchange rate is
investigated. The TVAR model captures
a non-linearity such as asymetric reactions of
remittances to shocks.
33
Cahier de Recherche 21
FIGURE 1: SCATTER PLOT OF REMITTANCES AND EXCHANGE RATE
2.2.
TVAR MODEL
The use of a nonlinear framework with regime switching determined by exchange rate was
motivated by the capacity of exchange rate to stimulate and to shorten remittances. The use of
monthly data is a relevant contribution in this context. We assume that « bad times » as periods of
an appreciation of exchange rate and a “good times” as periods of depreciation of exchange rate.
Bad times blunt purchasing power of remittances. Conversely good times increasing the
purchasing power of remittances. The value of remittances to Morocco might be influenced by the
appreciation of the Dirham against the Euro. Appreciation of a local currency erodes the
purchasing power of remittances to Morocco, and vice versa.
34
Cahier de Recherche 21
The threshold VAR can be specified as follows:
𝑌𝑡 = 𝐵1 (𝐿)𝑌𝑡−𝑙 + (𝐵 2 (𝐿)𝑌𝑡−𝑙 )𝐼[𝑄𝑡−𝑑 > 𝛾] + 𝑈𝑡 (1)
Where 𝑌𝑡 is a vector of endogenous variables (R and Q) and I is an indicator function that takes
the value of 1 if the value of exchange rate is higher than the threshold value 𝛾 and 0 otherwise.
𝐵1 (𝐿) and 𝐵2 (𝐿) are lag polynomial matrices. Q is the exchange rate, whereas d the delay
parameter is assumed to be less than or equal to lag l. Estimation of equation 1 can be done
directly by CLS (Conditional Least Squares). To estimate the TVAR
10
we use Conditional least
Square (CLS) technique which is implemented in “tsDyn” package in R software.
VAR LAG ORDER SELECTION
Information criteria such as AIC , HQ, SC and EPE are used to choose a lag length for the
unrestricted VAR-model. Max lag=10. Table 3 gives the optimal lag = 1.
TABLE 3: OPTIMAL LAG
lag lengh
1
2
3
4
5
6
7
8
9
10
AIC(n)
-14.07* -14.02 -14.00 -13.98 -13.91 -13.85 -13.87 -13.82 -13.77 -13.73
HQ(n)
-14.01* -13.92 -13.86 -13.79 -13.69 -13.58 -13.56 -13.47 -13.38 -13.30
SC(n)
-13.92* -13.77 -13.64 -13.52 -13.36 -13.19 -13.11 -12.96 -12.81 -12.67
FPE(n)
10
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
http://cran.r-project.org/web/packages/tsDyn/tsDyn.pdf
35
Cahier de Recherche 21
2.3.
EMPIRICAL RESULTS
Figure 3 shows the results of the impulse response functions from TVAR model. Results report
two regimes: the lower regime (exchange rate < 1€= 11.2048 MAD) and the upper regime for
(Q>11.1048). The impulse response functions show that the responses of remittances after a
shock in exchange rate are asymmetric. In the lower regime, the impact of devaluation of
exchange rate is positive. Conversely, in the regime 2, a positive shock in the exchange rate leads
to decrease of remittances. Hence, a movement in the change in the exchange rate causes two
effects namely: an income and substitution effect. If altruistic motivations dominate the self-interest
motivations, a policy of devaluation does not drain more remittances. When remittances for
investments are more important than those for consumption, in this situation, a policy of
devaluation can attract more remittances. The changes in remittances resulting from changes in
exchange rate depend on threshold value.
Devaluation of the exchange rate pushes migrants to send more money in the short run. This can
be explained by the substitution effect. The exchange rate is a way that could allow migrants to
compare their purchasing power between the host and the home country. The non-linearity of the
relationship proves the coexistence of two effects simultaneously (substituion and revenue).
However, other factors can play an important role in the variation of remittances, such as
economic conditions. The response of remittances does not happen immediately. This can be
explained by the coexistence of two types of behavior. The first is altruism in which remittances
are intended to meet the basic consumption needs of migrant families. This first type of
remittances caused by the needs of migrant families. In this case remittances should not be
influenced by other variables. The second case concerns remittances that are used for
investments. In this case, other factors may play an important role in the determination of
remittances.
36
Cahier de Recherche 21
FIGURE 2 : IMPULSE-RESPONSE FUNCTION
REGIME 1 (RESPONSE OF REMITTANCES) :
Q<11.2048
REGIME 2 (RESPONSE OF REMITTANCES) :
Q>11.2048
Threshold value: log(Q)=2.416342
Percentage of Observations in each regime: 57.9% 42.1%
37
Cahier de Recherche 21
CONCLUSION
A threshold value is estimated endogenously. The exchange rate of 11.2048 acts as a threshold
between a positive and a negative effect. Results can help policy makers. The Central Bank of
Morocco should take into consideration the exchange rate as a tool to increase remittances.
Indeed, reducing exchange rate volatility can stabilize the real value of remittances. The
government also can use the exchange rate policy in order to direct or influence remittances. It is
obvious that other economic factors may influence the volume of remittances. Remittance
decisions are complex. Indeed, remitting behavior varies depending on age, education, gender,
size of the household, etc. Remittances have arisen and given their large size, the government of
Morocco can use these to promote development. The government should establish policies
targeting maximization remittances. Finally, some variables such as interest rates, migrant stock,
etc. have not been considered due to their unavailability. However, despite these shortcomings we
believe that we have yielded interesting results that can be useful to policymakers.
38
Cahier de Recherche 21
REFERENCES
Acosta, P. A., Lartey, E. K.K., Mandelman, F. S., (2009) “Remittances and the Dutch disease,”
Journal ofInternational Economics, Elsevier, vol. 79(1), pages 102-116, September
Agunias, D. R., (2006) “Remittances and Development: Trends, Impacts, and Policy Options A
Review of the Literature.” Migration Policy Institute, Washington, D.C.
Amuedo-Dorantes, C., Pozo, S., (2004) “Workers' remittances and the real exchange rate: a
paradox of gifts”, World Development 32, pp. 1407–1417.
Aydas, S. T., Neyapti, B., Metin-Ozcan, V., (2004) “Determinants of Workers remittance: The Case
of Turkey”. Bilkent University Department of Economics Discussion Paper
Barajas, A., Chami, R., Hakura, D., Montiel. P., (2010) “Workers' Remittances and the Equilibrium
Real Exchange Rate: Theory and Evidence”. Department of Economics Working Papers 2010-14,
Department of Economics, Williams College.
Barro, D., Benninger, T., Bitang, A., Tiati, T., Gruhl. S., Koukoui. O., Mihai. A., (2007) “Migrations
internationales – une option de sortie par défaut ? ” Sous la direction de M. Jean-Jacques GABAS
et M. Jean COUSSY. Atelier international se Science-Po : Banque Mondiale. Rapport sur les
migrations.
Beja Jr., E. L., (2011) “Do international remittances cause Dutch disease?,” Migration Letters,
Migration Letters and The London Publishers, London, UK, vol. 8(2), pages 132-140, October.
Bettin G., Presbitero, F. A., Spatafora, N., (2014) "Remittances and vulnerability in developing
countries," Mo.Fi.R. Working Papers 93, Money and Finance Research group (Mo.Fi.R.) - Univ.
Politecnica Marche - Dept. Economic and Social Sciences.
Bouoiyour, J., (2006) “Migration, Diaspora et développement humain [Migration, Diaspora and
Human Development] ” MPRA Paper 37014, University Library of Munich, Germany.
39
Cahier de Recherche 21
Bouoiyour, J., (2013) "Les transferts des fonds des migrants marocains : Leviers de croissance et
du développement.” MPRA Paper 50537, University Library of Munich, Germany.
Bourdet, Y., Falck, H., (2006) “Emigrants’ Remittances and Dutch Disease in Cape Verde”.
International Economic Journal, 20: Volume
20, Issue 3, 2006 267-284.
Buch, C. M., Kuckulenz, A., Le Manchec, M-H., (2002) “Worker Remittances and Capital Flows.”
Kiel Working Papers 1130, Kiel Institute for the World Economy.
Cooper, Richard N. 1999. “Exchange Rate Choices.” Manuscript, Harvard University
Elbadawi, I. A., Rocha, R., (1992) “Determinants of Expatriate Workers' Remittances in North
Africa and Europe.” Working PaperWPS 1038, Country Economics Department, The World Bank,
Washington, DC.
El-Sakka, M. I. T., McNabb, R., (1999) “The
Macroeconomic Determinants of Emigrant
Remittances.” World Development, Elsevier, vol. 27(8), pages 1493-1502.
Faini, R., (1994) “Workers’
Remittances and the Real Exchange Rate: A Quantitative
Framework.” Journal of Population Economics, Springer, vol. 7(2), pages 235-45.
Faini, R., (2007)
“Migrations et transferts de fonds. Impact sur les pays d'origine. ” Revue
d’économie du développement, 21, issue 2, p. 153-182.
Fayad, G., (2011) “Remittances: Dutch disease or export-led growth?,” OxCarre Working Papers
057, Oxford Centre for the Analysis of Resource Rich Economies, University of Oxford.
Freund, C., Spatafora, N., (2008) “Remittances,
transaction costs, and informality,”
Journal of Development Economics, Elsevier, vol. 86(2), pages 356-366, June.
Ghosh, B., (2006) “Migrants’ Remittances and Development: Myths, Rhetoric and Realities.”
International Organization for Migration (IOM), and The Hague Process on Refugees and
Migration, Geneva.
Grabel, I., (2008) “The Political Economy of Remittances: What Do We Know? What Do We Need
to Know?” Political Economy Research Institute Working Paper series N° 184.
40
Cahier de Recherche 21
Gupta, P., (2005) “Macroeconomic Determinants of Remittances: Evidence from India.” IMF
Working Paper 05(224).
IMF International Monetary Fund (2005) “World Economic Outlook 2005”. Washington:
International Monetary Fund
Izquierdo, I., Montiel, P., (2006) “Remittances and Equilibrium Real Exchange Rates in Six Central
American Countries.” Department of Economics, Williams College.$
Joinern A., (2001) “Monetary Policy Effects in an Australian Bayesian VAR Model” Department of
Econometrics and Business Statistics Monash University
Köksal, N.E., (2006) “Determinants and Impact on the Turkish Economy of Reemittances.”
Université Paris I Panthéon – Sorbonne Ecole Doctorale d’Economie ‐ MATISSE (Unité mixte de
recherche 8595 CNRS).
Lartey, E.K.K., Mandelman, F., Acosta, P. A., (2012) “Remittances, exchange rate regimes, and
the Dutch disease: a panel data analysis.”
Review of International Economics, Wiley
Blackwell, vol. 20(2), pages 377-395, 05
Lianos, T. P., (1997) “Factors determining migrant remittances: The case of Greece.” International
Migration Review, 31(1), 72-87.
Lopez, H., Molina, L., Bussolo, M., (2007) “Remittances
and the real exchange rate,”
Policy Research Working Paper Series 4213, The World Bank.
Lucas, R. E. B., Stark, O., (1985) “Motivations to Remit: Evidence from Botswana.” Journal of
Political Economy, University of Chicago Press, vol. 93(5), pages 901-18, October.
Lueth, E., Ruiz-Arranz, M., (2007) “Are Workers' Remittances a Hedge Against Macroeconomic
Shocks? The Case of Sri Lanka. ” IMF Working Paper.
Makhlouf, F., (2013) “Transferts de fonds vers le Maroc, enjeux, comportement et impacts ».
Thesis, University of Pau (UPPA) France.
41
Cahier de Recherche 21
Makhlouf,F.,Chnaina, K., (2012) "Impact des transferts de fonds sur le taux de change réel effectif
en Tunisie," Working Papers 2011-2012_4, CATT - UPPA - Université de Pau et des Pays de
l'Adour, revised Jan 2012.
Makhlouf, F., Mughal, M., (2013) “Remittances,
Dutch Disease, and Competitiveness -
A Bayesian Analysis”. Journal of Economic Development, Chung-Ang Unviersity, Department
of Economics, vol. 38(2), pages 67-97, June.
Makhlouf, F., Naamane. A., (2013) "The
Impact of Remittances on Economic Growth.
The Evidence from Morocco," Working Papers 2013-2014_3, CATT - UPPA - Université
de Pau et des Pays de l'Adour, revised Sep 2013.
Miotti, L., Mouhoud, E.M., Oudinet, J., (2010) “ Determinants
and Uses of Remittances to
Southern and Eastern Mediterranean Countries: Insights from a New Survey,”
Development Working Papers 288, Centro Studi Luca d\'Agliano, University of Milano,
revised 30 Apr 2010.
Mongardini, J., Rayner,B., (2009) “Grants, Remittances, and the Equilibrium Real Exchange Rate
in Sub-Saharan African Countries.” IMF Working Paper No. 09/75.
Mughal, M., Makhlouf, F., (2011). “Volatility
of Remittances to Pakistan: What do the
Data Tell?” Economics Bulletin, vol. 31(1), pages 605-612.
Petri, M., Saadi-Sedik, T., (2006) “To Smooth or Not to Smooth - The Impact of Grants and
Remittances on the Equilibrium Real Exchange Rate in Jordan,” IMF Working Papers 06/257,
International Monetary Fund.
Rajan, R. G., Subramanian, A. n., (2005) “Aid
and Growth: What Does the Cross-Country
Evidence Really Show?” NBER Working Papers 11513, National Bureau of Economic
Research, Inc.
Ratha, D., (2007) “Leveraging Remittances for Development” Paper presented at the Second
Plenary Meeting of the Leading Group on Solidarity Levies to Fund Development, Oslo, February
6-7, 2007; and at the Brainstorming session on Migration and Development organized by the
42
Cahier de Recherche 21
Migration Policy Institute and the German Marshall Fund of the United States, Brussels, February
27, 2007. Euro Messo, 2006.
Roache, S. k., Gradzka, E., (2007) “Do Remittances to Latin America Depend on the U.S.
Business Cycle?” IMF Working Papers 07/273.
Russell, S. S., (1986) “Remittances from international migration: A review in perspective.”World
Development, 14(6), 677.
Sayan,S.(2006) "Business Cycles and Workers' Remittances: How Do Migrant Workers Respond
to Cyclical Movements of GDP at Home?, "IMF Working Papers 06/52, International Monetary
Fund.
Singer, D. A., (2008) “Migrant Remittances and Exchange Rate Regimes in the Developing
World.” Prepared for presentation at the Midwest Political Science Association annual meeting,
April 3-6, 2008.
Singh, R. J., Haacker, M., Lee, K.-W., (2009) “Determinants and Macroeconomic Impact of
Remittances to Sub-Saharan Africa.” IMF Working Paper 09/216. International Monetary Fund,
Washington, DC.
Sirkeci, I., Cohen, J. H., Ratha, D., (2012) “Migration and Remittances during the Global Financial
Crisis and Beyond.” ISBN (electronic): 978-0-8213-8827-3. World Bank.
Straubhaar,
T.,
(1986)
“The
Determinants
of
Remittances:
The
Case
of
Turkey.”
Weltwirtschaftliches Archiv, Vol. 122, No. 4, pp. 728–40.
Sy, M., Tabarraei, H., (2010) "Capital inflows and exchange rate in LDCs: The Dutch disease
problem revisited," PSE Working Papers halshs-00574955, HAL.
Vargas-Silva, C., (2008 ) “Are remittances manna from heaven? A look at the business cycle
properties of remittances.” The North American Journal of Economics and Finance, Elsevier, vol.
19(3), pages 290-303, December.
43
Cahier de Recherche 21
Vargas-Silva, C., (2009) “The Tale of Three Amigos, Remittances, Exchange Rate and Money
Demand in Mexico” Review of Development Economics, Wiley Blackwell, vol. 13(1), pages 1-14,
02.
Vargas-Silva,
C.,
Huang,
P.,
(2006)
“Macroeconomic
determinantsof
workers'
remittances: Hostversus home country's economic conditions.” Journal of
International Trade & Economic Development, Taylor and Francis Journals, vol. 15(1),
pages 81-99.
Yang, D., (2003) “Remittances and Human Capital Investment: Child Schooling and Child Labor in
the Origin Households of Overseas Filipino Workers.” Mimeo Havard University.
Yang, D., (2008) “International Migration, Remittances, and Household Investment: Evidence from
Philippine Migrants’ Exchange Rate Shocks,” The Economic Journal, Vol. 118, pp. 591-630.
44
Cahier de Recherche 21
Does Education Matter for the
Adoption of Information and
Communication Technologies (ICT)
in Developing Countries? Evidence
from Senegal
Mazhar MUGHAL
Professeur Groupe ESC Pau
IRMAPE
Barassou DIAWARA
45
Cahier de Recherche 21
ABSTRACT :
This paper empirically examines the role of workers’ human capital in determining the adoption of
ICT in developing countries based on firm level data. The empirical setting is Senegal in 2007. The
results of the probit model show that the managers’ education level does matter in the firms’
decision of whether or not to adopt ICT. Besides, the analysis highlights the significant role played
by the on-the-job training and the sales volumes in the probability of adopting new technologies.
Finally, the paper suggests that the workforce in the developing countries like Senegal needs to be
trained in ICT in order to raise the low levels of productivity, ultimately leading to economic growth
and lower poverty.
Key Words : Human capital, education, ICT, Senegal, workforce.
JEL Codes: I21, O10.
RESUME (IN FRENCH):
Cet article examine empiriquement le rôle du capital humain à disposition des firmes, dans les
pays en développement, sur l'adoption des technologies d'information et de communication (TIC).
L'analyse se base sur l'enquête des firmes au Sénégal en 2007.
Les résultats du modèle Probit montrent que le niveau d'éducation du manager joue un rôle
primordial dans la décision de l'entreprise d'adapter ou non les TIC. L'importance de la formation
au sain de l'entreprise et du volume des ventes est également soulignée. Enfin, l'article suggère
que la main-d'œuvre des pays en développement, tels que le Sénégal, doit être formée en TIC,
dans le but d'améliorer les bas niveaux de productivité, et d'aboutir finalement à la croissance
économique et à la baisse de pauvreté.
46
Cahier de Recherche 21
INTRODUCTION
The Efficient Consumer Response (ECR) concept was introduced in 1992 as a result of
competition from alternative store formats which highlighted major inefficiencies within the
supermarket industry and its supply chain. In order to survive, the US grocery industry leaders
formed a task force that took an initiative to study how to improve the performance of their supply
chain. The results of the study indicated that quick and accurate flow of information through the
supply chain enabled suppliers and distributors to anticipate demand requirements far more
11
accurately than current systems . The ECR initiative, therefore, transformed the supply chain from
a “push system” to a “pull system” where channel partners form new interdependent relationships
and where product replenishment is driven by point of sale (POS) data.
As the grocery industry changed, effects of these aforementioned events accrued and became
trends which eventually led to structural change. However, it still took time for a manufacturer to
move products from point A to point B. So while manufacturers were able to bypass many of the
qualitative functions that channel partners such as wholesalers performed, there still remained
physical and temporal activities that needed to be undertaken by someone. This fueled the growth
of the "partnership logistics" industry, which comprises third-party logistics providers,
12
transportation service companies, and public warehouses .
To follow up on these initiatives, manufacturers, retailers and wholesalers/distributors attempted to
establish a new spirit of cooperation and partnership. The challenge for the distributors and
retailers was reestablishing the value of the services they perform and leveraging the volume of
their independent customers, as the chain leverages the volume from its stores
13
Given these
conditions, retailers, distributors and manufacturers attempted to maximize the value they offer to
the customer – ECR enabled them to do this.
11
See Salmon Associates (1993).
12
See Sherman (1994).
13
See Sherman (1994), p. 20-24.
47
Cahier de Recherche 21
This paper examines the key initiatives surrounding ECR and discusses the primary issues
involved in its implementation. A brief literature review is presented in the second section while the
third section looks at the background of ECR and its components. The forth section presents a
framework and discussion and the conclusions are presented in the fifth section.
The rise of Information and Communications Technology (ICT) in the late twentieth century has
ushered in the age of information, the era of skill-biased technical change. This set of technologies
has revolutionized the business and trade environment by reducing the cost of communication,
making easy the acquisition of new production and managerial techniques, finding and
establishing commercial ties with distant customers and suppliers, making feasible for firms to
offshore and outsource chunks of production processes (Bayo-Moriones and Lera-López, 2007;
Hollenstein, 2004). Accessing and disseminating information has become possible with a few
mouse clicks, opening up a whole new world of advertisement, sales and business collaboration.
This has led to flexible production lines, savings of capital and labour hours, early adoption of
production and management techniques, lower production costs and improved product quality.
However, setting up and maintaining the ICT inside a firm requires skills which depend upon the
human capital available to the firm. Productive utilization of the ICT demands skills in the treatment
of information, data processing and man-machine-interaction, besides adequate computer skills
for the installation, operating and upkeep of the technology. This makes the presence of a trained
and skilled labour imperative (Arvanitis,
Schivardi and Trento,
2005; Bresnahan, Brynjolfsson and Hitt, 2002; Fabiani,
2005; Powell and Dent-Micallef, 1997).
Bayo-Morionesa and Lera-López (2007) analyse a survey of Spanish business establishments and
come up with a positive and significant association between the general level of employee
qualification and their use of ICT. Similarly, evidence from Italian manufacturing firms suggests
that an effective information technology use requires a set of complementary organizational
changes, implementation of which entails modification of the functional composition of the firm and
employment of skilled labour (Giunta and Trivieri, 2007). Caselli and Coleman (2001), in their
analysis of computer adoption of a worldwide panel find high levels of educational attainment to be
important determinants of computer-technology adoption. Other studies which highlight the
positive association between human capital and ICT adoption include
Gretton, Gali and Parham.
(2004), Black and Lynch (2000) and Doms Dunne and Troske (1997).
48
Cahier de Recherche 21
Although some literature does exist on the determinants of ICT adoption, few studies exist on the
role of level of human capital in the adoption of information technology by the corporate sector in
the developing countries. For instance, in a study of ICT adoption by Pakistani firms, Mughal and
Diawara (2011) find that on-the-job training, manager’s level of qualification and production
workers’ level of education are found to positively influence the use of emails, website and other
means of communication in Pakistani firms.
To the best of our knowledge, no study exists on this aspect of the African economies in spite of
the fact that ICT holds enormous potential in Africa (Wilson III and Wong, 2003). This paper
contributes to the literature by analyzing the relationship between the human capital employed in
the firms active in Senegal and the level and extent of ICT adoption prevalent in the country's
commercial environment. The West African country of Senegal is a useful study case because it is
fairly representative of other moderately open, low income economies of the region with few
natural resources and a predominantly agrarian economy. Its level of human capital accumulation
is above average for the region, and its geographical relative proximity to Europe and the United
States as well as its colonial heritage make it a representatively interesting case study with
probable policy implications applicable to other African countries with similar socioeconomic
characteristics.
Among the indicators of human capital, the paper not only considers formal education (such as
school, university or technical institute diploma) which provides general skills or theoretical knowhow that may contribute to labour productivity, but also on the job training which trains the
workforce in performing specific tasks through in-house or external skill-enhancement programs.
Given data limitations, the study mainly concerns with inter-firm diffusion, which can be defined as
the degree of technology penetration across the firms in a given time frame.
In this paper, we test the following hypothesis: Firms endowed with higher human capital make
better use of the ICT. For this, we make use of the World Bank’s Enterprise Survey undertaken in
Senegal in 2007, which especially focuses on different types of firms evolving in various sectors of
the economy. Given consistent data on individual firms’ human capital and usage of new
technologies, we employ a probit model to investigate the extent to which firms with higher human
capital adopt the ICT. The findings show that the human capital available to the firm, in particular
49
Cahier de Recherche 21
on-the-job training, has a positive and significant impact on the probability of using the ICT when
interacting with customers and suppliers.
The rest of the paper is organized as follows. Section 2 presents the evolution of ICT usage in
Senegal. Section 3 describes the empirical model, the choice of variables and the econometric
techniques used, followed by some key findings in section 4. Section 5 discusses the empirical
findings with possible conclusions and policy implications.
3. DIFFUSION OF NEW TECHNOLOGIES IN THE SENEGALESE ECONOMY
After its independence in 1960, Senegal inherited a relatively well-developed telecommunications
infrastructure which placed Senegal among Africa’s top ranking countries with regard to
information and communication technology.
Although Senegal enjoyed substantial economic advantages at the time of independence, other
African countries have also caught up in many areas since then. For example, relative to the
percentage of mobile cellular subscribers, Table 1 shows that in 2005, countries such as Gambia
(16.22%) and Mauritania (25.16%) had a larger percentage of cell phone holders than Senegal
(15.34%). However, since the early 2000s, the performance of Senegal in terms of use of ICT is
well above average for Sub-Saharan African and low income countries.
It is also note-worthy that the pattern of ICTs in Sub-Saharan Africa in general, and Senegal in
particular, is different from that of developed countries as shown in Table 1 (OECD countries). In
fact, for rich countries, most of the people already had access to internet (62%), personal
computers (63%) and telephones (85% for mobile phones) in 2005, reflecting a high and
widespread use of the ICTs. The significant difference with the developing countries such as
Senegal lies in their unexploited market potential.
Table 1. Selected ICT indicators in Senegal and the rest of the world
50
Cahier de Recherche 21
Yea
r
Seneg
al
C.
d'Ivoir
e
Gamb
ia
Maurita
nia
Niger
ia
SubSahar
a
MEN
A
Sout
h
Asia
Low
inco
me
OCDE
countri
es
Worl
d
Internet users (per 100 people)
199
0
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.30
0.05
199
5
0.00
0.00
0.01
..
..
..
0.01
0.02
..
4.12
0.80
200
0
0.40
0.23
0.92
0.19
0.06
0.51
0.86
0.48
0.16
32.88
6.74
200
5
4.79
1.04
3.80
0.67
3.54
2.17
10.4
8
3.65
2.19
61.54
16.2
2
Personal computers (per 100 people)
199
0
0.24
..
..
..
..
..
..
0.04
..
11.13
2.49
199
5
0.69
..
0.06
..
0.46
..
1.05
0.16
..
19.80
4.20
200
0
1.62
0.52
1.15
0.97
0.60
0.91
2.50
0.42
0.34
38.96
7.99
200
5
2.22
1.68
1.64
2.65
0.85
1.83
3.18
1.53
1.60
63.13
12.7
5
Telephone lines (per 100 people)
199
0
0.59
0.58
0.69
0.30
0.31
0.99
3.36
0.57
0.64
45.20
9.86
199
5
0.95
0.77
1.77
0.42
0.37
1.10
5.37
1.19
0.76
51.77
12.1
3
51
Cahier de Recherche 21
200
0
2.08
1.53
2.56
0.74
0.44
1.37
8.55
2.73
1.10
58.17
16.0
6
200
5
2.36
1.34
2.88
1.38
0.87
1.46
14.0
9
3.96
2.85
53.47
19.5
1
Mobile cellular subscriptions (per 100 people)
199
0
0.00
0.00
0.00
0.00
0.00
0.00
0.00 0.00
0.00
1.21 0.21
199
5
0.00
0.00
0.13
0.00
0.01
0.10
0.09 0.01
0.01
8.43 1.60
200
0
2.53
2.74
0.43
0.60
0.02
1.71
2.25 0.34
0.37
51.65
12.1
6
15.34 12.21 16.22
25.16
13.1
5
12.01
22.1
7.92
0
5.80
85.18
34.3
7
200
5
Notes: MENA stands for Middle East and North Africa
OECD stands for Organization for Economic Cooperation and Development
… means that data are not available
Source: World Bank (2009)
Figure 1 compares the telecommunications revenues in selected West African countries (including
Senegal). Although not the top-most country in terms of revenues generated by the
telecommunications sector, Senegal’s revenues from the provision of telecommunications services
are not only among the highest in the region, but are also constantly increasing since the 1990s.
Figure 1 shows the high returns associated with investments in telecommunications in Senegal.
52
Cahier de Recherche 21
Figure 1. Comparison of the telecommunications revenue (% GDP) among selected African
countries
9.00
Telecommunication revenues (%GDP)
8.00
7.00
Senegal
6.00
Burkina Faso
Cote d'Ivoire
5.00
Gambia, The
Mauritania
4.00
Nigeria
Sub-Saharan Africa
3.00
2.00
1.00
0.00
1990
1991 1992 1993
1994 1995 1996
1997 1998
1999 2000 2001
2002 2003 2004
2005
Years
Source: World Bank (2009)
A quick look at the use of ICT by firms in Senegal (Table 2) shows that the use of mobile phones is
common in most of the enterprises surveyed, whereas less than 40% of the firms surveyed use
emails (31.84%) and mainline phones (36.13%) in their interaction with suppliers and customers,
while only 10.24% use websites.
As regards the sectoral decomposition, while 100% firms in the machinery and equipment sector
use emails and websites in their interaction with customers and suppliers, not all enterprises in the
information technologies sector use emails (84%) and websites (87.5%). It is worth noting that
100% of the firms surveyed work with cellular phones and none of them uses e-mails and websites
(Table 2).
53
Cahier de Recherche 21
Table 2. Industries by type of ICT use
E-mail
Website
Cell phones
Mainline
phone
Food
38.55
9.64
..
..
Textiles
0.00
0.00
100.00
0.00
Garments
16.33
6.12
85.71
28.57
Chemicals
81.82
63.64
..
..
Plastics and rubber
50.00
0.00
100.00
100.00
Non metallic mineral
57.14
14.29
Fabricated metal prod
23.81
9.52
100.00
0.00
Machinery and equipment
100.00
100.00
..
..
Electronics
66.67
33.33
..
..
Construction
100.00
33.33
..
..
Wholesale
21.05
0.00
100.00
100.00
Retail
17.93
4.89
95.00
32.50
Hotels and restaurants
25.49
7.84
100.00
33.33
Transport
50.00
50.00
100.00
0.00
Information Technologies
84.00
12.00
87.50
75.00
Other manufacturing
31.91
9.57
100.00
37.50
..
54
Cahier de Recherche 21
Other Services
46.77
20.97
100.00
25.00
Total
31.84
10.24
94.96
36.13
Notes: The sample consists of 625 enterprises except for the cell phones and mainland phone
use which are based on 119 observations
… means that data are not available
Source: World Bank (2007)
In sum, there are substantial sectoral differences in the use of the ICT by Senegalese firms.
Consequently, it is worth examining the factors affecting firms’ decision on whether and how to
adopt new technologies. This paper focuses on one of them: the role of human capital.
4. EMPIRICAL STRATEGY
4.1.
ECONOMETRIC SPECIFICATION
We measure the impact of human capital (our main variable of interest) and firms’ characteristics
on the adoption of new technologies in Senegalese enterprises using the following general
relationship:
k
Yi   0  1 HUMAN    j X ij   i ,
(1)
j 2
where Yi is the dependent variable (a dummy variable) showing whether firm i has adopted ICT or
not, β0 is a constant, β1 and βj are the coefficients of explanatory variables HUMAN (human capital
indicator) and Xij, respectively, and εi is the error term.
In equation (1), Yi represents a set of dummy variables related to the use of technology which take
the value of 1 if firm adopts the new technologies in its interactions with clients and suppliers and 0
otherwise.
55
Cahier de Recherche 21
In equation (1), HUMAN refers to the set of education indicators considered in this study. One of
the education variables is on-the-job training, a dummy variable taking the value of 1 if the firm
offers formal training to its permanent employees and 0 otherwise. Another human capital variable
considered is the managers’ highest education level expressed in terms of grade completion
(primary education, secondary education, vocational training, graduate and post-graduate
degrees). We expect a significant positive role of manager’s qualification in the introduction of
technology. Technology adoption not only involves awareness of profitable investment in emerging
technologies, but also the understanding of the implementation process, the costs and challenges
involved as well as the requisite technical skills, and a qualified manager is more likely to possess
such knowledge base. Another human capital-related variable used is the average number of
years of schooling of the workforce taking part in the production process. The proportion of skilled
workers employed in the production process during the year 2006 is also used as an indicator of
workers’ skills and education. Adoption of new technology often takes place in firms with better
educated and skilled labour (Doms et al., 1997; Romijn, 1997). The last indicator of education
considered is the percentage of permanent production workers having received formal training.
Besides, in equation (1), Xij is a vector of variables controlling for the other characteristics of the
firm (see Table A1 in the appendix related to the summary statistics for the other variables in the
empirical analysis).
4.2.
DATA
Data have been taken from the World Bank (2007) Enterprise Surveys (WBES). The WBES
dataset is a stratified random sample of firms with a common questionnaire and sampling
methodology for all participating countries. The Surveys use standardized survey instruments and
a uniform sampling methodology to minimize measurement errors and yield data comparable
across the world. The World Bank Enterprise Survey in Senegal targeted establishments located
in Dakar, Kaolack, Saint-Louis and Thies in the following industries [according to International
Standard Industrial Classification (ISIC), revision 3.1]: all manufacturing sectors (group D),
construction (group F), retail and wholesale services (sub-groups 52 and 51 of group G), hotels
and restaurants (group H), transport, storage, and communications (group I), and computer and
56
Cahier de Recherche 21
related activities (sub-group 72 of group K). The completion rate (representativity of the survey) is
36% as 625 out of 1733 firms were included in the sample.
The data in this study concern at best 625 enterprises in various sectors, namely food, textiles,
garments, chemicals, plastics and rubber, non metallic mineral products, fabricated metal
products, machinery and equipment, electronics, construction, wholesale, retail, hotels and
restaurants,
transport
and
information
technology.
The
dataset
is
accessible
at
www.enterprisesurveys.org; detailed information on the sampling methodology is also available on
this website. Besides, the WBES database provides appropriate variables allowing estimating the
relationship existing between firm’s human capital (education or skills of the workforce) and the
adoption of ICT. Mean, standard deviation and the definition of the different variables used in the
regressions are summarized in Table A1 in the appendix.
The table indicates that the adoption of ICT by Senegalese firms is quite homogenous in various
sectors, with small standard deviations for the ITC-related variables. However, there are large
variations between the sectors for some human capital indicators (for example, the number of
skilled workers in 2006), as well as for some control variables such as sales in 2006 and firm
share in the local or national markets (Table A1 in the appendix).
4.3.
ESTIMATION METHOD
We adopt a probit estimation approach to assess the impact of human capital on the adoption of
ICT in developing countries. The choice of method is dictated by the nature of our dependent
variables which are all binary taking values of either one or zero (see Table A1 in the appendix).
Besides, probit is an appropriate method for studying the ICT-adopting behavior of firms because
enterprises either do or do not adopt new technologies, and a probit model is a statistical
procedure suitable for estimating the relationship between the dichotomous dependent variable
and a set of continuous explanatory variables.
Probit models transform a dichotomous dependent variable into a probability. The dependent
variable is hence categorical. Specifically, Yi in equation (1) is a discrete random variable that
assumes one of two possible values: 1 if firm adopts ICT during the surveyed year and 0 if it does
not. The independent variables may be either continuous or discrete, but they are assumed to be
57
Cahier de Recherche 21
non-stochastic. Therefore the probit estimation method basically tests the impact of human capital
(and other independent variables) on the probability of firms’ adoption of the new technologies, as
well as the probability of improvement in product quality and productivity as a result of ICT
adoption.
In addition, the education of the workforce is clearly endogenous with respect to the firm
introducing and implementing the ICT. However, the limitation of the data, both in its temporal
dimension and the number of variables included, does not allow us to handle the variable’s
potential endogeneity. However, without attempting to undermine the problem, it can be pointed
out that in the literature on the role of human capital on ICT adoption, endogeneity has not been a
major concern. For instance, studies such as Bartoloni and Baussola (2001) and Bayo-Morionesa
and Lera-López (2007) have not mentioned the possible endogeneity associated with the variable
“education of the workforce”. Some other studies (for example, Hollenstein, 2004 and Battisti et al.,
2007) attempted to tackle the problem related to the endogeneity and found that the robustness of
the results are not affected.
In the specific case of this study, except for the variable “on-the-job training”, human capital related
variables can be considered as exogenous. Manager’s qualification and average education level of
the production workers are most probably exogenous as managers and workers in Senegal
generally join the enterprises with their education levels already determined.
5. ESTIMATION RESULTS
5.1.
IMPACT OF EDUCATION
Manager’s qualification
The results related to the impact of the manager’s education level on the adoption of ICT are
presented in Table 3. Manager's qualification is positively associated with the probability of the firm
using emails and website and the adoption of internet (Table 6). It is worth-noting that the impact is
consistently significant for University education. The magnitude of the impact is highest for the use
of internet for communication where the marginal effect is 0.65. Higher degrees such as graduate
58
Cahier de Recherche 21
and postgraduate degrees are also important but only significant in the case of email use. Besides,
firms where the manager has a graduate degree (BSc, BA, etc) appear to be more likely to use
email and have a website. The results corroborate with the findings of past studies (e.g Correa et
al., 2010) which find the managerial education to be strongly and positively associated with web
use.
Results from Table 3 highlight the need of some university education for the manager for an
efficient adoption of ICT in Senegal. Compared to firms with managers without education, firms
where managers have some university education are more likely to adopt the ICT (the coefficient
varies between 17% and 65% depending on the type of ICT).
Production workers’ education
Results related to the impact of workers’ average education are presented in Table 4. As
compared with the “no education” status, firms where workers have 10-12 years of education at an
average are more likely to use emails and own a website. However, given the lack of statistical
significance of different coefficients, the education level of the workforce does not show a
significant impact on the adoption of ICTs in Senegal. A possible explanation may be that the
average education level of the workers does not play a major role
in the firm’s decision on
whether or not to adopt ICTs.
Share of skilled workers
The relationship between the ratio of skilled to unskilled workers and the probability of adopting the
new technologies is presented in Table 5. The findings show that the higher the proportion of
skilled workers relative to unskilled workers, the higher is the probability to use emails when
interacting with the customers and suppliers. In effect, the probability to use email is around 10%
and is statistically significant at 1%. The results imply that firms using email probably employ a
greater number of skilled workers compared to unskilled workers. However, this result can also be
due to the fact that firms intending to use the new technologies hire more skilled workers anyway.
The lack of a longitudinal survey data on the issue does not allow us to observe the recruitment
59
Cahier de Recherche 21
policy of the firms, thereby precluding any inference regarding the change in hirings prior to or as a
result of ICT adoption.
The above findings are in line with the works of Arvanitis (2005), Bayo-Moriones and Lera-López
(2007), Bresnahan et al. (2002) and Fabiani et al. (2005) for various developed countries, that
conclude that the presence of skilled workers fosters innovation and facilitates ICT adoption and
use at the firm level.
On-the-job training
Table 6 summarises the results related to the impact of staff training schemes. Findings show that
firms conducting on-the-job training are more likely to adopt the ICTs. In fact, the relationship
between the various indicators of ICTs and “on-the-job training” is positive and significant except
for the case of the use of internet for research. The effect varies with respect to the independent
variable chosen, and goes from 9% (for the marginal effect on website usage) to 48% (for the
marginal effect on the use of internet for delivery). Formal skill enhancement programs therefore
hold great importance in a firm’s decision to adopt new technologies. It may also be due to the fact
that firm’s planning to introduce the ICTs are more likely to train their staff to ensure an efficient
usage. This also points to the possibility that Senegalese workers do not possess prerequisite IT
skills and begin to use the technology only after receiving necessary training from the firm
60
Cahier de Recherche 21
Table 3. Effect of manager’s education level
(1)
(1’)
Email
Primary school
Started but did not complete secondary
Secondary School
(2)
(2’)
(3)
(3’)
(4)
(4’)
Website
High speed internet
Internet
communication
(5)
(5’)
for Internet
research
for
Param
eters
Marginal
effects
Parame Marginal
ters
effects
Param Marginal
eters
effects
Parame Marginal
ters
effects
Parame Marginal
ters
effects
0.40*
0.15*
0.09
0.01
0.35
0.12
0.35
0.10
-0.37
-0.08
(0.22)
(0.09)
(0.35)
(0.05)
(0.58)
(0.20)
(0.57)
(0.18)
(0.61)
(0.11)
0.24
0.09
-0.11
-0.01
0.13
0.04
0.28
0.08
-0.06
-0.01
(0.24)
(0.09)
(0.39)
(0.05)
(0.69)
(0.23)
(0.65)
(0.21)
(0.64)
(0.14)
0.42*
0.16*
-0.03
-0.00
0.69
0.25
0.39
0.12
0.01
0.00
(0.24)
(0.10)
(0.37)
(0.05)
(0.65)
(0.25)
(0.63)
(0.21)
(0.64)
(0.15)
61
Cahier de Recherche 21
Vocational Training
Some university training
Graduate degree (BA, BSc etc.)
MBA from university in another country
0.44
0.17
0.43
0.08
0.16
0.05
-0.39
-0.09
0.62
0.18
(0.29)
(0.12)
(0.40)
(0.09)
(1.09)
(0.38)
(1.09)
(0.21)
(0.81)
(0.28)
0.91***
0.35***
0.80**
0.17**
1.66** 0.59***
*
1.87***
0.65***
1.17*
0.38*
(0.29)
(0.10)
(0.37)
(0.11)
(0.65)
(0.18)
(0.66)
(0.18)
(0.60)
(0.23)
0.91***
0.35***
0.74*
0.16*
0.77
0.28
1.12
0.40
0.07
0.02
(0.30)
(0.11)
(0.39)
(0.11)
(0.87)
(0.34)
(0.79)
(0.30)
(0.70)
(0.17)
0.47
0.18
-0.28
-0.03
-
-
-
-
-
-
(0.64)
(0.25)
(0.77)
(0.07)
0.41***
0.22
0.03
0.92
0.34
0.64
0.21
0.39
0.10
(0.11)
(0.38)
(0.07)
(0.72)
(0.28)
(0.65)
(0.24)
(0.65)
(0.20)
Other postgraduate degree from university 1.07***
in this country
(0.33)
62
Cahier de Recherche 21
Other postgraduate degree from university 1.22***
in another country
Log sale in 2006
Age
Medium size (20-99 employees)
Large size (100 employees and more)
0.45***
0.42
0.07
0.93
0.35
-0.06
-0.02
0.58
0.17
(0.41)
(0.12)
(0.43)
(0.09)
(0.89)
(0.34)
(0.93)
(0.24)
(0.85)
(0.30)
0.15**
0.06**
0.23***
0.03***
0.57** 0.18***
*
0.67***
0.18***
0.25
0.06
(0.06)
(0.02)
(0.07)
(0.01)
(0.19)
(0.06)
(0.19)
(0.05)
(0.16)
(0.04)
-0.01
-0.00
0.01
0.00
-0.01
-0.00
-0.01
-0.00
-0.01
-0.00
(0.01)
(0.00)
(0.01)
(0.00)
(0.02)
(0.01)
(0.02)
(0.00)
(0.02)
(0.00)
0.30
0.12
0.32
0.05
2.17** 0.70***
*
1.52
0.54
1.20*
0.40*
(0.22)
(0.09)
(0.26)
(0.05)
(1.14)
(0.19)
(1.00)
(0.34)
(0.69)
(0.27)
0.25
0.10
0.38
0.07
-
-
-
-
-
-
(0.45)
(0.18)
(0.43)
(0.09)
63
Cahier de Recherche 21
Kaolack
Saint-Louis
Thies
Food
Garments
1.00***
-0.29***
-0.42
-0.04
-0.23**
1.02**
-0.49
-0.11
-0.82*
-0.13*
(0.28)
(0.05)
(0.47)
(0.04)
(0.73)
(0.57)
(0.10)
(0.75)
(0.07)
0.87***
-0.26***
-0.78*** -0.07***
-0.29***
1.48**
*
-1.02**
-0.19**
-0.39
-0.08
(0.24)
(0.05)
(0.49)
(0.02)
(0.82)
(0.66)
(0.07)
(0.63)
(0.10)
-0.12
-0.04
-0.05
-0.01
-0.25***
1.16**
*
-1.22**
-0.21***
-0.32
-0.06
(0.20)
(0.07)
(0.30)
(0.04)
(0.69)
(0.09)
(0.68)
(0.07)
(0.59)
(0.10)
1.05***
-0.31***
-1.19*** -0.10***
-
-
-
-
-
-
(0.28)
(0.06)
(0.33)
(0.02)
1.12***
-0.30***
-0.52*
-0.05*
-
-
-
-
-
-
(0.10)
(0.08)
64
Cahier de Recherche 21
Fabricated metal products
Wholesale
Retail
Hotels and restaurants
Other Services
(0.34)
(0.06)
(0.43)
(0.03)
-0.87**
-0.25***
-0.26
-0.03
(0.43)
(0.09)
(0.51)
(0.05)
-0.75*
-0.23**
-
(0.41)
(0.09)
0.88***
-0.28***
(0.27)
-
-
-
-
-
-
-
-
-
-
-
-
-
-0.45*
-0.05*
-0.80***
2.68**
*
-1.49*** -0.52***
-1.84*** -0.61***
(0.07)
(0.29)
(0.03)
(0.71)
(0.09)
(0.48)
(0.16)
(0.46)
(0.15)
1.03***
-0.29***
-0.58**
-0.06**
-
-
-
-
-
-
(0.31)
(0.06)
(0.39)
(0.03)
-0.67**
-0.21***
-0.14
-0.02
-
-
-
-
-
-
65
Cahier de Recherche 21
Other manufacturing
Constant
(0.29)
(0.08)
(0.30)
(0.03)
0.78***
-0.25***
-0.42*
-0.05*
(0.28)
(0.07)
(0.31)
(0.03)
-2.42**
-
-5.13*** -
(1.33)
(1.08)
Observations
R
2
Notes:
Standard
errors
in
-
-
-
-
-
-
8.19**
11.10**
*
-
-3.70
-
(3.40)
(3.36)
(2.74)
504
504
487
487
120
120
120
120
120
120
0.247
0.247
0.272
0.272
0.521
0.521
0.422
0.422
0.359
0.359
parentheses;
***,
**
and
*
means
significant
at
1%,
5%
and
10%
respectively.
66
Cahier de Recherche 21
Table 4. Effect of average education level of the production workers
(1)
(1’)
Email
0-3 years of education
4-6 years of education
7-9 years of education
10-12 years of education
Log sale in 2006
Age
Medium size (20-99 employees)
(2)
(2’)
Website
Parameter
s
Marginal
effects
Parameter Marginal
s
effects
-
-
0.00
0.00
(0.71)
(0.12)
-0.27
-0.09
0.15
0.03
(0.24)
(0.08)
(0.67)
(0.12)
-0.07
-0.02
0.40
0.08
(0.28)
(0.09)
(0.66)
(0.15)
0.05
0.02
-0.02
-0.00
(0.42)
(0.15)
(0.81)
(0.14)
0.30***
0.10***
0.16
0.03
(0.10)
(0.03)
(0.11)
(0.02)
-0.02*
-0.01*
-0.00
-0.00
(0.01)
(0.00)
(0.02)
(0.00)
0.04
0.01
0.66*
0.15*
(0.31)
(0.11)
(0.38)
(0.10)
67
Cahier de Recherche 21
Large size (100 or more employees)
Kaolack
Saint-Louis
Thies
Food
Garments
Fabricated metal products
Other manufacturing
Constant
-0.02
-0.01
0.63
0.15
(0.54)
(0.19)
(0.58)
(0.17)
-1.60***
-0.34***
-
-
(0.49)
(0.05)
-0.81***
-0.23***
-
-
(0.36)
(0.07)
0.11
0.04
-0.59
-0.08
(0.29)
(0.11)
(0.55)
(0.05)
-0.55*
-0.18*
-0.82**
-0.12**
(0.35)
(0.11)
(0.37)
(0.05)
-0.77**
-0.22**
-0.53
-0.07
(0.41)
(0.09)
(0.51)
(0.05)
-0.56
-0.17
0.11
0.02
(0.49)
(0.12)
(0.61)
(0.12)
-0.36
-0.12
-0.45
-0.07
(0.35)
(0.11)
(0.37)
(0.06)
-4.87***
-
-3.91*
-
(1.73)
(2.12)
68
Cahier de Recherche 21
Observations
R
2
254
254
201
201
0.246
0.246
0.254
0.254
Notes: Standard errors in parentheses; ***, ** and * means significant at 1%, 5% and 10%
respectively
Table 5. Effects of skilled workforce
(1)
(1’)
Email
Parameters
Age
Medium size (20-99 employees)
Large
size
employees)
(100
or
(2’)
Website
Marginal
effects
Parameters
Marginal
effects
0.10***
0.09
0.02
(0.08)
(0.03)
(0.08)
(0.02)
0.42***
0.16***
0.19
0.04
(0.13)
(0.05)
(0.13)
(0.03)
-0.01
-0.01
-0.00
-0.00
(0.01)
(0.01)
(0.02)
(0.00)
-0.29
-0.11
0.38
0.10
(0.37)
(0.14)
(0.44)
(0.12)
-0.10
0.42
0.11
Percentage of skilled to unskilled 0.25***
workers
Log sale in 2006
(2)
more -0.26
69
Cahier de Recherche 21
Saint-Louis
Thies
Food
Garments
Fabricated metal products
Other manufacturing
Constant
(0.66)
(0.24)
(0.69)
(0.21)
-1.28***
-0.40***
-
-
(0.51)
(0.10)
0.68*
0.27*
-0.61
-0.12
(0.37)
(0.14)
(0.61)
(0.09)
-0.52
-0.20
-0.65*
-0.13*
(0.43)
(0.16)
(0.40)
(0.07)
-0.89**
-0.31**
-0.41
-0.08
(0.51)
(0.15)
(0.56)
(0.09)
-1.22***
-0.37***
0.23
0.06
(0.67)
(0.13)
(0.68)
(0.19)
-0.49
-0.19
-0.47
-0.10
(0.45)
(0.17)
(0.43)
(0.09)
-7.44***
-
-4.36*
-
(2.41)
Observations
R
2
(2.34)
148
148
130
130
0.300
0.300
0.252
0.252
Notes: Standard errors in parentheses; ***, ** and * means significant at 1%, 5% and 10%
respectively.
70
Cahier de Recherche 21
Table 6. Effect of on-the-job training
(1)
(1’)
Email
Param
eters
Formal
training
(2)
(2’)
Website
(3)
High
internet
(3’)
(4)
(4’)
(5)
(5’)
speed Internet
for Internet
communication
delivery
(6)
for Internet
research
(6’)
(7)
(7’)
for Internet
purchasing
for
Margin
al
effects
Param
eters
Margina Param
l effects eters
Margin
al
effects
Parame Margin
ters
al
effects
Param
eters
Margin
al
effects
Param
eters
Margin Param
al
eters
effects
Margin
al
effects
on-the-job 0.32*
0.12*
0.49**
0.09**
1.03**
0.39**
1.31***
0.47***
1.62***
0.48***
0.62
0.19
1.55***
0.43***
(0.18)
(0.07)
(0.20)
(0.04)
(0.50)
(0.18)
(0.48)
(0.17)
(0.47)
(0.17)
(0.44)
(0.15)
(0.46)
(0.16)
Log sale in 2006
Age
0.20*** 0.08*** 0.23*** 0.03***
0.54*** 0.19***
0.51***
0.15***
-0.05
-0.01
0.25*
0.06*
0.41***
0.06***
(0.06)
(0.02)
(0.07)
(0.01)
(0.15)
(0.05)
(0.14)
(0.04)
(0.16)
(0.03)
(0.13)
(0.03)
(0.16)
(0.02)
-0.01
-0.00
0.01
0.00
-0.01
-0.00
-0.01
-0.00
-0.00
-0.00
-0.01
-0.00
-0.00
-0.00
(0.01)
(0.00)
(0.01)
(0.00)
(0.02)
(0.01)
(0.02)
(0.01)
(0.02)
(0.00)
(0.02)
(0.00)
(0.02)
(0.00)
71
Cahier de Recherche 21
Medium size
employees)
(20-99 0.28
(0.22)
Large size (100 or more 0.39
employees)
(0.44)
Kaolack
Saint-Louis
Thies
0.11
0.30
0.05
2.08*** 0.67***
1.27
0.46
1.84***
0.58***
1.38**
0.48**
1.48**
0.42**
(0.08)
(0.25)
(0.05)
(0.92)
(0.15)
(0.79)
(0.28)
(0.68)
(0.23)
(0.61)
(0.23)
(0.60)
(0.23)
0.15
0.28
0.05
-
-
-
-
-
-
-
-
-
-
(0.18)
(0.42)
(0.08)
-0.67**
1.01*** 0.29***
-0.06**
-0.28*** -0.72*
1.24***
-0.16*
-0.93*
-0.09*
-1.01**
-0.16** -0.22
-0.03
(0.27)
(0.03)
(0.74)
(0.10)
(0.97)
(0.05)
(0.76)
(0.06)
(0.74)
(0.08)
(0.05)
(0.48)
(0.09)
(0.59)
-0.07***
0.90*** 0.27*** 0.82***
-0.31*** -0.95**
1.39***
-0.20**
-0.20
-0.03
-0.51
-0.10
0.06
0.01
(0.24)
(0.05)
(0.47)
(0.02)
(0.70)
(0.08)
(0.65)
(0.08)
(0.56)
(0.09)
(0.58)
(0.09)
-0.18
-0.06
-0.17
-0.02
-0.29*** -1.37*** -0.25*** -0.73
1.25***
-0.08
-0.38
-0.08
0.02
0.00
(0.09)
(0.58)
72
Cahier de Recherche 21
(0.20)
Food
(0.30)
(0.04)
(0.62)
(0.09)
(0.63)
(0.06)
(0.75)
(0.05)
(0.52)
(0.09)
(0.56)
(0.08)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-0.05*
-
-0.79*** -1.61*** -0.57*** -
0.59**
-0.87**
-0.19**
-0.09***
1.13*** 0.33*** 0.99***
(0.27)
Garments
(0.06)
(0.31)
(0.02)
-0.07***
1.39*** 0.35*** 0.74***
(0.33)
Fabricated
products
(0.07)
(0.04)
(0.43)
metal -0.25
1.05*** 0.29***
(0.42)
Wholesale
Retail
(0.07)
(0.49)
0.98*** 0.27***
(0.41)
(0.07)
-
-
-0.42*
(0.02)
-0.03
(0.05)
-0.47*** -
73
Cahier de Recherche 21
0.99*** 0.31***
(0.25)
Hotels and restaurants
Other Services
Other manufacturing
Constant
(0.06)
2.75***
1.74***
*
(0.03)
(0.63)
(0.07)
(0.43)
(0.14)
(0.42)
(0.14)
(0.40)
(0.13)
(0.42)
(0.12)
-0.44
1.00*** 0.29***
-0.05
-
-
-
-
-
-
-
-
-
-
(0.30)
(0.03)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-3.66
-
-7.92*** -
(0.06)
(0.28)
1.63***
(0.37)
-0.08
0.66*** 0.21***
-0.01
(0.28)
(0.04)
(0.08)
(0.29)
-0.54**
1.04*** 0.31***
-0.06**
(0.26)
(0.03)
(0.06)
(0.29)
2.85***
5.02***
7.23***
-7.88*** -
0.95
(1.01)
(1.25)
(2.59)
(2.49)
(2.82)
(2.35)
(2.84)
74
Cahier de Recherche 21
Observations
R
2
506
506
489
489
121
121
121
121
121
121
121
121
121
121
0.220
0.220
0.255
0.255
0.494
0.494
0.407
0.407
0.400
0.400
0.325
0.325
0.376
0.376
Notes: Standard errors in parentheses; ***, ** and * means significant at 1%, 5% and 10% respectively
75
Cahier de Recherche 21
5.2.
IMPACT OF OTHER FACTORS
Various other variables affect the decision to adopt ICTs in Senegal. Among the most influential
factors is the firm sales. In fact, the results in Tables 3, 4, 5 and 6 show that the higher the volume of
the sales the higher is the probability to adopt the new technologies. This effect is significant for
various independent variables used in Table 3 (4 out of 5 regressions), Table 4 (1 out of 2
regressions), Table 5 (1 out of 2 regressions) and Table 6 (4 out of 5 regressions). The positive link of
sales with ICT adoption highlights the financial requirements for ICT adoption. Higher sales imply
increased revenues, which leads to the access to modern technology, and ultimately, to better use of
ICT.
The size of the company does not appear to exert much influence in ICT adoption, being statistically
non-significant in several regressions, similar to the firm’s age, which is mostly insignificant in the bulk
of the regressions.
We also study the impact of the companies’ regional distribution by introducing dummies for the
regions of Kaolack, Saint Louis and Thies (with Dakar as the default region). All the regional dummies
show a negative sign, meaning that firms installed in other regions (as compared to the ones
established in Dakar) are less likely to use the new technologies. These results can be explained by
the fact that most of the major high tech firms are situated in Dakar; for example 64.32% of the firms
surveyed are in Dakar (76% of those working in the IT sector are located in Dakar).
The impact of the type of industry is also taken into consideration in the analysis, with Information
Technology (IT) sector, which understandably appears to be the strongest adopter of ICT, taken the
reference group. Following sectors are considered: food, garments, fabricated metal products, other
manufacturing, wholesale, retail, hotels and restaurants, and other services. The findings show that
most industrial sectors are negatively associated with the adoption of ICT, showing the relatively low
importance that many Senegalese firms still attach to the ICT.
76
Cahier de Recherche 21
CONCLUDING REMARKS
This study is a first attempt to examine the role of human capital in the probability of the firms in
African countries adopting the new technologies in their interaction with customers and suppliers. The
empirical setting is 625 firms in Senegal surveyed in 2007. Our results are encouraging and show the
important role played by the human capital, sales and national market shares of the enterprises.
The findings show that besides on-the-job training, managers’ education level is positively and
significantly associated with the probability to use emails and websites when interacting with suppliers
and customers. These results imply that firms in which manager’s qualification is high or which run onthe-job training programs are more likely to use internet and/or emails when interacting with suppliers
and customers. Findings also show that average education level of the workers employed in the
production process is not significantly associated with the probability to adopt new technologies. In
sum, the study finds a strong role of human capital, which is pertinent in the context of developing
countries like Senegal which suffer from insufficient education attainment, poor skills and
consequently, low productivity. Managers in the developing country firms therefore need to possess at
least some university education in order to be able to efficiently introduce new technologies in the
Senegalese firms. The volume of sales also has a significant effect on the decision of firms to adopt
the ICTs in Senegal, suggesting the role of financial capabilities in the adoption and use of new
technologies.
This study used the 2007 cross-sectional survey data. Availibility of longitudinal data can help shed
more light on such open questions as whether Senegalese firms modify their hiring policy in
anticipation of technology adoption, is the lack of suitable workforce a hindrance to this end, and how
has increased use of ICT influenced the firms’ performance. Besides, there is a need to focus on the
importance of human capital in ICT adoption in specific industries, given that the relevance of
education may differ among different sectors. Conducting similar studies for a set of African or
developing countries can help us reach general conclusions on the role of human capital on ICT
adoption in African and developing countries.
77
Cahier de Recherche 21
REFERENCES
Arvanitis, S. (2005), “Computerization, Workplace Organization, Skilled Labour and Firm Productivity:
Evidence for the Swiss Business Sector”, Economic of Innovation and New Technology 14 (4), pp:
225-249.
Bartoloni, Eleonora and Baussola, Maurizio (2001), “The Determinants of Technology Adoption in
Italian Manufacturing Industries”, Review of Industrial Organization 19, pp: 305–328.
Battisti, J., Hollenstein, H., Stoneman, P., and Woerter, M. (2007), “Inter and Intra Firm Diffusion of
ICT in the United Kingdom (UK) and Switzerland (CH): An Internationally Comparative Study Based
on Firm-level Data”, Economics of Innovation and New Technology, vol. 16(8), pp: 669-687.
Bayo-Moriones, A. and F. Lera-López (2007), “A firm-level Analysis of Determinants of ICT Adoption in
Spain”, Technovation 27(6-7), pp: 352-366.
Black, S.E. and L. M. Lynch (2004), “What’s driving the New Economy? The Benefits of Workplace
Innovation”, The Economic Journal 114 (493), pp: 97-116.
Black S.E. and L.M. Lynch (2000), “The New Economy: Understanding the Role of Workplace
Practices”, NBER Working Paper series 7479.
Bresnahan, T., E. Brynjolfsson and L. M. Hitt (2002), “Information Technology Workplace
Organization, and the Demand for Skilled Labor: Firm-level Evidence”, Quarterly Journal of Economics
117 (1), pp: 339-376.
Card, D., F. Kramarz and T. Lemieux (1999), “Changes in the Relative Structure of Wages and
Employment: A Comparison of the United States, Canada and France”, Canadian Journal of
Economics 32 (4), pp: 843–877.
Caselli, F., and W. J. Coleman II (2001), “Cross-country Technology Diffusion: the Case of
Computers”, American Economic Review Papers and Proceedings, 91, pp: 328-335.
Correa, P. G., A. M. Fernandes and C. J. Uregian (2010), “Technology Adoption and the Investment
Climate: Firm-Level Evidence for Eastern Europe and Central Asia”, The World Bank Economic
Review 24(1) pp: 121-147.
Doms, M., T. Dunne and K. Troske (1997), “Workers, Wages and Technology”, The Quarterly Journal
of Economics 112 (1), pp: 253-290.
Fabiani, S., F. Schivardi and S. Trento (2005), “ICT Adoption in Italian Manufacturing: Firm-level
Evidence”, Industrial and Corporate Change 14 (2), pp: 225-249.
Giunta, A. and F. Trivieri (2007) “Understanding the Determinants of Information Technology Adoption:
Evidence from Italian Manufacturing Firms”, Applied Economics 39(10), pp: 1325-1334.
Gretton, P., J. Gali and D. Parham (2004), “The Effects of ICTs and Complementary Innovations on
Australian Productivity Growth”, In: OECD (Ed.), The Economic Impact of ICT: Measurement,
Evidence and Implications. OECD, Paris, pp: 105-130.
Hollenstein H. (2004), “Determinants of the Adoption of Information and Communication Technologies
(ICT) An empirical Analysis based on Firm-level Data for the Swiss Business Sector”, Structural
Change and Economic Dynamics 15(3), pp: 315-342.
78
Cahier de Recherche 21
Mughal, Mazhar & Diawara, Barassou, (2011) "Human capital and the adoption of information and
communications technologies: Evidence from investment climate survey of Pakistan,", Economics
Discussion Papers 2011-21, Kiel Institute for the World Economy, available at:
http://www.economics-ejournal.org/economics/discussionpapers/2011-21
Nelson R. and S. Winter (1982), An Evolutionary Theory of Economic Change, Harvard University
Press, Cambridge.
Powell, T.C. and A. Dent-Micallef (1997), “Information Technology as Competitive Advantage: the
Role of Human, Business and Technology Resources”, Strategic Management Journal 18 (5), pp: 375405.
Romijn, H., 1997. Acquisition of technological capacity in development: a quantitative case study of
Pakistan’s capital goods sector. World Development 25 (3), 359–377.
Wilson III, E. J. and K, Wong (2003), “African information revolution: a balance sheet”,
Telecommunications Policy 27, pp: 155-177.
World Bank (2007), Enterprise
http://www.enterprisesurveys.org/
Surveys,
The
World
Bank
Group.
Available
at
World Bank (2009), World Development Indicators Online, The World Bank Group. Available at
http://ddp-ext.worldbank.org/ext/DDPQQ/
79
Cahier de Recherche 21