cahier de recherche - esc pau

1

Cahier de Recherche 21

CAHIER DE

RECHERCHE

N°21

Décembre 2014

Institut de Recherche en Management et en

Pratiques d’Entreprise

The Groupe ESC PAU Institute for Research in

Management and Best Practices

2


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

PAR FARID MAKHLOUF P.25

DOES EDUCATION MATTER FOR THE ADOPTION OG

INFORMATION AND COMMUNICATION TECHNOLOGIES (ICT) IN

DEVELOPING COUNTRIES? EVIDENCE FROM SENEGAL

PAR MAZHAR MUGHAL, BARASSOU DIAWARA P.45

3


Transferts de fonds, capacité d’absorption et syndrome hollandais : cas du Maroc

Farid MAKHLOUF

Professeur Groupe ESC Pau

IRMAPE

4


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

5


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

en causant ce que l’on appelle le « syndrome hollandais1 » (Bourdet et Falck, 2006). Les effets macro-

économiques 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

sur le marché des changes2 (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.

2 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.

6


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

ajustements sur l’économie marocaine, en particulier par l’effet dépense3.

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).

7


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 1975-

2003 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 Période Méthode4 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.

8


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) +

Acosta et al. (2009) 1991 El Salvador

(DSGE)

+

Fajnzylber et López (2007) 1990-2003 Corrélation, transferts / Taux de change 8 pays LAC +

Singh et al. (2009) 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 Panel (20 pays) +

Fayad (2011)

1980 et

1990

Coupe transversale

IV (2OLS)

(*)

9


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

marocain, même si les travaux de Bouoiyour et al. (2004)5 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.

10


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.

Les interventions de la BAM6 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.

11


Tableau 2 : Résultats (GMM)

Table 2 MCO variable dépendante taux de change effectif réel

Taux de

Change

effectif

TNT V,Agriculture V,Industrie V,Service

Constante

6,470

(1.738)*

-0.055

(-0.210)

-7.833

(-2.942)**

1.645

(0.830)

-0.320

(-0.254)

Transferts de fonds

0.059

(0.397)

-0.279

(-2.64)***

-0.351

(-3.293)***

-0.03

(-0.487)

0.132

(2.620)*

PIB par tête

-1.322

(-2.934)***

0.488

(1.528)

0.930

(2.885)***

-0.09

(-0.396)

-0.251

(-1.645)

Masse Monétaire/PIB

0.344

(1.431)

-0.0762

(-0.447)

-1.105

(-0.615)

-0.006

(-0.047)

-0.04

(0.463)

Croissance (-1)

0.412

(1.474)*

0.100

(0.601)

0.327

(1.934)*

-0.097

(-0.774)

-0.048

(-0.601)

Ouverture

0.076

(0.281)

-0.545

(-2.83)***

-0.562

(-2.895)**

-0.145

(-1.007)

0.260

(2.831)***

Termes de l'échange

0.912

(2.095)**

-1.035

(-3.35)***

-0.635

(-2.041)**

-0.503

(-2.174)*

0.501

(3.402)**

R2 Ajusté 0.22 0.72 0.64 0.663 0.718

MMG

Taux de

change

effectif

TNT V,Agriculture V,Industrie V,Service

Constante

22.434

(3.493)***

-8.331

(-2.39)***

-21.739

(-5.880)

-6.672

(-3.704)***

12.709

(4.453)***

12


Transferts de fonds

-0.154

(-0.88)

-0.210

(-2.036)**

-0.513

(-4.312)***

0.157

(1.839)***

0.143

(2.443)**

PIB par tête

-2.889

(-4.052)***

1.310

(3.499)***

2.436

(5.674)***

0.369

(1.468)

-1.479

(-4.437)***

Masse Monétaire

1.100

(3.117)***

-0.331

(-1.942)**

-0.848

(-4.618)***

-0.314

(-2.893)***

0.577

(3.484)***

Croissance (-1)

1.200

(3.984)***

-0.335

(-1.429)

-0.157

(-0.781)

-0.296

(-3.265)***

0.268

(2.031)**

Ouverture

0.511

(2.087)**

-1.024

(-4.88)***

-0.677

(-4.312)***

0.286

(2.717)****

-0.104

(-0.680)

Termes de l'échange

0.216

(0.434)

-0.850

(-6.03)***

-0.447

(-2.945)***

0.775

(2.983)***

-0.303

(0.102)

*** significativité à 1%,**à 5%, * à 10%

( ) : t test

13


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 :

Où : θ paramètres et X les données.

Ou encore le paradigme bayésien peut être synthétisé dans le schéma 1 ci-après

dXff

XffXf

)|()(

)|()()|(

14


Schéma 1 : Paradigme bayésian

Où : π() la distribution a priori de paramètre et π(x/) la distribution de la densité x sachant le

paramètre . 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

instrumentale (Gibbs Sampler for Linear 'IV' Model7). 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

données / hypothèses

hypothèses

informations a priori

hypothèses /données 𝜋(𝜃/𝑋)

𝜋(𝑋/𝜃)

𝜋(𝜃)

15


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.

De manière plus précise, selon Rossi (2012)8 nous considérons la modélisation suivante:

),0(~),(

'

'

21

2

1

N

WXY

ZX

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

16


Les expressions des « a priori » sont données par suivant (Rossi, 2012) :

Nous allons considérer les mêmes valeurs a priori que celles de McCulloch et Rossi. En effet, nous ne

disposons pas d’a priori informatif9.

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.

, ),N(m~),( 1

A and V),IW(~

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

)A,N(m~ -1

17


Tableau 3 : Impact des transferts de fonds sur le taux de change – Variable instrumentale (PIB

du pays hôte)

Mean/Median SD

Intercept -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

Le taux de change est au certain. Un signe positif (négatif) est

équivalent à une appréciation (dépréciation). SD : standard deviation

18


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.

19


ANNEXES

Figure1 : Distributions a posteriori

1 : TOT ; 2 : OPEN, 3 : M2, 4 : GDPpercapita, 5 : Growth.

Distribution a posteriori (Transferts de fonds)

20


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:Addison-

Wesley.

http://www.tandfonline.com/loi/riej20?open=20#vol_20

http://www.tandfonline.com/toc/riej20/20/3

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.

22


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.p93-

10. 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).

http://ideas.repec.org/s/bla/reviec.html

http://ideas.repec.org/p/wbk/wbrwps/4213.html

http://ideas.repec.org/s/wbk/wbrwps.html


http://ideas.repec.org/p/tac/wpaper/2013-2014_9.html

http://ideas.repec.org/s/tac/wpaper.html

23


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.r-

project.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.



http://ideas.repec.org/p/nbr/nberwo/11513.html


http://ideas.repec.org/s/nbr/nberwo.html

http://cran.r-project.org/web/packages/bayesm/index.html

http://cran.r-project.org/web/packages/bayesm/index.html

24


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.

25


The Impact of Exchange Rate

Policy on Remittances in Morocco:

A Threshold VAR Analysis

Farid MAKHLOUF


IRMAPE / CATT

26


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

27


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 counter-

cyclical 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.

28


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.

http://en.wikipedia.org/wiki/Arab_Maghreb_Union

29


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 Authors

Amuedo-Dorantes and Pozo (2004) + Acosta et al. (2009) +

30


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

31


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

32


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.

33


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.

34


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.

35


The threshold VAR can be specified as follows:

𝑌𝑡 = 𝐵1(𝐿)𝑌𝑡−𝑙 + (𝐵2(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 TVAR10

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) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

10 http://cran.r-project.org/web/packages/tsDyn/tsDyn.pdf

36


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.

37


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%

38


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.

39


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.

40


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.

http://www.tandfonline.com/loi/riej20?open=20#vol_20

http://www.tandfonline.com/toc/riej20/20/3

http://ideas.repec.org/a/eee/wdevel/v27y1999i8p1493-1502.html

http://ideas.repec.org/a/eee/wdevel/v27y1999i8p1493-1502.html

http://ideas.repec.org/s/eee/wdevel.html

http://ideas.repec.org/a/spr/jopoec/v7y1994i2p235-45.html

http://ideas.repec.org/a/spr/jopoec/v7y1994i2p235-45.html

http://ideas.repec.org/s/spr/jopoec.html

http://ideas.repec.org/a/eee/deveco/v86y2008i2p356-366.html

http://ideas.repec.org/s/eee/deveco.html

41


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.

http://ideas.repec.org/s/bla/reviec.html

http://ideas.repec.org/p/wbk/wbrwps/4213.html


42


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



https://ideas.repec.org/p/tac/wpaper/2013-2014_3.html

https://ideas.repec.org/p/tac/wpaper/2013-2014_3.html

https://ideas.repec.org/s/tac/wpaper.html

http://ideas.repec.org/p/csl/devewp/288.html

http://ideas.repec.org/p/csl/devewp/288.html

http://ideas.repec.org/s/csl/devewp.html

http://ideas.repec.org/a/ebl/ecbull/eb-10-00500.html

http://ideas.repec.org/a/ebl/ecbull/eb-10-00500.html

http://ideas.repec.org/s/ebl/ecbull.html



http://ideas.repec.org/s/nbr/nberwo.html

43


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.

44


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.

http://ideas.repec.org/a/taf/jitecd/v15y2006i1p81-99.html

http://ideas.repec.org/a/taf/jitecd/v15y2006i1p81-99.html

http://ideas.repec.org/s/taf/jitecd.html

http://ideas.repec.org/s/taf/jitecd.html

45


Does Education Matter for the

Adoption of Information and

Communication Technologies (ICT)

in Developing Countries? Evidence

from Senegal

Mazhar MUGHAL


IRMAPE

Barassou DIAWARA

46


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.

47


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

accurately than current systems11

. 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,

transportation service companies, and public warehouses12

.

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 stores13

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.

48


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, 2005; Bresnahan, Brynjolfsson and Hitt, 2002; Fabiani,

Schivardi and Trento, 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).

http://www.sciencedirect.com.rproxy.univ-pau.fr/science?_ob=ArticleURL&_udi=B6V8B-4N74HV1-1&_user=1654077&_coverDate=07%2F31%2F2007&_alid=1113630854&_rdoc=3&_fmt=full&_orig=search&_cdi=5866&_sort=r&_docanchor=&view=c&_ct=8036&_acct=C000054100&_version=1&_urlVersion=0&_userid=1654077&md5=3bf16e25e775f471d52f0d68f46ea4e2#bib4







49


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 know-

how 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

50


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

51


Yea

r

Seneg

al

C.

d'Ivoir

e

Gamb

ia

Maurita

nia

Niger

ia

Sub-

Sahar

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

52


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

200

5 15.34 12.21 16.22 25.16

13.1

5 12.01

22.1

0 7.92 5.80 85.18

34.3

7

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.

53


Figure 1. Comparison of the telecommunications revenue (% GDP) among selected African

countries


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).

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Years

Tele

co

mm

un

icati

on

rev

en

ues

(%G

DP

)

Senegal

Burkina Faso

Cote d'Ivoire

Gambia, The

Mauritania

Nigeria

Sub-Saharan Africa

54


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

55


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


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:

,2

10 i

k

j

ijji XHUMANY

(1)

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.

56


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

57


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

58


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

59


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

60


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

61


Table 3. Effect of manager’s education level

(1) (1’) (2) (2’) (3) (3’) (4) (4’) (5) (5’)

Email Website High speed internet Internet for

communication

Internet for

research

Param

eters

Marginal

effects

Parame

ters

Marginal

effects

Param

eters

Marginal

effects

Parame

ters

Marginal

effects

Parame

ters

Marginal

effects

Primary school 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)

Started but did not complete secondary 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)

Secondary School 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)

62


Vocational Training 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)

Some university training 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)

Graduate degree (BA, BSc etc.) 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)

MBA from university in another country 0.47 0.18 -0.28 -0.03 - - - - - -

(0.64) (0.25) (0.77) (0.07)

Other postgraduate degree from university

in this country

1.07*** 0.41*** 0.22 0.03 0.92 0.34 0.64 0.21 0.39 0.10

(0.33) (0.11) (0.38) (0.07) (0.72) (0.28) (0.65) (0.24) (0.65) (0.20)

63


Other postgraduate degree from university

in another country

1.22*** 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)

Log sale in 2006 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)

Age -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)

Medium size (20-99 employees) 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)

Large size (100 employees and more) 0.25 0.10 0.38 0.07 - - - - - -

(0.45) (0.18) (0.43) (0.09)

64


Kaolack -

1.00***

-0.29*** -0.42 -0.04 -

1.02**

-0.23** -0.49 -0.11 -0.82* -0.13*

(0.28) (0.05) (0.47) (0.04) (0.73) (0.10) (0.57) (0.10) (0.75) (0.07)

Saint-Louis -

0.87***

-0.26*** -0.78*** -0.07*** -

1.48**

*

-0.29*** -1.02** -0.19** -0.39 -0.08

(0.24) (0.05) (0.49) (0.02) (0.82) (0.08) (0.66) (0.07) (0.63) (0.10)

Thies -0.12 -0.04 -0.05 -0.01 -

1.16**

*

-0.25*** -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)

Food -

1.05***

-0.31*** -1.19*** -0.10*** - - - - - -

(0.28) (0.06) (0.33) (0.02)

Garments -

1.12***

-0.30*** -0.52* -0.05* - - - - - -

65


(0.34) (0.06) (0.43) (0.03)

Fabricated metal products -0.87** -0.25*** -0.26 -0.03 - - - - - -

(0.43) (0.09) (0.51) (0.05)

Wholesale -0.75* -0.23** - - - - - - - -

(0.41) (0.09)

Retail -

0.88***

-0.28*** -0.45* -0.05* -

2.68**

*

-0.80*** -1.49*** -0.52*** -1.84*** -0.61***

(0.27) (0.07) (0.29) (0.03) (0.71) (0.09) (0.48) (0.16) (0.46) (0.15)

Hotels and restaurants -

1.03***

-0.29*** -0.58** -0.06** - - - - - -

(0.31) (0.06) (0.39) (0.03)

Other Services -0.67** -0.21*** -0.14 -0.02 - - - - - -

66


(0.29) (0.08) (0.30) (0.03)

Other manufacturing -

0.78***

-0.25*** -0.42* -0.05* - - - - - -

(0.28) (0.07) (0.31) (0.03)

Constant -2.42** - -5.13*** - -

8.19**

- -

11.10**

*

- -3.70 -

(1.08) (1.33) (3.40) (3.36) (2.74)

Observations 504 504 487 487 120 120 120 120 120 120

R2 0.247 0.247 0.272 0.272 0.521 0.521 0.422 0.422 0.359 0.359

Notes: Standard errors in parentheses; ***, ** and * means significant at 1%, 5% and 10% respectively.

67


Table 4. Effect of average education level of the production workers

(1) (1’) (2) (2’)

Email Website

Parameter

s

Marginal

effects

Parameter

s

Marginal

effects

0-3 years of education - - 0.00 0.00

(0.71) (0.12)

4-6 years of education -0.27 -0.09 0.15 0.03

(0.24) (0.08) (0.67) (0.12)

7-9 years of education -0.07 -0.02 0.40 0.08

(0.28) (0.09) (0.66) (0.15)

10-12 years of education 0.05 0.02 -0.02 -0.00

(0.42) (0.15) (0.81) (0.14)

Log sale in 2006 0.30*** 0.10*** 0.16 0.03

(0.10) (0.03) (0.11) (0.02)

Age -0.02* -0.01* -0.00 -0.00

(0.01) (0.00) (0.02) (0.00)

Medium size (20-99 employees) 0.04 0.01 0.66* 0.15*

(0.31) (0.11) (0.38) (0.10)

68


Large size (100 or more employees) -0.02 -0.01 0.63 0.15

(0.54) (0.19) (0.58) (0.17)

Kaolack -1.60*** -0.34*** - -

(0.49) (0.05)

Saint-Louis -0.81*** -0.23*** - -

(0.36) (0.07)

Thies 0.11 0.04 -0.59 -0.08

(0.29) (0.11) (0.55) (0.05)

Food -0.55* -0.18* -0.82** -0.12**

(0.35) (0.11) (0.37) (0.05)

Garments -0.77** -0.22** -0.53 -0.07

(0.41) (0.09) (0.51) (0.05)

Fabricated metal products -0.56 -0.17 0.11 0.02

(0.49) (0.12) (0.61) (0.12)

Other manufacturing -0.36 -0.12 -0.45 -0.07

(0.35) (0.11) (0.37) (0.06)

Constant -4.87*** - -3.91* -

(1.73) (2.12)

69


Observations 254 254 201 201

R2 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’) (2) (2’)

Email Website

Parameters Marginal

effects

Parameters Marginal

effects

Percentage of skilled to unskilled

workers

0.25*** 0.10*** 0.09 0.02

(0.08) (0.03) (0.08) (0.02)

Log sale in 2006 0.42*** 0.16*** 0.19 0.04

(0.13) (0.05) (0.13) (0.03)

Age -0.01 -0.01 -0.00 -0.00

(0.01) (0.01) (0.02) (0.00)

Medium size (20-99 employees) -0.29 -0.11 0.38 0.10

(0.37) (0.14) (0.44) (0.12)

Large size (100 or more

employees)

-0.26 -0.10 0.42 0.11

70


(0.66) (0.24) (0.69) (0.21)

Saint-Louis -1.28*** -0.40*** - -

(0.51) (0.10)

Thies 0.68* 0.27* -0.61 -0.12

(0.37) (0.14) (0.61) (0.09)

Food -0.52 -0.20 -0.65* -0.13*

(0.43) (0.16) (0.40) (0.07)

Garments -0.89** -0.31** -0.41 -0.08

(0.51) (0.15) (0.56) (0.09)

Fabricated metal products -1.22*** -0.37*** 0.23 0.06

(0.67) (0.13) (0.68) (0.19)

Other manufacturing -0.49 -0.19 -0.47 -0.10

(0.45) (0.17) (0.43) (0.09)

Constant -7.44*** - -4.36* -

(2.41) (2.34)

Observations 148 148 130 130

R2 0.300 0.300 0.252 0.252

Notes: Standard errors in parentheses; ***, ** and * means significant at 1%, 5% and 10%

respectively.

71


Table 6. Effect of on-the-job training

(1) (1’) (2) (2’) (3) (3’) (4) (4’) (5) (5’) (6) (6’) (7) (7’)

Email Website High speed

internet

Internet for

communication

Internet for

delivery

Internet for

research

Internet for

purchasing

Param

eters

Margin

al

effects

Param

eters

Margina

l effects

Param

eters

Margin

al

effects

Parame

ters

Margin

al

effects

Param

eters

Margin

al

effects

Param

eters

Margin

al

effects

Param

eters

Margin

al

effects

Formal on-the-job

training

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 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)

Age -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)

72


Medium size (20-99

employees)

0.28 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**

(0.22) (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)

Large size (100 or more

employees)

0.39 0.15 0.28 0.05 - - - - - - - - - -

(0.44) (0.18) (0.42) (0.08)

Kaolack -

1.01***

-

0.29***

-0.67** -0.06** -

1.24***

-0.28*** -0.72* -0.16* -0.93* -0.09* -1.01** -0.16** -0.22 -0.03

(0.27) (0.05) (0.48) (0.03) (0.74) (0.09) (0.59) (0.10) (0.97) (0.05) (0.76) (0.06) (0.74) (0.08)

Saint-Louis -

0.90***

-

0.27***

-

0.82***

-0.07*** -

1.39***

-0.31*** -0.95** -0.20** -0.20 -0.03 -0.51 -0.10 0.06 0.01

(0.24) (0.05) (0.47) (0.02) (0.70) (0.09) (0.58) (0.08) (0.65) (0.08) (0.56) (0.09) (0.58) (0.09)

Thies -0.18 -0.06 -0.17 -0.02 -

1.25***

-0.29*** -1.37*** -0.25*** -0.73 -0.08 -0.38 -0.08 0.02 0.00

73


(0.20) (0.07) (0.30) (0.04) (0.62) (0.09) (0.63) (0.06) (0.75) (0.05) (0.52) (0.09) (0.56) (0.08)

Food -

1.13***

-

0.33***

-

0.99***

-0.09*** - - - - - - - - - -

(0.27) (0.06) (0.31) (0.02)

Garments -

1.39***

-

0.35***

-

0.74***

-0.07*** - - - - - - - - - -

(0.33) (0.04) (0.43) (0.02)

Fabricated metal

products

-

1.05***

-

0.29***

-0.25 -0.03 - - - - - - - - - -

(0.42) (0.07) (0.49) (0.05)

Wholesale -

0.98***

-

0.27***

- - - - - - - - - - - -

(0.41) (0.07)

Retail - - -0.42* -0.05* - -0.79*** -1.61*** -0.57*** - -0.47*** - -

0.59**

-0.87** -0.19**

74


0.99*** 0.31*** 2.75*** 1.63*** 1.74*** *

(0.25) (0.06) (0.28) (0.03) (0.63) (0.07) (0.43) (0.14) (0.42) (0.14) (0.40) (0.13) (0.42) (0.12)

Hotels and restaurants -

1.00***

-

0.29***

-0.44 -0.05 - - - - - - - - - -

(0.30) (0.06) (0.37) (0.03)

Other Services -

0.66***

-

0.21***

-0.08 -0.01 - - - - - - - - - -

(0.28) (0.08) (0.29) (0.04)

Other manufacturing -

1.04***

-

0.31***

-0.54** -0.06** - - - - - - - - - -

(0.26) (0.06) (0.29) (0.03)

Constant -

2.85***

- -

5.02***

- -

7.23***

- -7.88*** - 0.95 - -3.66 - -7.92*** -

(1.01) (1.25) (2.59) (2.49) (2.82) (2.35) (2.84)

75


Observations 506 506 489 489 121 121 121 121 121 121 121 121 121 121

R2 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

76


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.

77


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 on-

the-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.

78


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.

79


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: 375-

405.

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 Surveys, The World Bank Group. Available at

http://www.enterprisesurveys.org/

World Bank (2009), World Development Indicators Online, The World Bank Group. Available at

http://ddp-ext.worldbank.org/ext/DDPQQ/

http://www.economics-ejournal.org/economics/discussionpapers/2011-21

cahier de recherche - esc pau

Documents