Page 1 of 19
European Journal of Applied Sciences – Vol. 10, No. 3
Publication Date: June 25, 2022
DOI:10.14738/aivp.103.12204. Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and
Demographic Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
Services for Science and Education – United Kingdom
Causality Test on Panel Data: An Application to the Study of the
Causality Between Agricultural Growth and Demographic Burst
in the Uemoa Area
ADEGNIKA Moutairou
Teacher-Researcher at the National School of Applied Economics and
Management (ENEAM) of the University of Abomey-Calavi
ABSTRACT
This article aims to analyze the main lines of the modern theoretical debate around
the interactions between demographic growth and agricultural growth in the
WAEMU space. It identifies the (none) causalities between certain demographic and
agricultural growth variables focused on an autonomous and intermediate
research field between the corpora of demography and agricultural growth theory.
The results of the application of the approach of Konya (2006) based on a test of
(none) causality in the sense of Granger (1969), show that the econometric results
obtained are ambiguous to the theoretical corpus which was constituted between
the strong population growth and weak agricultural growth in the WAEMU space.
According to the results of the various (no) tests, the causality between the
population surge and agricultural growth is far from being retroactive in certain
countries (Benin and Mali) unlike other countries in the area where we note the
existence of feedback between population growth and agricultural growth. Almost
all the countries of the UEMOA space seem to be part of the logic of the
populationists. Population growth in WAEMU countries remains a determining
factor for agricultural growth. It is the main lever of agriculture. This result
confirms the thesis of Jean Bodin (1530-1596).
Keywords: Population growth-Agricultural growth-(None) Causality-UEMOA space- Panel
JEL: C23, O41, O53, O55
INTRODUCTION
Population growth is still rapid in many parts of the WAEMU region. This population growth
seems to induce recurring problems of regional or local agricultural growth. The social,
economic and political issues in the WAEMU space are associated with the various demographic
trajectories whose adverse effects vary from one State to another in this space. This divergence
in the effects of population growth on agricultural growth is explained by the capacity of each
member country of the space to satisfy these basic food security and nutrition needs on the one
hand, but also by agricultural policy (crops cash crops/ and or food crops). Meeting food needs
and politics should be driven by strong health and education policies. In the UEOMA countries,
do the agrarian systems of the space reduce the food and nutritional insecurity affecting the
population, in both urban and rural areas? It is also about the capacity of the Member States of
the UEMOA space and local authorities to draw up and implement effective policies with a view
Page 2 of 19
282
European Journal of Applied Sciences (EJAS) Vol. 10, Issue 3, June-2022
Services for Science and Education – United Kingdom
to sustainable development (preserving the environment, preventing conflicts, ensuring
security goods and people, equipping and managing a growing number of agglomerations and
supporting the emergence of dynamic urban economies, etc.).
In the WAEMU area, the problems of population growth and agricultural growth are
particularly acute. The population growth rate is high there, 2.3% on average over the past ten
years. In addition, the current growth in foodstuffs is about 30% lower than in 1967. Low yields
and the scarcity of land is widespread there. These obstacles to the growth of agricultural
production are further reinforced by the fragmentation of agricultural holdings, poor land
management, a lack of technical and organizational progress in agriculture, and inappropriate
agricultural and economic policies. The food situation in WAEMU countries is particularly
delicate. It has deteriorated over the years despite some progress in agriculture, and the
demographic outlook suggests growing difficulties if significant agricultural policy measures
are not taken (Benson, 2004; Bruinsma, 2002; IFPRI, 2004; Paillard, 2010; Rosegrant, 2003).
In this alarming situation how to solve the resulting food security problem in this space. Several
methods are considered in this perspective, but it is clear that they have remained ineffective.
The first method advocated is to increase food production by 300% to provide barely adequate
diets for their projected 2 billion people by 2050.
The WAEMU space seems to be characterized by large-scale poverty and malnutrition, large
food deficits are observed in some countries in the space, a high and growing dependence on
the import of food products and aid dealer. Food products still represent around 70% of the
total value of agricultural production in this area, and it is estimated that demand in this area
will double by 2020.
Generalized population growth also concerns the agricultural labor force, but with variable
rates that contribute to widening the differences between countries in terms of the ratio of
dependents to labor force. The number of people to feed per worker, which was between 1.7
and 2.8 at the start of the period, increased from 2.3 to 5.4 forty years later. This corresponds
to different rates of urbanization and agricultural exodus. The group of Sahelian countries
(Mali, Niger, Burkina Faso) remained very rural and agricultural. Conversely, for the coastal
countries (Benin, Togo and Côte d'Ivoire) the relative place of agriculture in employment has
regressed and the number of people to feed per worker has increased from less than 3 to more
than 4. Senegal and Guinea Bissau followed intermediate trajectories, close to the overall
average Benoit-Cattin et al (2011). This amounts to an additional need of around 50 billion
dollars per year (at current prices). With a low level and stagnation of per capita income, large- scale poverty, the demand in the WAEMU space for high-value food products remains low.
Moreover, the area is experiencing fairly high demographic growth, around 3% on average. The
rural population represents more than half of the population in the space; it too is growing at
an average rate of 2%. This rural population constitutes the majority of the mass considered as
poor according to the criteria defined by the organizations of the United Nations. Agriculture is
its main economic activity in this space.
The different trends in GDP and population growth have led to even more contrasting trends in
per capita GDP by country, particularly for the most recent period considered by Maddison
Page 3 of 19
283
Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and Demographic
Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
URL: http://dx.doi.org/10.14738/aivp.103.12204
1973-2001. In the developing countries, only the Asian countries as a whole continued to
record strong growth in their GDP per capita during the period 1973-2001: 3.6% per year on
average, against 0. 9%per year for all Latin American countries, and +0.2% for all of Africa. The
insignificant growth of agricultural GDP per capita in this area between 1973 and 2001 is
attributable to several factors, among which we can cite: an international economic
environment less favorable to this area than in the previous period. The annual data available
between 1960 and 2013 for the major regions of the world confirm the negative impact of
strong demographic growth in sub-Saharan Africa on the increase in its average agricultural
GDP per capita, despite the recent recovery in the economic growth of the region. Agricultural
growth in the area was 6.5% in 2008 against an average of 4.4% during the period 2003-2007
and 3.2% in the 1990s. More than half of the countries in the region have reached the CAADP
target of 6% in 2008. These are Senegal (14.2%), Mali (10%), Burkina Faso and Niger (8.6%).
Given its weight (approximately 70% in agricultural value added), the crop production sub- sector remains the main driver of agricultural growth. With growth of 35.7% in Senegal, 15.7%
in Burkina Faso (11.6%), 81% in Niger, 84% in Benin. This performance recorded in 2008 is
attributable to measures to boost food production following the crisis of soaring food prices
and to the good distribution of rainfall in time and space during the agricultural campaign
2008/2009. However, there was a decline in agricultural growth in 2009. It stood at 4.8% in
2009 at the regional level. A third of the countries were able to achieve the 6% CAADP growth
target (compared to more than half in 2008).
Data on health, education, sanitation and other areas are not so good. Thus, the data available
at the WAEMU level on HIV/AIDS infections show a high prevalence in the countries of the
coastal region. These HIV infections are progressing thanks to the precariousness of social
conditions caused by the war and the appearance of refugee camps, particularly in Côte d'Ivoire,
accentuating promiscuity and prostitution. The movements of combatants, often accompanied
by rape and prostitution, are also likely to favor the progression of the prevalence rate. Indeed,
the prevalence rate is about 10.8% in Côte d'Ivoire and 7.2% in Burkina Faso. Senegal, Mali and
Niger have a prevalence rate of less than 2%, the lowest in the sub-region. The illiteracy rate
remains high in this part of Africa. In Niger it reached 84%. Mali, Burkina, Senegal and Benin
also have very high rates (between 60 and 80%). These rates are generally higher in rural areas
and for women. Statistics show a strong inequality between nations. For example, in 1999,
infant mortality varied from 50 in Mali to 252 in Niger. For a few rare countries, the infant
mortality rate increased between 1990 and 1999. This is the case of Côte d'Ivoire. Maternal
mortality remains a concern. The increase in the majority of cases of health expenditure
expressed in relation to GDP shows that the States are not indifferent to the health situation in
the sub-region. Life expectancy at birth increases very timidly in most cases. However, Ivory
Coast, Mali, Togo and Burkina are falling behind in terms of longevity. This situation can be
explained, among other things, by the economic crisis during the 1990s, and especially by the
persistence of diseases such as AIDS and malaria. Access to health services is high in Niger,
Senegal and Burkina with rates above 90% on the other hand, this rate is low in Benin (18%).
The latter even experienced a regression between 90-93 and 2000. Niger experienced the
greatest progress in this area because in 1990-93, only 30% of the Nigerien population had
access to health care. Access to sanitation remains a crucial problem for the entire sub-region
given the low rates observed, half of the countries being below 50%. Among the countries in
Page 4 of 19
284
European Journal of Applied Sciences (EJAS) Vol. 10, Issue 3, June-2022
Services for Science and Education – United Kingdom
this area, some have steadily increased the share of public resources earmarked for agriculture.
The Maputo ratio rose to 8% in 2008 in Togo. For the rest of the countries, we note a drop in
the ratio in Benin (from 8.2% between 1996 and 2000 to 6% over the period 2005-2008) and
its stagnation around 2% in Côte d'Ivoire. The work of Taondyandé et al (2012) indicates that
cereal production increased by 12.6% in 2008 in the West African region against an average
increase of 5.2% over the period 2000-2007. Its increase in 2008 is well above the trend
observed since the beginning of the decade in most countries. The largest increases were
recorded in Senegal where it almost doubled (99.3%), followed by Burkina Faso (41.1%), Niger
(25.8%) and Benin (9.4%). At the regional level, this increase in cereal production results from
the almost balanced increase in the supply of rice (13.9%), maize (13.5%) and millet/sorghum
(11.6%). However, when Nigeria is excluded, maize recorded the largest increase (28.3%),
followed by millet/sorghum (22.4%) and finally rice (19.8%). While the region is globally self- sufficient for maize, millet and sorghum, it remains structurally in deficit for rice, where the
rate of coverage of needs is approximately 35% in West African countries. In terms of
productivity, rice is the cereal speculation that recorded the most significant increase in yield
in 2008. In fact, the efforts made by the countries to increase the supply of rice following the
increase in its prices on the international market allowed an increase in its productivity
compared to its level of 2000-2007. With the exception of Côte d'Ivoire where it fell by 8.6%,
rice yield increased from 8% in Togo to 34% in Benin.
For cereal speculation that did not benefit from specific measures, yields also increased in some
countries. However, the climatic factor would be the main determinant of this increase. The
ReSAKSS-AO statistics (Annual Report on Agricultural Sector Trends and Prospects) in 2010
show that the supply of roots and tubers has also progressed beyond its recent trend. It
increased by 9% in the region against an average increase of 5.2% over the period 2001-2007.
This increase in production is the result of efforts to diversify the food supply of certain Sahelian
countries. Indeed, the increase in the production of roots and tubers in the region remains
lower than the regional trend in the main producing countries with the exception of Benin
where it increased by 54% compared to 2007. On the other hand, it tripled in Senegal
(+198.7%) and increased by 42.8% in Burkina Faso. The extension of cultivated areas entirely
explains the increase in supply in the main cassava-producing countries insofar as its yield fell
by 4.7% in Côte d'Ivoire, 1.6% in Togo and 0. 6% and remained constant in Benin. The increase
in plantings also explains the increase in yam production in producing countries with the
exception of Benin where its yield increased respectively by 4.3% compared to 2007.
The results recorded in 2009/2010 are globally below those of 2008. Cereal production at the
regional level remained almost constant (an increase of only 0.8%). It suffered a decline in 6
countries of the region. These are Niger (-27%), Burkina Faso (-17%), Senegal (-2%). The
supply of roots and tubers in the region increased by 5.3% in 2009 against 9% in 2008. The fall
in production in Côte d'Ivoire by 10.6% explains the decline in the rate of increase in regional
production compared to 2008. After having increased by 17% in 2008, the supply of legumes
(peanuts, dry beans, soybeans) in the area fell slightly by 0.1% in 2009 due to the drop in
production in Niger (-45.2% ), in Burkina Faso (-11.3%). It recorded an increase in Senegal
(+30.1%).
Page 5 of 19
285
Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and Demographic
Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
URL: http://dx.doi.org/10.14738/aivp.103.12204
Inflation is relatively low in WAEMU member countries. Over the 2000-2007 periods, it
remained below the regional convergence criterion of 3%. Currency instability could be the
cause. In 2008, there was a significant increase in the prices of agricultural products in all the
countries of the region. It oscillated between 8% in Côte d'Ivoire and Senegal to 28% in Togo.
This inflation is attributable to the transmission of higher international prices to local markets.
On average, the terms of trade were balanced over the 2000/2007 period in the region. In 2008,
despite the rise in the prices of local products, given the degradation of terms in most countries
of the region. The rise in producer prices did not make it possible to compensate for the loss of
purchasing power of farmers linked to the inflationary spiral of 2008. The current trend
between agricultural growth and population growth in the WAEMU space, achieve the second
goal of the SDGs? (Ending Hunger; Ensuring Food Security; Improving Nutrition and Promoting
Sustainable Agriculture). The challenges of reviving the agricultural sector in the face of
demographic growth in this space are stalling, insofar as in the WAEMU space, the budgets
allocated to agriculture remain low and below the Maputo agreements.
Faced with the multiple unexplained and unmetered constraints in the agricultural sectors of
the member countries of the space, the second objective of the SDGs seems to be mortgaged.
Among the constraints: agrarian and land reforms, the adoption of agricultural technical
innovations, water control, the perception of climate change with its adverse effects on
agricultural production, the impoverishment of infertile land and training agricultural
producers. Despite this precarious situation to achieve the second objective of the SDGs,
demographic growth and urbanization continue to increase at a high rate. Agricultural growth
and the various methods recommended to increase production and productivity remain
ineffective insofar as the strict causalities between agricultural growth and population growth
are not determined in an alarming and recurrent context. It is imperative in a context of
sustainable food security, to study the causality between agricultural growth and rural
population growth on the one hand, and agricultural growth and active population growth on
the other hand. Is there a retroactive causality between agricultural growth and population
growth in the WAEMU space?
Do rural and active demographic surges in the WAEMU space have recurring impacts on the
causality between agricultural growth and demographic surge? What is the relationship
between population growth and agricultural growth in the WAEMU space? How does
demographic growth affect agricultural growth in the WAEMU region? How does the
demographic transition theory apply to the current trends observed in the WAEMU space?
What is the criterion for comparing the neo-Malthusian thesis considering population growth
as a major problem for humanity, and the thesis which considers population growth as a neutral
or even positive factor for agricultural growth? How would you assess the relative urgency of
demographic concerns in the WAEMU space? This article therefore answers these multiple
questions which are real solutions to remedy and to confirm or invalidate the neo-Malthusian
and anti-Malthusian theses.
THEORETICAL AND EMPIRICAL LITERATURE REVIEW
For twenty-five years, the literature on the subject has increased steadily, without it being clear
where it begins or where it ends. It is true that the issue is broad; that more and more scientific
disciplines are involved in it and that above all there is growing agreement on the
Page 6 of 19
286
European Journal of Applied Sciences (EJAS) Vol. 10, Issue 3, June-2022
Services for Science and Education – United Kingdom
interrelationships between demographic phenomena and agricultural growth. Population
growth affects the degradation of agricultural resources. In most cases, the relationship
between these two elements is centered on the major scientific “theories”. In this situation,
what are the predominant doctrinal positions in the major international development
agencies? What does scientific field research confirm or invalidate, in well-specified contexts
and societies? The problem is more complex than it first appears. A school debate perhaps, but
above all an ideological conflict based on a different vision of the world (fear of the South, for
example) and which could lead to opposing political priorities: one would be based on family
planning, the other on development. If the positions of each side were clearly affirmed in the
1960s and 1970s, they are relativized today, but the root of the problem remains. Thus, the
pure and hard neo-Malthusian model, with its leader, Malthus (1978) explains the negative
impact of population growth on agricultural production. This concept of Malthus (1978) has
been globally and “logically” extended to the environment. With the finite world assumption,
the earth's physical and biological limits are the ultimate constraints to population growth and
socio-economic change.
In other words, it is the increase in populations (in the South, since those in the North are almost
stagnating) that is at the origin of the problems, which is the threat of tomorrow for the whole
world. For Ramade (1987), the major catastrophe that affects humanity and from which stems
most of the evils from which it already suffers or which threatens it is of intrinsic origin: it
comes from its anarchic reproduction with the consequence of an exponential increase in the
number of 'men. The Brundtland Report (1988) is more nuanced: "Poverty is both the effect
and the cause of global environmental problems", emphasizing inequalities, the relationship
between economic development and the environment, the responsibility of political national
and global economic system even if "the population explosion is a threat". The recent report of
the South Commission (1990) recognizes that "the rapid growth of the countries of the South
accentuates the pressure on natural resources, to varying degrees, depending on the
availability of cultivable land and land tenure systems", but the demographic pressure is only
one of seven factors affecting the environment, the others being land tenure, the type of
agricultural development, economic pressure from the North, the imperative of
industrialization and growth, the adoption of consumption habits using a lot of energy and
finally the exodus of rural populations towards the north. "There is a compelling need for action
to moderate population growth...because, although it may not always be the ultimate cause of
poverty, it can radically undermine a country's ability to value its human capital. For Brown
(1986), the extension of the population leads to a reduction in cultivable land per inhabitant,
overexploitation of the soil leading to its erosion, to a drop in its productivity and to a reduction
in food production. As demand increases, supply decreases. With the addition of droughts, as
in the Sahel in 1973-1974 and 1983-1984, we can end up with famine and an increase in
mortality.
In this same perspective, Hogan (1991) shows that the ongoing desertification is not the direct
consequence of population growth; it is the product of climatic accidents occurring in societies
with strong social inequalities which remove any alternative to the peasants. Authors such as
Paul and Anne Ehrlich (1968) have constantly repeated since the end of the 1960s that
humanity runs the risk of a collision with the natural world with the growth of the human
population could outweigh all the benefits of economic growth and the progress of modern
Page 7 of 19
287
Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and Demographic
Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
URL: http://dx.doi.org/10.14738/aivp.103.12204
science, resulting in the general ruin and devastation of the Planet. This neo-Malthusian
perspective has attracted much attention and provided the starting point for contemporary
debate on population growth. For the anti-Malthusianism, from the ambient pessimism of the
two previous models into a kind of optimism: by simplifying, man has always shown himself
capable of facing external threats and problems as he did yesterday. He has a great capacity for
adaptation and innovation. The population is only a completely secondary, sometimes even
favorable factor; the real factors are elsewhere: poverty, inequalities, inadequacy of
technologies, agricultural policies (priority to cash crops), land ownership, urban bias, wars,
and political regimes.
It is a position recalling that there are other emergencies than just family planning; it is also a
position that can justify any policy of non-intervention in fertility matters. The extreme
populationist thesis, defended in particular by Simon (1981), according to which there is no
population problem, almost denying environmental problems. There is no need to fear a
shortage of resources, because the larger a population, the greater its capacity for invention and
technological innovation (of substitute products). Demographic growth in the Third World can
be a factor of economic development in the long term, even if it is recognized that in the short
term it is a constraint. According to Boserup's (1965, 1981) thesis on agricultural development,
population growth or pressure is a stimulus, or even a necessary precondition for progress in
agriculture. The increase in rural densities, the progressive scarcity of land in relation to the
population lead to a more intensive use of land, requiring more work, resulting in increases in
productivity and a general evolution of the structures of production and power. . In a way, the
equation is reversed: it is under (demographic) constraint that technological progress can
occur. This “anti-Malthusian” argument, recognizing a positive role for the population, is
perhaps suitable for Western history or temperate countries, but does not seem generalizable
to the whole of the contemporary world. For example, a study (Pingali et al, 1984) covering 52
areas in Africa showed the importance of soil quality, rainfall and financial resources to make
the investments required for crop intensification following population pressure. . On the other
hand, in some regions (Nicoll, 1984), rural density: has increased without any subsequent
intensification of cultivation. Simplifying, on the one hand, one sacrifices sustainability or the
long term for maximum and immediate profit and pleasure, on the other hand, one overexploits
available natural resources in order to survive.
However, it is important not to rush the reasoning, by attributing in a privileged way, to use the
terminology of Shaw (1989), to an immediate and aggravating factor (the growth of the
population) what comes from basic factors (economic model, poverty, North-South
inequalities) which will be difficult and long to change. Commoner (1988) relativized the
importance of the demographic factor. The evolution of pollution (in different forms) in the
United States between 1950 and 1970 is to be attributed first to the technology used, then only
to the increase in consumption per capita and to that of the population. In an analysis this time
focusing on 65 developing countries between 1970 and 1980 and on three types of production
(cars, electricity, fertilizers and pesticides), he also arrives at a similar result: the weight of
demography, even if it is not negligible, is two to three times less than that of the technology
used. There is no close correlation between population growth rate and environmental
degradation. In summary, the theory attributing to population growth a primordial role in the
quality of agricultural growth is not confirmed by these quantitative studies. Cruse (1994)
Page 9 of 19
289
Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and Demographic
Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
URL: http://dx.doi.org/10.14738/aivp.103.12204
increase in the index corruption are major bottlenecks. Vimard et al (2011) examine the
relationships between demography, economic growth and social development in Africa. They
explain the main lines of demographic dynamics and show the specificity of the continent in the
general panorama of demographic transitions on a global scale. The authors place the debates
on population and development policies in the African context. They then analyze the general
relationships in Africa between population growth and development, highlighting the
specificity of the Maghreb and the role of the improvement of human capital in demographic
changes. On this basis, they then propose three possible demo-economic trajectories in the
medium term. In conclusion, emphasis is placed on the need to give development policies a
regional specificity. Kevin (2011) points out that Population growth has a proven influence on
food availability in Africa. This impact can be all the more detrimental as the population of
Africa will reach the figure of 2 billion people in 2050. It also notifies that in the areas of West
Africa, East and of the center will experience a demographic growth of 20% per year. He further
pointed out that among the fastest growing countries in the world are 10 African countries.
Over the next four decades the poorest people in Africa will be in the rural world. This
demographic growth will have its impact on the use of natural resources. These natural
resources will be depleted leading to more constraints for food production. Amadou SY (2014)
shows that in Mali, population growth has no significant short-term positive impact on
economic growth; it even contributes to increased food insecurity. Population growth could
have significant positive effects on economic growth since the investments will have the effect
of stimulating the production of companies leading to the demand for additional labor and
therefore the acceleration of economic growth and reduction of unemployment.
The work of Luc et al (2019) explains that the need to reduce the birth rate has been put
forward as the only possible population policy to fight against poverty. They showed that
population growth has a negative effect on food security. Their work also uses that the increase
in population densities has resulted in a double process of intensification of agriculture and
continuation of extensive practices. Similarly, the work of Putri et al (2019) showed that
demographic pressure on certain land carrying capacities. They analyze the correlation
between land pressure and food sufficiency in West Kalimantan. Their results confirm that land
pressure in West Kalimantan is mostly classified as safe, except for the city of Pontianak whose
population pressure is the most among all cities so that land carrying capacity is classified as
low. Factors of demographic growth and main activities of certain regions which affect land use
and consequently the environment. Adam et al (2019) use a farm household simulation model
to show that the interplay between human population growth and crop yields presents
challenges for agricultural production and farm household incomes in sub-Saharan Africa. They
note that increasing yield potential from more efficient use of livestock manure is one approach
to improving agricultural production and incomes in the face of impending population
pressures.
Our results suggest that, even without taking climate change into account, expected changes in
population density and crop prices in 2050 mean that crop production and income per person
could fall by 21% from 2013 values. if return potential and return spreads remain constant.
However, agricultural production and income per person could increase in 2050 by 8%
compared to 2013 values if (1) growth rates of potential yield increase by 1.13% each year and
for pulses increase by at least 0.4% each year, and (2) farmers use livestock manure more
Page 10 of 19
290
European Journal of Applied Sciences (EJAS) Vol. 10, Issue 3, June-2022
Services for Science and Education – United Kingdom
efficiently. Our prospective approach aimed at considering agricultural production at the scale
of agricultural household’s complements large-scale analyzes of the production dimension of
food security. Similarly, recent work by Mamy et al (2020) has shown that population growth
implies needs for access to food and nutritional security, health, education and employment for
newcomers. Their results reveal the existence of population growth above the country's
average but unevenly distributed a gradual increase in the quantity of cereals (maize, millet and
sorghum) per inhabitant linked to a faster increase in agricultural production population and
significant progress in terms of physical access to health and education infrastructure. They
explain that population growth poses new challenges related to food security and nutrition.
Akpan et al (2021) examine the relationship between total agricultural land use and population
growth rate from 1961 to 2018 in Nigeria. Their results revealed that agricultural and arable
crops experienced an increase in the rate of 0.62% and 0.72%, respectively per year, while the
growth rate of the total population was at 2.57%. In addition, urban and rural populations grew
at a rate of 4.75% and 1.67%, respectively. Moreover, agriculture and arable land use rates had
a significant positive correlation. From they reveal that rural population growth is lower than
urban growth rate, implying that rural population is deteriorating with its likely negative effect
on agricultural labor.
ANALYSIS METHOD AND TOOLS
The approaches of the econometric literature concerning the tests of (none) causality on panel
data do not yet present a great diversity. Nevertheless, the literature provides some approaches
or methods to test causality between two or more variables. Among these approaches we have:
Granger (1969-1980), Konya (2006), Dumitrescu and Hurlin (2011), Haugh-Pierce (1976-
1977) and Sims (1972). The basis of Granger's definition is the dynamic relationship between
variables. As noted, it is stated in terms of improving the predictability of a variable. In Granger,
temporal succession is central and one cannot discuss causality without taking time into
consideration (Sekkat, 1989). Causality is introduced into econometric analysis by Wiener
(1956) and Granger (1969). Originally, we find the formalization of the notion of causality in
physics, in particular in the work of Isaac Newton on the driving force (cause) and the change
of movement (effect). In this case, the notion of causality translates a principle according to
which if a phenomenon is the cause of another phenomenon, called "effect", then the latter
cannot precede the cause. However, its conceptual definition goes back to the speeches of
Aristotle or David Hume. Transposed into economics, the notion of causality takes on a specific
technical connotation. Indeed, if one variable caused another variable, then necessarily the two
variables must be correlated. Conversely, it is not enough for two variables to be correlated for
it to have causality (correlation is not causality). In this work, we are interested in the non- causality tests of David and Colin (2003), Hurlin and Venet (2001), Konya (2006) and
Dumitrescu and Hurlin (2011). Indeed, we present an econometric study of (non) causality in
the sense of Granger (1969) in a heterogeneous panel, based on the approaches of Konya
(2006); Dumitrescu and Hurlin (2011). This approach is used to test the existence of a causal
relationship between agricultural growth (Agricultural GDP), demographic pressure (rural
population) and the active population in the 8 WAEMU member countries. Our work takes into
consideration the approach of Konya (2006) to test the (non) causality between these two
variables. Also, since it is important to do preliminary tests (specification test, selection of lags,
stationary test) before performing a causality test, we implement some of these tests associated
with panel data, which are different from those considered in the case of time series. With
Page 12 of 19
292
European Journal of Applied Sciences (EJAS) Vol. 10, Issue 3, June-2022
Services for Science and Education – United Kingdom
=
�.: �!,+,$ = 0 ∀� = 1,2 ... ... �, �. � �+,# �� ����� ��� �+,# ����� ������� (1969)
�!: �!,+,$ ≠ 0 ∀� = 1,2 ... ... �, �. � �+,# ����� �+,# ����� ������� (1969)
Data source and specification of variables
The data used in this article comes from the FAO. The most widely used were those of the United
Nations Population Division, the World Bank and DHS and MICS Measurement (for
Demographic and Health Surveys, and Multiple Indicator Cluster Surveys). These data range
from 1960 to 2015. They concern the demography at the center of development trajectories in
the eight WAEMU countries (Benin, Burkina-Faso, Côte d'Ivoire, Guinea-Bissau, Mali, Niger,
Senegal, and Togo). These data obtained make it possible to measure the extent of population
growth on the agricultural growth achieved in this area. But this analysis also makes it possible
to understand the extent of the efforts that remain to be made, in the agricultural sector, to
really achieve the optimal agricultural growth covered by the demographic projections of this
WAEMU space. The results of the various causality tests and discussion.
Preliminary tests
First, we use the procedure of Hsiao (1986) to test the panel structure (homogeneous or
heterogeneous) underlying the data. This procedure is based on a general drafting of the panel
template, as follows:
�+# = �+ + �+�+,# + �+,# � = 1,2 ... ... . �, �� � = 1,2, ... ... �
Where: �+ denotes a vector of individual effects of dimension N, �+ denotes a regression
parameter vector of dimension (K, 1), with K the number of rectifiers in the model. �+,# are
independently and identically distributed, such that, �4�+,#6 = 0 , �4�+,#
) 6 = �0
) ∀ � =
1, ... �.i=1,...N.
First, we test the hypothesis of the global homogeneity of the parameters, which is written as
follows:
The Global Homogeneity Hypothesis =
�.: � = �+ �� � = �+ ∀ � = 1,2 ... ... �,
�!: ∃ �+ ≠ �, 12 3( 4 3) ���� �,� = 1,2, ... ... �.
(4)
If we cannot reject the hypothesis of global homogeneity, the model considered is written as
follows:
�+# = � + �5
�+,# + �+,# � = 1,2 ... ... . �, �� � = 1,2, ... ... � (5)
On the other hand, in case of rejection of the null hypothesis, we seek the source of the
heterogeneity by carrying out a second test of the homogeneity, but this time only for the
regression parameters( �+), the hypothesis is then worded as follows:
The hypothesis of the homogeneity of the regression parameters
=
�.: � = �+ ∀ � = 1,2 ... ... �,
�!: ∃ �+ ≠ �, ��� �,� = 1,2, ... ... �.
(6)
Page 13 of 19
293
Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and Demographic
Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
URL: http://dx.doi.org/10.14738/aivp.103.12204
If we cannot reject the null hypothesis, the model this time can be written as follows:
�+# = �+ + �5
�+,# + �+,# � = 1,2 ... ... . �, ��� � = 1,2, ... ... � (7)
On the other hand, if we can reject the null hypothesis, we test the homogeneity of the individual
effects, by considering the following hypothesis:
The hypothesis of the homogeneity of the regression parameters
=
�.: � = �+ ∀ � = 1,2 ... ... �,
�!: ∃ �+ ≠ �, ��� �,� = 1,2, ... ... �.
(8)
If we cannot reject the null hypothesis, then the model is written as follows:
�+# = � + �5
�+,# + �+,# � = 1,2 ... ... . �, �� � = 1,2, ... ... � (9)
On the other hand, if we can still reject the null hypothesis, then the model is globally
heterogeneous. The application of this procedure to the data of our study indicates the
following results:
Table 1: Results of the homogeneity test
Types of tests Related Fisher statistics P value
H0: Homogeneity of individual effects and
regression parameters
8.20 0.00264
H0: Homogeneity of regression parameters 1614,85 0,00000
H0: Homogeneity of individual effects 65,0552 6,55426e-064
Source : Results of our 2021 estimates
Based on these results in Table 1, we can reject all null hypotheses of homogeneity, which tells
us that the panel model associated with the data is heterogeneous and in the form of an
individual effect model. As Konya (2006) in his approach considered a heterogeneous panel,
we continue to apply this approach. In this case, the SUR system (Seemingly Unrelated
Regression), for the approach of Konya (2006), can be written as follows from a trivariate
model with two lags
(10)
⎩
⎨
⎧ �������!,# = �6,!# +X �6,!,$�������!#7$ +X �6+,$������!#7$ + X �6+,$������+#7$ +
$()
$(!
�6,+#
$()
$(!
$()
$(!
�������8,# = �6,+# +X �6,8,$�������8#7$ +X �68,$������8#7$ +X �68,$������8#7$ +
$()
$(!
�6,8#
$()
$(!
$()
$(!
��
(11)
⎩
⎨
⎧ ������!,# = �9,+# +X �9,!,$�������!#7$ + X �9!,$������8#7$ +X �9,8,$������8#7$ +
$()
$(!
�9,8#
$()
$(!
$()
$(!
������8,# = �9,8# + X �9,8,$�������8#7$ + X �9,8,$������8#7$ + X �9,8,$������8#7$ +
$()
$(!
�9,8#
$()
$(!
$()
$(!
Page 15 of 19
295
Adegnika, M. (2022). Causality Test on Panel Data: An Application to the Study of the Causality Between Agricultural Growth and Demographic
Burst in the Uemoa Area. European Journal of Applied Sciences, 10(3). 281-299.
URL: http://dx.doi.org/10.14738/aivp.103.12204
According to the results of this table 3, all the variables of our panel are stationary in level, so
we can now pass to the test of (non) causality by applying the approach of Konya, (2006) as
developed above by the method OLS. This test developed in Konya (2006) to analyze the causal
link between exports and the GDP of 27 countries has had several applications among the
authors. This is how authors like Khalil (2014) applied this test to analyze the causal
relationship between financial development and economic growth in developing countries. For
the case of this article, this test is applied by estimating an individual fixed-effect model (by
country) with two lags on the three explanatory variables (agricultural GDP, rural population,
active population) and performing for each estimated model by country, a test of nullity of the
parameters of the causal variables. Initially, agricultural GDP is regressed as a function of the
explanatory variables (agricultural GDP, rural population, active population, with two lags) and
using a Wald statistic, we test the nullity of the parameters of the lagged rural population and
labor force variables. In a second step, we resume the process while taking as explained variable
the rural population variable. The results of these various tests are presented in the following
table:
Table 4: Hypothesis test of (no) causality of the population push towards agricultural growth
Ho: Population growth does not cause agricultural growth in the sense of Granger
Country Stat. du test p. critique
Benin 28,92 0,000***
Burkina Faso 19,55 0,0001***
Ivory Coast 5,37 0,0244**
Guinea-Bissau 4,20 0,0203 **
Mali 25,02 0,000 ***
Niger 15,91 0,0002 ***
Senegal 9,12 0,0039 ***
Togo 19,39 0,0001 ***
Source: Our analysis results; Note that ***, **, *, respectively indicate significance at the 1%,
5%, 10% thresholds
According to the results of the test of non-causality of the population push towards agricultural
growth, we note that we can reject the hypothesis of non-causality at the threshold of 1% for
countries such as Benin, Burkina-Faso, Mali, Niger, Senegal and Togo and at the 5% threshold
for the Ivory Coast and Guinea-Bissau. Consequently, the population growth has influenced
agricultural growth in the countries of the study area over the period from 1960 to 2015. This
result indicates a strong influence of population growth on agricultural growth in the countries
of the WAEMU zone. This phenomenon is represented by the following graph, below.
This graph shows us the dynamics of the evolution of agricultural GDP and population in each
of the 8 WAEMU member countries. An analysis of this graph shows that agricultural GDP and
the population measurement variables (rural population and active population) evolve in
almost the same direction, but it is surprising to see that agricultural GDP evolves at a faster
rate than that of other population variables. This result is truly paradoxical because it fiercely
contradicts the Malthusian theory. But it should be remembered that in the context of this
article, the population is represented by the rural population and the active population and not
the population as a whole.