Download paper - African Development Bank

Accelerating Agro-Manufacturing to Feed Africa
By
John C. Anyanwu1
&
Mawuko Kponnou
Development Research Department
African Development Bank
Avenue Joseph Anoma
01 BP 1387
Abidjan 01, Cote d’Ivoire
Abstract
This paper documents stylized facts on agro-manufacturing, especially the food, beverages and
tobacco (FBT) manufacturing value added (MVA) in Africa and empirically analyzes its determinants
using data for a panel of African countries over the period 1990-2011. We also estimate for the sample
of Sub-Saharan Africa and North Africa during the same period. The analysis points to large
differences in sector shares both across countries at different levels of economic development. Using
two-stage least squares (2SLS) regression method, it finds that a large proportion of the cross-country
variation in FBT MVA can be accounted for by country characteristics, policy and institutional
variables. In particular, the paper finds that an inverted U-shaped relationship with real per capita
GDP. Key positive drivers for the entire continent include domestic investment rate, government
consumption expenditure, household consumption expenditure, social and political globalization,
dependence on oil, natural gas, coal and forest resources, arable land, and ICT
infrastructure/technology access. Major negative drivers are trade openness, domestic credit to the
private sector and population size. The policy implications and lessons of these results for increasing
FBT MVA and feeding Africa are discussed.
Keywords: Agro-Manufacturing, Food, Beverages, Tobacco, Manufacturing, Feed, Africa
JEL Classification: O14, O24, O40
1
Corresponding author: E-Mail: [email protected]
I.
Introduction
Agriculture can be used as a basis for manufacturing and hence contribute significantly to economic
transformation of African economies, just as it did earlier in many developed countries. In particular,
using agriculture as a basis for manufacturing, particularly by increasing agro-manufacturing and other
agribusiness, will not only feed the continent, reduce poverty, create jobs (especially for women and
the youth), generate foreign exchange, and speed up long-term structural change and the technology
and innovation needed for productivity growth, but also increase the demand (and prices) for farm
produce and hence the income of farmers. Stakeholders can make this happen by taking steps to
strengthen the links between agriculture and industrialization through agro-manufacturing. They can
also unlock the numerous benefits by implementing well-designed policies to overcome barriers that
prevent domestic players from emerging, reaching scale, and becoming globally competitive in agromanufacturing. Indeed, there is a perceived need to increase the rate at which the food processing
industry adds value to food products, in particular.
In addition, in recent years, much attention has been given to the questions of value-added activities
of the agro-manufacturing industry for the purposes of economic development. Another recent issue
relates to concerns on trade policies that had hitherto encouraged huge import of value-added food
and fiber products into African countries. Given the global importance of the agro-manufacturing
industrial sub-sector, special emphasis is given to Africa where the sector is the single largest
contributor to income and employment generation and plays a key role in dealing with the continent’s
challenge to achieve self-sufficiency in food production, create supply chains, improve agricultural
productivity and enhance the competitiveness of processed export products, reduce rural poverty and
foster economic advancement based on sound environmental management practices. As Otieno and
Mwangola (2006) argue, food industries have tactical and catalytic roles to play within agricultural-led
development strategies in Africa. This paper therefore explores the forces shaping value-added food,
beverage and tobacco industries (International Standard Industrial Classification (ISIC) codes 15 and
16) and their implications for feeding Africa.
This study is also important as it will help point the way towards the attainment of Sustainable
Development Goal (SDG) 9, which is to “build resilient infrastructure, promote inclusive and
sustainable industrialization and foster innovation”. This is more so since the attainment of this Goal
is a primary engine not only of feeding Africans, job creation, and economic growth but also of
investment flows, skills development, and technology transfer. In particular, Goal 9.2 aims at
promoting inclusive and sustainable industrialization and, by 2030, significantly raising industry’s share
of employment and gross domestic product, in line with national circumstances, and doubling its share
in least developed countries.
In developed economies, a sizeable body of literature has attempted to account for the broad set of
empirical regularities that characterize agro-manufacturing, emphasizing value addition, output and
employment both across countries and over time. This paper contributes to the literature in at least
two major dimensions. First, it provides important stylized facts on agro-manufacturing, especially the
food, beverages and tobacco component (a downstream agribusiness activity), in Africa. Second,
notwithstanding renewed theoretical and policy interest on this theme, the empirical evidence on the
proximate determinants of agro-manufacturing, particularly in Africa as a whole, remains scant. Our
paper adds to this literature by providing an empirical analysis of the policy and institutional drivers
of food, beverages and tobacco (FBT) manufacturing value added across a cross-section of 31 African
countries for which data are available.
The remainder of the paper is organized as follows. Section II presents a brief review of the literature,
while Section III reviews key stylized facts. Section IV presents the presents the model and data.
Section V presents and discusses the empirical results while Section VI concludes with policy
implications.
II.
Brief Review of the Literature
According to the FAO (1997), the agro-manufacturing or agro-processing industry is a subset of
manufacturing that processes or transforms raw materials and intermediate products derived from the
agricultural sector, including farming, forestry and fisheries. The International Standard Industrial
Classification (ISIC) also categorizes the following eleven divisions under the agro-processing
industry: food, beverages, paper and paper products, wood and wood products, textiles, wearing
apparel, furniture, tobacco, rubber products, footwear and leather and leather products.
For the purposes of this paper, the agro-manufacturing sector is limited in statistical terms by the
food, beverage and beverage manufacturing sub-sectors. The food system represents one of the most
significant components of the African economy. In particular, African countries’ food system provides
the population with a variety and widely available supply of food. It does so through a supply chain
of producers, manufacturers (processors), and distributors that provide food to consumers. Figure 1
illustrates the position and role of food manufacturers in the multilayered, dynamic, and multipurposed food supply chain.
The permeable borders of this dynamic food system connect it both to a global food system and to a
diverse, changing broader economy and society. One of the major characteristics of the agromanufacturing sector is its strong up- and down-stream linkages. Upstream, the agro-manufacturing
sector links to primary agriculture across a variety of farming models and products. Downstream, its
outputs are both intermediate products to which further value is added and final goods, which are
marketed for final consumption through wholesale and retail chains as well as a diverse array of fastfood franchises, restaurants, and pubs. Such link with agriculture makes agro-manufacturing critical
for feeding Africa, reducing hunger and mal-nourishment, poverty reduction, and job creation. Indeed,
according to the UNDP (2012), agro-food companies in Africa are critical in spurring rural
development, attracting foreign direct investment, providing enormous employment for women and
youth, providing market for financial institutions within the agribusiness setting, and contributing to
foreign exchange earnings. It also has tremendous role to play in reducing hunger and malnutrition,
ensuring year-round supply of food, and reduction of post-harvest losses. In addition, agromanufacturing industries play important roles in providing direct market access to producers, investing
in hard and soft infrastructure at rural and urban levels, transferring production and processing
technologies, processing and supplying food and food products to the population, and engaging in
direct export.
Below, we present a brief overview of the literature on the determinants of a key strand of agromanufacturing, namely food, beverages and tobacco, to motivate our analysis. Chenery (1960) finds
positive significant effect of income (per capita GNP in 1980 US dollars) on food and beverages as
well as tobacco manufacturing outputs but population (in millions) has no significant effect.
Figure 1: The Position and Role of Food Manufacturers in the Food Supply Chain
Food and
Beverage
service
Farm input
supply
Farm
production
Institutional
Buyers
First-Line
Handlers
Wholesale
and logistics
Manufacturers
Consumer
Retail Food
Stores
Food Banks
Source: Adapted from Institute of Medicine (IOM) and National Research Council (NRC) ((2015).
Tkalec and Vizek (2009) analyze the impact of macroeconomic policies on manufacturing production
in 22 manufacturing sectors of Croatia, using quarterly data from 1998:1Q to 2008:3Q. Their results
suggest that changes in fiscal conditions (especially government consumption expenditure), the real
effective exchange rate and personal consumption mostly affect low technological intensity industries
(food and tobacco; textiles and clothing; leather products; wood, wood products and cork; paper and
paper products, publishing, furniture and other manufacturing). Production in high technological
intensity industries (chemicals and chemical products; machinery and equipment; electrical and optical
products; transport vehicles, etc.) is, in general, elastic to changes in domestic investments, foreign
demand and fiscal policy. In particular, personal consumption expenditure is positively and
significantly correlated with output of food and beverages sector while real long-term interest rates
negatively and significantly correlate with it. Production of tobacco products is positively and
significantly correlated with personal consumption expenditure and real long-term interest rates but
government consumption expenditure is negatively and significantly correlated with it.
The European Commission (2009a, b) examines the key drivers of per capita value added in various
sectors for twenty-five EU countries. Those sectors are food, beverages, tobacco; textiles; clothing;
leather; wood and wood products; pulp and paper; publishing, printing, and reproduction; coke and
refined petroleum; chemicals; rubber and plastics; other non-metallic mineral production; basic metals;
fabricated metal products; machinery and equipment; office machinery and computers; electric
machinery; radio & television; medical instruments; motor vehicles and trailers; other transport
equipment; furniture and recycling; electricity, gas, and water supply; construction; and sale and repair
of motor vehicles. The findings indicate that exports and intermediate demand are the two most
important demand side manufacturing output drivers, while imports and government expenditure
have very little impact on growth in manufacturing. In New Member States fiscal deficits reduce
output growth in a number of industries. Other results show that real interest rates have a robust
negative correlation with manufacturing output growth. In particular, exports and intermediate
demand are positively and significantly correlated with food, beverages and tobacco manufacturing
value added while relative prices and short-term interest rates have significant negative correlation.
In studying the “Emerging Patterns of Manufacturing Structural Change”, Haraguchi and Rezonja
(2011) examine 18 manufacturing sectors (food and beverages; tobacco; paper; coke and refined
petroleum; fabricated metals; motor vehicles; furniture; textiles; wearing apparel; wood products;
printing and publishing; machinery and equipment; precision instruments; chemicals; rubber and
plastic; non-metallic minerals; basic metals; and electrical machinery and apparatus) for the period,
1963 to 2006. The results (both for small and large countries), when country-specific effects are
included, show universal effects of income in an inverted U-shaped form for all the sectors except for
the machinery and equipment sector that exhibit a U-shaped outcome. For the food and beverages
sector, in addition to the inverted U-shaped effect of income (real GDP per capita), natural resources
per capita is positively and significantly correlated with the sector for small countries but exhibits
negative significant relationship for large countries. And for tobacco products, in addition to the
inverted U-shaped income effect, both population density and natural resources per capita are
positively and significantly correlated with it for both small and large countries though these
(population density and natural resources per capita) have no effect in small countries.
Ahmed (2012), in a study of the Malaysia (1971-2000), finds that the factors affecting output growth
in Malaysian food industries are individual contributions of capital, labor, and materials, as well as the
combined contributions of the quality of these inputs expressed as total factor productivity growth.
His findings indicate the low quality of inputs into the food industries, which are input-driven rather
than total factor productivity growth-driven.
The above review shows that the few studies that had been carried out on the key drivers of agromanufacturing value added had been outside Africa. In contrast to these papers, we examine value
added in food, beverages and tobacco (FBT) manufacturing across African countries (all-Africa, SubSaharan Africa and North Africa), and account for its key determinants using a broader set of
fundamental as well as policy and institutional drivers. From a policy perspective, the results of this
paper will serve as a useful platform to formulate series of new agenda for agro-manufacturing,
especially food-manufacturing, policies in African countries.
III.
Stylized Facts
This section presents some stylized facts about manufacturing value added (MVA), with emphasis on
food, beverages and tobacco manufacturing value added as a share of total manufacturing value added.
The structure of manufacturing value added between 1990 and 2011 (during which data was available)
in Africa is shown in Figure 2. Among the various sectors, the food, beverages and tobacco a sector
producing essential consumer foods has a strong presence in the continent than in industrialized
countries. In industrialized countries, food manufacturing contributes well below 20 percent of the
total value added of the manufacturing industry, however, its share averages about 38 percent in Africa
between 1990 and 2011 – the highest single component of manufacturing value added.
Figure 2: The Structure of Total Manufacturing Value Added, 1990-2011
45.00
40.00
40.32
37.64
35.00
30.00
25.00
20.00
12.25
15.00
9.06
10.00
3.87
5.00
0.00
Food, beverages Chemicals MVA (%
and tobacco MVA
of Total MVA)
(% Total MVA)
Machinery and
transport
equipment MVA
(% of Total MVA)
Textiles and
clothing MVA (% of
Total MVA)
Other MVA (% of
Total MVA)
Source: Authors, using data from World Bank’s Online Database.
As Figure 3 demonstrates, FBT manufacturing as percentage of total manufacturing value added has
been consistently higher in Sub-Saharan Africa (SSA) than in North Africa. It averaged 42.5 percent
in SSA as against only 25.8 percent in North Africa between 1990 and 2011. However, both (and
hence Africa’s) have assumed a downward trend recently.
However, these sub-regional averages mask the country differences. For example, as Figure 4 shows,
Burundi has the highest average FBT MVA as percentage of total MVA at over 85 percent, followed
by Congo Republic at over 75 percent. Rwanda follows at about 75 percent.
Figure 5 shows a scatterplot of African countries on average FBT MVA and average real GDP per
capita. Countries that are in the southeast quadrant indicate those that they have experienced a low
real per capita GDP and relatively very high levels of FBT MVA. They include Burundi, Central
African Republic, Sierra Leone, Rwanda, Sudan, Malawi, and Uganda, among others. African countries
in the northeast quadrant have had relatively high real GDP per capita and relatively high FBT MVA.
This is particularly so for Gabon, Swaziland, and Congo Republic. Countries in the north-west
quadrant experienced high level of real per capita GDP but relatively low FBT MVA. It is not
surprising to find countries like Mauritius, South Africa, Botswana, Tunisia, Egypt, and Morocco, for
example, in this quadrant. This shows that countries with relatively higher incomes engage in very low
food manufacturing, underlining the commonly held view that the significance of the agro-industrial
sector decreases as countries become relatively more industrialized and diversified.
In the south-west quadrant, we find countries that have relatively low FBT MVA in spite of their low
real per capita income. They include Nigeria, Ghana, Kenya, Madagascar, and Niger. Another
interesting observation from Figure 5 is the inverted U-shaped relationship between economic
development and FBT MVA in Africa.
Figure 3: Food, Beverages and Tobacco MVA (% of Total MVA)
60
50
Percent
40
30
20
10
0
FBT MVA_Africa
FBT MVA_SSA
FBT MVA_North Africa
Source: Authors, using data from World Bank’s Online Database.
Figure 4: Africa - Top Ten Countries By Food, Beverages & Tobacco MVA, 1990-2011
Burundi
Congo, Rep.
Rwanda
Swaziland
Sierra Leone
Central African Republic
Sudan
Malawi
Uganda
Ethiopia
0
10
20
30
40
Source: Authors, using data from World Bank’s Online Database.
50
60
70
80
90
10000 15000 20000 25000
Figure 5: Africa: Scatter Plot of FBT MVA as Percentage of Total MVA, 1990-2011
Gabon
Mauritius
Algeria
South AfricaBotswana
Burkina
Faso
5000
Tunisia
Egypt, Arab Rep.
0
Madagascar
0
Swaziland
Morocco
Nigeria
Cote
d'Ivoire
Cameroon
Ghana
Zambia
Kenya
Senegal
Tanzania
Eritrea
Gambia,
The
Uganda
Niger
Ethiopia
Malawi
Congo, Rep.
Sudan
SierraAfrican
Leone Republic
Rwanda
Central
Burundi
20
40
60
Food,Beverages & Tobacco MVA as % of Total MVA
(mean) nygdppcapppkd
80
Fitted values
Source: Authors, using data from World Bank’s Online Database.
IV.
The Model and Data
4.1 The Model and Estimation Technique
Based on the above literature review and following the frameworks posited by Kochhar et al. (2006),
Jaumotte and Spatafora (2007), Nickell et al. (2008) and Dabla-Norris et al. (2013), the relationship
that we want to estimate can be written as:
FBTMVAit   0  1 log( rgdppcit )   2 log( rgdppc 2it )  3 ( X it )  i   i   it
(i  1,...., N ; t  1,....., T ),...............(1)
where FBTMVAit is the measure of food, beverages and tobacco MVA as percentage of total MVA
in country i at time t; 0 is the constant term; 1 is the elasticity of FBT MVA with respect to real per
capita GDP (international PPP) in 2005, rgdppc; 2 is the FBT MVA elasticity with respect to
quadratic real per capita GDP; X is the control variables, including government consumption
expenditure (as % of GDP), household consumption expenditure (as % of GDP), trade openness,
FDI inflows (as % of GDP), domestic credit to the private sector (as % of GDP), natural resource
rents as percentage of GDP (oil, mining, natural gas, coal, and forest), and arable land (as % of total
land area). Other control variables are total population (in log), age dependency (old), age dependency
(young), information and communications technology (ICT) accessibility (proxied by mobile phone
subscriptions (percent), secondary school enrolment (education), social globalization index, political
globalization index, and institutionalized democracy (polity2). In addition, λi and δi denote sub-
regional/country and year fixed effects, respectively, while εit is an error term capturing all other
omitted factors, with E(εit) = 0 for all i and t.
The dependent variable is the food, beverages and tobacco MVA as percentage of total MVA and the
estimations are carried out for Africa as a whole and for Sub-Saharan Africa and North Africa,
separately. Based on the theoretical and empirical literature, we use key drivers, including a range of
supply and demand factors as controls. Given the importance of income effects identified in the
theoretical literature (see (Chenery, 1960; Kuznets, 1971; Echevarria, 1997, 2000; Kongsamut et al.,
2001; Haraguchi and Rezonja, 2011), the estimates control for output per capita (log of GDP per
capita in constant PPP U.S. dollars) and its square (to account for non-linearities). A country’s stage
of development usually has the strongest influence manufacturing development (UNIDO, 2016).
Indeed, UNIDO (2016) states that as countries increase their GDP per capita, the share of low-tech
industries at low incomes rapidly declines while the shares of medium-tech and high-tech groups
increase. However, theory predicts that there is an inverted U-curve relationship between economic
development and share of manufacturing in value added. That is, the share of manufacturing in value
added tends to increase when developing countries start growing at low levels of per capita income.
It peaks at intermediate per capita incomes and later declines as services become more important at
high per capita incomes.
Included also is domestic investment ratio. The European Commission (2009b) notes that the average
investment ratio, used a proxy for the capital intensity, is expected to be positive as it reflects primarily
the neglected capital costs. According to Tkalec and Vizek (2009), high technological intensity
industries strongly react to changes in investments.
As Tkalec and Vizek (2009) note, government spending decreases output in industries characterized
by low, medium-low and high level of technological intensity. The opposite is true in medium-high
technological intensity industries, where an increase in public spending boosts production. Overall,
government consumption expenditure crowding out production in manufacturing industries is more
prevalent in industries requiring low and medium-low technological intensity like production of
tobacco. That is similar to the finding by the European Commission (2009b) that an expansionary
fiscal policy (government expenditure) in ten New Member States reduces manufacturing output.
Low technological intensity industries are somewhat elastic to changes in personal consumption
(Tkalec and Vizek, 2009). Thus, the impact of personal consumption on manufacturing output is
almost entirely limited to industries characterized by low technological efficiency. This may not be
surprising since these industries make products for final consumption.
Following the widely held view that globalization can facilitate technology transfer and contribute to
efficiencies in production, we include different globalization indicators. Two principal economic
globalization indicators included are international trade openness (measured as the ratio of exports
plus imports to GDP)( Matsuyama, 2009) and FDI inflows (as percent of GDP). FDI can provide
access to technology, to brand names, to global markets and has the potential to provide spillovers to
the domestic economy (UNIDO, 2016). FDI may affect FBT MVA through various mechanisms:
boosting productivity in the long run; filling expectations of demand increase; strengthening
competition and weakening oligopoly/monopoly elements; diffusing knowledge of new production
processes; stimulating the entry of firms in other sectors (horizontal linkages); and creating the right
conditions to enhance structural change. Also included are KOF’s indices of social globalization and
political globalization.
The degree of financial sector development is proxied by the ratio of domestic credit to the private
sector to GDP, which is posited to enable investment in higher productivity activities, greater
diversification, and risk sharing, and hence facilitate resource allocation across the economy (Levine,
2005).
Our estimates also include natural resources endowments by including the share of oil, mining, natural
gas, coal, and forest rents in GDP to account for the fact that a large fraction of economic activity in
resource-rich economies in Africa is subsumed by the rents from natural resources extraction. It is
posited that the endowment of abundant natural resources normally works against manufacturing
development, holding other conditions constant (Haraguchi and Rezonja 2011; UNIDO 2012).
UNIDO (2016) shows that high natural resource endowments do not have a positive effect on a single
industry, but they have particularly strong negative effects on electrical machinery and apparatus,
motor vehicles (for large countries) and chemicals, which are key in deepening and sustaining
industrialization from the upper middle-income stage. This is largely because exports of resource
commodities often lead to currency appreciation, making tradable manufacturing products less
competitive.
We include factor endowments, such as arable land (as percent of total land area), population, and
domestic investment (as percent of GDP) as a proxy of capital stock. Chenery and Taylor (1968) show
that a country’s population size tends to have overarching influence on economic structural change.
UNIDO (2016) shows that a larger population is generally conducive to manufacturing development
though there are differences in structural change within manufacturing between large and small
countries. Large countries, at higher incomes, tend to have a divergent pattern of thriving and other
industries, while in small countries, growth in most manufacturing industries slows at higher incomes.
By industry, small countries are likely to develop food and beverages much earlier than large countries,
but that industry’s growth in those countries is not as sustainable as in large countries: growth in food
and beverages starts slowing.
Our estimates include age dependency ratios (i.e., the non-working old and young populations as
fractions of the labor force) since they can affect labor supply, savings and consumption behavior.
The accessibility to ICT technology and infrastructure or service can influence value added in FBT
manufacturing by either facilitating or obstructing the reallocation of resources. To capture this, we
include telecommunications network as proxied by mobile phone subscriptions (as percent). An
increase in access to such ICT in the FBT sector can contribute to increase in FBT MVA by
eliminating relative price distortions and facilitating the reallocation of labor and other inputs, thereby
raising sector productivity.
The effect of educational attainment (see Lee and Wolpin, 2006) on FBT MVA is captured by
including the share of secondary education enrollment. Institutionalized democracy is represented by
polity2 and it is expected to be positively correlated with FBT MVA.
One possible problem with Equation (1) is that it assumes that all of the right-hand side variables in
the model — including per capita GDP — are exogenous to gender equality in primary school
enrolment. However, it is possible that real per capita GDP may be endogenous to FBT MVA. Reverse
causality may be taking place: real per capita GDP may be increasing FBT MVA, but FBT MVA may
also be affecting the level of real per capita GDP. Without accounting for this reverse causality, the
estimated coefficients in Table 2 may be biased. One way of accounting for possible endogenous
regressors is to pursue an instrumental variables approach. Therefore, to deal with this problem, we
also estimate the equation, instrumentalizing real per capita GDP variable with its four lagged levels,
using a two-step (IV) estimation method. Thus, our estimations (for Africa as a whole and for SSA
and North Africa) are done with IV-2SLS, including sub-regional/country and time (year) fixed
effects.
4.2
The Data
Data (1990 to 2011) for the variables in equation (1) are largely drawn from the World Bank’s WDI
Online database, except institutional democracy (polity2) from the PolityIV Project Online (2015) (see
also Marshall et al, 2016), and KOF’s indices of social globalization (comprising personal contacts,
information flows, and cultural proximity) and political globalization (comprising embassies in
country, membership in international organizations, participation in UN Security Council Missions,
and international treaties) developed by Dreher (2006, 2008). The descriptive statistics are presented
in Table 1. It reports the sample mean and standard deviation of the variables used in the estimations.
The data covers 31 African countries for which data is available. Appendix I presents the list of
countries included in the sample.
V.
Empirical Results
Table 2 shows the results when Equation (1) is estimated using the IV-2SLS estimation method with
sub-regional/country and year fixed effects. The consistency of the IV-2SLS estimators depends on
whether the instruments are valid in the gender equality in education regression. We examine this issue
by considering the tests of over-identifying restrictions. The no rejection of the null hypothesis implies
that instrumental variables are not correlated with the residual and are satisfying the orthogonality
conditions required. The IV-2SLS results pass the relevant tests. For example, the Sargan test of
overidentifying restriction fails to reject that the instruments are valid, i.e., not correlated with the
error term at conventional significance levels in the reported regression (p-value of 0.2923 in Africa
column, 0.2197 in the Sub-Saharan Africa column, and 0.1704 in the North Africa column).
In our results, the coefficient associated with the level of real GDP per capita is found to be positive
and statistically significant in both the overall Africa sample and in the Sub-Saharan and North African
samples. To test the hypothesis that real GDP per capita has a non-monotonic relationship with FBT
MVA, the squared real GDP per capita is included as an explanatory variable. The quadratic term is
negative in sign and significant in all three sample groups.
Thus, the results for Africa, SSA and North Africa provide evidence of a humped-shaped relationship
between real GDP per capita and food, beverages and tobacco MVA. This result suggests that
although higher levels of real GDP per capita are positively associated with FBT MVA in Africa as a
whole and in SSA and North Africa, the effect is not constant. Rather, for levels of real GDP per
capita above a certain point (US$4570), higher levels of real GDP per capita act to FBT MVA in Africa
as a whole, holding other factors constant. This relationship suggests that the marginal effect of real
GDP per capita exhibits decreasing returns for FBT MVA in African countries. Thus, this finding
supports Chenery (1960), Kuznets (1971), and Haraguchi and Rezonja (2011) assertion of a U-shaped
relationship between economic development and food manufacturing output.
Table 1: Descriptive Statistics of Main Regression Variables
Variable
Observations
Mean
Median
Standard Deviation
37.64
35.3
18.81
1106
7.82
7.65
1.02
1070
21.52
19.5
17.17
Government consumption
expenditure (%GDP)
Household consumption
expenditure (%GDP)
Trade openness
1043
16.05
14.7
7.67
1041
73.55
74.6
21.98
1091
76.55
64.1
49.18
FDI inflows (%GDP)
1100
4.02
1.7
10.14
Domestic credit to private sector
(%GDP)
Oil rent (%GDP)
1077
19.57
13
21.46
1103
6.12
0
14.74
Mining rent (%GDP)
1134
1.19
0
3.45
Natural gas rent (%GDP)
1120
0.54
0
2.10
Coal rent (%GDP)
1124
0.04
0
0.34
Forest rent (%GDP)
1109
6.71
0
8.46
Arable land (%of total land area)
1163
12.21
9.2
12.08
Log of population
1188
15.63
15.95
1.56
Age dependency ratio - Old
1188
6.38
5.9
1.58
Age dependency ratio - Young
1188
80.16
83.1
15.45
Mobile cellphone subscriptions
(%)
Secondary school enrolment ratio
1160
14.24
0.8
26.34
683
38.61
32.8
25.97
Social Globalization Index
1163
25.85
23.4
11.24
Political Globalization Index
1163
53.06
51.91
18.97
Institutional Democracy (Polity2)
1120
-0.001
-1.0
5.50
Food, beverages & tobacco (FBT)
MVA (% of Total MVA)
Log Real GDP per capita
Domestic investment (%GDP)
349
Source: Authors’ calculations, using data estimation data.
Domestic investment rate has positive and highly statistical significant effect on FBT MVA in allAfrica data and North Africa samples. As seen in Table 2, domestic investment rate is positive in sign
and significant at the 1 percent level in both cases. Our estimates suggest that, on average, a one
percent increase in the share of domestic investment rate will lead to about 0.68 percent and 0.74
percent increase in FBT MVA, respectively in all African and North African countries. However, its
effect is insignificant in the Sub-Saharan African estimation.
Government expenditure as a percent of GDP, as a government policy variable, increases FBT MVA
in all-Africa and SSA but reduces same in North Africa Our estimates suggest, for example, that a
one percentage point increase in government consumption expenditure is associated with increase in
FBT MVA by 1.08 percentage points in the whole of Africa – and by 0.83 in SSA countries. This
finding confirms that of Tkalec and Vizek (2009) for Croatia.
Table 2: IV-2SLS Estimates of the Determinants of FBT Manufacturing Value Added (% of MVA)
Variable
Africa
Sub-Saharan Africa
North Africa
Log Real GDP per capita
164.467 (3.87***)
117.145 (1.87*)
803.872 (3.56***)
Log Real GDP per capita squared
-10.722 (-4.13***)
-7.690 (-2.01**)
-50.687 (-3.76***)
Domestic investment (%GDP)
0.677 (3.45***)
0.020 (0.06)
0.737 (2.72***)
Government consumption expenditure
1.083 (3.29***)
0.834 (2.09**)
-1.658 (-1.73*)
(%GDP)
Household consumption expenditure
0.863 (5.61***)
0.703 (3.48***)
2.763 (10.49***)
(%GDP)
Trade openness
-0.507 (-5.56***)
-0.232 (-1.78*)
-0.834 (-8.89***)
FDI inflows (%GDP)
0.140 (0.45)
0.271 (0.74)
0.107 (0.36)
Domestic credit to private sector
-0.153 (-2.30**)
0.155 (-1.31)
0.344 (4.43***)
(%GDP)
Oil rent (%GDP)
2.117 (10.62***)
1.842 (6.90***)
5.244 (6.21***)
Mining rent (%GDP)
0.815 (1.57)
-0.434 (-0.55)
2.114 (3.46***)
Natural gas rent (%GDP)
1.683 (3.12***)
-15.872 (-1.20)
-1.423 (-2.69***)
Coal rent (%GDP)
2.325 (2.10**)
2.276 (1.69*)
Forest rent (%GDP)
0.872 (2.61**)
0.662 (1.68*)
-21.135 (-2.82***)
Arable land (%of total land area)
0.920 (6.89***)
1.049 (6.45***)
4.273 (3.57***)
Log of population
-11.548 (-4.74***)
-7.211 (-1.61)
0.344 (4.43***)
Age dependency ratio - Old
-1.499 (-0.96)
-7.969 (-2.81***)
Age dependency ratio - Young
-0.002 (-0.01)
-0.036 (-0.13)
Mobile phone subscriptions (%)
0.159 (2.44**)
0.344 (2.79***)
-0.040 (-0.66)
Secondary school enrolment ratio
0.029 (0.34)
-0.325 (-1.85*)
0.525 (4.75***)
Social Globalization Index
0.665 (4.45***)
1.300 (4.31***)
-2.662 (-6.29***)
Political Globalization Index
0.179 (2.35**)
0.205 (1.97**)
Institutional Democracy (Polity2)
-0.081 (-0.31)
-0.410 (-1.33)
3.676 (2.44**)
Constant
-494.419 (-3.31***)
-339.149 (-1.65*)
-3492.69 (-1.97**)
Time Dummies
Yes
Yes
Yes
Sub-region dummies
Yes
Yes
No
Country Dummies
No
No
Yes
R-squared
0.8802
0.8959
0.9912
Wald chi2
1140.66
991.34
4835.15
Prob > F
0.0000
0.0000
0.0000
N
155
115
43
Sargan test
2.46029 (p=0.2923)
1.5062 (p=0.2197)
1.8794 (p=0.1704)
Basmann test
1.7903 (p=0.4085)
0.96880 (p=0.3250)
0.22852 (p=0.6326)
Durbin test
4.53387 (p=0.1036)
1.61331 (p=0.4463)
0.683974 (0.7104)
Wu-Hausman test
1.67233 (p=0.1925)
0.51222 (p=0.6013)
0.03233 (p=0.9684)
Note: t-values are in parentheses; ***= 1% significant level; **=5% significant level; *=10% significant level.
Source: Authors' Estimations.
Household consumption expenditure as a percent of GDP is also positively and significantly
associated with FBT MVA in both all-Africa, SSA and North Africa samples – and more so for the
later. Our estimates suggest, for example, that a one percentage point increase in household
consumption expenditure is associated with increase in FBT MVA by 0.8s percentage points in the
whole of Africa. This finding also confirms that of Tkalec and Vizek (2009) for Croatia.
We investigated aspects of globalization that have been implicated as key drivers of FBT MVA. Our
results indicate that trade openness significantly reduces FBT MVA in all three sample groups (unlike
in Haraguchi and Rezonja, 2011). It does so at the 1 percent significant level in the all-Africa and
North Africa estimates. As UNIDO (2013) states, an increase of trade openness is a growth
opportunity for a country only if local resources can be deployed in adequate quantities to produce
goods for the external market. Also, domestic production capabilities have to be already in place in
order to respond to international competition, improve technology and exploit trade opportunities
from increased liberalization. We find that inward foreign direct investment (FDI) is not significantly
related to FBT MVA in any of the group estimations.
One of the innovative aspects of this paper is the inclusion of other aspects of globalization, namely
social and political globalization. Our results show that social and political globalization are positively
and significantly associated with FBT MVA in the all-Africa and SSA estimates. This shows that
international spread of personal contacts, information flows, and cultural proximity as well as African
countries’ establishment of embassies in foreign countries, membership in international organizations,
participation in UN Security Council Missions, and international treaties help to generate influences
that promote increased FBT MVA in the continent. However, social globalization is negatively and
significantly associated with FBT MVA in North African countries.
The credit variable has a negative and statistically significant effect on FBT MVA in the all-Africa
estimation but a positive and significant effect in the North Africa estimation. However, it has an
insignificant positive effect in the Sub‐Saharan Africa estimation. Haraguchi and Rezonja (2011) finds
that the availability of credit is only positive and significant for the lower end of the manufacturing
distribution. Our all-Africa result is not surprising given the problem of high interest rates in many
African countries. In addition, increases in financial depth have taken place without the requisite
development of lending expertise, mechanisms for monitoring, and supervisory and regulatory skills.
With respect to the role of natural resources, our results indicate that oil dependence has significant
positive association with FBT MVA in both all-Africa, SSA and North Africa estimations. Mineral
resource dependence has significant positive association with North Africa’s FBT MVA (contrary to
Rodrik and McMillan, 2011); natural gas dependence has significant positive association in the AllAfrica case but significant negative association with North Africa’s FBT MVA; dependence on coal
and forest resources have significant positive association with FBT MVA in all-Africa and SSA results.
Dependence on forest resources has significant negative association with FBT MVA in North African
countries.
The proportion of land that is arable is positively and significantly associated with a higher FBT MVA
in Africa as whole, in SSA and North Africa in line with Haraguchi and Rezonja (2011). Population is
negatively and significantly related with FBT MVA in the all-Africa estimation but has a strong positive
relationship with FBT MVA in North African countries. Age dependency ratio (old) is strongly
negatively related with the FBT MVA in Sub-Saharan Africa, conforming to the findings of Haraguchi
and Rezonja (2011).
Our results show that the coefficient of the ICT infrastructure/technology variable is positive and
highly statistically significant in the all-Africa and SSA estimations. In this case, we can conclude that,
except in North Africa, increases in access to ICT infrastructure/technology tend to lead to
improvements in FBT MVA. However, our ICT infrastructure/technology proxy is negatively
statistically insignificant in North Africa.
Secondary education enrollment is positive and statistically significantly correlated with FBT MVA in
North Africa but the reverse is true for SSA. Institutionalized democracy is positive in sign and
significant at the 5 percent level in North Africa. Thus, holding other variables constant, more
democratic countries tend to experience greater levels of FBT MVA in North Africa.
VI.
Conclusion and Policy Recommendations
Our empirical results can be summarized as follows:
(a) The coefficient associated with the level of real GDP per capita is positive and statistically significant
in both the overall Africa sample and in the Sub-Saharan and North African samples. Bt the quadratic
term of real GDP per capita is negative in sign and significant in these estimates. These provide
evidence of a hump-shaped relationship between real GDP per capita and FBT MVA in Africa;
(b) Domestic investment rate has positive and highly statistical significant effect on FBT MVA in
all-Africa data and North Africa samples;
(c) Government expenditure as a percent of GDP, increases FBT MVA in all-Africa and SSA but
reduces same in North Africa;
(d) Household consumption expenditure as a percent of GDP is also positively and significantly
associated with FBT MVA in both all-Africa, SSA and North Africa samples;
(e) Trade openness significantly reduces FBT MVA in all three sample groups. Social and political
globalization are positively and significantly associated with FBT MVA in the all-Africa and
SSA estimates but social globalization is negatively and significantly associated with FBT MVA
in North African countries;
(f) The credit variable has a negative and statistically significant effect on FBT MVA in the allAfrica estimation but a positive and significant effect in the North Africa estimation;
(g) Oil dependence has significant positive association with FBT MVA in both all-Africa, SSA
and North Africa estimations. Mineral resource dependence has significant association with
North Africa’s FBT MVA; natural gas dependence has significant positive association in the
All-Africa case but has significant negative association with North Africa’s FBT MVA;
dependence on coal and forest resources have significant positive association with FBT MVA
in all-Africa and SSA results; and dependence on forest resources has significant negative
association with FBT MVA in North African countries;
(h) The proportion of land that is arable is positively and significantly associated with a higher
FBT MVA in Africa as whole, in SSA and North Africa;
(i) Population is negatively and significantly related with FBT MVA in the all-Africa estimation
but has a strong positive relationship with FBT MVA in North African countries;
(j) Age dependency ratio (old) is strongly negatively related with the FBT MVA in Sub-Saharan
Africa;
(k) ICT infrastructure/technology variable is positive and highly statistically significant in the allAfrica and SSA estimations;
(l) Secondary education enrollment is positive and statistically significantly correlated with FBT
MVA in North Africa but the reverse is true for SSA; and
(m) Institutionalized democracy is positive in sign and significant at the 5 percent level in North
Africa.
What are the implications of these results for African countries? First, our results confirm that
prosperity (higher economic development of up to US$4570 per capita) promotes FBT MVA in
African countries. Therefore, African countries must take measures to increase their national incomes.
To increase per capita income, African countries must deepen macroeconomic and structural reforms
to increase their competitiveness, create increasing and more quality jobs and hence increase
participation in economic activity, dismantle existing structural bottlenecks to private and public
investment, and scale-up investments in hard and soft infrastructure to enhance local production and
regional integration. Others are structurally transform the economy for increased trade
competitiveness in knowledge-intensive manufacturing, and increase productivity, especially in
agriculture, through creating incentives and opportunities for the private sector and increasing
government support to small farm holders in terms of finance, formalization of land ownership, and
technical advice.
Second, given our finding that domestic investment rate increases FBT MVA in most of Africa,
achieving higher levels of investment as its effectiveness must remain an active goal of governments
in Africa. A key challenge, for African countries, therefore, is to mobilize increased resources for high
domestic investment. Successful promotion of investment will require actions and measures at the
national and regional level as indicated earlier. Further efforts should also be made to improve the
efficiency and effectiveness of public institutions, if these are to serve as genuine partners with the
private sector. Sustainable domestic investment needs increased human capital investment to enhance
the health and welfare of populations and generate the skills required in a competitive global
environment.
Third, given our finding that government consumption expenditure increase FBT MVA in all of
Africa, , achieving government expenditure effectiveness must remain an active goal of governments
in Africa. Adoption of high level best practice principles to inform the development of these processes
will help African governments achieve this. Those broad principles should include the following key
elements: a nationally coordinated approach to the development of significant strategic projects and
programs; and the promotion of competitive markets. Others relate to decision-making based on
rigorous cost-benefit analysis to ensure the highest economic and social benefits to the nation over
the long term; a commitment to transparency at all stages of the decision-making and project
implementation processes; and a public sector financial management regime with clear accountabilities
and responsibilities. At the same time, efforts to reform the fiscal system for consolidation by both
the executive and legislative arms of government are imperative to reduce government consumption
expenditure to avoid wastes, corruption and crowding out resources for public sector investment and
gender equality. In addition, public spending on education (as well as on health and other human
capacity), when targeted at women, especially the poor, can produce a quadruple dividend, increasing
FBT MVA and feeding Africa in the short run and increasing the chances for women and the youth
as well as household in general to access formal jobs and thus break free from poverty trap.
Fourth, to make globalization work for increased FBT MVA, local resources need to be deployed in
adequate quantities to produce goods for the external market. In addition, domestic production
capabilities have to be put into place in order to exploit trade opportunities, increase response to
international competition, and improve technology.
Fifth, for domestic credit to work for the FBT MVA and feed Africa, lending rates reduction is
imperative while developing the requisite lending expertise, mechanisms for monitoring, and
supervisory and regulatory skills of operators of the African financial system.
Sixth, for long-term FBT MVA development and feeding Africa, African countries with abundant
natural resources need prudent institutions to manage revenues from resource exports so as to avoid
undue currency appreciation and underinvestment in physical and human capital. Indeed, efficient
management of natural resources in Africa requires actions throughout the value chain. In particular,
a new natural resources management framework is needed for better governance, sectoral linkages,
economic growth and human, capacity and infrastructure development – with strong parliamentary
legislation, oversight, and representation throughout the resources value chain. Given that oil, gas and
mineral resources are non-renewable resources, it is vital to negotiate more beneficial and transparent
contracts with oil/mining Multinational Corporations operating in Africa, and ensure that these
companies do not evade taxes. For greater returns to African countries in terms of royalties/rents, for
example, the governments should engage in auctions for oil/mining rights. In this regard, international
financial institutions like the African Development Bank have a critical role to play in helping these
countries acquire the much-needed capacity not only to negotiate beneficial contracts but also for
effective management of natural resource revenues. Other measures to promote efficient and effective
allocation of public expenditure include promoting high levels of transparency, ensuring that the
political system has a centralized system of financial authority and control, and the legislation of a
‘fiscal constitution’ that imposes ceilings (and perhaps also floors) on public spending from resource
revenues.
Seventh, there is need to complement formal education with technical and vocational education and
training (TVET). TVET builds on formal education to deliver specialized technical training and calls
for a skilled workforce capable of operating state-of-the-art technologies. TVET programs have to
overcome three main challenges, including the enhancement of its public reputation, proper and
effective coordination among various agencies, and effective monitoring, evaluation and feedback
mechanism. But both formal education and TVET call for explicit incentives such as vocational or
engineering scholarships and demand-driven courses to train workers in the technical standards in the
FBT manufacturing sector. In addition, skill policies have to be aligned with Africa’s broader
socioeconomic development agenda. This requires strong coordination between stakeholders engaged
in policy-making, both public and private sectors.
Lastly, the promotion of effective democracy will help in the design of policies friendly to FBT MVA
development. This requires political will, commitment, good governance (including the control of
corruption, transparency and accountability, the rule of law, government effectiveness, and political
stability), inclusive development, collaborative spirit to formulate and faithfully implement the
requisite policies, strategies, plans and collective action as well as the institutional changes needed for
increased gender equality in education. Other critical measures to promote democracy in Africa
include promoting and maintaining effective rule of law, deepening macroeconomic and structural
reforms and increasing investments to raise national income, and implementing greater economic and
political inclusion, especially in North Africa. Democracy will thrive and will be sustained and stable
when there is the willingness to lose (contestation) and when there are capacities to challenge and
enforce the rules of the game. Contestation means that parties are able to win but are willing to lose.
In other words, opposition parties have to be able to compete effectively with incumbents, with the
credible potential to hold incumbents accountable while voters and parties must be willing to lose
elections. Also, laws must be effectively enforced. This means that a sturdy, thriving, durable and
stable democracy requires a government with the capacity to enforce both the rules of the game and
the policies produced through those rules against violation or nullification either by abusive agents of
the government itself or by private actors, whether common criminals, would-be warlords or the
military.
References
Ahmed, E. M. (2012), “Malaysia’s Food Manufacturing Industries Productivity Determinants”, Modern
Economy, 3, 444-453.
Chenery, H. B. (1960), “Patterns of Industrial Growth”, The American Economic Review, 50(4), 624-654.
Chenery, H. B. and Taylor, L. (1968), “Development Patterns: Among Countries and over Time”, The
Review of Economics and Statistics, 50(4), pp. 391–416.
Dabla-Norris, E., Thomas, A., Garcia-Verdu, R. and Chen, Y. (2013), Benchmarking Structural
Transformation Across the World, IMF Working Paper WP/13/176.
Dreher, A. (2006), “Does Globalization Affect Growth? Evidence from a new Index of
Globalization”, Applied Economics, 38, 10: 1091-1110.
Dreher, A., Gaston, N. and Pim Martens, P. (2008), Measuring Globalization - Gauging its Consequence,
Springer, New York.
Duarte, M., and Restuccia, D. (2010), “The Role of the Structural Transformation in Aggregate
Productivity,” Quarterly Journal of Economics, Vol. 125, No. 1, February, pp.129-173.
Echevarria, C. (1997), “Changes in Sectoral Composition Associated with Economic Growth,”
International Economic Review, Vol. 38, No. 2, Ma, pp. 431-452.
Echevarria, C. (2000), “Nonhomothetic Preferences and Growth,” Journal of International Trade and
Economic Development, Vol. 9, No. 2, June, pp. 151-172.
European Commission (2009a), European Competitiveness Report 2008, European Commission,
Luxembourg.
European Commission (2009b), Sectoral Growth Drivers and Competitiveness in the European Union,
European Commission, Luxembourg.
Food and Agriculture Organization (FAO), (1997), The State of Food and Agriculture, FAO, Rome.
Available at:
http://www.fao.org/docrep/w5800e/w5800e00.htm.
Haraguchi, N. and Rezonja, G. (2011), Emerging Patterns of Manufacturing Structural Change,
Development Policy and Strategic Research Branch Working Paper 04/2010, UNIDO, Vienna.
Institute of Medicine (IOM) and National Research Council (NRC) (2015), A framework for assessing
effects of the food system, The National Academies Press, Washington, DC.
Jaumotte, F., and Spatafora, N. (2007), Asia Rising: A Sectoral Perspective, IMF Working Paper
WP/07/130, International Monetary Fund, Washington DC.
Kochhar, K., Kumar, U., Rajan, R., Subramanian, A. and Tokatlidis, I. (2006), India’s Pattern of
Development: What Happened, What Follows? IMF Working Paper WP/02/22, International
Monetary Fund, Washington DC.
Kongsamut, P., Rebelo, S. and Xie, D. (2001), “Beyond Balanced Growth,” Review of Economic Studies,
Vol. 68, October, pp. 869-82
Kuznets, S (1971), Economic Growth of Nations: Total Output and Production Structure, Harvard University
Press, Cambridge, Mass.
Lee, D. and Wolpin, K. I. (2006), “Intersectoral Labor Mobility and the Growth of the Service Sector,”
Econometrica, Vol. 74, Issue 1, pp. 1-46.
Marshall, M. G,. Ted Robert Gurr, T. R. and Jaggers, K. (2016), Polity IV Project: Political Regime
Characteristics and Transitions, 1800-2015, Center for Systemic Peace, Vienna.
Matsuyama, K. (2008), “Structural Change,” in The New Palgrave Dictionary of Economics, Second
Edition, Durlauf, S. N. and Blume, L. E.(Eds.), Palgrave Macmillan, New York.
McMillan, M., and Rodrik, D. (2011), “Globalization, structural change and productivity growth,”
Chapter 2 in Making Globalization Socially Sustainable, Bacchetta, M. and Jansen, M. (Eds.),
International Labour Organization and World Trade Organization, Geneva, Switzerland.
Nickell, S., Redding, S. and Swaffield, J. (2008), “The Uneven Pace of Deindustrialisation in the
OECD,” The World Economy, Vol. 31, Issue 9, pp. 1154-1184.
Otieno, W. A. and Mwangola, A. (2006), Why Africa has Fallen Short of Building Dynamic
Agroprocessing Capabilities: Constraints, Options and Prospects, ATPS Special Paper Series No. 29,
ATPS, Nairobi.
Polity IV: Regime Authority Characteristics and Transitions Datasets, 1800-2015.
http://www.systemicpeace.org/inscrdata.html
Tkalec, M. and Vizek, M. (2009), “The Impact of Macroeconomic Policies on Manufacturing
Production in Croatia”, Privredna kretanja i ekonomska politika 121/2009, 61-92.
United Nations Development Program (UNDP)(2012), The Roles and Opportunities for the Private Sector
in Africa’s Agro-Food Industry: UNDP African Facility for Inclusive Markets, UNDP, Johannesburg.
United Nations Industrial Development Organization (UNIDO) (2013), Sustaining Employment Growth: The
Role of Manufacturing and Structural Change, UNIDO, Vienna.
United Nations Industrial Development Organization (2016), The Role of Technology and Innovation in Inclusive
and Sustainable Industrial Development, UNIDO, Vienna.
World Bank (2016), World Development Indicators 2016, World Bank, Washington, DC.