Chapter 4: Differences in Intersectoral Linkages and its Implications for sub-Saharan Africa: Evidences
4.3. Results
The variables are computed using an Excel spreadsheet designed by Itaki (2014)14. Furthermore, the initial form of orbit analysis applies to two or multiple variables if they are time series but not to a panel data. To construct a cross-sectional data, first, orbit analysis is replicated individually for each country and then, the results of this process are compiled together to construct a new datasettreated with STATA®.
The cross-sectional data take the form: where i is the individual dimension and t is the time dimension. A series of descriptive and graphical analysis are conducted with the new dataset in order to highlight the dynamics of the GDP aggregates, and to draw implications for policy discussions.
For the res
agriculture, hunting, forestry, and fishing (ISIC
A-(ISIC-
-restaurants and hotels (ISIC G-(ISIC-I).
4.3.1. Sectoral decompositioninpercentage share of GDP
The following table summarises the main activities by sector. The data based on 38 countries from the UN database reported in Table 2 show that on average, for the last decade, more precisely 2000-2013, the share of agriculture in percentage of GDP accounted for 25.31 percent; mining 22.35 percent; manufacturing 9.18 percent;
construction 4.03 percent, wholesale 12.26; transport 6.35 and other activities 20.51.
Sorted according to its weight, the table ranks agriculture at the first position, followed by mining, other activities, wholesale, manufacturing, transport and construction.
Table 2 - GDP Sectoral Decomposition SSA, in percentage, 1970-2013
GDP Components 1970-1979 1980-1989 1990-1999 2000-2013
Agriculture 31.06 28.88 28.62 25.31
Mining 17.50 17.94 18.60 22.35
Manufacturing 9.62 9.89 9.66 9.18
Construction 4.24 3.58 3.53 4.03
Wholesale 12.28 13.03 13.33 12.26
Transport 5.34 5.73 5.73 6.35
Other activities 19.95 20.95 20.53 20.51
Source: Author. Compiled data fromUNSD (2014)based on a data of 38 SSA countries
When looking at the data expressed in percentage share of GDP, agriculture is playing an important role. However, a large share of GDP, however, does not imply that agriculture is playing a leading role in the economy among other variables such as mining or other activities, this, type of relations is addressed in orbit analysis presented as follows.
4.3.2. Results of the orbit analysis
Three types of outputs are delivered by the Excel computation: 5-, 7- and 9-year moving averages. The results of the 9-year moving average are then used to reconstruct a new set of data in STATA® for a descriptive analysis. In addition to the main SSA, data on 38 individual countries were also compiled in a dataset, according to different attributes,
such as the sub-region cluster, geographic conditions, the fact of being part of the New Alliance project, political freedom and the business environment.
The graphical output from the Excel spreadsheet is reproduced in Figure 9 below, which reports the output of the 9-year moving average and its summary statistics are reported in the appendix. Then, extreme points,
value of each sector, are plotted to look how they changed over time. It shows how agriculture interacted with other industries. Accordingly, agriculture will be used as the key variable inthis analysis.
The results of the output table of the 9-year moving average are showing the broad picture, how the data interacted among each other across time. Nevertheless, with multiple variables, the figure presented below might be slightly problematical for the interpretation. Hence, for a more detailed analytical purpose, descriptive statistics using Figure 9 to understand this concept of intersectoral linkages.
Figure 10 shows the scores recorded by each sector on average to compare how each variable outperformed during the period of study. This graph compares two data.
First a weighted average of the aggregated data of SSA and second the individual analysis.
Figure 10 - Weighted results of orbit analysis, decomposition 1972-2011
On average, manufacturing (3.25797) was leading in the sense of orbit analysis followed by agriculture (3.1436), other activities (3.12602), mining (3.05734), transport (2.93066), wholesale (2.068545) and construction (2.4075). In other words, manufacturing exerts a stronger pulling force that heralds changes among other variables.
To understand the changes over the studied period, the first (non-missing value) and last value of the score recorded by each sector are depicted below showing the extreme points i.e. the initial and final period (here 1973-2011). This study points out
3.1436 3.05734 3.25797
2.4075
2.68545
2.93066 3.12602
Mean of 9-years Moving average Sub-Saharan Africa
1-Agriculture 2-Mining
3-Manufacturing 4-Construction
5-Wholesale 6-Transport
7-Other
two types of kick-starter. Type 0 (1972 kick-starter) considers all the time span of the period of analysis and type I (2011 kick-starter) which is the highest score at a time t within a period of analysis.
Figure 11 - First and last scores of each sector, GDP decomposition 1973-2011
In Figure 11, the first graph (on the left side) exhibits that in the initial period (1973), other activities, construction, agriculture, manufacturing, and transport are the kick-starters and its sequential followers. On the right hand side of the same figure, manufacturing, mining, other activities, agriculture, construction, transport and wholesale sectors are respectively ranked in the same manner in the final period.
The results of the disaggregated data according to geographical decomposition are slightly different from time series type. Therefore, the average score of the first year will therefore be applied to the analysis. The data are also presented according the sub-regional category, namely: Western, Southern, Eastern and Middle Africa.
3.16667
2.33333 3.16667
3.5
2 3.16667
3.66667
First Non-Missing Points
2 3
4.4
1.8 1.4
1.8 2.4
Last Non-Missing Points
SSA aggregated
1-Agriculture 2-Mining
3-Manufacturing 4-Construction
5-Wholesale 6-Transport
7-Other
Table 3 - Leading-following relations across countries, SSA 1972-2011
Rank Western Africa Southern
Africa Eastern Africa Middle Africa
1 Manufacturing
(3.077) Agriculture
(3.336) Mining
(3.083) Manufacturing (3.109)
2 Agriculture
(3.032) Transport
(3.053) Agriculture
(3.066) Agriculture (3.012)
3 Wholesale
(2.962) Mining
(2.989) Transport
(3.012) Wholesale (3.010)
4 Mining
(2.926) Wholesale
(2.934) Manufacturing
(2.961) Transport (2.926)
5 Construction
(2.916) Construction
(2.844) Construction
(2.865) Mining
(2.825)
6 Transport
(2.893) Manufacturing
(2.783) Wholesale
(2.837) Construction (2.789)
7 Other
(2.808) Other
(2.675) Other
(2.792) Other
(2.705) 1972 kick-starters
1 Construction
(3.240) Transport
(3.750) Construction
(3.712) Construction (3.190)
2 Agriculture
(3.219) Agriculture
(3.625) Agriculture
(3.212) Agriculture (2.976)
3 Manufacturing
(3.146) Mining
(3.167) Mining
(2.924) Wholesale (2.929)
4 Transport
(3.010) Wholesale
(2.792) Manufacturing*
(2.833) Transport (2.667)
5 Wholesale
(2.958) Manufacturing*
(2.708) Wholesale*
(2.833) Other
(2.548)
6 Mining
(2.802) Other*
(2.708) Transport
(2.803) Manufacturing (2.524)
7 Other
(2.625) Construction
(2.250) Other
(2.682) Mining
(2.357) 2011 kick-starters
1 Agriculture
(2.000) Manufacturing
(2.250) Mining
(2.341) Mining
(2.339)
2 Construction
(1.977) Mining
(2.125) Manufacturing
(2.045) Agriculture (2.089)
3 Manufacturing
(1.969) Other
(2.000) Transport
(2.034) Transport (2.071)
4 Transport
(1.875) Construction
(1.938) Other
(1.807) Manufacturing (2.054)
5 Mining
(1.852) Agriculture
(1.844) Agriculture
(1.693) Wholesale (1.786)
6 Wholesale
(1.742) Transport
(1.500) Construction
(1.614) Other
(1.536)
7 Other
(1.703) Wholesale
(1.469) Wholesale
(1.591) Construction (1.536) Degree of variations
1 Other
(0.714) Agriculture
(0.857) Agriculture
(0.838) Construction (0.774)
2 Transport
(0.683) Construction
(0.731) Wholesale
(0.664) Agriculture (0.762)
3 Mining
(0.679) Transport
(0.724) Mining
(0.648) Mining
(0.729)
4 Wholesale
(0.669) Mining
(0.675) Other
(0.642) Other
(0.688)
continued
5 Agriculture
(0.663) Manufacturing
(0.666) Construction
(0.633) Manufacturing (0.637)
6 Construction
(0.650) Wholesale
(0.659) Transport
(0.614) Transport (0.624)
7 Manufacturing
(0.619) Other
(0.571) Manufacturing
(0.607) Wholesale (0.603) Source: Author. Notes: score of the ranking points in parentheses.* indicates variables recording the same scores.
The degree of variation indicates the intensity of the changes of one variable across time measured by the standard deviation. The degree of variation helps to understand the deviation of the leading and following variables. The results in Table 3 indicate that across the region, on average:
Manufacturing was a leading sector in Western and Middle Africa; agriculture in Southern Africa and finally, mining in Eastern Africa. For the four regions, with the exception of Southern Africa, agriculture is playing the role of a following variable.
Among the 1972 kick-starters, the variables recording highest scores at the beginning of the period of analysis across the region are construction in Western, Eastern and Middle Africa, whereas in the Southern region, it was led by transport.
The variables that are leading at the end of the period of study, or 2011 kick-starter are agriculture in Western Africa, manufacturing in Southern Africa, and mining in Eastern and Middle Africa.
The variables that recorded the great degree of variation or the sector of activity that has changed actively across time are other activities in Western Africa agriculture in Southern and Eastern Africa and construction in Middle Africa.
Nonetheless, comparing the value of each sector at the beginning and the end of the period of study is not enough to appreciate the changes across time. For the rest of
other sectors of the economy.
4.3.3. Sector-to-sector analysis
The
sector-to-leading-following relations and major turning points among the two. Various graphs compiled around the 9-year moving average output of the orbit analysis are used to support this step in which the scores are compared in a time series with the differences
a. Agriculture and manufacturing (ISIC-D)
The result of the 9-year moving average is showing that over the period of the study, manufacturing and agriculture scored respectively, on average, at the first and second position, followed by other activities, mining, transport, wholesale, and finally by construction. Figure 12 depicts the interactions between agriculture and manufacturing in the region for 1973-2011.
In the initial period, both variables are recording the same scores and are among -s
started to thrive. Indeed, for the region of SSA, the major turning point took place in 1992-onward where manufacturing took the lead over agriculture whose progress was heading towards a downward movement as depicted by the impulse bar in the same figure.
The agricultural sector is also the most dynamic regarding the density of changes in the overall economy. According to the implicit correlation tests agriculture and manufacturing show a weak positive correlation. However, the resultant force of the
manufacturing sector makes it one of the leading industries in the region followed by agriculture.
Figure 12 - Leading-following relations between agriculture and manufacturing in SSA
b. Agriculture and other activities (ISIC J-P) In percentage share of GDP, for the period
2000-occupies the third position after the mining industries. In the initial value of the orbit 72
kick-Likewise, in the final value of the same analysis, it ranks at the third position.
Across time, as highlighted in Figure 13, the sector was exceeded by agriculture in 1993.
Then after this period, it again saw a steady growth, leaving agriculture lagging behind.
The impulse bar is also revealing two major periods with respect to agriculture and
1970 1980 1990 2000 2010
year
Agriculture Manufacturing
Agriculture vs. Manufacturing
9 year moving average - 1970-2011
other activities. First, 1976-1993 in which agriculture was leading and second, 1993-2011 when other activities took the lead over agriculture. Looking at the summary statistics, on average, this sector is positioned at the third place after manufacturing and agriculture. Over this period of study, it accounted for 2 to 4.22 points. This sector was also one of the most dynamic after agriculture accounting for greater variability over the years.
Figure 13 - Leading-following relations between agriculture and other activities in SSA
c. Agriculture and mining (ISIC C-E)
The mining curve saw a scattered progression over time. Independently from its evolution vis-à-vis agriculture, a booming period of the mining sector could be noticed from the 1990s.
Figure 14 shows the interactions of the scores between agriculture and mining as well as their differences across time, here expresses as an impulse bar, representing the
1970 1980 1990 2000 2010
year
Agriculture Other
Agriculture vs. Other
9 year moving average - 1970-2011
period between 1973 and 2011. The major turning point appeared in 1992 when mining surpassed agriculture. For this period, mining held leading position as exhibited by the impulse bar. Mining activities in percentage share of GDP represent the second industry in the region. The results of the orbit analysis show that on average, mining ranked at the fourth positionover 1972-2011 (see: Figure 10).
Figure 14 - Leading-following relations between agriculture and mining in SSA
d. Agriculture and transport, storage and communication (ISIC-I)
Transport, storage and communication has on average scored at the fifth position. This sector saw a scattered progress from 1970 to the mid-2000s. The impulse bar is revealing two important periods with respect to agriculture and transport. The first period is 1973-1991 in which agriculture was leading in the sense of orbit analysis, over transport and second, 1991-onward when transport took the lead over agriculture. This leading-following relation is showninFigure 15.
1970 1980 1990 2000 2010
year
Agriculture Mining
Agriculture vs. Mining
9 year moving average - 1970-2011
Figure 15 - Leading-following relations between agriculture and transport in SSA
e. Agriculture and wholesale, retail trade, restaurants and hotels (ISIC G-H) The wholesale, retail trade, restaurant and hotels recorded a steady development between 1970 and 1980. It started to decline significantly after the 1980s. On average, this sector ranks at the sixth position after transport and accounted for 1.44-3.77 with a variation 0.62 point over 1970-2011. Similar to the other sectors of the economy, a turning point also occurred in 1991 in this sector in which a major convergence could be noticed in terms of ranking points. The only difference with the other industries is the fact that it tumbled around 2010.
Figure 16 exhibits the relations and differences across time between agriculture and wholesale over the period 1973-2011.
1970 1980 1990 2000 2010
year
Agriculture Transport
Agriculture vs. Transport
9 year moving average - 1970-2011
Figure 16 - Leading-following relations between agriculture and wholesale in SSA
Source: Author
The relations between agriculture and wholesale saw up and down movements across time. Wholesale lagged behind agriculture, according to their differences across time, except for the period between 1991 to 1999 when wholesale was leading over agriculture.
f. Agriculture and construction (ISIC-F)
The c 1972 kick-starters As it can be noticed from
Figure 17, it soon saw a declining phase, following the agriculture trend line. On average, it accounted for 1.66-3.55 point with a variation of 0.51 overall the period covered by the study. The construction sector ranks at the seventh position on average, following the wholesale sector.
Furthermore, the difference shows that agriculture has been leading over
1970 1980 1990 2000 2010
year
Agriculture Wholesale
Agriculture vs. Wholesale
9 years moving average - 1970-2011
kick-after agriculture. The impulse bar is also revealing that the differences in score across time were greater in 1972-1991. However, these differences started to narrow from 1992 to the present.
Figure 17 - Leading-following relations between agriculture and construction in SSA
After the brief presentation of the results of orbit analysis, the following section presents the discussions around the question posited at the beginning of this chapter.