Structural reforms are required in food systems to enhance their contribution to the Agenda 2030 for sustainable development (Webb et al., 2020). Indeed, the overarching goal of the United Nations Food Systems Summit 2021 (UN, 2021) is to catalyze action for the transformations needed in food systems. However, evidence highlights that benefits in one area may determine costs in others (Mehrabi et al., 2020). Thus, the design of sustainable agricultural policies requires balancing multiple goals (Barret et al., 2020) and broader assessments of the potential outcomes and their consequences (Piñeiro et al., 2020). The failure to analyze ex-ante the potential effects of policy interventions could result in the preclusion of positive synergies given the predominance of unintended negative effects.
Livestock has been acknowledged as one of the most complex subsectors of model agriculture (Msangi et al., 2014). Traditionally, livestock policy analysis has focused on assessing the behavior of the system as decision-making units operating in isolation. An advantage of this approach is that it permits the assessment of the direct effects of a policy change on a specific production system. One of the limitations is that it does not allow examination of how these changes affect the rest of the economy. This is a relevant pitfall in the context of the food systems approach, in which the interdependencies between different agents are primary features. To address this gap, the Food and Agriculture Organization (FAO) of the United Nations developed the Livestock Policy Simulation Model (LPSM).
The LPSM is a recursive dynamic open (CGE) model designed for country-level ex-ante policy analysis (Fig 1). LPSM is “recursive” as the model is solved sequentially and repeatedly, one period at a time; “dynamic” as the economy’s progression is modeled over time; “open” as the model allows for trade across borders. These characteristics allow the model to simulate the direct and indirect effects of policy change or economic shocks in the sector. This study aims to introduce the LPSM, illustrating its potential to conduct ex-ante livestock policy analysis, in support of sectoral planning processes. By employing Ethiopia as a case study, we examined the effects of increasing livestock total factor productivity (TFP).
Ethiopia is one of the countries with the largest livestock population in the world (FAO, 2021). Projections suggest that by 2050, the population will double, and GDP will triple, substantially increasing the demand for livestock products (FAO, 2019). Thus, increasing the sector’s productivity will be fundamental to ensuring that economic gains will not come at an environmental cost. To address this challenge, the government has embarked on a sectoral planning exercise to enhance the livestock sector’s sustainable transformation (Federal Democratic Republic of Ethiopia, 2016).
The results show that an increase in productivity growth by 2030 will have several positive effects on economic growth (0.7%), employment generation (0.78%), household consumption (0.82%), and poverty reduction (0.18pp); however, there might be also negative environmental effects associated with a total increase in GHG emissions (5.9%). These results have important policy implications, showing that the achievement of some SDG targets could conflict with accomplishing others. The analysis illustrates the capacity of LPSM to integrate economic, social, and environmental dimensions while quantifying spillover and trade-offs.
The Economic Structure of Ethiopia’s Livestock Sector
This section describes the economic structure of Ethiopia’s livestock sector based on information derived from the country’s social accounting matrix (SAM) for 2016. As the LPSM is livestock-specific, these statistics disaggregate livestock production to the species/product level, with crops, industry, and services comprising the remainder of total economic activity. These data provide an overview of economic functioning and the contributions of different sectors. Table 1 presents a summary of the main descriptive statistics calculated from the SAM. The value-added or GDP contribution of the agriculture sector is 36.68% (Table 1), demonstrating Ethiopia’s heavy economic reliance on the sector. In comparison, Kenya’s agricultural value-added for the same year was 26.31%, and Uganda's was 28.82% (World Bank, 2020). Of agriculture’s total value-added, livestock contributes 16.5% (6.07% of GDP), with the remainder attributable to crops. As such, the livestock sector is a small but significant component of the total value-added.
Table 1 Economic structure of the Ethiopian livestock sector
|
Sectoral structure %
|
Sectoral factor intensity %
|
Sectoral demand structure %
|
Commodity
|
VA
|
EMP
|
EXP
|
EXP-O
|
IMP
|
IMP-D
|
Unskilled
|
Semi-skill
|
Skilled
|
Capital
|
Land
|
Intermed
|
Priv Cons
|
Gov Cons
|
Inventory
|
Exports
|
Beef Cattle
|
1.56
|
1.21
|
3.04
|
0
|
0.02
|
0.41
|
65.3
|
4.3
|
1.6
|
8.3
|
20.4
|
18.5
|
63.8
|
0
|
0
|
0
|
Dairy Cattle
|
2.43
|
1.9
|
4.71
|
0
|
0
|
0
|
65
|
4.1
|
1.6
|
8.1
|
21.1
|
28
|
72
|
0
|
0
|
0
|
Sheep
|
0.65
|
0.52
|
1.31
|
33.54
|
0
|
0
|
67.5
|
4.2
|
1.8
|
8
|
18.5
|
4.1
|
64.2
|
0
|
-2.1
|
33.8
|
Goats
|
0.62
|
0.49
|
1.24
|
0
|
0
|
0
|
67.5
|
4.3
|
1.8
|
8
|
18.5
|
0
|
97.2
|
0
|
0
|
0
|
Poultry
|
0.32
|
0.25
|
0.61
|
0.02
|
0.01
|
0.76
|
64.6
|
4.3
|
1.7
|
8.5
|
21
|
22
|
49.9
|
0
|
0
|
0
|
Other Livestock
|
0.49
|
0.43
|
0.88
|
1.33
|
0
|
0.01
|
49.1
|
2
|
1.1
|
39.1
|
8.8
|
35.9
|
50.1
|
0
|
0
|
1.3
|
Meat
|
0.09
|
0.53
|
0.07
|
17.65
|
0
|
0.21
|
13.8
|
2.2
|
1
|
83.1
|
0
|
27.1
|
55.3
|
0
|
0
|
17.6
|
Dairy
|
0.03
|
0.46
|
0.05
|
0.03
|
0.1
|
5.21
|
28.9
|
5.9
|
1.9
|
63.3
|
0
|
9.8
|
90.1
|
0
|
0
|
0.1
|
Other Food
|
2.05
|
3.3
|
2.2
|
1.24
|
5.28
|
27.09
|
18
|
3.7
|
1.8
|
76.5
|
0
|
29.5
|
69.7
|
0
|
0
|
0.8
|
Services
|
42.06
|
41.52
|
20.56
|
7.16
|
11.14
|
7.35
|
16.3
|
6.6
|
16.4
|
60.8
|
0
|
49.3
|
28.6
|
15.4
|
0
|
6.6
|
Crops
|
30.61
|
24.26
|
59.24
|
6.31
|
5.06
|
5.21
|
64.7
|
4.2
|
1.9
|
8.3
|
20.8
|
10
|
82.9
|
0
|
0
|
7.1
|
Industry
|
19.09
|
25.13
|
6.09
|
1.11
|
78.38
|
44.68
|
32.8
|
14.1
|
8.9
|
44.2
|
0
|
32.8
|
12.5
|
0
|
0
|
0.7
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
VA: Sectoral share in value-added (GDP contribution); EMP: Employment Share; EXP: Export share; EXP-O: domestic sectoral supplies between exports and domestic sales; IMP: import share; IMP-D: domestic sectoral demand between exports and domestic sales output; Unskilled: low skilled labor factor intensity; Semi-skill: semi-skilled labor factor intensity; Skilled: skilled labor factor intensity; Capital: capital factor intensity; Land: land factor intensity; Intermed: intermediate sectoral demand; Priv Cons: private consumption sector demand; Gov Cons: government consumption sector demand; Inventory: inventory changes; Exports: Export sector demand.
The high level of disaggregation of Ethiopia’s SAM allows for a detailed analysis of sectoral factor intensity. Factor intensity reflects the relative importance of each production factor within a sector. All production factors are represented, namely, labor (disaggregated by skill level), capital, and land. In terms of factor intensity, livestock along with crops use the highest share of labor; however, the disaggregation of employment depending on skill level indicates that the greatest proportion is unskilled. The ratio of unskilled to semi-skilled to skilled labor for cattle, sheep, goats, and poultry is approximately 65%, 4.3%, and 1.7%, respectively. Livestock operates on average at a higher capital intensity than crops (13.3% vs. 8.3%) and a slightly lower land factor intensity (18.1% vs. 20.8%). However, with crops having a larger GDP component, each factor’s total economic share is significantly larger, for example, 4.8% of employed labor in livestock vs. 24.26% in crops.
Turning to the demand structure, livestock has a significantly higher intermediate use share than crops (18.03% vs. 10%). Among livestock production systems, poultry, dairy cattle, and beef cattle have larger forward linkages with other activities (>18%). In other words, these production systems have a higher multiplier effect and, as a result, promoting this production system leads to greater second-round effects. The private (or household) consumption profile varies by species, with goats (97.2%) having the highest private consumption fraction followed by dairy (90.1%), crops (82.9%), dairy cattle (70%), sheep (64.2%), and beef cattle (63.8%). Sheep and other livestock were the only animals exported. Most of this export value is in the form of live animals, followed by hides and skins (Hailemariam et al., 2009).
Baseline scenario
This section covers the model’s projected macroeconomic indicators for the baseline scenario. The baseline scenario represents a business-as-usual projection to 2030 without any policy changes. Given that the intent is to build such a scenario, the baseline should maintain the economy’s initial macroeconomic structure. Moreover, to facilitate the presentation and analysis, the baseline scenario assumptions are as simple and transparent as possible.
We use available information and estimates to generate a plausible picture for the country’s economic development for the base year period (2016–2020). Drawing on projections from the International Monetary Fund World Economic Outlook (IMF, 2020), we impose an average GDP growth rate of 7.4% for the analysis period. In the baseline scenario, GDP growth is imposed by endogenously calibrating the TFP. In addition, we assume that government consumption, government to household transfers, and domestic and foreign government net financing grow at the same rate as GDP.
Taxes are fixed at their base-year rates, which means that they grow at the same pace as the overall economy. For the total population and working-age population (15-64 years), we use projections from the 2019 World Population Prospects of the UN Population Division (UN, 2019). Overall, average annual growth rates for the period 2021-2030 are similar across the different indicators (at an average of 7.5% per year during the period 2016-2030). GDP growth is sufficient to bring about relatively rapid employment expansion; the employment underutilization rate is reduced from 15.5% in 2016 to 12.6% in 2030. In turn, real wages grow at a rate of 2.2% per year on average.
In per-capita terms, household consumption grows at a rate of 4.3% per year, more rapidly than government consumption, and in step with private investment. Consequently, there is a significant decrease in the poverty rate from 23.3% in 2019 to 10.0% in 2030. This poverty drop is particularly significant for rural households. At the sectoral level, the GDP growth rate is close to aggregate GDP growth. For livestock, growth is lower because of slow land supply growth and low-income elasticities of demand.
Policy simulations
We simulate the effects of an exogenous 10% increase in livestock TFP between and 2021-2030. This is a rise in the level of output (meat, milk, eggs) while keeping the level of inputs (land, labor, capital) constant. We assume that the productivity level of intermediate inputs remains constant. In particular, based on FAOSTAT (2021) historical data, we assume that yields per production system will remain stable (Annex 1). As shown in Table 1, livestock production systems differ in size; thus, to conduct a meaningful comparison, we calculated the semi-elasticities of selected results with respect to changes in the value-added of the targeted production system. The results below present the effect of the shock on the livestock sector as a whole, and for each livestock production system separately, as a deviation from the business-as-usual simulation by 2030.
By 2030, the national poverty rate is reduced by 0.18 pp compared to the base scenario (Fig 2). In the species-specific scenarios, poverty is reduced by less, but at least by 0.01 pp relative to the base scenario. The scenario that yields the largest poverty reduction targets dairy cattle. This result is explained by dairy’s higher value-added share compared to other production systems. As observed, the employment level increases for all production systems, particularly for beef and dairy cattle. The household consumption results closely follow the effect on poverty reduction and employment generation, reflecting the effect that an income increase, particularly among the poorest, has on livestock product demand.
Overall, we find positive macroeconomic effects (Fig 2). Therefore, indicators such as GDP and private consumption grow on average at rates higher than those recorded in the base scenario. For instance, by 2030, the level of GDP is 0.7%, and private consumption is 0.8% higher than in the base scenario. The slope of the simulations suggests that the effect varies between the medium and long term among the production systems. For example, whereas for beef cattle, the effect of the shock on GDP remains in the long term, for dairy cattle, it stabilizes, while for goats’ decreases.
The effect on major macroeconomic aggregates tends to be more favorable when the production system is directly or indirectly export-oriented. Table 1 shows that beef cattle is indirectly export-oriented through the meat sector. However, dairy cattle and dairy products are not exported. In turn, given the assumption that Ethiopia is a global market price taker, an increase in export products' domestic supply has a lower price decrease effect in the domestic market.
Other macroeconomic indicators, such as investment, exports, and imports, also show a positive impact (Fig 3). However, the effect depends on which production system is shocked. A TFP increase in dairy cattle generates the largest gains in terms of private (household) consumption, which is partially explained by the fall in household food basket relative price. A TFP increase in beef cattle, in turn, has the highest effect on exports, reflecting the international trade orientation of meat production; nevertheless, sheep is the production system with the largest increase in exports, when controlling for its size (see Table 1).
Enhancing production system output increases not only the system’s demand and supply but also downstream and upstream activities along the food supply chain (Fig 4). In the case of beef cattle, for example, a 10% TFP change leads by 2030 to an increase in cattle supply of 5.7% and a rise in meat output of 7.8%. In addition, increased livestock production also generates additional demand for products such as cereals and trade services, which further stimulates growth in other manufacturing and service sectors.
We found an increase in the total equivalent emissions for all production systems (Fig 5). As observed, the result of the combined effect of increasing TFP in all production systems is an increase of 5.9% in total livestock emissions by 2030. The production system with the highest increase was beef cattle, with a change of 3.5% over that of the base scenario, followed by dairy cattle with a 1.3% increase. These results are explained by the different species’ emission coefficients, technological changes, the effect of the different production systems’ sectoral output, and increasing productivity’s economic rebound effect.
The emission coefficient for beef cattle is the largest; thus, an increase in the output leads to relatively higher emissions than other production systems. As shown in Annex 1, during the last two decades, the carcass weight per animal has remained constant, suggesting that an increase in supply will mainly occur through an increase in head numbers. Beef cattle are the most export-oriented production systems through their input-output relationship with the meat sector; thus, an increase in output does not lead to an equivalent reduction in domestic prices, generating indirect incentives to expand production further. In other words, the beef cattle results reflect the combination of a relatively large emission coefficient with a relatively large output increase.