the transforming value chain of ethiopia’s “orphan” tef crop · 2019-07-30 · the...
TRANSCRIPT
Vol.:(0123456789)1 3
Planta (2019) 250:769–781 https://doi.org/10.1007/s00425-019-03224-0
REVIEW
The transforming value chain of Ethiopia’s “orphan” tef crop
Fantu Bachewe1 · Mekdim Dereje Regassa2 · Bart Minten1 · Alemayehu Seyoum Taffesse1 · Seneshaw Tamru3 · Ibrahim Worku Hassen4
Received: 3 September 2018 / Accepted: 20 June 2019 / Published online: 3 July 2019 © Springer-Verlag GmbH Germany, part of Springer Nature 2019
AbstractMain conclusion Despite the relatively little attention given to tef, the value chain is quickly transforming and is expected to further do so in the near future.
Abstract Tef is called an “orphan” crop in Ethiopia as it receives relatively little attention from the Ethiopian government and from international donors. Given the low yields of tef compared to other crops, it is often viewed as a low-priority crop and relatively little is known about the value chain of tef. We fill some of this knowledge gap in this paper. We illustrate tef’s importance in Ethiopia’s food systems and the rapid changes upstream, midstream, and downstream in its value chain. We show that tef production and productivity is rapidly increasing and that tef markets are improving over time. More specifically, using a growth decomposition analysis, we find that while the expansion of land and labor use have been important sources of growth in tef production, the relative contributions of modern input use and agricultural extension have been increasing over time. We also show that tef has greater economic potential, with comparatively more of it consumed by the better-off segments of the population, indicating that its importance is likely to grow over time as income grows in the country. Using reasonable assumptions on income growth, urbanization, and commercialization, we estimate that national tef consumption and marketed output will increase by about 250 and 300%, respectively, over a 20-year period.
* Bart Minten [email protected]
Fantu Bachewe [email protected]
Mekdim Dereje Regassa [email protected]
Alemayehu Seyoum Taffesse [email protected]
Seneshaw Tamru [email protected]
Ibrahim Worku Hassen [email protected]
1 International Food Policy Research Institute, Addis Ababa, Ethiopia
2 University of Bonn, Bonn, Germany3 Licos-University of Leuven, Leuven, Belgium4 Harvard University, Cambridge, USA
1 Compared to other cereals, tef is hardy and able to withstand adverse weather conditions and is therefore considered a lower risk crop (Fufa et al. 2011). Tef is grown at middle elevations between 1800 and 2200 meters above sea level and in regions where there is adequate rainfall. These characteristics, together with it being easy to store, seemingly explain the sustained importance of tef in Ethiopia. Its grain is mainly used for making injera but tef is also valued for its fine straw, which is used for animal feed as well as for other purposes, such as mixing with mud for building. .
Introduction
While tef (Eragrostis tef) is an important crop in Ethio-pia, it is often called an “orphan” crop (Fufa et al. 2011). Tef receives little attention from international agricultural
research centers (the Consultative Group of International Agricultural Research or CGIAR). Primarily, this is because tef falls outside the mandate of these centers that are struc-tured along major international crops such as maize, wheat, and rice. The Ethiopian government has also devoted little attention to the crop, likely because of its desire and empha-sis to ensure enough food supply in the country. Given the low yields of tef compared to other crops, it is often viewed as a low-priority crop.1 Given these reasons, there is rela-tively less knowledge about the production, marketing, and consumption patterns of tef.
We fill this knowledge gap in this analysis and contrib-ute to the international literature by focusing on three main questions. First, we aim to understand what the importance of tef is in production, markets, and consumption and how
770 Planta (2019) 250:769–781
1 3
it compares with other crops in Ethiopia’s agricultural econ-omy. Second, given the rapid transformation happening in Ethiopia (World Bank 2015), we assess the nature of trans-formation in each of these segments of the tef value chain, focusing upstream on productivity increases (relying on a growth decomposition analysis), midstream on changes in marketing and processing margins, and upstream on evolv-ing patterns in consumption. Finally, we model expected changes in tef consumption and marketed surplus in the near future using reasonable assumptions on income growth, urbanization, prices, and marketization.
A value chain approach is used in the paper to analyze the tef crop upstream, midstream, and downstream. In contrast to similar studies which mostly rely on case studies or rapid rural appraisal methods, access to large-scale data sets at each level of the value chain is used the consequent advan-tage being that assessments of the functioning of the tef value chain at the national level can be made with reasonable confidence. Tef is found to be the most important cash crop in the country and a major source of income for many poor farmers. The results of the growth decomposition analysis show that expansion of land and labor has been important to explain production growth of tef and that the relative con-tribution of modern inputs and agricultural extension has been increasing over time. We also find that tef’s importance is likely to grow over time given the income growth in the country with rapid increases estimated for national tef con-sumption and marketed output in a 20-year period. Appro-priate and sustained attention to this important orphan crop in Ethiopia is, therefore, required and warranted.
The structure of the paper is as follows. In Sect. 2, we discuss the data sets and methods that we use to analyze the transformation in the tef value chain. Transformation in the upstream (production), midstream (marketing), and downstream (consumption) segments of the value chain is discussed in Sects. 3, 4, and 5 respectively. Section 6 dis-cusses simulations of future demand and of commercial mar-ket development. We finish with the conclusions in Sect. 7.
Methodology
Tef output growth decomposition
This study uses a modified Solow (1957) growth account-ing method to decompose growth in tef output. The method decomposes growth in output into changes in inputs used, total factor productivity (TFP), and exogenous factors that contribute to changes in tef output. Central to the method is the assumption that there exists a functional relationship between tef output and inputs used and other factors that influence tef production. Suppose, in a given year, t, tef out-put is given as:
where Q is the quantity of tef produced and T(t) stands for the cumulative effect of technical change.2 Nine production inputs are included: farming labor ( Lt ), capital ( Kt ), area of land cultivated to tef ( At ), chemical fertilizers ( Ft ), improved seeds ( Mt ), irrigation ( It ), agro-chemicals ( Pt ), extension ( Et ), and service sector outputs (e.g., transportation and banking services) used by tef producers ( St ). Exogenous factors that affect tef output, but are not directly put into production, such as infrastructure and returns to changes in the scale of tef production, are represented by the vector Zt.
Differentiating both sides of (1) with respect to time and dividing that by Q gives:
where ΔQ ( �Q∕�t ) stands for the time derivative of output. Similarly, ΔJ , ΔT(t) , and ΔZ stand for time derivatives of the nine inputs, technical change, and exogenous factors, respectively. Equation (2) can be rearranged as:
To get from Eqs. (2) to (3), we use wJ = (�Q∕�J) × (J∕Q) , where wJ . is the relative share of input J in crop output. Equation (3) indicates that the contribution of a given input/factor to output growth that occurred between periods t and t + 1 is determined by how much its use changed, and its share between t and t + 1. Exogenous factors, Zt , are rep-resented by the last two expressions in (3): ΔRTS stands for changes in returns to scale (RTS), and ΔRR stands for changes in infrastructure, which we proxy by rural roads. � and � represent the rate of change in output per unit of change in RTS and rural roads, respectively.
If time series data on tef output, shares and changes in factors used in tef production, length of rural roads, and esti-mates of � , � , and ΔRTS are available ΔT(t)∕T(t) or changes in total factor productivity (TFP) can be estimated using Eq. (3). We follow Carlaw and Lipsey (2003) to estimate the effect of returns to scale, as the excess of the sum of shares of factors put into production over 1, which occurs if returns to scale are constant, or � =
∑
J wJ − 1 . Moreover, ΔRTS is given as the excess of the sum of payment to inputs used in production weighted by the rate of change in the
(1)Qt = T(t)f(
Lt,Kt,At,Ft,Mt, It,Pt, St,Et, Zt)
,
(2)
ΔQ
Q=
ΔT(t)
T+∑
J
�Q
�J
ΔJ
Q+
�Q
�Z
ΔZ
Q
where J ∈ [L,K,A,F,M, I,P, S,E],
(3)
ΔTFP =ΔT(t)
T=
ΔQ
Q−∑
J
wJ
ΔJ
J− �ΔRTS − �ΔRR.
2 Equation (1) assumes neutral technical change that shifts the pro-duction function without affecting the marginal rates of substitution (Solow 1957). For a review of the growth accounting method, see also Carlaw and Lipsey (2003).
771Planta (2019) 250:769–781
1 3
inputs ( ∑
J wJΔJ∕J ) over the rate of growth or expansion in the tef production ( f ) or ΔRTS =
∑
J wJΔJ∕J − f , where the minimum of growth in cultivated area and labor, inputs indispensable in crop production, is used as a proxy for f. Finally, we replace � by Zhang and Fan’s (2004) estimate of elasticity of TFP with respect to rural roads (0.042).
Tef production in Ethiopia was likely affected by other factors. For instance, farmers’ use of organic fertilizer and improved land and water management practices—that the Government of Ethiopia invested in expanding—likely con-tributed to growth in tef output. Furthermore, crop produc-tion in Ethiopia is largely rainfed, implying that weather con-ditions affect changes in crop outputs. Insofar as these and other factors are not included in the analyses, corresponding estimates of changes in TFP also include the effects of those factors.
Data source and coverage
The descriptive as well as the growth accounting analysis uses data on output and inputs in tef production obtained from Central Statistics Agency (CSA) annual reports (Central Statistical Agency 2005a, b, c). We chose these data, because they are the only nationally and regionally representative series that are systematically available over time. The quality of CSA data has been challenged by some authors (e.g., Dercon and Hill 2009; Gollin 2011; Mandefro and Jerven 2015), and they are, nevertheless, considered to be among the best in Africa. Moreover, the regularity and consistency of the underlying surveys should allow big pic-ture overviews and analysis in this area. We will, however, also use alternative datasets as appropriate.
Data on agricultural production are collected annually through an agricultural sample survey implemented by the CSA (Central Statistical Agency 2017a, b). This survey is typically fielded in more than 2000 enumerations areas and more than 40,000 farmers are visited. For example, in 2010/11, almost 45,000 agricultural household were inter-viewed. The survey collects data, among others, on area allo-cation, production levels, yields, use of harvest, and land management practices. The sample is set up in such a way that the results are representative of the regional and zonal levels. The annual data from these surveys for the period 2003/04 until 2016/17 are used in the analysis.
The consumption analysis relies on the Ethiopian House-hold Consumption and Expenditure Survey (HICES) data set from the past four rounds, i.e., 1995/96, 1999/00, 2004/05 and 2010/11. These data were also collected by the Central Statistical Agency (CSA). In total, 11,678, 17,320, 21,560 and 27,831 households were interviewed over the four peri-ods, respectively. The survey contains detailed information on consumption and expenditures of both food and non-food items. These data are explored to analyze trends in the
consumption of different food categories over these periods.3 The paper focuses on quantity consumed and expenditures on the three varieties of tef (white, mix, and red) and of injera, a spongy flatbread, the main national dish in Ethio-pia (as well as in Eritrea). It is to be noted that there have been some differences in data collection methods over the years, and some caution in the interpretation over time is warranted (Stifel and Woldehanna 2014). In addition to the HICES dataset, the retail price data set of the CSA is used to account for inflation over this period (Central Statistical Agency 2017a).
Price data used in this study are collected as part of the CSA’s Consumer Price Survey (CSA 2017b). These data were collected in 116 urban retail markets in all regions of Ethiopia. The number of markets in each region is approxi-mately proportional to the region’s share of the total urban population to ensure a sufficient degree of national repre-sentativeness. Thirty-two markets are surveyed in Southern Nations, Nationalities, and Peoples (SNNP) region, 24 in Oromia, and 20 in Amhara (the three biggest regions), while 12 markets are surveyed in Addis Ababa (by far the largest urban center with officially more than 3 million residents). The smaller regions include only a handful of markets. We also use prices from producer price surveys that CSA con-ducts in a large number of districts in the country (CSA 2017c).
For the growth decomposition analysis, we further rely on rural road data obtained from the National Bank of Ethi-opia (National Bank of Ethiopia (NBE) 2017). We proxy changes in intermediate service sector inputs by changes in tef output, which assumes that changes in intermediate service sector input use are proportional to growth in tef out-put. Unlike data on output and input use levels, which often are available annually, data on factor shares are infrequent. We derive factor shares for labor, capital, land, chemical fertilizer, pesticides, and services from the 2009/10 Social Accounting Matrix (SAM) of Ethiopia (Engida et al. 2011). We complement the latter by the relative output elasticities with respect to quantity of improved seeds, irrigation, and area covered by extension, obtained by estimating a crop production function using data from the Agricultural Growth Program (AGP) of Ethiopia baseline survey. The AGP data set used in the analysis pertains to the 2010/11 main agricul-tural season and includes about 7000 households, sampled to represent 8 million households residing in Tigray, Amhara, Oromiya, and SNNP regions.
3 Note that since a different cleaning procedure was followed before statistics were calculated, there are slight differences with the national estimates.
772 Planta (2019) 250:769–781
1 3
Transformation upstream: tef production
Background
Tef is estimated to be the most important crop in Ethio-pia’s agricultural economy using a number of indicators. In 2016/17, it was estimated by the Central Statistical Agency that tef made up 21% of all the cultivated area by private smallholders in the Meher season, covering about 3 million hectares, and that it was grown by 7 million farmers.4 As there are 17.4 million (grain) farmers in Ethiopia in total, this implies that 40% of all these farmers grow tef. The sec-ond most important crop is maize at 15% of all cultivated land, followed by sorghum accounting for 13%.
While the tef area has grown by 50% over the last decade, from 2 million hectares in 2003/04 to 3 million hectares in 2016/17, the share of tef in total area cultivated has stayed relatively stable over time. It was 21.3% in 2003/04 and it stayed at that level in 2016/17. The share of other crops also remained surprisingly stable over time, with seemingly no important diversification in Ethiopia’s agricultural economy during the Meher season to date. For example, the share of cereals in total area cultivated during the Meher season was as high as 75% in 2003/04, and although it declined over time, it still remained as high as 72% in the year 2016/17.
Tef is the most important single crop area-wise in Ethio-pia, but its importance in agricultural output is lower. This is due to the relatively low yields of tef compared to most other crops and especially other cereals. The total national production of tef in 2016/17 (5.0 million tons) was lower than maize (7.8 million tons), but higher than wheat (4.5
million tons) and sorghum (4.7 million tons) (CSA 2017a). The average yield of tef in that year was 1.66 tons/ha, less than half the yield of maize (3.68 tons/ha).
On the other hand, prices paid per kg of tef are considera-bly higher (Table 1). The total value of tef output in 2016/17 amounted to 2.9 billion USD, the highest level in the country for a single crop (Table 1).5 Maize, wheat, and sorghum are valued at 1.9, 1.8, and 1.6 billion USD, respectively. Overall, tef makes up 31% of the total value of cereal output.
Within the cereal sector, tef is the most commercialized crop, with an estimated 30% of the production sold (CSA 2017b). The value of commercial surplus of tef—that part of production that is sold—in 2016/17 was estimated to be 850 million USD. Tef makes up half of the value of total commercial surplus of the cereal sector and equals the com-mercial surplus of all other cereals combined (Table 1). The cash receipts of farmers from the sales of tef are 27% higher than that from coffee and almost triple that from sesame (Table 1). Noting that coffee and sesame are the two major exports of the country, tef is, thus, by far the most important cash crop in the country.
Evidence of tef production growth
In exploring the associates of growth in tef production, the agricultural year 2003/04 is used as the base year. As Fig. 1 illustrates, tef production has increased by almost 200% over
Table 1 Production and commercial surplus of cereals, coffee, and sesame 2016/17 Source: authors’ calculations, CSA agricultural sample surveys, 2016/17
Quantities (million tons) Price Values (Billion USD)
Production Market surplus (Birr/kg) Production Market surplus
Tef 5.02 1.49 11.03 2.87 0.85Barley 2.02 0.25 6.95 0.73 0.09Wheat 4.54 0.84 7.71 1.81 0.33Maize 7.85 0.87 4.75 1.93 0.21Sorghum 4.75 0.49 6.42 1.58 0.16Finger millet 1.02 0.12 6.57 0.35 0.04Oats/‘Aja’ 0.05 0.01 7.03 0.02 0.00Rice 0.14 0.03 9.70 0.07 0.02Total cereals 25.38 4.10 – 9.34 1.71Coffee 0.47 0.23 55.03 1.34 0.67Sesame 0.27 0.19 32.41 0.45 0.31
5 To value production, the following methodology was used. The median retail price collected by CSA in all the surveyed markets in the country over the period July 2013–June 2014 was calculated. The prices of the following types in the CSA’s data set were used as an approximation of the price of the product: mixed tef, white wheat, white barley, white maize, white sorghum, millet, oats, and imported rice. The exchange rate of January 2014 of 19.33 Birr per USD to convert Birr is used. For calculations of commercial surplus, the number published for year 2013/14 (CSA, 2014b) is relied upon.
4 Taffesse et al. (2012) show that smallholders generate 95% of the total production of the main crops in Ethiopia, and that 97% of the total crop production is in the Meher season. The Belg season is, therefore, relatively less important.
773Planta (2019) 250:769–781
1 3
the last 13 years, and a combination of both area and yield increases of 52% and 97%, respectively. Yield growth has, thus, been the main contributor to production growth. Dur-ing the first half of the decade, yield and area growth were at similar levels (for the period 2003 until 2007). The gap between the two growth rates has, however, becomes wider since. Figure 1 also shows that the number of tef farmers has increased significantly over the period considered: an increase of 53%. As the area increase was of similar magni-tude as the number of tef farmers, there has been no signifi-cant change in the average area of tef cultivated per farmer over this period.
We use complementary data sources to check the robust-ness of these estimates.6 We present evidence from three groups of complementary methods and data sources. First, CSA data on growth in tef yields are compared with those obtained from other large-scale rural household survey data sets collected over the last decade (for a detailed discus-sion of these data sets, see Bachewe et al. 2018). While the surveys cover relatively large parts of the country, caution is required in comparing yields over time, since the surveys
were fielded in different areas and periods, and for different purposes. However, growth rates in yields of tef calculated using these surveys are generally similar to the growth rates obtained from CSA data (Table 2).
Second, we use data from the Ethiopian Rural Household Survey (ERHS), a unique longitudinal household survey that contains information on agriculture, consumption, assets, and income of almost 1500 households in 15 villages across the country surveyed for over a decade (1994–2009). The ERHS data set indicates improvements in tef yields, but the yield levels were lower than the corresponding values from CSA data, however.
Third, focus groups in the 304 kebeles (villages) in the Agricultural Growth Program (AGP) baseline survey were asked to compare average yields in their community in 2011 relative to 10 years earlier. Focus groups included at least five people knowledgeable about the community, such as community leaders, kebele chairpersons, elders, religious leaders, and teachers. Growth in white and black tef yields averaged 20 and 17%, respectively, in total over the 10-year period. These growth rates were, therefore, lower than for data of CSA.
To further understand the growth in tef yields, we look at changes in the adoption of modern inputs over time. Since the early 1990s, Ethiopia has implemented several cereal intensification programs promoting the adoption of mod-ern technologies. At the center of these strategies was the push for adoption of chemical fertilizer and improved seed packages by smallholders. Figure 2 shows to what extent the share of fertilized tef area has changed in recent years. While this share stood at 60% in 2005, it increased to 80% in 2016. Similar changes have been noted in the adoption of pesticides as well (Fig. 2). While the adoption of improved seeds is not easy to measure, Minten et al. (2016a, b), how-ever, document the rapid adoption of the improved Quncho variety in well-connected areas. The Government of Ethio-pia has further made large investments on an agricultural
0
50
100
150
200
250
300
350
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Inde
x (1
00=y
ear 2
003)
Area Holders Produc�on Yield
Fig. 1 Changes in tef production, area, yield, and number of pro-ducers (private peasant holders, Meher season), 2003–2016 Source: authors’ calculations, CSA agricultural sample surveys, 2003/04–2016/17
Table 2 Estimates from alternative data sets of annual tef yield growth in Ethiopia, %
Source: authors’ calculations
Survey Period Annual tef yield growth (%)
CSA 2003–2016 5.5Ad hoc surveys(8 surveys)
2009–2014 4.7
ERHS 2004–2009 1.7
0
10
20
30
40
50
60
70
80
90
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
% o
f tef
are
a
Share fer�lized Share pes�cide Share extension package
Fig. 2 Changes in fertilizer and pesticide use and in access to exten-sion (2004–2017, 3-year moving average) Source: authors’ calcula-tions, CSA agricultural sample surveys, 2003/04–2016/17
6 It is important to keep in mind that most of these data sets use farmers’ recall in measuring output and acreage, while the CSA uses crop-cuts.
774 Planta (2019) 250:769–781
1 3
extension system focused on the provision of advisory and training services led by frontline development agents. Ethiopia, thus, achieved one of the highest extension agent-to-farmer ratios in the world. This increasing coverage by extension agents is also shown in Fig. 2.
Therefore, these complementary data sources illustrate significant changes in access to extension, adoption of mod-ern inputs, and consequent yield and production growth in the tef subsector, as shown earlier using CSA data. However, documented growth rates from these sources are generally also slightly lower. Differences in survey methods may have contributed to these deviations. CSA statistics on yields are based on data collected through crop-cut and recall methods, while the others rely on recall data.
Sources of tef production growth
Table 3 shows the results of the growth accounting analysis (for a summary of the data used in this analysis, see Annex 1). The second column of the table provides changes in tef output starting from 2004/05. The remaining columns provide the contribution of inputs and exogenous factors to growth in crop output. The table indicates that annual growth in tef output averaged 7.8% during 2004/05–2017/18 (6th row), while it also indicates tef output grew more rap-idly in the first half of the period than in the second. Out of the 7.8% growth in output, 2.0% points was accounted for by labor (column 3), while expansion in cultivated land accounted for 0.7% points. A further 1.2, 1.1, and 0.4% points of tef output growth originated from increases in uses of chemical fertilizer, improved seeds, and area covered by extension packages, respectively.
The table also provides a basis for exploring growth in tef output resulting from changes in TFP. Annual change in TFP accounted for about 1.9% points of the 7.8% growth in tef output. The contributions of TFP were slightly higher during the second half of the period relative to the first half, possibly reflecting the relatively lower growth rates in labor and land used in tef production (Fig. 3). Results of the growth accounting analyses not only show the importance of labor and land in the tef output growth recorded in the last 15 years, but also indicate the growing importance of modern inputs and agricultural extension. The contribution of labor and land in the first half of the period was about 42% of the growth in tef output (columns 3 and 5 in the 8th row) recorded during the same period. The contribution of labor and land in the second half of the period declined to 23% (columns 3 and 5 in the 9th row). In contrast, the con-tribution of modern inputs (fertilizer, improved seeds, and pesticides) and agricultural extension, which accounted for 33% growth in tef output during the first half, increased in
Tabl
e 3
Con
tribu
tion
of in
puts
, oth
er fa
ctor
s, an
d TF
P to
tef o
utpu
t gro
wth
, %
Sour
ce: a
utho
rs’ c
ompu
tatio
n
Ave
rage
s dur
ing
the
perio
dTe
f out
put
Labo
rC
apita
lLa
ndFe
rtiliz
erIm
pr. s
eeds
Pesti
cide
sIr
rigat
ion
Exte
nsio
nSe
rvic
esRe
t. to
scal
eRu
ral r
oads
∆ T
FP(1
)(2
)(3
)(4
)(5
)(6
)(7
)(8
)(9
)(1
0)(1
1)(1
2)(1
3)(1
4)
Gro
wth
(%)
200
4/05
–10/
119.
482.
920.
021.
051.
381.
230.
060.
250.
210.
020.
230.
341.
77 2
011/
12–1
7/18
6.70
1.02
0.01
0.51
1.16
0.91
0.06
0.05
0.50
0.02
0.10
0.05
2.30
200
4/05
–17/
187.
781.
990.
020.
701.
221.
100.
060.
100.
360.
020.
160.
201.
86Re
lativ
e co
ntrib
utio
n (%
) 2
004/
05–1
0/11
100.
030
.80.
211
.114
.613
.00.
62.
62.
20.
22.
43.
618
.7 2
011/
12–1
7/18
100.
015
.20.
17.
617
.313
.60.
90.
77.
50.
31.
50.
734
.3 2
004/
05–1
7/18
100.
025
.60.
39.
015
.714
.10.
81.
34.
60.
32.
12.
623
.9
775Planta (2019) 250:769–781
1 3
importance to 40% in the second half (columns 6–10 in the 8th and 9th rows, respectively).
Transformation midstream: marketing
Tef value chains are found to be relatively unsophisticated, short, and in contrast to common perception, fairly well organized (Minten et al. 2016a, b). At farm level, there are no interlinked transactions with buyers of the produce (which is often seen in other countries, especially in more developed value chains), the role of credit is minor, and most of the transactions are cash transactions. Farmers obtain a relatively high share of the final retail price (on average 80%), and the importance of distress sales is low.
Over the last decade, changes in the market structure of tef seem to have been linked with economic growth, urbanization, improved roads (and, consequently, decreased transport costs, and possibly competition in better and big-ger fleets), and greater access to ICT (mobile phones are universally used by brokers and traders, striking deals, and bypassing wholesale markets entirely). These changes in tef markets have given rise to predicted outcomes of considera-ble market improvements. For example, increasing urbaniza-tion, increased supply, and income growth have led to more quantities traded and greater economies of scale, and thus to lower margins overall. In addition, access to better price information has contributed to a more efficient marketing system and, consequently, lower overall margins (Minten et al. 2014).
Monthly price series for tef have been collected over the last 15 years by the Central Statistical Agency (CSA) at the retail, producer, and milling level. By comparing these prices, the evolution of urban–rural marketing and process-ing margins over time can be analyzed. The increasing com-petition between mills, as mentioned by retailers (Minten
et al. 2014), seems to have led to a significant reduction of the ratio of milling charges over tef retail prices over the last 15 years (Fig. 3). These margins have dropped on average to half the level of 15 years earlier.
By comparing producer prices from neighboring major production zones in Addis to retail prices in Addis Ababa and looking at trend lines, we see that the share of the pro-ducer in the final retail prices increased from a level of between 74 and 78% in 2001 to between 84 and 86% in 2017 (Fig. 4).7 Despite having the highest prices, white tef shows the lowest producer-to-retail ratio, indicating significantly higher marketing costs than other types of tef. Moreover, large variability of these margins is seen over time with a significant decrease in shares of wholesale and producer in the final retail prices in 2009 and 2010. One contributing factor might have been that the government eliminated fuel price subsidies in October 2008 as the share of producers in the final retail prices started to reduce significantly from that time for all three types of tef. Nevertheless, it is not imme-diately clear what has been driving this variability and these differences between different types of tef, and thus, further research appears necessary.
0
.01
.02
.03
.04
.05
2001
m7
2002
m1
2002
m7
2003
m1
2003
m7
2004
m1
2004
m7
2005
m1
2005
m7
2006
m1
2006
m7
2007
m1
2007
m7
2008
m1
2008
m7
2009
m1
2009
m7
2010
m1
2010
m7
2011
m1
2011
m7
2012
m1
2012
m7
2013
m1
2013
m7
2014
m1
2014
m7
2015
m1
2015
m7
2016
m1
2016
m7
2017
m1
ratio: milling charges over (white) teff price fitted values
Fig. 3 The ratio of milling charges over (white) tef retail prices in Addis Ababa, 2001–2017 Source: authors’ calculations from CSA price data
.5.6
.7.8
.91
2001
m7
2002
m1
2002
m7
2003
m1
2003
m7
2004
m1
2004
m7
2005
m1
2005
m7
2006
m1
2006
m7
2007
m1
2007
m7
2008
m1
2008
m7
2009
m1
2009
m7
2010
m1
2010
m7
2011
m1
2011
m7
2012
m1
2012
m7
2013
m1
2013
m7
2014
m1
2014
m7
2015
m1
2015
m7
2016
m1
2016
m7
2017
m1
white producer Linear (white producer)mix producer Linear (mix producer)red producer Linear (red producer)
Shar
e in
reta
il pr
ice
Fig. 4 Share of producer in the final retail prices of tef in Addis Ababa, 2001–2017 Source: authors’ calculations from CSA price data
7 Nominal prices of tef have been rising substantially over the period considered, linked to general inflation in the country. However, when we look at the real prices of tef over the last decade (2007–2017), tef prices in the beginning of 2017 were at similar levels as in 2007. We saw a significant hike for the period 2009–2010 with prices almost 50% higher than a year earlier, seemingly driven by foreign exchange issues, rapid general inflation because of monetary expansion, and the global food crisis. However, prices came down since. While prices rose again in the years 2015 and 2016, real price levels in 2017 were similar to 2013 and 2007.
776 Planta (2019) 250:769–781
1 3
Overall, despite large variability, the results show that the shares of urban–rural marketing and milling in final retail prices have declined significantly over a 15-year period, seemingly indicating improved marketing systems.8
Transformation downstream: tef consumption
This section of the paper considers consumption patterns and their changes over time using four rounds of HICES data sets, covering the period from 1996 until 2011. To ensure comparability over time in this analysis, expenditures are deflated using the national Consumption Price Index (CPI) and values are expressed in constant 1996 Birr, implying a lowering of nominal prices from later HICES rounds
Table 4 Food consumption and real per capita expenditures, by category
Source: authors’ calculations based on HICES, CSA
Real per capita expenditures (Birr/capita/year)
1996 2000 2005 2011
Birr Share (%) Birr Share (%) Birr Share (%) Birr Share (%)
White tef 16 2.6 18 3.0 17 2.6 18 2.5Mixed tef 22 3.6 28 4.6 18 2.8 19 2.6Red tef 31 5.0 32 5.2 22 3.4 19 2.6Injera 5 0.8 5 0.8 9 1.4 34 4.7Total tef 73 11.9 82 13.4 67 10.3 90 12.4Wheat 44 7.2 53 8.7 57 8.8 56 7.7Barley 28 4.6 23 3.8 41 6.3 18 2.5Maize 63 10.2 70 11.5 40 6.2 57 7.8Sorghum 43 7.0 37 6.1 52 8.0 37 5.1Five major cereals 251 40.8 265 43.4 257 39.7 258 35.5Other cereals, pulses, and oilseed 87 14.1 89 14.6 85 13.1 121 16.6Vegetables, fruits, roots, and tubers 100 16.3 132 21.6 96 14.8 143 19.7Animal products 61 9.9 63 10.3 73 11.3 94 12.9Other foods 115 18.7 59 9.7 136 21.0 112 15.4Total food 615 100.0 610 100.0 648 100.0 727 100.0
Consumption (kg/capita/year) kgs Share (%) kgs Share (%) kgs Share (%) kgs Share (%)
White tef 5 2.2 6 2.0 7 2.2 8 2.3Mixed tef 8 3.5 11 3.7 8 2.5 9 2.5Red tef 11 4.8 13 4.3 10 3.1 10 2.8Injera 2 0.9 3 1.0 3 0.9 7 2.0Total tef 25 10.9 31 10.3 27 8.3 29 8.2Wheat 21 9.2 25 8.3 30 9.3 25 7.1Barley 14 6.1 10 3.3 24 7.4 10 2.8Maize 34 14.8 43 14.3 25 7.7 51 14.4Sorghum 17 7.4 23 7.7 32 9.9 28 7.9Five major cereals 111 48.5 132 44.0 138 42.6 143 40.5Other cereals, pulses, & oilseed 30 13.1 28 9.3 27 8.3 29 8.2Vegetables, fruits, roots, & tubers 54 23.6 114 38.0 105 32.4 114 32.3Animal products 15 6.6 14 4.7 17 5.2 17 4.8Other foods 20 8.7 12 4.0 37 11.4 49 13.9Total food 229 100.0 300 100.0 324 100.0 353 100.0
8 As the government is conscious of local price increases by open-ing tef export markets, it has currently only allowed a limited number of commercial farmers to start producing tef to fulfil the increasing export demand. Ethiopia imports significant quantities of wheat each year—partly provided as food aid—but it has been shown that the country as whole is a net agricultural exporter in value terms (Minten et al. 2019).
777Planta (2019) 250:769–781
1 3
through a division by this CPI. Quantities consumed per capita were calculated, as well.9 The results of this exercise are presented in Table 4. They illustrate a number of inter-esting findings.
First, some important general shifts in the food basket are noted over time. Overall, the share of cereals in total food expenditures is declining. While the share made up 43.4% of expenditures in 2000, it had declined to 35.5% 10 years later. Most growth in the non-cereal food categories was recorded in the ‘other food’ category that grew from 9.7 to 15.4% from 2000 to 2011. There is also an increasing importance of animal products over time. While the share of animal products is still relatively low, it has grown from 9.9% of the food basket in 1996 to 12.9% in 2010. These patterns are a reflection of Bennett’s law that describes a relative decline in starchy staples and an increase in animal proteins with income (Bennett 1941).
Second, the most important crop within the cereal expen-ditures of the food basket is tef. It accounted for 12.4% of food expenditures in 2011. This compares to 7.8% for maize, 7.7% for wheat, and 5.1% for sorghum. Over time, some minor shifts within the consumption of cereals are observed. For example, the share of expenditures of sorghum in food expenditures was 6.1% in 2000 and 8.0% in 2005, but it declined to 5.1% in 2011. Compared to 2000, the share of maize in food expen-ditures has decreased as well. Sorghum and maize are both characterized by low income elasticities and it seems that the growing average incomes in Ethiopia might lead consumers to retreat from consuming these crops (Berhane et al. 2012).
Third, some important changes are noted as well within the tef category. While red tef made up 5.0% and 5.2% of food expenditures in 1996 and 2000, respectively, this share declined to half that level in 2011. Expenditures on red tef in 2011 were only 2.6% of all food expenditures. Expendi-tures on white tef were consistently lower than those on red and mixed tef in 1996 and 2000, but they were at an equal level in 2011. Hence, there is a notable shift away from the cheap red tef to the more expensive white tef. The most important change within the tef expenditures is, however, the quick emergence of injera as an important food item in the food basket. It represented 4.7% of expenditures in 2011, a significant increase compared to 0.8% of total food expen-ditures in 1996 and 2000. This seems to follow the pattern that as consumers become richer and opportunity costs of women’s time in the household are on the rise, ready-to-eat foods become more readily part of the consumption basket (Kennedy and Reardon 1994; Dibley et al. 1995).10
Fourth, total tef consumption over the years has remained at similar levels. It was as high as 31 kgs per capita in 2000, dropped to 27 kgs in 2005, and then increased again to 29 kgs in 2011. Within the tef categories, the same trends are seen as in tef expenditures. The quantities consumed of white tef and injera are on the rise. White tef consumption increased from 5 kgs in 1996 to 8 kgs in 2011 and injera consumption increased in the same period from 2 to 7 kgs. On the other hand, red tef (from 13 kgs in 2000 to 10 kgs in 2011) and mixed tef (from 11 kgs in 2000 to 9 kgs in 2010) consumption decreased over time.
Significant research has been conducted in agricultural economics to understand the link between income and food consumption patterns. The parameters resulting from such research are important as they allow for economic modeling to assess impacts on consumption of food policy changes, as well as projecting food requirements in the future. There have been significant methodological advances. While demand for food items was previously analyzed in single-equation models, these estimates often led to inconsistencies in parameters when total food baskets were considered. To address this issue, a methodology called the Almost Ideal Demand Systems (AIDS) was developed. This method is widely used to estimate parameters as part of complete food demand systems (Deaton and Muellbauer 1980). Researchers have also tried to improve their understanding of transforming food systems in economic development and associates with changes in consumption (e.g., Reardon and Timmer 2007).11 Comparing the differences in consumption patterns of richer and poorer households is often indica-tive of how transformation of food systems will shape food economies in a particular country.
In an effort to understand these patterns in Ethiopia, all households in the HICES survey of 2011 were ranked by wealth quintile, from the poorest quintile 1 to the richest quintile 5. The shares of different consumption categories were then calculated. As expected (e.g., Bouis and Haddad 1992; Bouis 1994; Subramanian and Deaton 1996; Pingali 2007), strong differences in the composition of consump-tion baskets are seen over poverty quintiles. The five major cereals make up 37.7% of the poorest quintile and this sur-prisingly only declines minimally for the richest quintile to 32.2% (Table 5). Notably, the share is relatively stable for the poorest three quintiles and drops off only for quintile 4 and 5, suggesting that transformation in the food basket has only started to occur in the richest two quintiles. The food budget for animal foods for the richest households comprises 20.4%,
10 However, some caution in the interpretation of injera consumption might be required given that rural consumption seems low and given that injera can be made from different cereals..
11 A number of food consumption patterns can be distinguished with increasing income and economic development: 1. processed and ready-to-eat foods take off; 2. cereals become less important; 3. the share of high-value crops such as fruits and vegetables, dairy and ani-mal products, and fish in food consumption baskets increase.
9 To convert injera to kgs of tef, it was multiplied by a conversion factor of 0.325.
778 Planta (2019) 250:769–781
1 3
yet this is only 7.9% for the poorest ones. As for cereals, the higher consumption of animal products is especially noted for the richest quintiles 4 and 5.
The share of cereals in total food consumption is rela-tively stable over poverty quintiles, but there are, however, large differences within this category by poverty quintile. Maize and sorghum are the two cereals that are typically consumed more by the poor than by the rich. 13.5% of all food expenditures of the poor go towards maize. This compares to 2.5% for the rich. These numbers are 7.4 and 1.8%, respectively, in the case of sorghum. The consumption of barley is low overall, but its share also decreases when incomes increase. There is relatively little variation in wheat, based on the poverty level; however, it is consumed slightly more by the rich.
Tef consumption also shows a distinctive pattern by pov-erty level. It increases consistently over poverty quintiles and tef only makes up 6.9% of the food expenditures of the
poorest quintile. By contrast, tef comprises 18.8% of all food expenditures of the richest quintile. Tef is, therefore, clearly a preferred food of the rich. Moreover, within the tef category, there are a number of other further distinctive pat-terns. Red tef makes up 28% of the tef expenditures for the poorest households (quintile 1), but this decreases to 10% for the richest ones. On the other hand, expenditures on injera drop from 48% for the richest quintile to 38% for the poorest one, likely driven by the significantly higher prices per kg or per calorie for injera as seen in Table 5.
In a more complete quantitative approach, Tafere et al. (2010) use an AIDS model to estimate income elasticities from the HICES data of 2004/05. As predicted from the previous tables, animal products have the highest income elasticity of all the food product categories considered (Table 6). A doubling of income leads to a 172% increase in animal expenditures in urban areas and a 198% increase in rural areas. Animal products are, therefore, an economically
Table 5 Share of expenditures by poverty quintile, 2011. Source: authors’ calculations based on HICES, CSA
Q1 (poorest) Q2 Q3 Q4 Q5 (richest) Total
Food White tef 1.1 1.6 2.2 2.7 4.1 2.5 Mixed tef 1.3 1.9 2.5 3.0 3.6 2.6 Red tef 2.0 2.5 3.2 3.0 1.9 2.6 Injera 2.6 2.6 3.0 4.3 9.1 4.7 Total tef 6.9 8.6 10.8 13.0 18.8 12.4 Wheat 6.6 7.0 8.0 8.4 7.9 7.7 Barley 3.2 3.3 2.9 2.3 1.3 2.5 Maize 13.5 11.1 9.4 6.3 2.5 7.8 Sorghum 7.4 6.8 6.5 4.9 1.8 5.1 Five major cereals 37.7 36.8 37.6 35.0 32.2 35.5 Other cereals, pulses, and oilseeds 15.7 17.2 17.2 17.3 15.5 16.6 Vegetables, fruits, roots, and tubers 22.9 21.4 18.9 19.4 17.4 19.7 Animal products 7.9 9.0 10.0 13.2 20.4 12.9 Other foods 15.8 15.6 16.3 15.2 14.5 15.4 Total food 100.0 100.0 100.0 100.0 100.0 100.0
Table 6 Own price and income elasticity of demand for selected food items, 2004/05. Source: Tafere et al. (2010)
Urban Rural
Own price elas-ticity
Income elasticity Own price elas-ticity
Income elasticity
Tef − 0.92 1.10 − 0.92 1.20Wheat − 1.00 0.78 − 0.94 1.19Maize − 0.93 0.37 − 0.70 0.82Sorghum − 0.93 − 0.36 − 0.71 0.51Pulses and other cereals − 0.88 0.90 − 1.03 0.74Animal products − 0.91 1.72 − 0.94 1.98Fruit, vegetables and other root
crops− 0.99 1.22 − 1.01 1.18
Other foods − 0.92 0.66 − 0.92 0.92
779Planta (2019) 250:769–781
1 3
superior product (e.g., Delgado et al. 1999; Delgado 2003; Berhane et al. 2012). Tef also shows high income elas-ticities, i.e., 1.2 in rural and 1.1 in urban areas, indicating that a doubling of income increases expenditures by 120% and 110%, respectively. On the other hand, other cereals show much lower elasticities. Sorghum even has a nega-tive income elasticity in urban areas, indicating that it is an economically inferior commodity in this urban environment. When households become richer, the consumption of such goods is reduced. The importance of sorghum as a food is, therefore, likely to reduce, and the importance of tef is likely to increase with the rise in income over time, as Ethiopia becomes wealthier and more urbanized.
Simulations of future demand and of commercial market development
We use the estimated income elasticities to evaluate how demand for tef products might evolve in the future. We do so by integrating expected population dynamics, differenti-ating between urban and rural areas, relying on population projections by the World Bank. We further assume a uniform annual income growth of 3% and no real price increases. Figure 5 shows, under these assumptions, the evolution in demand for tef for rural and urban areas. The results show that the demand for tef will increase by 250% between 2011 and 2030. We see especially rapid growth in urban demand for tef. While urban consumption of tef made up 43% of total tef consumption in 2011, this is expected to increase to 55% in 2030.
Next, we look at the development of commercial markets for local consumption. We take as a base the share of pur-chases in total tef consumption from the HICES data in 2011 for rural (22.5%) and urban (93.9%) areas. We assume an increase in market use in rural areas over time by 10% points (to 32.5%)—an increase that is partly driven by improved rural infrastructure in the future given the government’s
large investments in this area. Figure 5 shows the results of these simulations.
The results illustrate the rapid growth in urban markets as commercial urban tef demand will increase by 350% compared to the situation in 2011. While the share of urban markets made up already 76% of the commercial value of tef in 2011, this is expected to slightly increase to 78% of the nation-wide value of commercial markets. These data, there-fore, indicate to what the extent the rural–urban tef value chains will grow in the future. Rural commercial markets are expected to grow by 304%. Overall, a growth of 340% is expected by 2030. This type of growth will have enormous implications for different agents involved in these markets, such as traders, transporters, wholesalers, retailers, with these numbers expected to increase rapidly in the future.
Conclusions
The analysis in this paper illustrates the importance of tef in Ethiopia’s agricultural and food economy as well as important transformations in the value chain. Upstream, it is estimated that tef is Ethiopia’s most important single crop, making up 21% of its cultivated area during the most important Meher season. Tef is grown by 7 million farmers. The production value of the crop was estimated in 2016/17 to be as high as 2.8 billion USD. Moreover, it is shown that tef is the most important cash crop in the country. The value of income generated for farmers from tef is estimated to be more important than from coffee, the most important export product of Ethiopia; and the value of the commercial sur-plus of tef is as high as the commercial surplus of all the other cereals in the country combined. Over the last decade, significant increases have occurred in the production of tef. Using a growth decomposition method, we find that expan-sion of land and labor use in tef production has been impor-tant to explain growth, but we also show that the relative
Fig. 5 Evolution of consump-tion and commercial market development for tef (2011–2030) Source: authors’ calcula-tion based on HICE, CSA
0
5
10
15
20
25
30
2011 2030
Cons
ump�
on -B
illio
n 20
11 B
irrrural urban
0
5
10
15
20
25
30
2011 2030
Com
mer
cial
mar
ket-
Billi
on 2
011
Birr
rural urban
780 Planta (2019) 250:769–781
1 3
contribution of modern input use, agricultural extension, and TPF to tef production growth is increasing over time.
Midstream, we note improved domestic marketing. Despite large variability, the shares of urban–rural mar-keting and milling in the final retail prices have declined significantly over a 15-year period. These changes seem to have been linked with improved roads (and consequently decreased transport costs, and possibly competition in better and bigger fleets) and greater access to ICT (mobile phones are universally used by brokers and traders, striking deals, and bypassing wholesale markets entirely).
Downstream, the last decade has seen a shift from the least preferred yet lower cost foods, to costlier and more pre-ferred food options. Within the tef sector, ready-to-eat injera and the more expensive white tef are on the rise, while the cheap red and mixed tef are on the decline. Tef is consumed more by urban households than by rural households. Tef is further characterized by high income elasticities, evaluated at 1.10 in urban areas and 1.20 in rural areas. These figures demonstrate that tef is an economically superior commod-ity, implying that with an increase in income, this leads to a disproportional increase in tef consumption, and, moreover, that tef will continue to be a product that is consumed in greater quantities by the rich, in mostly urban areas, than by the poor. The lower consumption by the poor is partly explained by the high prices of tef which are typically twice as high as the cheapest cereal, i.e., maize.
Despite these improvements, the transformation of the tef production and marketing systems is still at an early stage of agricultural development (e.g., Reardon and Tim-mer 2007). At the production level, the number of farm-ers who use improved varieties is still low, the quantities of chemical fertilizers that are being used are still below the recommended levels, and mechanization, which is hap-pening quickly in other emerging economies, is still mostly absent (Minten et al. 2014). In addition, very little verti-cal integration is observed, as well as minimal coordination
mechanisms between tef production and marketing. Mid-stream and downstream of the value chain, little evidence of up-scaling of trade is seen, nor of modern retail or of branding. Typically, these can be observed as agricultural market development gets underway (Reardon et al. 2012).
Ethiopia’s economy is transforming, people’s incomes are increasing, and as consumers purchase more pre-ferred foods, they consume more tef. Using reasonable assumptions on income growth, urbanization, prices, and marketization, we estimate that national tef consumption and commercial markets will increase by about 250 and 300%, respectively, over a 20-year period. However, such increasing demands come at a price. There is pressure on the government and other stakeholders to change agricul-tural and food policies and their implementation, as well as pressure on the smallholding traditional farmers to modify their practices to adapt to a changing consumption pat-tern, with new farming techniques that increase yield and minimize waste. Tef is emerging to become a crop that is more productive and renowned for its gluten-free and rich nutrient constitution. Nevertheless, the time is ripe for the emergence of tef from what was once thought “a threat to food security that should be abandoned”, to a crop that deserves more recognition locally and globally.
Author contribution statement All authors contributed equally to this work.
Appendix tables for growth accounting analyses
See Tables 7 and 8.
Table 7 Summary of data used in growth accounting Sources: aCentral Statistical Agency (2005a); bCentral Statistical Agency (2005b); cCentral Statistical Agency (2005c); dNational Bank of Ethiopia (NBE) (2017)
Variable 2004/05 2007/08 2010/11 2012/13 2014/15 2017/18
Number of farm holders (millions)a 4.9 5.6 6.1 6.3 6.5 6.8Crop area (million hectares)a 2.1 2.6 2.8 2.7 3.0 3.0Number of draft cattle (millions)c 9.7 11.5 12.9 13.4 14.3 15.1Chemical fertilizer (000 MT)b 80 132 191 210 250 323Improved seeds (000 MT)b 0.4 0.6 1.0 1.4 1.8 2.0Pesticides applied area (000 ha)b 523 781 966 1116 1281 1483Irrigated area (000 ha)b 8 10 12 11 11 10Extension covered area (000 ha)b 367 393 541 666 953 1083Rural roads (00 kms)d 18.4 23.9 29.2 32.6 33.0 33.9Tef output (000 MT)a 2026 2777 3483 3765 4751 5283
781Planta (2019) 250:769–781
1 3
References
Bachewe F, Berhane G, Minten B, Taffesse AS (2018) Agricultural transformation in Africa? Assessing the evidence in Ethiopia. World Dev 105:286–298
Bennett MK (1941) International contrasts in food consumption. Geogr Rev 31:365–374
Berhane G, McBride L, Tafere K, Tamiru S (2012) Patterns in foodgrain consumption and calorie intake. In: Dorosh P, Rashid S (eds) Food and agriculture in Ethiopia: progress and policy challenges. Uni-versity of Pennsylvania Press, Philadelphia, pp 190–218
Bouis HE (1994) The effect of income on demand for food in poor countries: are our food consumption databases giving us reliable estimates? J Dev Econ 44(1):199–226
Bouis HE, Haddad LJ (1992) Are estimates of calorie-income elastici-ties too high?: a recalibration of the plausible range. J Dev Econ 39(2):333–364
Carlaw K, Lipsey RG (2003) Productivity, technology, and economic growth: what is the relationship? J Econ Surv 17(3):457–495
Central Statistical Agency (2005a–2018a) Agricultural sample survey: volume I—report on area and production of crops (Private Peas-ant Holdings, Meher Season). Federal Democratic Republic of Ethiopia, Addis Ababa
Central Statistical Agency (2005b–2018b) Agricultural sample survey: volume III—report on farm management practices (Private Peas-ant Holdings, Meher Season). Federal Democratic Republic of Ethiopia, Addis Ababa
Central Statistical Agency (2005c–2018c) Agricultural sample sur-vey: volume II—report on livestock and livestock characteristics (Private Peasant Holdings, Meher Season). Federal Democratic Republic of Ethiopia, Addis Ababa
Central Statistical Agency (2017a) Country and Regional Level Con-sumer Price Indices. Federal Democratic Republic of Ethiopia, Addis Ababa
Central Statistical Agency (2017b) Consumer Price Survey. Federal Democratic Republic of Ethiopia, Addis Ababa
Central Statistical Agency (2017c) Producer Price Survey. Federal Democratic Republic of Ethiopia, Addis Ababa
Deaton A, Muellbauer J (1980) An almost ideal demand system. Am Econ Rev 70:312–326
Delgado CL (2003) Rising consumption of meat and milk in devel-oping countries has created a new food revolution. J Nutr 133(11):3907S–3910S
Delgado CL, Rosegrant MW, Steinfeld H, Ehui SK, Courbois C (1999) Livestock to 2020: the next food revolution (vol 61). Intl Food Policy Res Inst, Washington
Dercon S, Hill RV (2009) Growth from agriculture in Ethiopia: iden-tifying key constraints. Paper prepared as part of a study on Agri-culture and growth in Ethiopia by the Department of International Development (DFID)
Dibley D, Boughton D, Reardon T (1995) Processing and preparation costs for rice and coarse grains in urban Mali: subjecting an Ipse Dixit to empirical scrutiny. Food Policy 20(1):41–50
Engida E, Tamru S, Tsehaye E, Debowicz D, Dorosh P, Robinson S (2011) Ethiopia’s growth and transformation plan: a computable general equilibrium analysis of alternative financing options. ESSP II Working Paper 30, Addis Ababa
Fufa B, Behute B, Simons R, Berhe T (2011) Strengthening the tef value chain in Ethiopia. Mimeo, Addis Ababa
Gollin D (2011) Crop production in Ethiopia: assessing the evidence for green revolution. International Growth Centre, London
Kennedy E, Reardon T (1994) Shift to non-traditional grains in the diets of East and West Africa: role of women’s opportunity cost of time. Food Policy 19(1):45–56
Mandefro H, Jerven M (2015) GDP growth and poverty decline in Ethiopia: who owns the numbers?. Addis Standard, Addis Ababa
Minten B, Tamiru S, Stifel D (2014) Structural transformation in cereal markets in Ethiopia. J Dev Stud 50(5):611–629
Minten B, Tamru E, Engida E, Tadesse K (2016a) Feeding Africa’s cities: the case of the supply chain of tef to Addis Ababa. Econ Dev Cult Change 64(2):265–297
Minten B, Tamru S, Engida E, Tadesse K (2016b) Transforming staple food value chains in Africa: the case of tef in Ethiopia. J Dev Stud 52(5):627–645
National Bank of Ethiopia (NBE) (2017) Annual report 2016–2017. Federal Democratic Republic of Ethiopia, Addis Ababa
Pingali P (2007) Westernization of Asian diets and the transformation of food systems: implications for research and policy. Food Policy 32(3):281–298
Reardon T, Timmer CP (2007) Transformation of markets for agricul-tural output in developing countries since 1950: how has thinking changed? In: Evenson RE, Pingali P (eds) Handbook of agricul-tural economics, vol 3. Elsevier Press, Amsterdam, pp 2808–2855
Reardon T, Chen K, Minten B, Adriano L (2012) The quiet revolution in staple food value chains: Enter the dragon, the elephant, and the tiger. Asian Development Bank, International Food Policy Research Institute, Mandaluyong City, Philippines
Solow RM (1957) Technical change and the aggregate production func-tion. Rev Econ Stat 39:312–320
Stifel D, Woldehanna T (2014) Utility-consistent poverty in Ethiopia, 2000–11: welfare improvements in a changing economic land-scape, WIDER working paper 2014/125
Subramanian S, Deaton A (1996) The demand for food and calories. J Political Econ 104:133–162
Tafere K, Taffesse AS, Tamru S (2010) Food demand elasticities in Ethiopia : estimates using household income consumption expenditure (HICE) survey data, ESSP Working Paper 11. Addis Ababa, Ethiopia
Taffesse AS, Dorosh P, Gemessa SA (2012) Crop production in Ethio-pia: regional patterns and trends. In: Dorosh P, Rashid S (eds) Food and agriculture in Ethiopia: progress and policy challenges. University of Pennsylvania Press, Philadelphia, pp 53–84
World Bank (2015) Ethiopia’s great run: the growth acceleration and how to pace it. Report no. 99399-ET. Washington DC: World Bank
Zhang X, Fan S (2004) How productive is infrastructure? A new approach and evidence from rural India. Am J Agric Econ 86(2):492–501
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Table 8 Factor shares used in growth accounting analyses
Sources: aEngida et al. (2011); bAuthors’ computations using Engida et al. (2011) and the AGP baseline survey data set from 2011
Factor Factor shares in Ethiopia SAM of 2009a
Factor shares used in analysesc
Labor 0.713 0.643Capital 0.004 0.004Land 0.216 0.202Fertilizer 0.061 0.057Improved seeds 0.039Pesticides 0.004 0.004Irrigation 0.025Extension services 0.024Services 0.002 0.002Total 1.000 1.000