Technical Efficiency of Teff Farms Controlling for Neighborhood effects in Ethiopia

doi:10.21203/rs.3.rs-30863/v1

Download PDF

Research

Technical Efficiency of Teff Farms Controlling for Neighborhood effects in Ethiopia

https://doi.org/10.21203/rs.3.rs-30863/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Teff can be called the lifeblood of the majority of Ethiopians. However, Teff production is constrained by farmers’ own and neighbor farmers’ socioeconomic, demographic, intuitional, and farm characteristics that decrease Teff productivity. To fill this gap, a combination of modern technologies with a better level of farm efficiency becomes indispensable. Hence, this study aims to examine the technical efficiency of Teff farms controlling for neighborhood effects in Ethiopia using 858 Teff farmers. The stochastic frontier translog production and Spatial Lagged Explanatory (SLX) models are estimated using two waves of panel data from the Ethiopian socioeconomic survey. The unobserved heterogeneity has been controlled for by a correlated random effect estimation. The results show that the average technical efficiency of Teff farm is 78.5%, indicating that farmers have about 21.5% likelihood for improving their Teff farm efficiency, averagely. Besides, the estimated correlated random effect of the spatial lagged explanatory model indicates that gender, age, education, family size, community participation, planting method, average neighbors' community participation, average neighbors' planting method, average neighbors' mineral fertilizer application, and average neighbors' use of improved seed were found to significantly affect the level of technical efficiency of Teff farms in the study area. Moreover, we show that controlling for the neighborhood effects in the estimation of technical efficiency of Teff production statistically and significantly changes the inferences. Therefore, neighborhood effects can be the best alternative platform for farm knowledge transmission and use as a development actor if properly managed and organized.

Other Economics

CRE estimation

Ethiopia

Neighborhood effect

SLX model

Technical efficiency

Teff

Translog production function

Table 1: Summary statistics of covariates used in the stochastic frontier model and SLX models

Variables	Mean	SD*	Min	Max
Inputs for Teff production
Yield (ton ha^-1)	1.21	0.27	0.64	2.05
Land (ha.)	1.24	0.75	0.20	3.50
Seed (kg ha^-1)	30.77	9.58	10.86	50.99
Fertilizer (kg ha^-1)	86.29	82.94	0	312.70
Labor (Man Day ha^-1)	154.81	77.20	40.26	399.15
Oxen (number)	2	0.80	1	4
Sources of technical inefficiency
Farmers’ own characteristics
Gender ( = 1 if the household is male-headed)	0.87	0.34	0	1
Education of the household head (years of schooling)	2	2.95	0	13
Participation in extension service program (= 1 if they participated)	0.56	0.49	0	1
Community participation (= 1 if they participated)	0.76	0.43	0	1
Credit service (= 1 if they used)	0.31	0.46	0	1
Age of the household head (years)	50	13.41	22	99
Use of improved seed (= 1if they applied)	0.2	0.4	0	1
Planting method (=1 if they applied row planting)	0.12	0.33	0	1
Livestock holding (TLU)	4.27	4.31	0	30
Household size (number)	8	2.98	1	16
Application of mineral fertilizer (= 1 if they applied)	0.75	0.43	0	1
Distance to the nearest market (km)	67.8	42.09	2.9	201.6
Soil fertility of the plot (= I if the plot is fertile )	0.38	0.48	0	1
Average neighbor farmers’ characteristics
Neighbor farmers’ use of improved seed (= 1if they applied)	0.31	0.28	0	1
Neighbor farmers’ level of community participation (= 1if they participated)	0.74	0.27	0	1
Neighbor farmers’ use of mineral fertilizer (= 1if they applied)	0.73	0.32	0	1
Neighbor farmers’ extension service provision (= 1if they participated)	0.54	0.35	0	1
Neighbor farmers’ level of education (years of schooling)	2.16	1.83	0	10
Neighbor farmers’ Teff planting method (= 1if they applied row planting)	0.12	0.18	0	1

*SD indicates Standard Deviation, pooled data (N = 1716)

Source: Own analysis based on ESS (2014 and 2016)

Table 2: Time-varying model of stochastic frontier based on Battese and Coelli (1992)

Input variables	Estimated value	Robust Std. Err.	Z	P > \|z\|
lnLand	.1206475	.0352202	3.43	0.001^***
lnLabor	.2853783	.0617842	4.62	0.000^***
lnFertilizer	.0076977	.0037747	2.04	0.04^**
lnSeed	.4310508	.0970771	4.44	0.000^***
lnOxen	.0197815	.0617685	0.32	0.749
lnLand*lnLabor	-.0378225	.0169698	-2.23	0.026 ^**
lnLand*lnFertilizer	.0005562	.0011528	0.48	0.629
lnLand*lnSeed	-.0081779	.0253831	-0.32	0.747
lnLand*lnOxen	.0014157	.0144446	0.10	0.922
lnLabor*lnFertilizer	.0012113	.0020587	0.59	0.558
lnLabor*lnSeed	-.0183248	.0339553	-0.54	0.589
lnLabor*lnOxen	-.0209723	.0229249	-0.91	0.360
lnFertilizer*lnSeed	-.0043527	.0027864	-1.56	0.118
lnFertilizer*lnOxen	-.0021991	.0014632	-1.50	0.133
lnSeed*lnOxen	.0259242	.0351486	0.74	0.461
time (t)	.000335	.0000676	4.96	0.000^***
t*lnLand	-.000017	.0000152	-1.12	0.264
t*lnLabor	-.0003638	.0000406	-8.96	0.000^***
t*lnFertilizer	-4.93e-06	1.52e-06	-3.23	0.001^***
t*lnSeed	.0003362	.0000497	6.76	0.000^***
t*lnOxen	2.52e-06	.0000241	0.10	0.917
β₀(Constant)	-2.265807	.3205598	-7.07	0.000^***
/Lnsigma²	-5.30692	.0672164	-78.95	0.000^***
/ilgtgamma (γ)	-1.287981	.3307949	-3.89	0.000^***
µ	.2426185	.1206865	2.01	0.044^*
Ƞ (eta)	.3099621	.132728	2.34	0.020^*
sigma²	.0049572	.0003332
Gamma	.2161947	.0560547
sigma_µ²	.0010717	.000325
Sigma_v²	.0038855	.0002473
Wald chi²(19)	14001.90	0.0000^***
Log pseudo-likelihood	2078.0198

*** => significant at 1% level; ** => significant at 5% level; and * => significant at 10% level; N = 1716; number of panel groups = 858; t = 2 years. The dependent variable is a logarithm of Teff yield (lnYield). Source: Own analysis based on ESS (2014 and 2016) data

Table 3: Technical efficiency scores and Teff yield gap analysis due to technical inefficiency

Variable	Mean	SD*	Min	Max
Actual Yield (ton ha^-1)	1.213	0.266	0.641	2.05
Technical Efficiency	0.785	0.017	0.679	0.925
Potential yield (ton ha^-1)	1.516	0.325	0.799	2.545
Yield gap (ton ha^-1)	0.303	0.064	0.107	0.495

* = Standard Deviation, N = 1716, number of panel groups = 858, and t = 2 years

Source: Own analysis based on ESS (2014 and 2016) data

Table 4: The results of the CRE estimation of the SLX model

Parameters	Without neighborhood effects estimate (robust std. error)		With neighborhood effects estimate (robust std. error)
	Estimated value	SE⁺	Estimated value	SE⁺
Farmers own farm characteristics
Constant	0.5528113 ***	0.0171147	0.4341563***	(0.0125214
Gender ( = 1 if the household head is male)	-0.0105174 ***	0.0027761	-0.0051681***	0.0018365
Education (years of schooling)	-0.0026602***	0.0004304	-0.0017335***	0.0003007
Age (years)	-0.0059423***	0.0006466	-0.0030702***	0.0004516
Age^2	0.0000341 ***	5.94e-06	0.0000187***	4.17e-06
Family size (number)	-0.0040373***	0.0007878	-0.002803 ***	0.0004972
Livestock holding (TLU)	-0.0007244	0.0007324	-0.0004491	0.0004836
Extension (= 1 if they participated)	-0.0098264*	0.0051281	-0.0032904	0.0033652
Credit (= 1 if they used)	0.0032793	0.0045804	0.00066	0.0031472
Community participation (= 1 if they participated)	-0.0274244***	0.0055334	-0.0093939**	0.0037042
Improved seed (= 1if they applied)	-0.015873***	0.0045018	-0.0011051	0.0029046
Fertilizer usage (= 1 if they applied)	0.0006612	0.0067365	-0.0071943	0.0044972
Soil fertility ( = 1 if the plot is fertile)	-0.0188108***	0.0039822	-0.0056838 **	0.0027743
Planting method (=1 if they applied row planting)	-0.0402027***	0.00468	-0.0084955**	0.0034658
Distance to market (km)	-0.0000763	0.0000804	-0.0000805	0.000051
Distance to market^2	1.76e-08	4.31e-07	1.92e-07	2.65e-07
Average Neighbors’ farm characteristics
Z_Ext (neighbors’ extension service participation )	-	-	-0.0067002	0.0080697
Z_FertilUsage (neighbors’ fertilizer usage)	-	-	-0.0220136*	0.0122593
Z_ comParticipation (neighbors’ community participation )	-	-	-0.0483096***	0.0083671
Z_Improvseed (neighbors’ improved seed usage)	-	-	-0.1205222***	0.0045714
Z_Education (neighbors’ education)	-	-	0.0004379	0.0004519
Z_Planting (neighbors’ row planting usage)	-	-	-0.0448689***	0.0083998
Sigma_u	0.01395214		0.01382005
Sigma_e	0.0065134		0.00651933
Rho	0.82105975		0.81797673
Wald chi2	2136.00***		4172.46***
R²-within =	0.2904		0.7023
R²-between =	0.3439		0.5176
R²-overall =	0.2226		0.6457

*** => statistically significant at the 99% confidence level; ** => statistically significant at the 95% confidence level; * => statistically significant at the 90% confidence level; + => robust Standard Error; the dependent variable is technical inefficiency score; N = 858; and t = 2 years; neighborhood effect variables are denoted with a leading “Z_”

Source: Own analysis based on ESS (2014 and 2016) data

Table 5: Correlation matrix between Xit and Zit variables

Variables	Average neighbor farmers’ socioeconomics characteristics (neighborhood effects )
farmers’ own socioeconomics characteristics	improved seed	Community participation	fertilizer	extension	education	planting method
Gender	-0.0076	0.0427	0.0042	-0.0066	-0.0012	-0.0122
Education	0.0273	0.0377	0.0196	-0.0794	0.2945	0.0498
Age	0.0798	-0.0059	0.0741	0.0323	0.0309	0.0783
Credit	-0.0264	0.0320	0.2347	0.1865	-0.0090	-0.0172
Community participation	0.1087	0.2354	-0.0044	0.0579	0.0627	0.0632
Improved seed usage	0.1697	0.0585	-0.0361	0.0237	0.0498	0.1616
Planting method usage	0.2073	0.0392	-0.0102	0.0144	0.0410	0.2272
TLU	-0.0145	-0.1136	-0.0720	-0.0749	-0.0091	-0.0198
Household size	0.0893	0.0562	-0.0289	-0.0190	0.1188	0.1039
Fertilizer application	-0.0253	-0.0215	0.4397	0.3123	0.0234	0.0005
Distance market	-0.0610	0.0079	-0.2477	-0.2808	0.1251	-0.0409
Soil fertility of the plot	0.1221	0.0347	-0.0307	-0.0035	-0.0208	0.1262

Source: Own analysis based on ESS (2014 and 2016) data

Table 6: VIF test

Variable	VIF	1/VIF
Age	42.51	0.023525
Age^2	42.45	0.023556
Dismkt	10.83	0.092337
Dismkt^2	10.54	0.094855
Z_Ext	2.12	0.471089
Z_FertilUsage	2.08	0.479717
Ext	1.61	0.619550
FertilUsage	1.60	0.625637
Z _Improvseed	1.51	0.663607
Z _Planting	1.40	0.715134
Z _Education	1.31	0.765789
Z _comParticipation	1.29	0.773285
Planting	1.26	0.790800
Education	1.26	0.791895
Family size	1.25	0.799077
Improvseed	1.24	0.809642
ComParticipation	1.23	0.812060
Credit	1.10	0.905402
SoilFertility	1.07	0.935783
TLU	1.07	0.936682
Gender	1.05	0.954114
Mean VIF	6.18

Source: Own analysis based on ESS (2014 and 2016) data

Table 7: Hausman test

	Coefficients
	(b) FE	(B) RE	(b-B) Differences	Sqrt (diag(v_b – v_B)) S.E.
Age	-.0439849	-.0030344	-.0409505	.0024591
Age^2	-1.89e-06	.0000186	-.0000204	.0000224
HHSIZE	-.0002243	-.0030917	.0028675	.0002794
TLU	-.0001077	.0000597	-.0001674	.0001735
Ext	-.0021317	-.0046552	.0025235	.0009718
Credit	-.0003685	.0002427	-.0006112	.0008419
SWC	-.0030219	-.0114155	.0083936	.0010455
Improvseed	.0047258	-.0014693	.0061951	.0008747
FertilUsage	.0006042	-.0052613	.0058655	.0014233
Soilfertility	.0004982	-.0067517	.0072499	.0006704
SowingMethod	.0058268	-.0113201	.0171469	.0010461
Z _Ext	-.00032	.0007485	-.0010685	.0028309
Z _FertilUsage	-.0019523	-.0043786	.0024262	.0044736
Z _ comParticipation	-.001068	-.0300698	.0290018	.0026985
Z _Improvseed	.0004267	-.1170868	.1175136	.0034047
Z _Planting	-.0017314	-.0408359	.0391045	.0023141
b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(16) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 1591.51
Prob>chi2 = 0.0000

Source: Own analysis based on ESS (2014 and 2016) data

renamedeb91b.docx

Download PDF

Reviews received at journal
29 Apr, 2022
Reviewer #2 agreed at journal
23 Dec, 2020
Review #1 received at journal
09 Sep, 2020
Reviewer #1 agreed at journal
10 Aug, 2020
Reviewers invited by journal
03 Aug, 2020
Editor assigned by journal
23 May, 2020
Submission checks completed at journal
22 May, 2020
Editor invited by journal
22 May, 2020
First submitted to journal
20 May, 2020

You are reading this latest preprint version

Technical Efficiency of Teff Farms Controlling for Neighborhood effects in Ethiopia

Status:

Version 1

Abstract

Figures

Full Text

Tables

Supplementary Files

Status:

Version 1