Adoption Analysis of Maize

Factors Determining the Adoption of New Maize Varieties in the Hills of Nepal

DR. SUDARSHAN PRASAD REGMI

Dr. Y. S. Parmar University of   Horticulture and Forestry,

 Nauni, (Solan) –Himachal Pradesh, India

Abstract

Maize is the second most important food crop grown in Nepal. It is cultivated predominantly under rainfed conditions in the hills. Recent studies suggested that one of the major reasons of the nearly-stagnating productivity growth is due to slow adoption of improved high-yielding varieties. The main objective of this paper was to identify factors that determine the probability and extent of adoption of improved maize. The study is based on the interview of randomly selected 316 farmers in Sindhupalchowk and Ramechhap districts in the hills of the Central Development Region. Two econometric models, Probit and Tobit, were used to identify factors that influence farmers’ decision-making in the adoption of new technology. The results indicate that age and education of farmers, number of contacts with extension agent and quality of land significantly increase the probability of adoption of new varieties. Tobit results indicate that a literate Tamang farmer living within 5 km of motorable road is estimated to adopt new variety in 0.22 ha of land while an illiterate farmer of the same ethnic group living beyond 5 km from road is estimated to adopt new variety in 0.09 ha only. The findings suggest that people living in remote areas are least benefited from technology generation and dissemination. Hence, the public sector needs to put extra effort to create awareness and offer opportunities to those disadvantaged groups to improve their livelihood and increase productivity at the national level.

Key Words

Adoption, Probit, Tobit, Ethnicity, Maize, Socioeconomic

Introduction

Maize is the second most important staple food crop cultivated predominantly under rainfed conditions in the hills of Nepal. The area distribution of maize by agro-ecological zones indicates that the mid-hills occupy 70 percent; followed by the terai & inner-terai (22 percent); and the high-hills (8 percent). Despite the efforts initiated to increase the production of maize since the introduction of High Yielding Varieties (HYVs) during 60s’, progress has not been very encouraging in the past. The statistics published by the Ministry of Agriculture and Cooperatives (MOAC) and Central Bureau of Statistics (CBS) show that the maize production growth in the hills remained lower than population growth until 1997. The trend was changed for the better since then and the production growth was 2.92% per annum, which was higher than population growth of 1.97% per annum during 2000-2004 (Figure 1). However, a substantial part of this growth was due to growth in area under the crop. There is little scope of maize area expansion without deteriorating the environment as almost all the cultivable area in the hills is already under cultivation. Therefore, any increase in maize production must largely come through the adoption of high-yielding varieties.

A wide gap among the potential yield (5 ton /ha), attainable yield (3.5 ton/ha) and actual yield (1.9 ton/ha) has been observed (Tiwari, et al., 2005).This indicates that there is a wide room to get the most out of the potential yield. The adoption of modern varieties and associated technologies seem to offer an opportunity to substantially increase output and income. However, in the past, there had been limited success in the adoption of HYV and component technologies of maize cultivation due to several socio-economic and institutional constraints (Paudyal et. al., 2001). The present paper sought to identify the factors that determine the adoption of improved HYV maize in two hill districts of Nepal.

Materials and methods

The study was conducted in Ramechap and Sindhupalchowk districts, located in the hills of the Central Development Region. Of the two sample districts, Ramechhap represented an area with better market access and Sindhupalchowk, an area where NGOs are active. A random sampling procedure was used to select 316 households from 21-village development committees for interview.

The data at the district and combined level were analyzed using Probit (Goldberger, 1964) and Tobit (Tobin, 1958) models. The binary Probit model was used to estimate the marginal effect of the explanatory variables on adoption of improved maize varieties. Tobit (censored normal regression) model was used to estimate the probability of adoption and marginal change in the area planted to improved maize varieties with change in the value of explanatory variables. The coefficients of the regression models were estimated through the maximum likelihood estimation (MLE) techniques using the LIMDEP-7 (Greene, 1995) econometric package.

Results and discussion

Out of the 316 sample households, 176 (56%) reported having cultivated improved maize varieties in the past. Reported adoption rate was higher in Ramechhap (57%) than in Sindhupalchowk (54%) district. One of the reasons of higher level of adoption in Ramechhap might be its better access to market. Among the different ethnic groups, the Brahmin were the ones to adopt most. More than 68% of Brahmins adopted improved technology, followed by Majhi (67%) and Newar (62%). A lower proportion (44%) of Dalit (lower caste) households adopted improved technology.

The results of probit analyses for district and combined data are presented in Table 1. The coefficients for age, land type and education (literacy) of farmer were positive and significant for Ramechhap. Distance from the road had a negative significant effect on the adoption of improve maize, indicating that as the distance from the road head increases, the probability of adoption would decrease. Coefficients for total area cultivated and family size were positive and significant for Sindhupalchowk district whereas coefficients for both of the variables were non-significant for Ramechhap. The combined data analysis shows that education of farmer, number of extension contact and type of land has positive significant impact on the adoption of improved maize. The other positive variables for adoption were total number of training taken, total area of the farmer, membership of an organization and borrowing.

Table 1: Probit model estimates of coefficients and marginal effects for determinants of improved maize cultivation in two hill districts of Nepal, 2004.

Variables	Ramechhap		Sindhupalchowk		Combined
	Coefficient	Marginal effect	Coefficient	Marginal effect	Coefficient	Marginal effect
Intercept	-1.26**	-0.49**	-0.56	-0.22	-0.81**	-0.32**
Age	0.03**	0.01***	-1.0E-03	-3.0E-04	0.01	3.0E-03
Literacy @	0.64**	0.25**	0.38	0.15	0.38**	0.15**
Family size	-0.08	-0.03	0.11**	0.04**	-0.07**	-0.03**
Distance from Road	-0.03*	-0.01*	-0.03	-0.01	-0.01	-4.1E-03
Annual income	-8.8E-07	-3.4E-07	1.6E-06	6.5E-07	6.4E-07	2.5E-07
Total trainings	0.03	0.01	0.03	0.01	0.10	0.04
Member of organization @	0.48	0.19	0.23	0.09	0.13	0.05
No. of Extension contact	0.02	0.01	0.02	0.01	0.02**	0.01***
Total Area	-0.37	-0.14	1.17***	0.46***	0.27	0.11
Loan borrowed, @	0.02	0.01	0.03	0.01	0.01	2.2E-03
Types of Land @	0.78***	0.30***	0.10	0.04	0.45***	0.18***
Likelihood ratio	36.09***		29.98***		40.57 ***
Number of observation	156		160		316

Note: @ = Dummy variable, *, ** and *** indicate significance at 0.1 level, 0.05 level, and 0.01 level, respectively.

The sample households were sub-divided into four ethnic groups, viz. Brahmin-Chhetri; Newar; Tamang-Majhi; and Dalit; and were used as dummy variables in the Tobit model to analyze the effect of ethnicity in the adoption of technology. The rest of the variables used in the Tobit model were similar to those used in the Probit model. The result shows that the coefficients derived from the Tobit model with ethnicity as dummy variable for combined data had similar signs with that of the Probit model except for borrowing (Table 2). However, they differ in the district level analysis. None of the ethnicity dummy was significant, indicating that there is no significant difference on the adoption behavior of farmers belonging to different ethnic groups if the technology is available to all of them.

Table 2: The Tobit model estimates of index function and marginal effects for determinants of improved maize cultivation in two hill districts of Nepal, 2004.

Variables	Ramechhap		Sindhupalchowk		Combined
	Coefficient	Mar. effect	Coefficient	Mar. effect	Coefficient	Mar. effect
Constant	-0.35**	-0.21**	-0.30**	-0.19**	-0.25**	-0.15**
Age	0.01***	3.6E-03***	1.2E-03	8.0E-04	2.4E-03*	1.4E-03*
Literacy @	0.22***	0.13***	0.17***	0.11***	0.16***	0.10***
Family size	-0.03***	-0.02***	-0.01	-0.01	-0.02***	-0.01***
Distance from Road	-0.01**	-0.01**	-0.01	-0.01	-0.01***	-0.01***
Annual income	1.0E-06**	6.3E-07**	1.0E-06	6.5E-07	1.0E-06***	6.5E-07***
Total no. of training	0.02	0.01	0.02	0.01	0.03	0.02
Member of organization @	0.07	0.04	0.03	0.02	0.02	0.01
No, of extension contact	0.01**	2.9E-03**	8.0E-04	5.0E-04	0.01***	3.2E-03***
Total area	0.10*	0.06	0.31***	0.20***	0.21***	0.13***
Loan borrowed @	0.04	0.02	-0.02	-0.01	-4.0E-04	-2.0E-04
Type of Land	2.4E-03	1.4E-0	0.06	0.04	0.03	0.02
Tamang_ Majhi @	0.08	0.05	0.03	0.02	0.04	0.03
Newar @	-0.05	-0.03	0.04	0.03	-0.02	-0.01
Dalit @	0.23	0.14	-0.10	-0.06	0.04	0.02
Sigma	0.28***		0.26***		0.28***

Note: @ Dummy variable; *, ** and *** indicate significance at 0.1 level, 0.05 level, and 0.01 level, respectively.

The Tobit model coefficients were further analyzed to estimate the probability of adoption and area to be cultivated with improved maize varieties as affected by literacy of the farmer; distance of the farm from nearest road; and ethnicity. The probability of adoption and estimated expected area to be cultivated with improved maize is consistently higher for literate farmers and those living up to 5 km from the road than for illiterate farmers and those living farther from road (Table 3). The combined result for two districts shows that a literate Tamang/Majhi farmer that lives within 5 km from road has 73% probability of adoption and is expected to cultivate improved maize in 0.22 unit of land compared to an illiterate farmer of the same ethnic group living beyond 5 km from road.  Such a farmer has 45% probability of adoption and is expected to cultivate improved maize in 0.09 units of land.

Table 3: Probability of adoption and expected area under improved maize by literacy, distance from road and ethnicity in two hill districts of Nepal, 2004.

District	Ethnicity	Literate				Illiterate
		≤ 5 km from road		>5 km from road		≤ 5 km from road		>5 km from road
		Probab.	Exp.area	Probab.	Exp.area	Probab.	Exp.area	Probab.	Exp.area
Ramechhap	Tamang/Majhi	0.85	0.31	0.81	0.27	0.60	0.15	0.54	0.12
	Newar	0.72	0.21	0.67	0.18	0.42	0.08	0.36	0.07
	Dalit	0.95	0.45	0.93	0.41	0.79	0.25	0.74	0.22
Sindhupalchowk	Tamang/Majhi	0.72	0.20	0.67	0.17	0.47	0.09	0.41	0.07
	Newar	0.75	0.21	0.69	0.18	0.50	0.10	0.43	0.08
	Dalit	0.55	0.12	0.48	0.10	0.29	0.05	0.24	0.04
Combined	Tamang/Majhi	0.73	0.22	0.67	0.18	0.52	0.12	0.45	0.09
	Newar	0.66	0.17	0.59	0.14	0.44	0.90	0.63	0.16
	Dalit	0.72	0.21	0.66	0.18	0.51	0.11	0.44	0.09

Concluding remarks

The education of a farmer and his contact with extension agents are identified to be driving forces in the adoption of new technology. Technology adoption can be speeded up by creating awareness among the farmers. Informal education campaign and hands-on training at the grassroots level should therefore be initiated/ strengthened in the remote areas. Distance from the road had a negative relationship with technology adoption, an evidence of the unavailability of the seed and component technology which is primarily due to transportation costs and lack of awareness. Community-based seed production could be a vehicle for faster dissemination of improved maize in the far-flung communities. Investment in education and infrastructure will no doubt have a positive impact on technology adoption and the livelihood of people residing in the remote hills. The NGOs and private sector need to be mobilized for the delivery of inputs, including much needed credit to remote areas as government effort alone will not be able to achieve this goal in reasonably short period of time.

Acknowledgment

This study was supported through the small grants project under the Hill Maize Research Project, Nepal. We express our profound gratitude and sincere appreciation to all farmers in  the study area, and to Dr. Carlos Urrea, (Former HMRP Leader), Dr. NP Rajbhandari (Former HMRP Adjunct Scientist), Dr. DP Sherchand (Former NMRP Coordinator); TUKI Sunkoshi, Sindhupalchowk; and District Agriculture Development Offices in Ramechhap and Sindhupalchowk.

References

Central Bureau of Statistics (CBS) 2003. Statistical Year Book of Nepal, 2003. CBS, Kathmandu.

Greene, W.H. 1995. Limdep Version 7.0 Manual. Econometric Software Incorporated.

Goldberger, A.S. 1964. Econometric Theory, Wiley, New York.

Mc Donald, J. F. and R. A. Moffitt. 1980. The use of Tobit Analysis. Review of Economics and Statistics, LXII (May): 318-321

Ministry of Agriculture and Cooperatives (MOAC) 2004. Statistical Information on Nepalese Agriculture. MOAC, Kathmandu.

Paudyal K.R., J.K. Ransom, N.P. Raqjbhandari, K. Adhikari, R. Gerpacio and P. Pingali, 2001. Maize in Nepal: Production Systems, Constraints and Priorities for Research, Kathmandu, NARC and CIMMYT.

Tiwari, TP, J. Ransom, C. Urrea, F. Gonzalez, G. Ortiz-Ferrara, N.P. Rajbhandari, K. Adhikari and D.P. Sherchan. Improving the Dissemination of new Maize Technologies to Small Scale Farmers in Nepal. Paper prepared for 9ARMW presentation 5-9 September, 2005.

Tobin, J. 1958. Estimation of Relationship for Limited Dependent Variables. Econometrica, 26. 26-36.