Productivity and Efficiency Analysis of Maize under Conservation Agriculture in Zimbabwe

Please download to get full document.

View again

of 20
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report



Views: 6 | Pages: 20

Extension: PDF | Download: 0

Productivity and Efficiency Analysis of Maize under Conservation Agriculture in Zimbabwe Kizito Mazvimavi a, Patrick V Ndlovu b, Henry An c and Conrad Murendo d Selected Paper prepared for presentation
Productivity and Efficiency Analysis of Maize under Conservation Agriculture in Zimbabwe Kizito Mazvimavi a, Patrick V Ndlovu b, Henry An c and Conrad Murendo d Selected Paper prepared for presentation at the International Association of Agricultural Economists (IAAE) Triennial Conference, Foz do Iguaçu, Brazil, August, Copyright 2012 by [Mazvimavi K., Ndlovu P.V., An H. and Murendo C.]. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies. a Head, Impact Assessment Office, International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Hyderabad India b Graduate Student, Department of Resource Economics and Environmental Sociology, University of Alberta, Edmonton, AB, Canada c Assistant Professor, Department of Resource Economics and Environmental Sociology, University of Alberta, Edmonton, AB, Canada d PhD Student, Department of Agricultural Economics and Rural Development Georg-August-University Göttingen, Göttingen, Germany Productivity and Efficiency Analysis of Maize under Conservation Agriculture in Zimbabwe Kizito Mazvimavi 1, Patrick V Ndlovu 2, Henry An 3 and Conrad Murendo 4 Abstract This study sought to evaluate the performance of conservation agriculture (CA) technologyessentially comparing productivity and efficiency levels in maize production in CA and conventional farming. The analysis is based on a three year panel sample of smallholder farming households and employing a stochastic production frontier model compare productivity and technical efficiency between CA and conventional farming. Study results indicate that CA technology is implemented in relatively smaller plots than conventional farming (0.36ha compared to 0.85ha) but has a significant contribution to total maize production, on average 50% of output share. Output elasticities indicate positive responses for labor and seed in CA, and negative responses in conventional farming. On the other hand, there are negative responses to land and draft in CA. Fertilizer has a greater positive response in CA than in conventional farming. Overall returns to scale are similar for CA and conventional farming (0.84 and0.89 respectively). There is evidence of technical progress in CA for the three year panel period. Technical progress has been land-saving but seed and fertilizer-using in CA, while land-using and seed-saving in conventional farming. Joint frontier estimates indicate that farmers will produce 39% more in CA compared to conventional farming. Technical efficiency levels are generally the same (about 68%) for both technologies. Two-thirds of farmers achieve efficiency scores in the 60-80% range both CA and conventional farming technologies. These results show significant yield gains in CA practices and significant contributions to food production. CA is landsaving, and this is an important issue for land constrained farmers because they can still have viable food production on smaller area. But high labor demands in CA present some problems in adoption, particularly for the poorer farmers. Key words: Conservation agriculture, productivity, efficiency, technical change I. Introduction Maize production is an important component of food security and livelihood for smallholder farming communities of Zimbabwe. The maority of smallholder farmers grow maize primarily for subsistence using conventional farming technology based on ox-drawn plow for tillage purposes. The challenge in Zimbabwe s smallholder agricultural sector is to raise the productivity of the staple cereal as a way of solving food insecurity problems. The per capita maize production is steadily declining, and this has been attributed to significant decline in yields over the years from 1500 kg/ha in the early 1990s to 500kg/ha after 2000 (Government of Zimbabwe, 2002). Similar to most parts of sub-saharan Africa, agricultural productivity levels in Zimbabwe have fallen due to land degradation as a result of many years of erosive cultivation, declining soil fertility as farmers fail to replenish soil fertility (Mano, 2006). 1 Head, Impact Assessment Office, International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Hyderabad India 2 Graduate Student, Department of Resource Economics and Environmental Sociology, University of Alberta, Edmonton, AB, Canada 3 Assistant Professor, Department of Resource Economics and Environmental Sociology, University of Alberta, Edmonton, AB, Canada 4 PhD Student, Department of Agricultural Economics and Rural Development Georg-August-University Göttingen, Göttingen, Germany 1 The response to this food crisis in Zimbabwe has been the wide scale relief distribution of food aid and direct agricultural input assistance without an exit strategy for sustaining some of the new technologies promoted within the context of relief aid (Rohrbach et al, 2005; DFID, 2009). As part of these relief and recovery programs, research and development initiatives have seen the introduction of a specific set of technology options that aim to improve and stabilize crop yields while preserving soil and water, while using precision methods to apply inputs. These set of technology options is defined as conservation agriculture (Thierfelder and Wall, 2010; Twomlow et al., 2008). But, the key to a prolonged increase in agricultural production is to improve productivity, which can be achieved through better technology and efficiency. In Zimbabwe there has been maor investments and policy drive towards CA as a way of improving productivity through efficient use of production inputs, improved management, timeliness of operations and conserving the soil. However, in the past increase in land productivity has come from intensification of agricultural production and the adoption of yield enhancing technologies especially modern high yielding varieties and fertilisers. Higher efficiency gives subsistence farmers the opportunity to produce more output using the current level of inputs especially land which is limited in supply. Gains in output through productivity growth have become increasingly important in Zimbabwe as opportunities to bring additional virgin land into cultivation have significantly diminished in recent years. So far there is no empirical evidence to show that CA can indeed lead to efficiency gains which can increase productivity that is crucial for improving livelihoods of smallholder farmers in Zimbabwe. The few studies that have assessed the effect of CA adoption on production efficiency (Solis, 2005, Oduol et al., 2011, Musara et al., 2012) have used cross sectional data. The studies have concluded that adoption of CA practices push smallholder farmers closer to their production frontier and an improvement of human capital variables such as access to extension and education can significantly reduce inefficiencies. Given the nature of CA and the fact that agronomic benefits from soil improvement are only realised in the long term, the use of panel data is more appropriate for a realistic assessment of impact. Through monitoring farmers who have adopted CA over time, the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) has created a database upon which this study will be based. The paper is structured as follows: Section II is a review of the literature on productivity measurement and section III develops the theoretical and econometric model for estimating productivity impacts. Section IV describes the data used in the study including sample selection issues. Section V is a discussion of diagnostic and model specification issues in the econometric model. Section VI reports the maor empirical findings. The summary follows in the last section. 2 Section II. Literature review CA practices in Africa A comprehensive review of conservation CA practices in Zimbabwe, and other Southern African countries is given by Mazvimavi (2011). CA in Zimbabwe is largely practiced by smallholder farmers using small farm implements such as the hand hoe to create planting basins. Though specifications may vary, CA technologies typically involve agricultural management practices that prevent degradation of soil and water resources and thereby permit sustainable farm productivity without environmental degradation (Haggblade et al., 2004; Wysocki, 1990; ECAF, 2002). Farmers and agencies working to improve farm productivity have experimented with a broad range of these soil and water conservation technologies that are collectively known as CA. Tsegaye et al., (2008) assess the impacts of conservation agriculture on land and labor productivity in Ethiopia. Their study analyzes the adoption of the different components of CA and finds that the initial decision to adopt CA is influenced by regional location, family size, access to extension, and formal education. They also find a positive relationship between land productivity and use of CA components such as herbicide application. Hassane et al., (2000) evaluate the impact of planting basin, and use of fertilizer and manure on millet crops in Niger. Their study finds that over a five year period from 1991 to 1996, farmers experienced yield gains of up to 511%. Similarly, significant yield gains are also noted in a study in Zambia by Haggblade and Tembo (2003) who note that farmers who dug planting basins and applied crop residues and fertilizer achieved 56% yield gains in their cotton fields and 100% yield gains in their maize fields. Gowing and Palmer (2008) examine evidence of CA benefits amongst small-scale farmers in Africa and conclude that CA does not overcome constraints on lowexternal-input systems. They note that CA will deliver the productivity gains that are required to achieve food security and poverty targets only if farmers have access to fertilizers and herbicides. They further asset that adoption of CA by small-scale farmers is likely going to be partial as opposed to full adoption. While there is evidence of CA gains in the literature, there are also studies that present a sharply contrasting assessment of CA impacts. Giller et al. (2009) suggests that empirical evidence is not clear and consistent on CA contributions to yield gains. Their study notes concerns that include decreasing yield in CA, increased labor requirements when herbicides are not used, a shift of the labor burden to women, and problems with mulching requirements due to its shortage or competing use as livestock feed. They also note that there are many cases where adoption of CA was temporary and only lasted for the course of active promotion of the technology by NGOs and research but was not sustained beyond that. 3 Technical Efficiency and Productivity Growth The measurement of technical efficiency and productivity growth is an area of study that has attracted the interest of a number of researchers since the work of Farrell in 1957 (Farrell, 1957). Technical efficiency is ust one component of overall economic efficiency, i.e. producing maximum output given the level of inputs employed (Kumbhakar and Lovell 2000). Efficiency change essentially contributes to productivity growth. Efficiency can be considered in terms of the optimal combination of inputs to achieve a given level of output, that is input-orientation efficiency, or the optimal output that could be produced given a set of inputs, that is output-orientation efficiency. Productivity assessment is often associated with measurement of technical change. The work of Battese and Coelli (1988, 1992, 1995) has made notable contributions on measurement of production efficiency using stochastic production frontier approach. Khumbakar and Lovell (2000) proposes an econometric method that is based on a primal approach where shifts in the production frontier are due to technical change. It is often important to interpret results of efficiency and productivity analysis in the context of the time period analyzed, and also consider issues such as the degree of sample homogeneity, output aggregation, and use of different methodologies in the analytical process. Total factor productivity growth is defined as growth in output that is not explained by change in inputs. Following this definition and assuming that production is not always on the frontier, change in productivity can be decomposed into two separate components: a) movements towards or away from the frontier due to changes in technical efficiency; and b) shifts in the frontier due to the effect of technological innovations or progress. Effects of scale changes can also be incorporated in this measure (Coeli et al.,2005) Parametric and non parametric approaches A non-parametric approach to frontier, the Data Envelopment Approach (DEA) was developed by Charnes, Coopers and Rhodes (1978). The parametric approach was developed simultaneously by Aigner, Lovell and Schmidt (1977) and Meeusen and van den Broek (1977) who proposed the stochastic frontier production function. Both approaches are used in empirical work. However, a weakness associated with the DEA approach is that all deviations from the frontier are associated with inefficiency. In agriculture this assumption is restrictive considering that production is variable due to factors such as weather, pests and diseases. The stochastic production frontier on the other hand allows for error in measurement. 4 Section III: Development of theoretical and econometric model This study uses a stochastic production frontier to estimate productivity and technical efficiency. To estimate technical efficiency, a oint frontier is used since this is a comparative analysis of two technologies. Data for the two technologies is pooled so that technical efficiency predictions are derived from the same data. This is based on discussions by Battese et al., (2004) on comparing different groups in technical efficiency estimation. OLS regressions and Stochastic frontier models The study will use OLS regressions to model maize production and retrieve output elasticities and returns to scale associated with CA and conventional farming. Two separate models are estimated (for CA and non CA) using a structural form indicated in the translog production function in equation 1. y it 0 J 2 x it k x itxkit tt ttt tx itt 1 2 J J k where yit is the log of the output produced, the subscript i = 1, 2,, N denotes households in the panel data, t = 1, 2,, T are time periods, and, k = 1, 2,, J are the inputs used, represented by vector x in farm production. Technical change is neutral with respect to inputs if, and only if, t = 0, and absent if, and only if, t=tt=t=0. The panel stochastic frontier model to predict technical efficiency is given in equation 2, with the same specification as equation 1 except that the error term is composed of two independent elements: vit ~ iid N(0, v 2 ) is the random noise error component and uit 0 is the technical inefficiency error component. In the econometric estimation, a oint panel is used, pooling observations for CA and conventional farming, and incorporating a dummy variable to control for these technologies. y it 0 J 2 x it k x it xkit tt ttt t x itt vit 1 2 J J k The inefficiency effects model provides some explanations for the variations in efficiency levels among farmers. Following the stochastic production frontier model in Equation 2, it is assumed that the inefficiency effects are independently distributed and uit arises by truncation at zero of the normal distribution with mean, t, and variance,, where t is defined by 2 u M 0 mzmt tt m1 (3) t where z is a vector of farm specific inefficiency related variables (m=1,, M), at time period t, and coefficient are unknown parameters to be estimated. Since the dependent variable in the inefficiency model is a measure of inefficiency, a positive J J e it u (1) it (2) 5 sign on a parameter indicates a negative efficiency effect. A one stage approach that uses a maximum likelihood estimator is used to estimate the production and inefficiency effects simultaneously. Variables for the direct factors of production are land (A), labor (L), draft animals (K), fertilizer (F), and seed (S). The output (Y) for the production function is maize produced in kgs. Land is total cultivated area in hectares. Labor is total farm labor available in the household, expressed in male adult equivalent units. Variables hypothesized to be explanatory factors of technical efficiency include ; gender (dummy variable taking the value of 1 if male headed household, zero if female headed), age and education of household head, asset endowments, and access to draft power. A time variable is included to estimate the effect of time on technical efficiency. Land and labor are also included in the efficiency model. The model used in the study assumes time varying technical efficiency, using a truncated frontier model. Section IV: Data This study makes use of ICRISAT panel data from household surveys collected since 2008 in Zimbabwe. The panel study aimed to examine CA adoption practices including labor allocation, technology adoption determinants, and productivity impacts observing the same farmers in successive seasons of real CA practice in non experimental setting. The study makes comparison of CA technology with alternative conventional farming practices for the same households (i.e. a household practices both technologies). The data was collected in 15 rural districts in Zimbabwe. Table 1 shows the average number of households interviewed in the full survey sample and the selected sample (used for this particular study). During the panel period, there were incidences of attrition as some households could not be re-interviewed in successive seasons of the surveys. As a consequence, the panel data used in this study is un-balanced. This may open doors to some econometric problems associated with attrition bias. A possible solution to attrition bias is to use dynamic panel data models. However, this study does not tests for attrition bias nor make use of dynamic panel data models. There were instances where some farmers did not produce maize in particular seasons, or where the maize crop was completely wiped out by drought. As a result, this study makes use of a sub sample of the original panel household sample. This sub sample considers maize producing households and excludes observations where no maize was produced. Further details on sample selection are discussed in the proceeding sub-section. 6 Table 1. Survey sample and selected sample for study Technology Survey sample Selected sample CA Conventional Combined CA Conventional Combined Total observation Source: ICRISAT Conservation Agriculture panel data Table 2 gives some descriptive statistics of the production variables and factors hypothesized to explain technical efficiency in maize production. Output refers to total maize produced in kilograms. In this study, aggregation of output from different plots, as well as aggregation of inputs is done. Aggregation is used in this case by making implicit assumptions on seperability. Table 2. Summary Statistics for factors of production and efficiency factors Production variables Efficiency variables Year Output Area Labor Draft Seed Fertilizer Gender Age School Experience Ill Assets CA average Conventional average Average average Source: ICRISAT Conservation Agriculture panel data The efficiency variables include gender, age, education level, farming experience of the household head, and presence of chronically ill persons in the household (dummy variable proxy for impact of HIV/AIDS, named Ill). Gender and illness are proportion of households (multiply by 100 to express as percentage). Asset endowments are expressed as an index which captures information on the availability of farming implements e.g. plows, cultivators, hoes, in a household. In general there is not a lot of variation in input use for the thre
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!