Impact of land fragmentation, farm size, land ownership and crop diversity on profit and efficiency of irrigated farms in India

Land Use Policy 31 (2013) 397–405

Contents lists available at SciVerse ScienceDirect

Land Use Policy journal homepage: www.elsevier.com/locate/landusepol

Impact of land fragmentation, farm size, land ownership and crop diversity on profit and efficiency of irrigated farms in India A.V. Manjunatha a,b,∗ , Asif Reza Anik c , S. Speelman d , E.A. Nuppenau a a

Institute of Agricultural Policy and Market Research, Justus Liebig University Giessen, Senckenbergstr. 3, 35390 Giessen, Germany Agricultural Development and Rural Transformation Centre, Institute for Social and Economic Change, Bangalore 560072, India Department of Agricultural Economics, Bangabandhu Sheikh Mujibur Rahman Agricultural University, Salna, Gazipur 1706, Bangladesh d Department of Agricultural Economics, Ghent University, Coupure Links 653, 9000 Ghent, Belgium b c

a r t i c l e

i n f o

Article history: Received 21 November 2011 Received in revised form 25 June 2012 Accepted 5 August 2012 Keywords: Land fragmentation Farm size Land ownership Crop diversity Profit efficiency Irrigated farms India

a b s t r a c t In this article, we analyze the impact of land fragmentation, farm size, land ownership and crop diversity on farm profit and efficiency of 90 groundwater irrigated farms in the hard rock areas of South India. As we hypothesize that these variables may impact both, farm profit and efficiency in alternative ways, we develop four different stochastic frontier and inefficiency effect models by shifting some of these variables from the inefficiency model into the profit function. The underlining reason is to know the impact of different combination of these structural variables on farm profit and efficiency. Our analysis shows that there exist high levels of inefficiency among the sample farms. Among the variables influencing efficiency, the most pronounced effects are observed with land fragmentation, land ownership and crop diversity. Land fragmentation is positively and significantly associated with inefficiency, whereas land ownership and crop diversity is negatively and significantly associated with inefficiency. In addition land fragmentation can also have a significant negative effect on farm profit. We further find that smaller farms appear to have lower inefficiencies than larger farms due to the more efficient use of inputs by the former category. Importantly, when a farmer owns a failed well, this also contributes to the inefficiency, since well failure increases cost of irrigation. Further the average profit efficiencies are higher for unfragmented farms, large farms, owner operated farms and farms with a diversified cropping pattern as compared with their counterparts. Knowledge on the factors influencing farm profit and efficiency is crucial for policy makers and extension agents for improving efficiency levels of the groundwater irrigated farms especially in water scarce regions of the country. © 2012 Elsevier Ltd. All rights reserved.

Introduction In agrarian economies, land reforms, especially land redistribution can play a pivotal role in reducing poverty and land inequity and India is no exception to this. The country’s land redistribution policy has to some extent been successful in reducing the poverty and land inequity (Jha et al., 2005; Mearns, 1999). Overtime, in the process of land reform, there has been an increase in land fragmentation, a decrease in farm size, an increase in land degradation and redistribution of resource ownership (Jha et al., 2005; Mearns, 1999; Niroula and Thapa, 2005, 2007; Rahman and Rahman, 2008). The seminal work of Schultz’s on efficient peasants states that peasants in poor countries maximize returns from their input use since they are rational decision makers (Schultz, 1964). After this

∗ Corresponding author. Tel.: +91 9448402848; fax: +91 80 23217012. E-mail addresses: [email protected] (A.V. Manjunatha), [email protected] (A.R. Anik), [email protected] (S. Speelman), [email protected] (E.A. Nuppenau). 0264-8377/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.landusepol.2012.08.005

work, several scholars have researched on the impacts of land fragmentation and farm size on efficiency. Literature argues for three main factors for boosting up land fragmentation in India. These are law of inheritance of paternal property, absence of a progressive tax on inherited land and underdeveloped land market (Ghatak and Roy, 2007; Niroula and Thapa, 2005). As a consequence the country is inter alia experiencing a decline in farm size and an increase in the number of operational holdings. For instance, the agricultural land per capita declined from 0.41 ha in 1980 to 0.31 ha in 2009 (FAO, 2012). Between 1990–91 and 2000–01, the number of operational holdings increased from 106.64 million to 119.93 million, while the operational farm size has reduced from 1.57 ha to 1.33 ha (Government of India, 2010). This is the opposite from the trend in East Asian countries like Japan and South-Korea. In these countries the farm size is getting larger and the number of operational holdings are reduced. In these countries, a gradual shift of labor force from agriculture to non-agricultural sectors is experienced due to the availability of relatively attractive income opportunities outside

398

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

the agricultural sector. This has resulted in land consolidation and encouraged commercialization, which has helped the farms towards more efficient utilization of labor and other inputs, thereby improving the farm efficiency (Niroula and Thapa, 2005). In contrast, in India, land fragmentation has discouraged commercialization, thus promoting inefficiency in agricultural production (Jha et al., 2005; Parikh and Nagarajan, 2004). Some of the contributing factors to inefficiency are: sub-optimal application of farm inputs, excess traveling time, loss of productive land due to bunding or hedging, difficulties in use of modern technology as well as problems with monitoring (Jha et al., 2005; Parikh and Nagarajan, 2004). Land fragmentation could drive farmers towards intensive agricultural practices such as continuous farming and monocropping, resulting in deterioration of land quality. This increases production cost and lowers land productivity (McPherson, 1982; Ram et al., 1999). Land fragmentation might, however, also drive towards crop diversification, which may act as a risk reducing strategy especially in the areas suffering from labor scarcity, natural disasters and successive droughts (Niroula and Thapa, 2005; Tan et al., 2006). In addition diversification can contribute to improvements in soil fertility, if crops with different nutrient requirements use soil resources more optimally. Farmers also apply fertilizers differently to different crops and ultimately some sort of balanced application of fertilizer is ensured (Hazra, 2000). Identifying options for improving efficiency of irrigated farms has become a challenging task for the policy makers and researchers. It is reported that, for example, the inefficiency of water use in agricultural sector is to the extent of 60–70% in India, indicating tremendous potential for increasing water use efficiency (Palanisami, 2009). In a densely populated country like India, the scope of increasing agricultural production by increasing total farm land and water use is costly and almost reached saturation and hence the most likely option here to meet the country’s food requirement is to increase productivity by increasing efficiency. While most literature on efficiency analyzes the impact of land fragmentation, farm size, and land ownership on farm efficiency for a single crop (Bardhan, 1973; Lau and Yotopoulos, 1971; Ali and Flinn, 1989; Coelli et al., 2002; Rahman, 2003; Rahman and Rahman, 2008; Alam et al., 2011), this should be extended to multiple crops. In reality farmers grow one or more crops during a year or within a season depending on their available resources. They do not decide about resource allocation considering the profit from a single crop, rather they think about the whole farm and allocate their resources accordingly. This is widely discussed in literature about dynamic modeling of farm allocation decisions (Dogliotti et al., 2003; Miranowski and Orazem, 1994; Hopper, 1965). There are also studies on impacts of land fragmentation or land ownership or farm size or crop diversification on productivity or efficiency, but to our best knowledge there are no studies considering all these aspects together in a single region in a general way and particularly in India. In addition, there are no reported studies in India similar to our approach (profit frontier and inefficiency effect model) for analyzing the above mentioned issues. In this paper, we particularly focus on the impact of land fragmentation, farm size, land ownership and crop diversity on farm profit and efficiency, using farm level survey data from 90 groundwater irrigated farms in India. Four different stochastic Cobb–Douglas profit frontier and inefficiency effect models are developed, which is explained in detail in the next section. The remainder of the paper is organized in three sections: Section “Research methodology” describes the study region and data collection followed by the methodology of the stochastic profit frontier approach. Section “Results and discussion” provides research findings and analysis of those findings and the final section “Concluding remarks” offers conclusions and policy implications.

Research methodology The study area and data collection The present study was undertaken in one of the taluks1 of the Eastern Dry Zone (EDZ) of Karnataka, which lies in the hard rock region of south India. The EDZ is a semi-arid region, characterized by insufficient rainfall (784 mm), lack of water recharge and lack of access to perennial rivers (Chandrakanth et al., 2004; Government of Karnataka, 2006). In this zone, groundwater supplies more than 90% of the total irrigated land. Overtime, effects of overdraft has led to well failure, decline in water tables and decline in groundwater outputs, thus constraining agriculture. Despite improvements in groundwater extraction and use of new technologies (Manjunatha et al., 2011), water scarcity is exacerbated by population growth and growing effective demand for groundwater intensive agricultural production. As a result of water overdraft, the region is declared as an over-exploited zone by the department of mines and geology, meaning that the extraction of groundwater exceeds more than 85% of recharge (Government of Karnataka, 2006). Using multistage sampling techniques, primary data needed for the study were collected from 90 groundwater irrigated farms belonging to 10 villages of Malur taluk for the agricultural year 2007–08. The villages where water scarcity was most severe were purposively selected and this was done considering the secondary data on water availability from the Karnataka state water resources department as well as upon consultation with water resource experts. The total geographical area of the taluk is 645 km2 with a population of 0.207 million. For their livelihood the residents here mainly rely on agriculture and wage employment (Government of Karnataka, 2006). Groundwater is the major source of irrigation in the taluk contributing approximately 99% (17,956 ha) to the total irrigated area (Government of Karnataka, 2006). In the study region, vegetable production dominates over cereals and other food crops. The main agricultural products of the region are tomato, potato, beetroot, carrot, beans, cauliflower, banana, flowers, etc. The vegetables produced are mainly sold at the local market. The remaining vegetables are supplied to the Bangalore city, which is only 46 km away from the region. Informal groundwater markets emerged in the region because of the low groundwater table, high initial investments necessary to construct bore wells and the high risk of well failure (Manjunatha et al., 2011). Land fragmentation is a severe problem in the Malur taluk. For instance the number of operational holdings has increased from 38,627 in 2000–01 to 43,047 in 2005–06, while the per-capita operational holdings decreased from 1.15 ha to 1.07 ha during the same period (Government of India, 2012). Also, the region is experiencing soil degradation due to the intensive cropping activity (high value agriculture) throughout the year (personal communication). Unfortunately, there are no reported estimates for the degraded land as such, but some researchers have used proxies such as reduction in agricultural land and increase in waste land over the years to indicate the severity of degradation. The region experienced an increase of waste land to the total land from 27.88% in 1998 to 38.42% in 2002 as well as a decrease in usable agricultural land from 40.95 to 22.56% during the same period (Ramachandra and Uttam Kumar, 2004). Considering the complexity of the interlinked issues and the competition for limited resources, especially land and water, the selected study region was regarded suitable to analyze the issue of land fragmentation, farm size, land ownership and crop diversification. Using a structured and pre-tested questionnaire, detailed

1

A taluk is an administrative unit in India. A taluk consists of several villages.

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

information was elicited from the irrigated farms. The following aspects were covered in the questionnaire: (a) general information about the family size, land holdings, land fragmentation, land ownership (own land or rented land), education level of the farmer, irrigation wells, water sharing, crop diversity, etc. and (b) specific information regarding inputs (seeds, labor, water, manure and chemical fertilizers, plant protection chemicals) and outputs, i.e. quantities of different farm products and their corresponding prices. The expert knowledge of the local village institutions was used while collecting information on water use and prices of output. Empirical models Inefficiency in a production process can be analyzed by three measures: technical, allocative and scale efficiency (Coelli, 1996): Technical efficiency gives the capacity of a farm to achieve the highest output with a given level of inputs. Technical efficiency can be decomposed into pure technical efficiency and scale efficiency. The former relates to the most efficient scale of operation of the farm in terms of average productivity. The later is estimated by separating the scale effect from the technical efficiency. Allocative efficiency reveals the capacity of the farm to apply the inputs in optimal quantities at given prices. Economic efficiency finally is the combination of technical and allocative efficiency and it is also sometimes termed production efficiency (Coelli, 1996; Wang et al., 1996). Profit efficiency is used for the present study since it incorporates technical, allocative and scale efficiency and any error in the production choice converts into lower profits for the farmer (Ali and Flinn, 1989; Wang et al., 1996). While estimating stochastic frontiers, it is important to check theoretical consistency, flexibility, and choice of the appropriate functional form (Sauer et al., 2006). Hereby, an alternative to the widely used Cobb–Douglas production function is the translog production function. Compared to Cobb–Douglas, a translog function is more flexible, because Cobb–Douglas imposes a severe prior restriction on farms’ technologies by limiting the production elasticities to being constant and the elasticities of substitution to unity (Wilson et al., 1998). Unfortunately our dataset did not allow us to test the translog function. This because our sample size is not large enough to enable estimating the high number of coefficients, e.g., for the interaction variables, in the translog function. Moreover Cobb–Douglas is still widely used and a good number of articles can be found using Cobb–Douglas production function in Indian agriculture because of its advantages in computation and interpretation (Parikh and Nagarajan, 2004; Puran et al., 2011). In our study, we applied stochastic profit functions and inefficiency models in order to analyze the impact of land fragmentation, farm size, land ownership and crop diversity on profit and efficiency of groundwater irrigated farms in India. Description and measurement techniques of these and other explanatory variables used in different models are available in Table 1. Any stochastic frontier model is a combination of two models. For instance, in our case the first model is a profit function. The additional component here is the error term which comes from another regression known as the inefficiency effect model. In this second model different variables, which are assumed not to effect profit directly but do in an indirect way through influencing farm efficiency, are included. While developing frontier models one has to be sure about which variable to include in which model. This was quite difficult for land fragmentation and crop diversity. However, as it is theoretically reasonable we assumed that they have an effect on both farm profit and efficiency. Finally to address all different possibilities we have developed four different models. The models are: model 1 is the basic model which only considers input costs in the profit function and land fragmentation, farm size, land ownership, crop diversity and other relevant farm characteristics

399

in this inefficiency effects model. Such model is conceptualized by assuming that both, land fragmentation and crop diversity, affect farm efficiency, but not farm profit; model 2 tests whether the crop diversity and land fragmentation affect farm profit by incorporating the previous two variables in the profit function along with all input costs. This model assumes that crop diversity and land fragmentation affect farm profit but have no affect on farm inefficiency and hence are dropped from the inefficiency model; the next model, model 3, assumes that farm profit might be affected by land fragmentation, but not by crop diversity. Here land fragmentation is incorporated in the profit function and the crop diversity variable is in the inefficiency model; and in model 4 it is to know whether crop diversity affects profit efficiency. In this model crop diversity is placed in the profit function along with all other input costs. Farm size is not included in the profit function because the analysis is carried out on per acre basis and its impact can be know in all the four models in the inefficiency part. Our model approach is somewhat similar to the approach of Rahman and Rahman (2008) who analyzed effects of land fragmentation and resource ownership on productivity and efficiency in Bangladesh rice farming using a translog production function. The general form of the Cobb–Douglas profit frontier for the ith farm is defined as: ln i = ˛0 +

J 

˛j ln xij + vi − ui

(1)

j=1

where the dependent variable i is the annual profit of the ith farm (INR) which is estimated by adding the profit from all crops grown in a year by a farmer, xi ’s are the cost of different production inputs such as seed, water, labor (hired and own) and fertilizers (manure and chemical fertilizers). The inputs costs are estimated separately for each crop grown in a farm and later are added for all crops in a year for each farmer. Crop diversification (using the Herfindahl index) and land fragmentation (using a dummy) are also included in some of the models estimated. The Herfindahl index represents crop diversification/specialization and is estimated as the summation of all squared area shares occupied by crop/s in total cropped area. The value of this index varies from zero to one. It takes the value of one when there is full specialization and approaches to zero when there is full diversification (Rahman, 2009). The error component vi from Eq. (1) is assumed to be identically and independently distributed (i.i.d.) as N(0, v2 ). This is the usual assumption generating a normally distributed error that represents the random shifts in the frontier. Unlike vi , ui is a nonnegative, unobservable random variable that captures the technical inefficiency of the observations and is assumed to be distributed independently of the normally distributed error term (vi ). The technical inefficiency model enables us to know variables which may affect farm efficiency. The technical inefficiency equation can be written as: ui = ı0 +

K 

ıd zik + ωi

(2)

k=1

where zi ’s are the explanatory variables that explain inefficiency. The list of inefficiency variables used in different models include: the Herfindahl index of crop diversification, a dummy for land fragmentation, a dummy for failed/non-functioning well/s, a dummy for leased-in land, the ratio of agricultural labor to the farm size and a dummy of farm category (small and large). ωi is the unobservable random error assumed to be independently distributed with a positive, half-normal distribution. The profit efficiency of irrigated farms is assumed to be influenced by different input costs and other farm characteristics.

400

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

Table 1 Summary statistics of variables on per acre basis. Variable

Description of the variables

Mean (std. dev.)

Profit

Gross return minus total cost of the whole farm divided by total farm land (INR/acre) Per acre cost of seed/seedlings purchased from the market (INR/acre) Per acre cost of water estimated by amortized cost method (INR/acre) Per acre cost of both family and hired labor (INR/acre) Per acre cost of organic and chemical fertilizers (INR/acre) Dummy: 1 = farmer apply pesticides; 0 = otherwise Dummy: 1 = farmer has fragmented land (more than one land fragment); 0 = otherwise The index value approaches to one when there is full specialization and approaches to zero when there is full diversification Dummy: 1 = the farmer has failed well; 0 = otherwise Dummy: 1 = farmer has a leased land; 0 = otherwise Ratio of agricultural labor to family size Dummy: 1 = small farmer with up to 2.5 acres; 0 = large farmers with more than 2.5 acres

13153.9 (7572.4)

Seed cost Water cost Labor cost Fertilizer cost Plant protection chemicals Land fragmentationa Herfindahl index of crop diversification Failed wellsb Leased-in land Labor ratio Farm category

2548.7 (1674.0) 2152.1 (1069.6) 4124.4 (1964.9) 4565.5 (2298.7) 0.822 (0.384) 0.344 (0.478) 0.553 (0.279) 0.578 (0.497) 0.122 (0.329) 0.660 (0.223) 0.744 (0.439)

Notes: During the data collection period the 1 Indian Rupee (INR) = 0.02481 US$ = 0.01602D ; one acre = 0.4 ha. a A farmer is considered to have fragmented land when he owns more than one piece of agricultural plot. b Failed wells are usually abandoned wells which are unable to yield sufficient water for growing a crop or no water at all.

We used maximum likelihood estimates for all parameters of the stochastic frontier and the inefficiency models can be simultaneously obtained by using the computer program STATA 8, which estimates the variance parameters that are expressed in terms of:  2 = v2 + u2 and  = u2 / 2 . Here,  is the ratio of the variance of farm-specific technical efficiency to the total variance of output and has a value between zero and unity, where if  = 0, inefficiency is not present and if  = 1, there is no random noise. If  is not significantly different from zero, the variance of the inefficiency effect is zero and the inefficiency variables are entered directly into the model (Battese and Coelli, 1995). Technical efficiency (TEi ) of the ith farm is the ratio of the observed output for the ith farm, relative to the potential output defined by the frontier function, where the input vector xi is given. Given the specifications of the stochastic frontier models, the technical efficiency of ith farm can be shown to be equal to: TEi =

yi exp(xi ˇ − ui ) = = exp(−ui ) exp(xi ˇ) exp(xi ˇ)

(3)

The technical efficiency of a farm is between zero and unity and is inversely related to the inefficiency effect. Results and discussion Socio-economic profile and descriptive statistics From the inputs, fertilizer is the major cost component followed by labor cost, seed cost and water cost. The majority of the farms (82.2%) use several plant protection chemicals (PPCs). The rest of the farms (17.8%) does not use any PPCs because these farms grow only mulberry (Table 1). It is to be noted that mulberry leaves are fed to silk worms for production of silk and hence the use of PPCs could reduce the cocoon/silk yield. One third of the sample farms have fragmented land (Table 1). These farmers have more than one agricultural plot at different locations within the same village or in the neighboring villages. Mainly as a result of fragmentation, redistribution of ownership is observed for land, labor, water, bullock/machine power which shows the interlocked nature of the resources. Similar redistribution of ownership of resources such as land, family labor and draft animals is also observed in Bangladesh (Rahman and Rahman, 2008). The average irrigated area is less than a hectare (0.77 ha). The cropping pattern for the sample farms largely depends on the availability of functioning wells. The overdraft of water has led to initial

and premature failure of irrigation well caused by the cumulative well interference, thereby increasing cost of irrigation and decreasing efficiency. About 58% of the farms have failed wells ranging from 1 to 3, indicating a high probability of well failures in the region (Table 1). Crop diversification is a common practice in the study region, with 70% of the sample farms cultivating several crops. But the average Herfindahl index of crop diversification is 0.553 indicating the presence of moderate level of crop diversification (Table 1). Farmers diversification strategies mainly include growing different vegetables within a season or across seasons or combination of mulberry plus vegetable crop/s. Farmers diversify for various reasons. A possible reason is that due to land fragmentation they own land of varying quality. Alternatively they might diversify to cope with the production and price risks. The positive effect of land fragmentation on crop diversification was reported in the studies of Niroula and Thapa (2005) in south Asia and Tan et al. (2006) in China. Land leasing is not a common practice in the area and only 12% of the farms have leased-in land. About 74% of the farms are small farms, owning only 53% of the total irrigated land, whereas large farms (26%) own 47% of the land holdings showing a skewed distribution of land. Such skewed distribution is similar to the findings of Fujita and Hossain (1995), Sharma and Sharma (2006) and Meinzen-Dick (1997) in Bangladesh, India and Pakistan, respectively. In addition they have pointed out skewness in distribution of functioning wells which is also observed in our region. Family size is an indicator of work force available for farming. The share of agricultural labor to the family member is 0.66. The rest (0.44) are either children or sick or old or work in off-farm activities (Table 1).

Hypotheses testing and variance parameters The null hypothesis, H0 = ı0 = ı1 = . . . = ı8 = 0, states that there is no inefficiency effect in farm profit. As the calculated value using the likelihood-ratio test is higher than the tabulated value, this hypothesis is rejected for all the four models at a 1% level of significance, implying that there are significant technical inefficiency effects in farm profit (Table 3). Sauer et al. (2006) suggested to check for two different regularity conditions in the frontier model. These are: (a) monotonicity, i.e. positive marginal products with respect to all inputs (∂y/∂xi > 0) and thus non-negative output elasticities; and (b) diminishing marginal productivity (∂2 y/∂xi2 < 0) with respect to all inputs, i.e. the marginal products, apart from being positive should be decreasing in inputs. When the monotonicity

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

401

Table 2 Regularity conditions checks. Variables: INR/acre

Seed cost Water cost Labor cost Fertilizer cost

Diminishing marginal productivity (∂2 y/∂xi2 < 0) for every input

Monotonicity (∂y/∂xi > 0) for every input Model 1

Model 2

Model 3

Model 4

Model 1

Model 2

Model 3

Model 4

3.575 0.631 0.510 0.655

2.999 0.585 0.142 0.669

3.643 0.779 0.331 0.477

3.137 0.645 0.144 0.806

−0.054 −0.001 −0.0001 −0.0002

−0.0475 −0.0001 −0.00004 −0.513

−0.0546 −0.0006 −0.0001 −0.0001

−0.049 −0.0005 −0.00004 −0.0002

conditions are violated, the duality theory no longer prevails (Barnett, 2002). Results in Table 2 clearly demonstrate that both these conditions are fulfilled for all the inputs in all the four models. The estimated values of  in all the models are significantly different from zero, and support the use of an inefficiency model. The corresponding variance-ratio parameter * imply that at least 60% of the differences between the observed and maximum frontier profit in all the models is due to the existing differences in efficiency levels among farms. Further, the null hypothesis of no technical inefficiency effects in the profit function is rejected since the p value is significant in all the models. This implies that significant level of technical inefficiency effects are present in all the models.

Econometric results and discussion Table 3 presents the maximum-likelihood estimates for the parameters of the Cobb–Douglas stochastic profit function and technical inefficiency effect model. Tables 4 and 5 present the estimated farm profit efficiencies. The mean profit efficiency for the farms is lowest in model 2 where both the land fragmentation and crop diversification variables are included in the profit function. Profit efficiency is highest in model 1, where only the production factor’s costs are included in the profit function. The average mean efficiency for model 1 is 93.2% implying that on an average a farm can increase its profit by about 7.3% [(100 − 93.2) × 100/93.2] by improving efficiency. Possibility to increase profit by increasing efficiency is 43.3% in model

Table 3 Maximum-likelihood estimates for parameters of Cobb–Douglas stochastic profit function and inefficiency effect model. Variables Profit function Seed cost

Model 1

Dummy of land fragmentation

0.238** (0.110) 0.153 (0.100) 0.228* (0.132) 0.101 (0.143) 0.130 (0.674) –

Herfindahl index of crop diversification

–

Constant

3.610*** (0.674)

Water cost Labor cost Fertilizer cost Dummy of plant protection chemicals

Technical inefficiency predictors Dummy of land fragmentation Herfindahl index of crop diversification Dummy of failed wells Dummy of leased-in land Labor ratio Dummy of farm category Constant Model diagnostics  2 = u2 + v2  = u2 /(u2 + v2 ) * Log likelihood function Null hypothesis: no inefficiency effects in the profit function (p-value) Total number of observations

2.668*** (1.131) 3.198*** (1.014) 1.392** (0.683) 1.590*** (0.652) −5.036*** (2.227) −0.432 (0.282) −2.890 (1.481)

Model 2 0.200* (0.117) 0.043 (0.109) 0.233 (0.167) 0.093 (0.156) 0.292 (0.268) −0.177** (0.081) −0.441** (0.200) 5.589*** (0.826)

Model 3 Coefficient (std. error) 0.242** (0.123) 0.099 (0.117) 0.166 (0.145) 0.125 (0.167) 0.053 (0.329) −0.079 (0.066) – 4.594*** (0.837)

Model 4 0.209* (0.125) 0.043 (0.130) 0.281 (0.164) 0.103* (0.162) 0.194 (0.328) – −0.348** (0.183) 4.626*** (0.815) 0.250* (0.133) –

–

–

– 0.139* (0.081) 0.370*** (0.106) −0.204 (0.160) 0.009 (0.094) 0.693*** (0.153)

1.593*** (0.612) 0.495* (0.277) 0.771*** (0.281) −0.171 (0.501) −0.186 (0.232) −1.291 (0.848)

0.093*** (0.019) 0.852*** (0.167) 0.677 −1.920 0.022

0.092*** (0.005) 0.809*** (0.000) 0.606 −15.249 0.007

0.156*** (0.028) 0.813*** (0.106) 0.613 −13.264 0.057

0.144*** (0.011) 0.838*** (0.133) 0.653 −19.326 0.056

90

90

90

90

Notes: The likelihood ratio test was significant at 99% significance level. * is equal to /[ + (1 − )/( − 2)] (Coelli et al., 1998). * 90% significant level. ** 95% significant level. *** 99% significant level.

0.140 (0.142) 0.425*** (0.159) −0.234 (0.284) 0.032 (0.151) 0.154 (0.230)

402

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

Table 4 Profit efficiency scores. Variables

Model 1

Model 2

Model 3

Model 4

Mean Std. deviation Minimum Maximum

0.932 0.155 0.268 0.995

0.698 0.165 0.204 0.999

0.808 0.146 0.281 0.955

0.734 0.132 0.340 0.939

2, which is the highest among all the models. Improvement in efficiency will result in an increase in farm profit of around 23.8% and 36.2% in model 3 and model 4, respectively. In all the models farms exhibit a wide range of variation in profit efficiency ranging from approximately 20 to 100% (Table 4). In all the models, the positive signs of all the input variables (seed cost, water cost, labor cost and fertilizer cost) in the profit functions are in line with the literature. As our main motivation for developing these models is to know the impact of land fragmentation, farm size, land ownership and crop diversification, we will mainly focus our discussion on the results for these variables. In model 1 and model 4, the land fragmentation variable is positively and significantly associated with profit inefficiency, whereas in model 2 and model 3, the same variable is negatively associated with farm profit (Table 3). This is obvious because land fragmentation has resulted farms to incur additional costs due to loss of land, wastage of inputs, requirement of more traveling time, loss of labor’s working hour due to shifting from one land to another land and increased water conveyance costs for transferring water from the water source to different piece of land. Our results are similar with the findings of Schultz (1953), Bardhan (1973), Parikh and Nagarajan (2004), Raghbendra et al. (2005) and Rahman and Rahman (2008). These researchers found inverse relationship between land fragmentation and efficiency due to sub-optimal application of inputs. In model 1, model 3 and model 4 the farms with fragmented land have significantly lower efficiency levels than their counterparts who operate in a single piece of land. Only in the model 2 fragmented land owners operate with relatively higher efficiency level, but the difference here is insignificant (Table 5).

In India, crop diversification became a popular practice with the advent of modern technologies during the green revolution in the late sixties and early seventies. Currently crop diversification is still followed because of benefits, primarily economic objective and sometimes motivated by government policies (e.g., Technology mission on oil seeds). Usually farmers shift from low remunerative crops to high remunerative crops. Further crop diversification widens the choice of crop production in a specific area so that production related activities are enhanced and risk is also reduced (Hazra, 2000). In model 1 and model 3, the Herfindahl index of crop diversification is positively and significantly associated with inefficiency, whereas it is negatively and significantly associated with farm profit in model 2 and model 4 (Table 3). A positive sign of this index in the inefficiency model means that profit inefficiency is positively associated with specialization, implying that crop diversification significantly improves profit efficiency. A negative sign of this index in the profit function indicates that farm profit is negatively associated with specialization. Crop diversification, therefore, significantly improves farm profit. Farms diversifying their crops exhibit relatively higher efficiency levels in all the four models than their counterparts who do monoculture (Table 5). Our results here are consistent with the findings of Ram et al. (1999) in India, Nel and Loubser (2004) in South Africa and Rahman (2009) in Bangladesh. These scholars have found beneficial impact of crop diversification on productivity and efficiency. It is to be noted that crop specialization might be a better option for the short run to get higher profits and efficiency, but in the long run, this could lead to severe environmental consequences such as water depletion, loss of soil fertility, water logging and salinity. For instance, shift from pulses, oil seeds and cotton to rice–wheat cropping pattern in north western states of India is a perfect example of this tragedy (Hazra, 2000). As mentioned before, leasing-in and leasing-out is not widely practiced in our study area in spite of land scarcity and land fragmentation. There is a fear of claim of ownership by the tenants in case of longer period contracts and hence agricultural land in the area is generally leased for a maximum period of 2 years. The leased-in land variable is positively associated with the inefficiency which implies that farmers having leased-in land are less efficient

Table 5 Distribution of farm profit efficiency according to land fragmentation, farm category, land ownership and crop diversity. Category

Land fragmentation No land fragmentation Land fragmentation Farm category Large farms Small farms Land ownership Farmers with leased-in land Farmers without leased-in land Crop diversity No Yes * ** ***

90% significant level. 95% significant level. 99% significant level.

Profit efficiency scores (std. dev.) Model 1

Model 2

Model 3

Model 4

0.987 (0.002) 0.826 (0.042)***

0.515 (0.020) 0.530 (0.034)

0.832 (0.014) 0.763 (0.035)*

0.763 (0.013) 0.667 (0.029)***

0.882 (0.045) 0.949 (0.016)*

0.522 (0.034) 0.582 (0.020)

0.781 (0.038) 0.818 (0.016)

0.711 (0.033) 0.743 (0.015)

0.946 (0.0145) 0.831 (0.081)***

0.536 (0.018) 0.405 (0.046)***

0.831 (0.014) 0.645 (0.054)***

0.755 (0.013) 0.590 (0.048)***

0.846 (0.059) 0.957 (0.012)**

0.541 (0.047) 0.514 (0.018)

0.653 (0.043) 0.853 (0.011)***

0.718 (0.043) 0.739 (0.013)

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

than their counterparts who produce on own land. This variable has significant impact in all the four models (Table 3). The differences of mean efficiency scores between the farms with leased-in land and farms which only cultivate on their own land is significant in all the models. In every case the owner-operators has higher levels of efficiency (Table 5). Coelli et al. (2003) and Rahman (2003) also observed positive association between efficiency and land ownership. According to Rahman (2003), one possible reason might be the relatively poor quality of land that is generally rented to tenants which is also true in our case (personal communication). Additionally tenants might also not be able to perform farm operations timely because they have less control over land and water. They moreover apply relatively lesser quality inputs due to their poor financial status. By promoting land consolidation, which can be in an indirect way, leasing-in and leasing-out of land may have positive impacts on farm production (Niroula and Thapa, 2005). The dummy of failed wells is positively associated with inefficiency in all the models. The coefficient of this variable is significant in model 1, model 2 and model 3 (Table 3). Such relations imply that farmers owning failed bore well/s operate at lower efficiency levels than their counterparts who only own functioning well/s. Such relationship is consistent with the economic theory since failed wells contribute to the cost of irrigation, thereby contributing to inefficiency. The indiscriminate nature of water pumping by the farms from a shared aquifer has led to declining water table and failure of irrigation wells. Due to declining water tables farmers have to pump deeper and hence have to bear higher installation cost than earlier. When irrigation wells fail (run dry), the farmer has two options to continue farming. Either he has to invest in an additional irrigation well or he has to depend on neighbors for water. Both choices increase the cost of irrigation for the farmer (Manjunatha et al., 2011). The farm category does not have a significant impact in any of the models. The negative sign in model 1 and model 3 indicates relatively higher inefficiencies of large farms, while the positive sign in model 2 and model 4 represents relatively higher inefficiencies of small farms (Table 3). The computed average efficiency scores imply that small farms are relatively more efficient than large farms in all the models (Table 5). This may be due to the fact that small farms use the resources in a better way as compared to the larger farms. In addition small farms devote more time to farming since it is more likely to be their main income source, which may not be the case for the large farms as they also derive income from off-farm activities. The results are consistent with the results of Schultz (1964), Carter (1984), Lau and Yotopoulos (1971) and Chand et al. (2011). These authors reported that small farms are more efficient than large farms since small farms utilize the resources to the fullest extent for achieving higher productivities. However Battese (1992) and Ram et al. (1999), Alam et al. (2011) reported higher efficiencies with large farms due to usage of agricultural innovations. The labor ratio variable has a negative sign in all the four models (Table 3). Such negative associations imply that farms with relatively low labor ratio (ratio of agricultural labor to family size) have lower levels of profit efficiency. These farms have to rely less on hired labor. This is not surprising because productivity of family labor is obviously higher than that of hired labor. Such results are consistent with the findings of Rahman and Rahman (2008) who reported positive contribution of ownership of labor and other inputs to efficiency.

Concluding remarks The study used stochastic profit frontier and inefficiency effect models to analyze the impact of land fragmentation, farm size, land

403

ownership and crop diversification on the profit and efficiency of 90 groundwater irrigated farms in south India. Results demonstrate that land fragmentation has a reciprocal relationship with farm profit and efficiency. Small farms are found to be more efficient than large farms due to their relatively efficient use of resources. Usually large farms are also facing the problem of land fragmentation more than the small farms. These results argue for some policy actions for reducing land fragmentation particular in the case of large farms where there is wastage of resources which is contributing negatively to efficiency. As such interventions of increasing farm size may lead to higher farm profits and efficiency. Land consolidation might be a solution here. But land consolidation might be a challenging task in India especially due to high population growth rate, inheritance of paternal property and lack of enough off-farm income generating activities (Parikh and Nagarajan, 2004; Jha et al., 2005; Niroula and Thapa, 2005, 2007). Considering the Indian context the most likely option to be achieved in the shortest possible time is creating off-farm employment opportunities. We do not undermine or neglect importance of actions regarding land consolidation, but as this is related with religious, socio-cultural and socio-economic aspects, actions regarding this need time and strong political motivation. For instance in order to overcome the adverse effects of land fragmentation, the Indian government allocated INR 5 million for a period of 5 years since 2000. However the government has noticed a complete failure except in few states where there is less variation in the land quality (Jha et al., 2005; Niroula and Thapa, 2005). With increasing off-farm employment opportunities people will be more and more involved in non-agricultural sectors which will decrease their reliance on land. At present, in rural areas, agriculture is the main income source for the farmers and even if they have some off-farm income sources they rely less on those. But if these sources can offer enough income, people may leave agriculture and this will create a scope for the remaining farmers to increase their farm size. Even if the people in non-agricultural sources do not sell their land, they may lease it to other farmers. It was evident from the study of Tan et al. (2006) in China that land fragmentation has reduced by 15% and average land size has increased by 13% with the availability of off-farm income activities in Jiangxi Province. Usually credit institutions focus more on large farms considering the higher repayment capacity and risk bearing ability. Due to this small farms have to rely on informal credit markets where the interest rates are up to 40% per annum. Additionally banks provide loans for income generating activities mostly on own land but not to buy the agricultural land (personal communication). Therefore, the farmers, especially the small ones need some special credit facilities to be able to increase their farm size. The present credit policy especially collateral requirement and interest rate has to be revised facilitating the small farms through differentiating them from large farms. Steps should also been taken to reduce complexity in Indian land market which results in high transaction cost and time requirement (Mearns, 1999). The easy and flexible policies for land transactions such as leasing-in, leasing-out, buying and selling of land could also help to reduce the land fragmentation. As such policies have also proved to be successful in China and Japan (Tan et al., 2006; Niroula and Thapa, 2005). Further, the extension agents should play an active role in dissemination of suitable sustainable technologies to the different category of farmers aiming for further reducing the inefficiency. Crop diversification is considered as a risk reducing option (production and market) in the study region. The production risk is mainly from insufficient and uncertain water availability, whereas the market risk is due to the huge price fluctuations, especially in the vegetable market. The Indian government should encourage region-specific policies regarding crop diversification through extension services and credit facilities, as such practices could

404

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405

improve efficiency. Even crop rotation as a part of diversification could further reduce the production risk attributed from the droughts, pests and diseases (Nel and Loubser, 2004), which is considered as a profitable and sustainable farming practice. More research may be undertaken to identify suitable crop diversifying strategy in general and crop rotation in particular. Land ownership is positively associated with profit efficiency. Currently in India, there are several policies for land or landless farmers and very few policies encouraging tenant farming. The Indian government should frame specific policies for facilitating tenant farms with adequate inputs and technology which might be an option to keep pace with the efficiency levels of owner operated farms. We conclude that problems such as land reform, labor, water and land scarcity, production and price risk, absence of enough nonfarm employment opportunities and underdeveloped land market are responsible for land fragmentation, reduced farm size, redistribution of land ownership and crop diversification that imposed varying impacts on farm profit and efficiency. Such information is crucial for the policy makers and extension agents for reducing the inefficiency levels in irrigated agriculture in India. Further our study is at micro-level case study involving cross sectional data of groundwater irrigated farmers, results and implications apply to problematic regions in eastern dry zone of Karnataka, India or regions with similar settings in India. Additional research involving more cross sectional data or panel data can further give more clarity of the impact of structural variables on farm profit and efficiency. Acknowledgements Mr. Manjunatha and Dr. Speelman are funded as doctoral fellow and postdoctoral fellow by the German Academic Exchange Service and the Research Foundation Flanders, respectively. We are grateful to two anonymous referees for helpful comments on an earlier version of this paper. All errors remain ours. References Alam, M.J., Huylenbroeck, G.V., Buysse, J., Begum, I.A., Rahman, S., 2011. Technical efficiency changes at the farm-level: a panel data analysis of rice farms in Bangladesh. African Journal of Business Management 5, 5559–5566. Ali, M., Flinn, J.C., 1989. Profit efficiency among Basmati rice producers in Pakistan Punjab. American Journal of Agricultural Economics 71, 303–310. Bardhan, P.K., 1973. Size, productivity and returns to scale: an analysis of farm level data in Indian agriculture. The Journal of Political Economy 81, 1370–1386. Barnett, W.A., 2002. Tastes and technology: curvature is not sufficient for regularity. Journal of Econometrics 108, 199–202. Battese, G.E., 1992. Frontier production functions and technical efficiency: a survey of empirical applications in agricultural economics. Agricultural Economics 7, 185–208. Battese, G.E., Coelli, T.J., 1995. A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empirical Economics 20, 325–332. Carter, M.R., 1984. Identification of the inverse relationship between farm size and productivity: an analysis of peasant agricultural production. Oxford Economic Papers 36, 131–145. Chand, R., Prasanna, P.A.L., Singh, A., 2011. Farm size and productivity: understanding the strengths of smallholders and improving their livelihoods. Economic and Political Weekly 46, 5–11. Chandrakanth, M.G., Bisrat, A., Bhat, M.G., 2004. Combating negative externalities of drought: a study of groundwater recharge through watershed. Economic and Political Weekly 39, 1164–1170. Coelli, T.J., 1996. A guide to FRONTIER Version 4.1: a computer program for stochastic frontier production and cost function estimation. Working Paper. Centre for Efficiency and Productivity Analysis, University of New England, Armidale. Coelli, T.J., Rahman, S., Thirtle, C., 2002. Technical, allocative, cost and scale efficiencies in Bangladesh rice cultivation: a non-parametric approach. Journal of Agricultural Economics 53, 607–627. Coelli, T.J., Rahman, S., Thirtle, C., 2003. A stochastic frontier approach to total factor productivity measurement in Bangladesh crop agriculture, 1961–92. Journal of International Development 15, 321–333. Coelli, T.J., Rao, D.S.P., Battese, G.E., 1998. An Introduction to Efficiency and Productivity Analysis. Kluwer Academic Publishers, Boston.

Dogliotti, S., Rossing, W.A.H., van Ittersum, M.K., 2003. ROTAT, a tool for systematically generating crop rotations. European Journal of Agronomy 19, 239–250. FAO, 2012. FAOSTAT. Food and Agriculture Organization, Rome, Italy, http://faostat.fao.org/ Fujita, K., Hossain, F., 1995. Role of the groundwater market in agricultural development and income distribution: a case study in a northwest Bangladesh village. The Developing Economies 33, 442–463. Government of India, 2010. Agricultural Statistics at a Glance. Department of Agriculture and Cooperation, http://agricoop.nic.in/Agristatistics.htm Government of India, 2012. Agricultural Census. Department of Agriculture and Cooperation, http://agcensus.nic.in/ Government of Karnataka, 2006. Karnataka at a Glance. Department of Economics and Statistics, Bangalore, India. Ghatak, M., Roy, S., 2007. Land reform and agricultural productivity in India: a review of the evidence. Oxford Review of Economic Policy 23, 251–269. Hazra, C.R., 2000. Crop diversification in India. In: Papademetriou, M.K., Dent, F.J. (Eds.), Crop Diversification in the Asia Pacific Region. FAO RAP Publication. Food and Agricultural Organisation, Rome, http://www.fao.org/docrep/003/x6906e/x6906e00.htm Hopper, W.D., 1965. Allocation efficiency in a traditional Indian agriculture. Journal of Farm Economics 47, 611–624. Jha, R., Nagarajan, H.K., Prasanna, S., 2005. Land fragmentation and its implications for productivity: evidence from Southern India. Working Paper. Australia South Asia Research Centre, Canberra, Australia. Lau, L., Yotopoulos, J., 1971. A test for relative efficiency and application to Indian agriculture. American Economic Review 61, 94–109. Manjunatha, A.V., Speelman, S., Chandrakanth, M.G., Van Huylenbroeck, G., 2011. Impact of groundwater markets in India on water use efficiency: a data envelopment analysis approach. Journal of Environmental Management 92 (11), 2924–2929. McPherson, M.F., 1982. Land fragmentation: a selected literature review. Development. Discussion Papers, Harvard Institute for International Development, Harvard University, pp. 4–8. Mearns, R., 1999. Access to land in rural India. Working Paper. The World Bank, Washington, DC. Meinzen-Dick, R., 1997. Groundwater markets in Pakistan: participation and productivity. Research report-105, International Food Policy Research Institute, Washington, DC. http://pdf.usaid.gov/pdf docs/PNABZ947.pdf Miranowski, J., Orazem, P., 1994. A dynamic model of acreage allocation with general and crop-specific capital. American Journal of Agricultural Economics 76, 385–395. Nel, A.A., Loubser, H.L., 2004. The impact of crop rotation on profitability and production risk in the Eastern and North Western Free State. Agreko 43, 101–111. Niroula, G.S., Thapa, G.B., 2005. Impacts and causes of land fragmentation, and lessons learned from land consolidation in South Asia. Land Use Policy 22, 358–372. Niroula, G.S., Thapa, G.B., 2007. Impact of land fragmentation on input use, crop yield and production efficiency in the mountains of Nepal. Land Degradation and Development 18, 237–248. Palanisami, K., 2009. Water markets as a demand management option: potentials, problems, and prospects. In: Saleth, R.M. (Ed.), Promoting Irrigation Demand Management in India: Potentials, Problems, and Prospects. International Water Management Institute, Colombo, pp. 47–69, http://www.iwmi.cgiar.org/Publications/Other/PDF/NRLP%20Proceeding-3.pdf Parikh, K., Nagarajan, H.K., 2004. How important is land consolidation? Land fragmentation and implications for productivity: case study of village Nelpathur in Tamil Nadu. Occasional Paper. Department of Economic Analysis and Research, NABARD, India. Puran, M., Manjunatha, A.V., Bauer, S., Mirza, N.A., 2011. Technical efficiency and environmental impact of Bt cotton and non-Bt cotton in north India. AgBioForum 14, 164–170. Raghbendra, J., Hari, K.N., Subbarayan, P., 2005. Land fragmentation and its implications for productivity: evidence from southern India. ASARC Working Papers 2005/01. Australian National University, Canberra, Australia. https://digitalcollections.anu.edu.au/bitstream/1885/43002/2/WP2005 01.pdf Rahman, S., 2003. Profit efficiency among Bangladeshi rice farmers. Food Policy 28, 487–504. Rahman, S., Rahman, M., 2008. Impact of land fragmentation and resource ownership on productivity and efficiency: the case of rice producers in Bangladesh. Land Use Policy 26, 95–103. Rahman, S., 2009. Whether crop diversification is a desired strategy for agricultural growth in Bangladesh? Food Policy 34, 340–349. Ram, K.A., Tsunekawa, A., Sahad, D.K., Miyazaki, T., 1999. Subdivision and fragmentation of land holdings and their implication in desertification in the Thar Desert, India. Journal of Arid Environments 41, 463–477. Ramachandra, T.V., Uttam Kumar, 2004. Geographic resources decision support system for land use, land covers dynamics analysis. In: Proceedings of the FOSS/GRASS Users Conference, Bangkok, http://gisws.media.osakacu.ac.jp/grass04/viewpaper.php?id=5 Sauer, J., Frohberg, K., Hockmann, H., 2006. Stochastic efficiency measurement: the curse of theoretical consistency. Journal of Applied Economics 9, 139–165. Schultz, T.W., 1953. The Economic Organization of Agriculture. McGraw Hill, New York.

A.V. Manjunatha et al. / Land Use Policy 31 (2013) 397–405 Schultz, T.W., 1964. Transforming Traditional Agriculture. Yale University Press, New Haven. Sharma, P., Sharma, R.C., 2006. Factors determining farmers’ decision for buying irrigation water: study of groundwater markets in Rajasthan. Agricultural Economics Research Review 19, 39–56. Tan, S., Heerink, N., Qu, F., 2006. Land fragmentation and its driving forces in China. Land Use Policy 23, 272–285.

405

Wang, J., Cramer, G.L., Wailes, E.J., 1996. Production efficiency in Chinese agriculture: evidence from rural household survey data. Agricultural Economics 15, 17–28. Wilson, P., Hadley, D., Ramsden, S., Kaltsas, L., 1998. Measuring and explaining technical efficiency in UK potato production. Journal of Agricultural Economics 48, 294–305.