Enhancing agricultural intensification through contract farming: evidence from rice production in Senegal

Abstract

Background

Agricultural intensification is important in increasing agricultural production and productivity. Especially in developing countries faced with a rapidly growing population and a concomitant low agricultural productivity and food insecurity, intensification is considered very important to boost productivity and feed an ever-growing population. However, intensification remains low in developing countries owing to constraints such as inexistent or imperfect input and output markets, and weak public institutions. The institution of contract farming (CF) can address these constraints and may thus contribute to agricultural intensification. However, how CF contributes to agricultural intensification has not been well explored. In this study, we explore the role of CF towards agricultural intensification in rice production in Senegal.

Methods

To estimate the effects of CF on agricultural intensification proxied by fertilizer use, expenditure on fertilizer, improved rice varieties, and tractor use, we employ a high-frequency unbalanced panel data. We estimate a correlated random effects model to address time-invariant unobserved heterogeneity related to participation in contract farming.

Results

Our results show that CF is positively correlated with fertilizer use, fertilizer expenditure, tractor use, suggesting that CF contributes towards agricultural intensification.

Conclusion and policy implications

The institution of CF can play an important role in spurring agricultural intensification in developing countries. Policy makers, especially in developing countries can harness the potentials of CF to increase agricultural intensification and transform the agricultural sector. However, the negative environmental effects of such intensification should be considered and contracts should be better designed to rather promote sustainable intensification.

Introduction

Over 2 billion people are food insecure while 700 million live below the poverty line [28, 73]. Most of them constitute smallholder farmers living in sub-Sahara Africa (SSA) [21]. Reversing such trends and achieving SDGs 1 and 2 of no poverty and zero hunger, respectively, warrants agricultural transformation marked by increased agricultural production and productivity [32, 50, 70]. To increase agricultural production, smallholder farmers especially in SSA have generally relied on expansion of agricultural lands and extensive use of labor [70].

The viability of such strategies is being threatened by factors such as population growth, environmental concerns, and structural transformation. For example, population growth has led to the conversion of arable land to non-agricultural purposes such as infrastructural development [68]. This trend is particularly common in West Africa Hermann et al. [33]. Additionally, converting forests and grasslands to agricultural land is associated with high environmental costs such as loss of biodiversity [34]. The rapid structural transformation, characterized by the strong emergence of an off-farm sector, which attracts labor, also suggests that the amount of labor available to agriculture is reducing [74]. At the same time, these changes put more pressure on the agricultural sector that has to feed an ever-growing population. To address these challenges, agricultural intensification has emerged as an important solution [11, 16, 52, 70].

Agricultural intensification involves increasing agricultural production and productivity with little or no expansion of the area under cultivation [50]. It is argued to be an important facet of agricultural transformation, which also constitutes aspects of commercialization. Agricultural intensification involves the use of modern inputs such as improved varieties and fertilizers and mechanization technologies such as tractors. Improved seed varieties are usually characterized by their yield-increasing potential and resistance to environmental stressors, and may thus lead to increased productivity especially when combined with fertilizers. The use of agricultural machinery such as tractors also reduces the demand for agricultural labor for activities such as land preparation, allowing farmers to time properly their farming operations [38]. Moreover, it is argued that mechanization increases the profitability of other inputs [40]. Intensification may also encompass the use of irrigation technology. The availability of irrigation technologies reduces the reliance on rainfall, allowing farmers to produce multiple times in a year, maximizing the output from their lands. In fact, Suri and Udry [65] argue that irrigation and fertilizers are key technologies for agricultural intensification due to their ability to cancel out natural differences in the supply of water and farmland. With low productivity, a rapidly growing population and rising food insecurity, SSA is ideal for intensification.

However, agricultural intensification remains quite low in SSA. The region constitutes the lowest in terms of fertilizer use [16, 61, 65]. The region also constitutes the lowest in terms of mechanization and irrigation [65]. Such low intensification has been attributed to a host of binding constraints such as lack of finance [25, 34, 62], inexistent or imperfect markets for inputs and outputs [18, 34], and institutional factors such as weak and inefficient public extension services [18, 47]. Addressing these constraints may increase farmers’ access to modern production methods and lead to intensification. An institution capable of addressing such constraints is contract farming (CF).

Contract farming is a pre-planting arrangement between a buyer-typically an agroprocessor/supermarket- and a farmer or producer. This arrangement provides the farmer with specific characteristics to support production in exchange for agricultural produce that meet certain characteristics [55, 67]. CF can take many forms, but they essentially provide financial or production support to farmers. Hence, CF can contribute to agricultural intensification in many ways. CF may reduce financial constrains by providing farmers with cash which may allow them to access modern inputs such as fertilizers and improved seeds, hence contributing to intensification. Some CF arrangements also provide modern inputs such as fertilizers and improved seeds as well as technical support allowing farmers to intensify their production. Also CF shares the risks between producers and buyers allowing farmers to invest in agricultural production through improved technologies [7]. By linking farmers to modern value chains, CF provides a ready market for farm produce.

Accordingly the institution of CF has been widely used to improve agricultural value chains especially in Africa [62, 63]. Indeed, a large body of literature shows that CF benefits smallholders in terms of income and food security [5, 7, 12, 30, 42, 57, 63]. However, it is not clear whether these benefits result from intensification. Since CF provides price premiums, such increase in income may have resulted from price premiums rather than increased agricultural productivity through intensification. While increases in price premiums maybe important, their benefits may not be as high compared to intensification since they may only raise farmers’ income for the time being. In addition to raising smallholder income through increased productivity, intensification is key towards structural transformation [4]. Hence, most agricultural policies aim at promoting agricultural intensification. Therefore understanding how CF contributes to intensification offers new avenues to spur agricultural intensification and agricultural transformation. Moreover, understanding the role of CF on intensification provides clearer mechanisms through which contracts operate and can aid in better contract designs.

In this study, we evaluate the relationship between participation in CF and intensification, proxied by the use of modern inputs such as inorganic fertilizers, and improved rice varieties as well as use of mechanization such as the tractor. We employ a high frequency three-wave pseudo-panel data for our analysis. We estimate correlated random effects model to address potential time-invariant unobserved heterogeneity associated with participation in CF.

Our study contributes to the literature on agricultural intensification in different ways. First, we provide novel and supporting insights on how CF can lead to agricultural intensification through the use of fertilizers, improved seeds, and mechanization. To our knowledge, only Mano et al. [40] have examined the role of CF on agricultural intensification. However, they focus on adoption of agronomic practices and use of tractors without looking at improved seeds, which may constitute one of the most important aspects of intensification. Relatedly, Arouna et al. [7] highlight that CF may lead to increased yields through fertilizers and pesticides but they do not consider any form of mechanization. We go beyond these studies by looking at these three important aspects of intensification, fertilizer use, improved seeds, and tractor use, which may be complementary. For example, mechanization has been shown to improve the profitability of other inputs such as improved seeds and fertilizers [40]. These indicators are also argued to reflect agricultural intensification [16, 34, 52].

Our second contribution comes from the nature of our data. Existing studies [7, 40] use one period cross-sectional datasets. While the use of cross-sectional datasets may not be a problem in itself, it may limit our understanding of the relationship between CF and intensification, which is arguably a dynamic process that can be properly understood under different periods [16]. Moreover, it is difficult to estimate causal effects using cross-sectional data. The use of panel data allows us to understand properly how CF can lead to intensification as well as better estimate causal effects. Beyond using a panel dataset, we also provide new evidence from a different production environment, which increases the external validity of other studies. Although we focus on Senegal, the challenges in rice production such as limited mechanization reflect those in other West African countries such as Cote d’ivoire, Nigeria, and Benin where CF is also an important institution in rice production [40, 42, 62]. Hence, our results may also be applicable in such settings.

Lastly, we focus on CF arrangements in rice, an important staple crop. Focusing on rice is important because rice is instrumental for the food security of West Africa [8, 62]. According to Soullier et al. [62], West Africa is the second highest importer of rice. Given the strategic role rice plays in food security in West Africa, it is important to understand how its production can be improved. Hence, understanding if CF leads to intensification contributes to the broader discourse of improving food security. Additionally, rice is one of the few cereals in which yields can be boosted through intensification [24]. Moreover, most studies on CF have focused on specialty/high-value crops such as oil palm [21, 31, 59], sugarcane [71], and fruits and vegetables [5, 22, 57]. However, staple crops such as rice are argued to have broader welfare effects [43]. Hence, understanding whether CF leads to intensification of a staple crop as rice lends more evidence to policy actions aimed at simultaneously improving productivity and reducing poverty.

More generally, our study contributes to broader debate on agriculture and food security. While agriculture is already a key sector in terms of employment creation and food supplies, agricultural intensification has been tipped to be important for food security, income, and structural transformation [4, 20, 27]. Particularly, the intensification of a crop like rice, which has enormous food security implications in Africa cannot be overemphasized. Intensification can increase agricultural production and productivity, increasing the quantity of own-produced food, and consequently improve food security. Increased production and productivity may also increase households’ market participation, leading to increased income. In Ghana, for instance, Addison et al. [1] show that the uptake of improved rice seeds and fertilizers improved farmers’ incomes. Such income can increase market access and increase household food security. Hence, studying the role of CF in increasing intensification is important not just for rice production systems but for broader goals such as food security, agricultural development, and structural transformation.

The remainder of the paper proceeds as follows. In section two, we provide the context of the study. Section three provides a conceptual framework. Section four presents the data and methodology. In section five, we present and discuss the results while section six concludes and highlights some policy implications.

Study context

This study focuses on Senegal, a West African country where rice is a major staple [66]. Although rice is an essential part of the Senegalese diet, the country has mostly relied on imports due to failure of local production to meet local demand [8, 62]. Importing over 1 million tons of rice per annum and spending a whooping 654 M USD, the country ranks as one of major importers of rice in Africa [29, 62]. Besides importing to fill the demand gap, most Senegalese also prefer imported rice for both culinary reasons and its relatively cheaper price [62]. Realizing that investing in the rice value chain could reverse such trends, the government has implemented a series of interventions aimed at boosting local rice production. Between 2009 and 2014, the Senegalese government implemented two national rice sufficiency programs, which involved professionalization of the rice value chain and provision of rice millers [63]. Despite these interventions, rice production remained very low marked by a reduction in yields [26].

To further boost production, the intensification of rice production system emerged as a policy option (Africa [2, 19, 37]). This intensification, which could increase yields with more efficient input use comprises of use of irrigation facilities, improved seeds, chemical or organic fertilizers as well as machinery such as tractors [37], Africa [2]. These interventions were implemented in the Senegal river valley area (SRVA) which is a major rice production basin, accounting for over 80% of national rice production [64, 69]. The SRVA area offers a unique landscape for such intensification. The availability of ample irrigation water means that farmers can produce rice multiple times in a year. This is also made possible by the availability of short-term high-yielding and climate-resistant varieties like the Sahel 108. The availability of tractor would also reduce labor costs, ensuring that farmers properly time their operations.

Despite these opportunities, intensification remains very low in the area. This low intensification has been attributed to a host of factors such as limited finance by farmers [25, 62], Africa [2], limited availability of machinery which affects the timing of farming operations like land preparation [66] and limited institutional support to farmers [63]. For example, MacCarthy et al. [41] argue that less than 10% of the available land is irrigated, reducing farmers’ abilities to engage in multiple cultivation within the same year. Other studies have also cited limited use of fertilizer and machinery in rice production [41], Tanaka, [66]. To simultaneously address these challenges and boost intensification, CF schemes emerged in the area.

Two main types of contracts were introduced in the area [63]. On the one hand are production contracts between farmers and a private miller. In addition to cash, such contracts also provide farmers with production support such as tractor services to engage in land preparation and threshing machine to prepare rice upon harvesting [63]. The farmers in turn repay through rice paddy commensurate to the value of the loans they received. Independent farmers can sign such contracts. On the other hand are marketing contracts, which provide mainly cash support to farmers. Such contracts involve three parties: farmer organizations, millers, and the agricultural bank [63]. The contract is on a collective liability basis where the agricultural bank provides loans to farmer organizations and the organizations distribute money to farmers. The farmers then repay their loans by supplying paddy to millers who pay them through banks. This ensures that the bank directly deducts the liabilities of the farmer organizations. Failure to repay loans implies that farmer organizations are not eligible for more loans in the following season.

While such contracts are expected to boost agricultural intensification in the rice sector, there is little evidence to support such. Rather, studies in the area have looked at the welfare implications of such contracts. For instance, Ndip and Sakurai [49] show that participation in CF arrangements in the SRVA is associated with improved diet quality. Their study, however, fails to distinguish the types of contracts. Soullier and Moustier [63] also analyzed the welfare impacts of such contracts providing mixed evidence. Their results suggests that marketing contracts are not profitable while production contracts are. This may provide suggestive evidence that production contracts that provide support to farmers such as fertilizers and machinery services may have thus boosted their production. However, whether such contracts indeed boosted rice intensification has not been empirically studied. This study aims to fill this gap. Understanding whether CF indeed boosts intensification is important given that CF arrangements have been criticized on both social grounds such as increased demand for labor [54, 55, 58].

Conceptual framework

Agricultural intensification holds potentials to transform the agricultural sector of developing countries. First proposed by Boserup [17] and accentuated by Ruthenberg [60], intensification is seen as important strategy to improve agricultural productivity amidst rapid population growth, which may lead to reduction in arable land and a concomitant increase in the demand of food. Intensification hinges on the use of modern farm inputs to boost agricultural productivity [16]. For farmers to harness the benefits of intensification, there is a need for an enabling environment that will grant farmers access to improved inputs as well as machinery. An enabling environment would constitute, for example, readily available and accessible input and output markets, which may increase farmers’ access to inputs. Public extension services may also constitute part of the enabling environment by providing information about modern production methods using modern inputs. The big question then is how to provide an enabling environment. Policy options such as subsidies and improved public extension are plausible. However, in many developing countries, public extension systems are inefficient and markets are imperfect or weak [18, 47]. An institution such as CF is capable of producing such an enabling environment.

CF arrangements may take various forms but are usually categorized as production and marketing contracts [14, 55]. These contracts usually have distinct features, although some features may overlap. However, the features of CF arrangements suggest that it may lead to agricultural intensification.

Production contracts usually involve the provision of inputs such as improved seeds, fertilizer, pesticides, and machinery services to farmers [13, 36]. This may directly engender agricultural intensification as farmers can use such inputs to boost agricultural productivity. In Benin, Arouna et al. [7] show that this is the main mechanism through which CF participation increases farmers yield. In other words, CF may have led to agricultural intensification. The availability of such inputs may also induce the expansion of cultivated area if the initial area under cultivation was small. Still in Benin, Arouna et al. [7] further show that participation in CF increases the area under cultivation. Contracting firms also provide machinery services such as tractors, which allow farmers to perform crucial farm operations. Technical support provided by CF may also involve training farmers on various aspects of modern production. In Madagascar, Minten et al. [46] show that CF arrangements trained farmers on the production of organic manures, which led to their increased use on farmers’ plots. The provision of inputs such as seeds and fertilizers and the availability of machinery service suggests that CF may lead to complementary use of modern technologies, which may increase productivity.

Marketing contracts on their part usually provide farmers with cash loans during the production season [36, 55]. Such loans may allow farmers to buy improved seeds, fertilizers and hire machinery services, which may also lead to intensification. Unlike production contracts, the marketing contracts allow farmers to be responsible for their own production by selecting inputs and services they deem necessary, as long as it allows them to meet their own end of the contract. The features of both types of contracts suggest that they may be important for agricultural intensification. While production contracts may directly affect intensification, marketing contracts may work indirectly.

CF may also engender intensification indirectly. A salient feature of CF arrangements is that they provide a guaranteed output market, the absence of which may deter farmers from investing in agricultural production since they are not sure of selling their produce [9, 25]. Even if farmers have the ability to intensify production by investing in fertilizers and machinery, the absence of markets nay make them more risk-averse, discouraging such investments. By providing a guaranteed output market CF reduces farmers’ risk-averseness allowing them to invest in modern inputs.

However, CF may not always lead to intensification due to issues such as input and cash diversion. Although CF arrangements may provide farmers with inputs, the farmers may divert such inputs to other crops, which may not lead to intensification. In extreme cases, the farmers may even sell such inputs to cater to other immediate household needs such as sickness of a household member or provision of food. This is even worse in the case where farmers are provided cash as in the case with marketing contracts. Farmers may divert the cash and may end up not using it for agriculture-related reasons. In such cases, CF may not lead to intensification.

Data and methods

Farm survey

The data for this study comes from a field survey conducted in the Podor and Dagana departments of the SRVA, which is argued to be the main rice production area in Senegal [66, 69]. The survey was implemented in 2021 by trained enumerators. A multistage sampling procedure was employed. In the first stage, 180 farmer organizations were randomly selected from 3304 organizations. Farmer organizations were chosen because the contracts in this area are awarded on a collective liability basis through such organizations, although farmers cultivate individual plots. However, farmers may also sign independent contracts with rice millers. From these organizations, rice-producing households were randomly selected which served as the sample of the study.

Although the survey was done in 2021, information about rice production and other variables was obtained for different production seasons. A key advantage of the SRVA is the existence of multiple cropping seasons within a year which allows farmers to cultivate rice several times within the same year [25]. The respondents were asked about rice production in each of the seasons from 2019 to 2021 (Dry cold season 2018–2019, Dry hot season 2019, Winter 2019, dry cold season 2019–2020; dry hot season 2020, Winter 2020; dry cold season 2020–2021, and dry hot season 2021). However, we only focused on three seasons because most farmers cultivated rice in these seasons (dry hot season 2019, dry hot season 2020, and dry season 2021) and we could thus get complete information on most of the variables. We use this seasonal data to construct a pseudo-panel data of three waves. Given that some data points were missing for some of the households, we end up with an unbalanced panel data of 547 observations over three seasons which we use for the main analysis. It is important to note that the sample in each season varies. For each production season, the farmers were asked about the use of contracts and inputs for rice production such as chemical fertilizers, expenditure on fertilizers, rice varieties planted, and use of tractors for any activity. Since these variables change over seasons, it introduces variation in the data.

Measurement of key variables

Our measures of intensification include use of chemical fertilizer, expenditure on fertilizer, use of improve rice varieties, and use of tractor. Other studies have used similar variables as indicators of intensification [16, 52]. They are measured as follows. We measured chemical fertilizer based on the quantity used. Households were asked the amount of chemical fertilizer used for rice cultivation. Our measure of chemical fertilizer is the total amount in kilograms used by households over the production season. We also measured expenditure on fertilizer based on the amount of money farmers spent on fertilizer during the production season. The use of improved seeds is measured as a dummy variable which, takes a value of one if the farmer uses Sahel 108 and zero otherwise during the production season. We focus on Sahel 108 because of its popularity in the area. The variety is also known to be short-seasoned and climate-resilient (Africa [3]). Hence, its use by farmers may lead to improved rice yields. Tractor use is also measured as a dummy variable, which takes a value of one if the farmer uses tractor for at least one farm activity for rice production during the production season. Because these farming activities are usually time and labor demanding and may thus delay planting, the use of tractor helps to overcome such problems. Hence, tractor use may be an important aspect of intensification. Our treatment variable of interest, contract farming, is measured as a dummy variable. We also measure it by seasons. It takes a value of one if a farmer uses any type of contract in the season considered and zero otherwise.

Empirical framework

Estimation strategy

We are interested in estimating the effect of CF on agricultural intensification measured through inorganic fertilizer use, expenditure on fertilizer, use of improved seeds, and use or tractor. To do this, we estimate models of the form

$$I_{{{\text{it}}}} = \beta_{0} + \beta_{1} {\text{CF}}_{{{\text{it}}}} + X_{{{\text{it}}}} \beta_{2} + \overline{{X_{i} }} \beta_{3} + \varepsilon_{{{\text{it}}}}$$

(1)

where i and t index a household and the time (season) respectively, I is the vector of our intensification variables, CF is the contract farming status of a household. The vector X includes household time-variant characteristics such as the household size, educational level, marital status, age of household head, livestock ownership, other income and off-farm income, $\overline{{X_{i} }}$ is the average of the time-variant characteristics. Our parameter of interest is $\beta_{1}$ which gives the relationship between CF and the intensification variables. A positive sign of this coefficient suggests that CF is positively associated with agricultural intensification. Ideally, Eq. (1) can be estimated using a random effects but this may result in biased estimates due to endogeneity. Endogeneity may arise from measurement error, reverse causality, and unobserved heterogeneity [12], Wooldridge et al. [72]. It is possible that unobserved factors that affect a household’s decision to participate in CF will simultaneously affect intensification. For example, highly entrepreneurial and motivated farmers who are likely to participate in CF are more likely to use fertilizers and improved seeds to boost their yields. Hence, even without CF, they will normally use more fertilizers. In addition, there are possibilities of measurement errors. Since most responses especially about the use of contracts is based on recall data, farmers may misreport their involvement in contracts in past seasons introducing some measurement errors. Reverse causality may arise from the fact that farmers who practice intensive agriculture are more likely to join CF schemes since such schemes are usually selective [10]. This may also bias our estimates. To reduce the bias in our estimates, such endogeneity issues must be addressed. Since we cannot fully address endogeneity, we rather address time-invariant unobserved heterogeneity associated with CF.

Identification strategy

To address time-invariant unobserved heterogeneity and reduce the bias in our estimates, we explore the panel nature of our data. The two most widely used panel data estimators are random (RE) and fixed effects (FE) whose difference lies in the way the unobserved heterogeneity is treated Wooldridge et al. [72]. The random effects model assumes the unobserved heterogeneity is uncorrelated with the observed covariates and can thus be added to the error term. However, this is a very strong assumption. As already discussed, unobserved heterogeneity is likely to be correlated with CF in our case and RE will produce biased estimates. The fixed effects estimator may be used. The fixed effects model on its part assumes that the unobserved heterogeneity is correlated with the observed covariates but can be eliminated by averaging out the variables. While this may eliminate unobserved heterogeneity and reduce the bias in our estimates, it relies mostly on within variation in the variables. Hence, with little variation in the variables overtime as in our case, the estimates are less efficient [44]. Moreover, the use of fixed effects to estimate non-linear models may lead to biased and inefficient estimates owing to the incidental parameter problem. To overcome these problems, we utilize the correlated random effects (CRE) estimator also known as the pseudo-fixed effects or the Mundlak device. While quite similar to the fixed effects, this estimator eliminates unobserved heterogeneity by modelling it as a linear projection of the averages of time-variant variables. Hence, we include the time averages of time-variant variables in our regressions $\left( {\overline{{X_{i} }} } \right)$. We also include seasonal dummies in our regressions. The CRE is also advantageous over the fixed effects in that it produces less bias estimates when there is little variation in the variables [48]. We, therefore, employ the correlated random effects estimator. For the non-linear models such as use of tractor and improved seeds, we used the linear probability model in the CRE framework due to its non-reliance on functional form and ease of interpretation [6]. It is worth noting that the CRE estimator only deals with time-invariant unobserved heterogeneity. To the extent that time-variant unobserved heterogeneity affects participation in CF, we cannot fully claim causal identification. Also, reverse causality issues cannot be addressed with the CRE estimator. In summary, we do not appropriately deal with the endogeneity of contract farming, hence we interpret our estimates as associations. This is one of the major limitations of the study. To fully address endogeneity issues, the CRE framework could be combined with a control function in a two-stage approach. However, this requires the use of an instrumental variable, which we lack. Hence, we could not employ such a method.

It is worth noting that some household report zero chemical fertilizer and expenditure in some seasons. To retain them for analysis, we perform an inverse hyperbolic sine transformation on the outcome variable following Bellemare and Wichman [15].

Results and discussion

Sample characteristics

The characteristics of the sampled households are presented in Table 1. Starting with our outcome variables, households use an average of about 1000 kg of inorganic fertilizer and spending about 189,000FCFA on fertilizer. About 89% of sampled households use a tractor. This is relatively large given that smallholder farmers in SSA are generally known for limited use of agricultural machinery [65]. In this case, contracting firms provide farmers with tractor services which may explain the high value. In terms of improved seeds, about 70% of the sampled households used the rice variety Sahel 108. This variety is known for its high yielding potential and climate resilience (Africa [3]). In the full sample, about 18% of households engage in contract farming. Although CF is argued to benefit households, its adoption is relatively still very low. This may be due to the fact that contracts are selective with wealthier farmers more likely to be selected [10, 45].

Table 1 Sample characteristics

Most of the sampled households are headed by males who are mostly married and have attained at least primary education. Household heads are about 57 years of age. Only about 17% of households earn off-farm income suggesting that agriculture is the main activity in the area. Besides crop production, households also produce livestock and hold assets worth over 230,000 FCFA.

In Table 2, we do a mean comparison of contract and non-contract farmers. There is no significant difference in fertilizer use between the two groups. However, contract farmers expend more on fertilizers than non-contract farmers do. Non-contract farmers are more likely to use tractors compared to contract farmers but there is no difference in their use of improved rice varieties. These results are mixed, and do not provide any clear pattern as to whether contract farming may lead to intensification. However, such results may not be very insightful since both observed and unobserved factors may affect the relationship between participation in CF and intensification. To better understand the relationship between CF and intensification such factors must be controlled for. We do so in our econometric analysis.

Table 2 Mean comparison

Econometric results

Table 3 presents the econometric results of contract farming on intensification variables. For the columns 1–4, we do not control for seasonal dummies while we control for them in columns 5–8. The results suggest that CF is positively correlated with chemical fertilizer use. That is farmers that participate in contract farming use more fertilizer. Our results corroborate those Arouna et al. [7] who show that participation in CF increases yields through increased fertilizer use. This finding implies that CF may play an important role in intensifying rice production systems. Increasing fertilizer use is an important aspect towards intensification in rice production [24]. Inorganic fertilizers provide critical nutrients such as nitrogen and phosphorous and play an instrumental role in improving rice yields in SSA [35]. Additionally, fertilizer use may also boost the productivity of other inputs such as improved seeds.

Table 3 Correlated random effects estimates of contract farming on agricultural intensification

The results also show that participation in CF is positively correlated with expenditure on fertilizer. This implies that the loans that farmers receive from contract schemes are actually used to acquire fertilizer for rice production. An often cited problems with CF arrangements is the diversion of cash or inputs for other purposes [36, 53, 55]. That is, farmers may use cash obtained for the contracted crop to produce other crops or spend on another household needs. This kind of problem has been shown to affect the performance of CF schemes. It seems that is not the case in our study area. Our findings suggest that farmers actually use the cash to purchase fertilizer. In a typically low fertilizer-use setting, this further implies that farmers are also interested in intensification for higher rice yields.

We now turn to use of improved rice varieties. Although the estimates are not statistically significant, they suggest a positive correlation between CF participation and the likelihood of using improved rice varieties. It is worth mentioning that improved rice variety here is Sahel 108, which is known for its short gestation period and climate resilience. A plausible explanation for this statistically insignificant result is that many farmers in the area are already using such improved varieties, hence the difference between contract farmers and non-contract farmers is too small to show any detectable effects.

Mechanization is another important aspect of intensification. Turning to our mechanization variable, the results suggest that participation in CF increases the likelihood of using a tractor. Our results contrast those by Mano et al. [40] who showed that CF reduced the likelihood of using tractor among contract farmers. They argued that the tractor-service provider lacked experience in rice production and hence delayed with providing the service to farmers. This limited farmers’ access to tractor services. In our case, since production is organized collectively and contract loans are mostly awarded to farmer organizations, such problems are averted, increasing farmers’ access to tractor services. Tractors are used in rice production for multiple activities during land preparation. The delay of any of these activities, may affect the timing of other farm operations such as transplanting, with considerable yield implications. Hence, the use of tractor is an important aspect towards agricultural intensification. This is arguably one of the most important benefits of CF given that smallholder farmers in Africa have limited access to machinery. Improved mechanization has also been shown to offer numerous welfare benefits [38].

So far, our results suggest that participation in CF may induce agricultural intensification in rice production systems. Increased expenditure on fertilizer and fertilizer use in low-fertilizer use settings may lead to substantial increases in yields. Increasing access to mechanization such as tractor may also reducing farm labor needs while properly timing farm operations. In a production system such as rice where each activity must be properly timed, the use of machinery may play important role in intensification. By improving access to three important aspects of intensification, the institution of CF can be used to improve agricultural intensification in staple chains. Intensification of especially rice systems can lead to increased yields and food security [56].

It is important to note that intensification has its own downsides, which are mostly related to the environment. For example, extreme use of inorganic fertilizer may result in loss of important soil fauna, contamination of water bodies, and a host of other environmental issues [23]. The focus on specific varieties promotes monocultures and may lead to biodiversity loss. Tractors use fuel, which may contribute to the emission of greenhouse gases, which pose serious threats to the environment. As such, focusing on intensification without considering the environmental effects may offset the benefits of CF in the long-run. Especially in a world already challenged by a serious environmental crisis, policy discourses should be more oriented towards sustainable intensification, which simultaneously boost agricultural productivity, while protecting the environment. While it may be argued that that the institution of CF contributes to environmental degradation through its role in intensification, it is also important to note that CF may also be used to push for sustainable intensification. This can be achieved through well-designed contracts, which are environmentally oriented. For example, contracts may oblige farmers to combine both organic and inorganic fertilizers and discourage the use of practices that pose environmental issues. Additionally, incentives can be provided to farmers who implement sustainable practices. Contract farming arrangements can thus focus more on sustainable intensification. This, however, does not take away the important role the institution can play in boosting agricultural intensification in developing countries.

While contract farming may play an important role in boosting agricultural intensification, it is important to use the institution cautiously owing to its numerous social issues. For example, CF may exacerbate gender inequality at the household level by reducing women’s agency and decision-making power [39]. The ownership of crucial resources such as land are mostly a prerequisite for contracts and given that men have more access to such resources, they are most likely to engage in CF arrangements and may thus exacerbate gender inequality. Additionally, farmers with such resources are mostly the wealthiest, implying that only such farmers are likely to engage in CF arrangements, exacerbating rural inequality [51]. Contract farming may also lead to complete dependence by the farmers on the contracting firm. In case the contracting firm fails to buy the products, the farmers may become stranded. While these issues can be addressed through proper contract designs, policy makers should consider them while leveraging the institution of CF as a way of intensifying agricultural production.

Conclusion and policy implication

In this study, we evaluate the role of CF in improving agricultural intensification in rice production system in Senegal. Using a high frequency unbalanced panel data, we estimate correlated random effects model to deal with unobserved heterogeneity. We measure agricultural intensification by quantity of chemical fertilizer used, expenditure on fertilizer, use of improved rice varieties, and use of tractor services. Our results reveal that participation in CF is positively correlated with our indicators of intensification, suggesting that CF can play an important role towards agricultural intensification. Particularly, the use of machinery such as the tractor may reduce labor needs and allow farmers to properly time their farming operations, which may lead to increased yields.

Our findings hold a few policy implications. First, they support the notion that institution of CF can be used to drive agricultural intensification in staple chains in developing countries. Although our results suggest so, the attention of CF has rather focused on its welfare benefits towards farmers such as increasing their income, food security, and diets. Little attention has been given on its intensification effects, although it is passively mentioned in the literature. Policy makers can therefore harness the potentials of CF towards improving agricultural intensification. CF arrangements can be scaled to cover other important staple crops. Secondly, and relatedly, mechanization services can be made more available to farmers through CF. This may be particularly important given the reduction of farm labor owing to rural–urban migration. Lastly, designing more CF schemes that may directly provide services to farmers such as improved varieties, fertilizers and machinery may fast track agricultural intensification in developing countries.

It is also important to point out some limitations of our study. In this study, we have mostly focused on a small area in Senegal. Although it is argued that the challenges faced by the rice sector in Senegal are similar to those faced in other West African countries, it is not clear whether CF may engender intensification in other settings. Broader studies covering wider geographical areas are needed to ascertain the intensification effects of CF. Particularly, cross-country studies may be instructive. Also, we have only focused on rice, a staple crop. It is not clear if CF arrangements in other crops will produce similar effects. Second, we use a relatively short panel of three waves. It would be more interesting to use longer waves of panel data to ascertain if indeed CF leads to intensification over the long run. Future studies can leverage longer panel datasets to this end. Additionally, we use recall data to construct our panel data. This may lead to potential misreporting by farmers. Relatedly, we do cannot fully claim causality since we have only dealt with time-variant unobserved heterogeneity and not endogeneity of CF. Studies with more rigorous designs may, therefore, be necessary to clearly establish causal links between CF and intensification. Lastly, CF arrangements can come in various forms and features, although generally classified as production and marketing contracts. It would be interesting to look at how specific contracts may lead to intensification.

Despite these limitations, our results suggest that CF can play an important role towards agricultural intensification. Faced with a rapidly growing population, which is reducing the available arable land and a concomitant increase in food needs, developing countries can leverage the institution of CF to transform their agricultural production systems towards increased production and productivity, which may also improve food security.