Climate-induced increase in temperature and rainfall variability severely threaten the agricultural sector and food security in the Indian state of Odisha. Climate-smart agricultural (CSA) practices, such as crop rotation and integrated soil management, help farmers adapt to climate risk and contribute to a reduction in greenhouse gas (GHG) emissions. Therefore, this paper examines the impact of CSA practices on yield and income in vulnerable semi-arid districts of Odisha—Balangir, Kendrapara, and Mayurbhanj. We use primary survey data from 494 households collected in 2019–2020, using a multi-stage stratified sampling approach and structured questionnaire. Propensity score matching (PSM) and the two-stage least square method (2SLS) have been used to analyze the impact of CSA on income and productivity. Two instrument variables, namely distance to the extension office and percentage of adopters in a village, are used to control self-selection bias and endogeneity in our model. Both models show a positive and significant impact of the adoption of CSA on farmers’ productivity and income. The study sheds light on the significant contribution of CSA practices in fostering sustainable income growth amid environmental challenges. Overall, our results suggest that small and marginal farmers of Eastern India, a highly environmentally vulnerable area, can significantly improve their income and productivity by adopting CSA technology. Hence, policymakers should scale the adoption of CSA technology through effective extension programs.
Hinweise
The original online version of this article was revised due to a retrospective Open Access cancellation.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
Climate-induced natural disasters such as cyclones, flash floods, and landslides led to crop damage across over 36 million hectares of land in India between 2016 and 2021, resulting in approximately $3.75 billion in financial losses for farmers (Hindu 2022). Projections indicate that a 1.5 °C temperature increase could amplify annual damages from river flooding by about 49%, and cyclone-induced damages are expected to rise by 5.7% (Hindu 2022). Given this state of heightened vulnerability, adopting climate-smart agriculture (CSA) technology emerges as a potent strategy because it is likely to enhance productivity, reinforce resilience against climatic shocks, and reduce greenhouse gas emissions while ensuring food and income security for farmers (FAO 2011; Lipper et al. 2014; Rahut et al. 2021; Aryal et al. 2022; Tanti et al. 2022). CSA practices yield a dual advantage for private and public stakeholders and play a pivotal role in alleviating poverty among rural communities (Pretty 2008; Branca et al. 2021). A multitude of studies highlight the triple-win dynamic inherent in CSA practices: enhanced agricultural productivity and crop yields; promotion of environmental sustainability; and, lastly, socio-economic gains, bolstering the livelihoods of farmers in both developing and developed nations (Xiong et al. 2014; Challinor et al. 2014; Mungai et al. 2016; Makate et al. 2017; Lan et al. 2018).
There has been a growing interest in CSA practices and their impact on agricultural productivity and income (Mizik 2021; Shahzad and Abdulai 2021). Sain et al. (2017) demonstrate that various CSA practices, including the use of heat- and water-tolerant maize varieties and pest- and disease-resistant bean varieties, conservation tillage, agroforestry, and irrigation, have yielded a favorable financial return for Guatemalan farmers. Mango et al. (2018) ascertain that implementing CSA practices, such as small-scale irrigated farming and improved seed varieties, substantially increases agricultural revenue and net income. Makate et al. (2019) confirm that the joint adoption of multiple CSA practices has a more positive impact on productivity and income for smallholder farmers than single adopters. Sardar et al. (2021) conducted interviews with 420 farmers across three agroecological zones in Pakistan and discovered that farmers adopting multiple CSA show a 48% increase in farm revenue per hectare compared to non-adopting farmers. Another notable example of CSA’s positive effects on farmers’ livelihoods is from Tanzania, where Tripathi et al. (2022) have found that intercropping maize and beans with wide inter-row spacing significantly boost production and annual income. A study in Ghana by Agbenyo et al. (2022) shows that smart irrigation techniques, crop insurance, and organic fertilizers increase household income by 11%.
Anzeige
CSA practices such as mulching and trench-building have shown the potential to boost biodiversity and biocontrol, contributing to sustainable land management and carbon sequestration (Branca et al. 2021). In certain southern African contexts, such as Malawi, Mozambique, and Zambia, Mutenje et al. (2019) explore the cost-benefit analysis of implementing a blend of strategies encompassing soil conservation, crop diversification, improved seed varieties, and water conservation, all of which yielded positive economic and environmental outcomes. Similarly, Branca et al. (2021) highlight the economic advantages of minimum soil disturbance (MSD) farming over traditional tillage-based methods in Malawi and Zambia. In Ethiopia, Zerssa et al. (2021) highlight the multiple benefits of Integrated Nutrient Management (INM), agroforestry, and water-smart adoption techniques, which encompassed increased income, productivity, carbon sequestration, reduced greenhouse gas emissions, and enhanced resilience to climate change. In India, many farmers have incorporated CSA techniques, such as superior crop varieties, laser land levelling, and zero tillage, leading to an increase in farm production and lower production costs (Lopez-Ridaura et al., 2018). Furthermore, implementing soil-enhancing practices, such as regular soil bunds, has decreased crop failure risk (Kumar et al. 2020). Overall, these collective findings emphasize the transformative potential of CSA practices in enhancing agricultural sustainability and environmental resilience.
Employing methodologies such as endogenous switching regression (ESR), propensity score matching (PSM), and marginal treatment effects (MTE), past studies have illustrated the various positive impacts of CSA adoption on food security, agriculture production, and farm income across different regions. Fentie and Beyene (2019) use data from Ethiopia, affirming the positive relationship between row planting and agricultural income and food security. Habtewold (2021) conduct research in Ethiopia and highlighted that multidimensional poverty reduction is achievable by synergizing the row planting method and chemical fertilizers. Awotide et al. (2022) conduct a study in Mali and found that the most poor farmers can reap the higher benefits from CSA adoption. Bazzana et al. (2022) employ an agent-based model to investigate the influence of CSA on rural households’ welfare. Shahzad and Abdulai (2021) find that CSA practice adoption improves food security in Pakistan by enhancing dietary diversity and reducing poverty among households. In Pakistan, Ali and Rahut (2018) find a positive impact of a CSA, i.e., land laser levelling, on water use, crop yield, and household income. Pal and Kapoor (2020) demonstrate that CSA practices ensure better income and food security in semi-arid regions of India. Finally, Agarwal et al. (2022) conclude that the adoption of CSA practices had a range of benefits in that Indian state of Bihar, including a reduction in out-migration by 21% and a narrowing of the knowledge gap between genders, highlighting the multifaceted advantages of CSA practices in the context of rural development.
Odisha, an eastern Indian state, has been suffering from the growing adverse impact of climate change, such as floods and droughts, resulting in unpredictable agricultural yields (Mishra et al. 2016). This climatic uncertainty has posed various challenges to farmers, including crop losses, poor harvests, and unpaid bank loans, with some even resorting to suicide (Pattanayak and Mallick 2016; Mohanty and Lenka 2019). Rice production, especially in the rainfed lowlands of Odisha, faces a critical issue of flash floods that submerge rice plants for 10 to 15 days (Dar et al. 2017). Moreover, inconsistent rainfall and delayed monsoons in inland districts also lead to paddy crop failures. In response to these challenges, Odisha’s agricultural landscape is transforming from traditional farming methods to more adaptive climate-smart agriculture practices (Tanti et al. 2022). This shift involves embracing various CSA strategies, such as adjusting planting schedules, diversifying crops, rotating crops, using drought-resistant seeds, and implementing smart soil management techniques (Sahu and Mishra 2013). Given the current state of affairs, this study seeks to address the central research question: How does the adoption of CSA practices influence paddy yield and farm income in rural Odisha?
The present study examines the impact of two CSA practices on crop yield and farm income of small and marginal farmers in Odisha. The study adds valuable insights to the existing body of knowledge on the economic advantages of climate adaptation (Rahut and Ali 2017; Guntukula 2020; Branca et al. 2021; Rahut et al. 2021). The contribution of the study is threefold. First, this study examines the adoption of major climate-smart agricultural (CSA) practices prevalent in the study area, specifically crop rotation and integrated soil management. Second, the study offers a thorough impact evaluation of the adoption of CSA in the study area. Assessing the impact on income and yield provides a comprehensive understanding of the welfare of farmers who adopt CSA practices in comparison to those who do not. Third, two prime impact evaluation methods, propensity score matching (PSM) and two-stage least squares (2SLS), are employed to address unobserved selection bias, and to control for endogeneity issues by accounting for both observable and unobservable heterogeneities, thereby yielding more robust estimates.
Anzeige
The subsequent sections are structured as follows: The materials and methods in Sect. 2 outline the study area, sampling methodology, variable selection, and the econometric techniques employed for estimation. Sections 3 and 4 present and discuss the results, encompassing descriptive and regression analyses. The paper culminates in Sect. 5, which presents concluding remarks, policy recommendations, and limitations of the study.
2 Materials and methods
2.1 Study area
Odisha, an eastern Indian state, was chosen as the study area due to its predominant agricultural economy, where 18% of the GDP originates from agriculture, rendering it particularly susceptible to climate change. Approximately 83% of its population resides in rural areas, with 61.8% of its 17.5 million inhabitants employed in the agricultural sector. The state encompasses ten distinct agroclimatic zones, experiencing frequent climatic shocks such as droughts, erratic rainfall, cyclones, and floods within the same year. In addition, issues like sea inundation and salination are emerging concerns along its coastal areas, compounding its vulnerability to climate change.
Three districts, Balangir, Kendrapara, and Mayurbhanj, were selected for the study. Figure 1 shows the political boundaries and geographical location of the study area. Balangir, a landlocked district, is particularly susceptible to droughts and features a predominantly hot and humid climate throughout the year. The district relies heavily on monsoon rainfall between June and August, with minimal perennial irrigation coverage. Only 10% of cultivated land is connected to irrigation during the Kharif season and 4.6% during the Rabi season. Crops like rice, various grains (black gram, green gram), cotton, sesame, and oilseeds (groundnut and sunflower) are predominant, and crop diversification is employed to combat drought. Kendrapara, situated along the eastern coastline, faces frequent flash floods and cyclones. Its primary crops include paddy, groundnut, sunflower, green gram, and black gram. Notably, jute serves as the district’s main cash crop. Mayurbhanj, a district characterized by a hot and humid climate with moderate drought conditions, is primarily inhabited by tribal communities. Farming here leans heavily on indigenous knowledge-based climate-smart agriculture (CSA) practices. Rainfed agriculture prevails, supplemented by various irrigation projects. A mere 15% of cultivable land is irrigated during the Kharif season and 5.3% during the Rabi season. Major crops encompass paddy, maize, wheat, pulses, and oilseeds.
Fig. 1
Study area. Note: The map shows the administrative boundaries of study areas and the location in the State of Odisha
×
2.2 Survey and sampling design
This study used cross-sectional data obtained from 494 farmers in Odisha during the 2019–2020 period. The acquisition of primary data followed a multi-stage stratified sampling approach. Three districts that are exceptionally prone to climate risks with high vulnerability were selected.
Based on consultation and recommendations provided by Government officials from Odisha’s Ministry of Agriculture, two blocks each were chosen from each district based on their vulnerability to climatic conditions. Further, 34 villages were randomly selected from 6 blocks. Finally, employing the random walk method, 494 households were chosen across all the selected villages. A visual representation of the sampling process is presented in Fig. 2. For data collection, a comprehensive questionnaire was administered, along with interviews. The questionnaire encompassed questions about household characteristics, perceptions of climate change, encounters with climate-induced shocks, decisions regarding adopting specific practices, utilization of government extension services, income, and production for the given year. Prior to the final survey, the questionnaire underwent pre-testing and validation procedures.
Fig. 2
Sampling flowchart
×
The sample size is determined using the statistical formula proposed by Arkin and Colton (1963).
In Eq. (1), \(n\) represents the required sample (385) for undertaking the study; N represents the total number of households (4,209,660) in the study area; Z represents the confidence level at a 95% level, with a value of 1.96; \(p\) represents the estimated population proportion, set at 0.5; and \(\varvec{d}\) is used as the error limit, set at 5% (0.05). The minimum required sample size to conduct this study is 385. However, we collected data from 550 households. After cleaning the data, we used 494 samples for the analysis.
2.3 Data description
2.3.1 CSA practices adopted in the study areas
The agricultural practices adopted by farmers in the study area encompass a range of climate-smart agriculture (CSA) techniques, such as integrated soil management (ISM), crop rotations, green manure, agroforestry, high-yielding variety seeds (HYV), drought-resistant seeds, crop diversification, and micro-irrigation. However, we have specifically chosen two CSA practices, i.e., integrated soil management (ISM) and crop rotation based on their significant impact on enhancing agricultural resilience to climate change, promoting sustainable soil health and ensuring farm productivity. These selected practices were deemed significant for their potential based on the initial scoping field study observations, focused group discussion, and consultation with field extension officers for the impact evaluation in the study area. These two practices are used as binary variables in the model. The farmers who adopted the practices are termed “adopters,” and those who did not are termed “non-adopters.”
These CSA practices are widely practiced by farmers in the study area, considering the prevailing climate change conditions. Crop rotation involves the sequential cultivation of different crops in a field over time to enhance soil quality and mitigate pest and disease issues (Ball et al. 2005; Zhao et al. 2020). This method finds extensive use in Odisha, particularly in the Balangir District. Its core objectives encompass augmenting crop yields, sustaining soil fertility, and reducing reliance on synthetic inputs (Pritchard et al. 2013). ISM denotes a comprehensive approach that integrates practices for the prudent utilization of organic fertilizers, chemical fertilizers, and pesticides (Bationo et al. 2007; Chen et al. 2011). ISM contributes to the sustainable enhancement of soil structure, nutrient content, and biological activity (Killham 2011). Consequently, it fosters optimal plant growth and agricultural productivity while minimizing adverse environmental repercussions.
2.3.2 Farm income and paddy yield
This study considers total farm income and crop yield from the previous year as the dependent variables. We calculate total farm income by summing up the income reported by the participants from the kharif and rabi season crops, including paddy, maize, cotton, and vegetables annually. Importantly, secondary or off-farm income from various sources is excluded from this calculation. To enhance the distribution of the data, address outliers within the sample, and facilitate coefficient interpretation in percentage terms, we transformed the dependent variable into a natural logarithm. This aims to create a smoother data distribution and facilitate a more meaningful interpretation of the coefficients. The second dependent variable in the study is the total paddy yield, which we determined by dividing the annual paddy production by the overall cultivated land size. The cultivated land size for paddy was computed by summing up both the total owned land holdings and the land leased for paddy cultivation.
2.3.3 Other covariates
The research draws upon existing literature to identify key independent variables. These encompass household characteristics, farming experience, and the education level of the household head. Moreover, extension-related variables, including access to government extension services, membership in Self-Help Groups (SHGs), access to credit, and access to subsidies, are also integrated into the analysis. We have constructed two indexes—agricultural mechanization index and assets index—and incorporated them as explanatory variables. The “assets index” includes ownership of durable goods and dwelling conditions. The “agricultural mechanization index” is constructed based on agricultural implements and machinery availability. These indexes were developed using principal component analysis (PCA). Detailed information regarding the creation of these indexes is provided in Tables 8, 9, and 10 in the Appendix.
2.4 Econometric specifications
2.4.1 Propensity score matching
The propensity score matching (PSM) model represents a well-recognized impact evaluation technique in contemporary research. Recent studies, including Sellare et al. (2020); Vanderhaegen et al. (2018); Akoyi and Maertens (2018); and Ali and Rahut (2018), have employed the PSM model. This technique mitigates the self-selection bias that arises due to observable characteristics. It accomplishes this by pairing a subset engaged in adoption activities with a subset that lacks similar observable attributes. The objective is to ensure that the influence of self-selection bias stemming from observables is minimized in the results. The comparison involves adopters and non-adopters based on shared assistance (Becker and Ichino 2002). In the initial stage of the PSM, a logit model is utilized to regress the adoption status of each farmer against factors that potentially influence the choice to adopt CSA practices. Propensity scores for each observation are derived from this initial regression. These scores, ranging from 0 to 1, indicate a farmer’s propensity to adopt CSA practices. A higher score signifies a higher likelihood of adoption, while a score closer to 0 implies a lower likelihood.
In the subsequent stage of PSM, balanced groups are formed based on their estimated propensity scores. Various matching techniques, including kernel matching, k-nearest neighbor matching, and Mahalanobis distance matching, are used to achieve this balance. We utilized the first three techniques to estimate the average treatment effect on the treated (ATT). ATT gauges the disparity between an alternative outcome in which the same households do not implement CSA practices and an outcome observed among households that adopted these practices. A t-test is performed to investigate the difference between the matched treated and untreated observations within the PSM model to ascertain statistical significance. This evaluation determines if the outcomes from the matched treated and untreated groups are statistically significant. A statistically significant and positive difference indicates successful treatment implementation. Addressing potential bias inherent in collected data involves utilizing sample probabilities from a logit model in the first stage. Subsequently, treatment and control groups are established based on these probabilities.
2.4.2 Two-stage least square method
The possible problems that arise in an impact evaluation study regarding agricultural development programs are endogeneity, information flow and market efficiency, government control, and cyclical development of farmers (Birkhaeuser et al. 1991). Among them, endogeneity is a significant issue in the impact study. The endogenous variable correlates with the error term (Wooldridge 2013). The endogeneity problem could be dealt with through various approaches. The instrumental variables (IV) technique is considered the most efficient in the impact evaluation paradigm (Murray 2006; Burgess et al. 2016; Cawley et al. 2018). The first assumption is that there should be a significant positive correlation between the IV and endogenous variables. Secondly, IV should not correlate with the dependent variable and error term (valid). The 2SLS method follows two stages of estimation. A predicted value is generated in the first stage by regressing the IV with an endogenous variable. Then, in the second stage, the predicted value is used as an exogenous variable and regressed with the outcome variable (Gujarati 2003). We use the following reduced-form equation for y2; the first stage of the 2SLS method is to use the reduced form of the equation, which reflects the endogenous variable as the dependent variable.
where \({y}_{2}\) is our endogenous regressor (climate-smart agriculture practices), \({\gamma }_{k }\) is our estimated coefficient parameter to estimate, \({z}_{k}\) are our instrumental variables (\({z}_{1}\) = distance to extension office, \({z}_{2}\) = percentage of multiple adapters in the village), X represents the vector of explanatory variables, and \({\nu }_{2}\) is the error term. The requirement of a positive correlation between \({z}_{k}\) and \({y}_{2}\) confirms that the instruments rightly impact the endogenous regressor. From the first equation, the predicted values \(\widehat{y}\) are generated and substituted in the structural equation model 2. Our second-stage structural equation follows the following structure:
where \({y}_{1}\) denotes the dependent variable (total agricultural income and paddy productivity), which is also an unbiased estimator. \({\beta }_{j}\) is the estimated coefficient parameter, \({\widehat{y}}_{2}\) is our replaced endogenous variable, X is the vector of all other explanatory variables, and u1 is our error term. Post-estimation tests have been conducted to validate the model’s and the instruments’ significance. The multivariate Cragg-Donald Wald F test was conducted to determine the significant impact of instrumental variables on an endogenous variable. The significant F-statistic validates the instruments as strong predictors of endogenous variables (Stock et al. 2002; Cawley et al. 2018).
3 3. Results
3.1 Descriptive statistics
Table 1 presents the variable description and descriptive statistics of the variables used in the empirical models, respectively. The detailed description and sources of variables have been incorporated in Appendix Table 8. The mean annual farm income for the farm households in the study area amounts to INR 92,810.1 According to the situation assessment of agricultural households and land and livestock holdings of households in rural India, 2019, conducted by the National Statistical Office (NSO), the average monthly gross income of a farmer in Odisha was INR 5112. This income is ranked second lowest in India. The household income distribution exhibits notable heterogeneity, driven by uneven land ownership. The mean paddy yield is recorded as 16.5 quintals per acre, ranging from 7.5 to 31.2 quintals per acre. Farmers utilizing improved seed varieties and possessing favorable land for cultivation tend to achieve greater productivity than those relying on traditional seed varieties and rainfed agriculture. Approximately 62% of farmers are engaged in integrated soil management practices, while 59% have adopted crop rotation as part of their agricultural strategies.
Table 1
Variable definition and descriptive statistics
Variables category
Variable names
Description of the variables
Mean
Std. dev.
Min
Max
Dependent
Total agriculture income INR (log)c
Converted to log
11.021
0.964
7.783
13.067
Total paddy productivity (quintal per acre)c
Quintals per acre
16.513
6.464
7.500
31.25
Endogenous
Crop rotationd
If adopted = 1, 0 otherwise
0.587
0.493
0
1
Integrated soil management d
If adopted = 1, 0 otherwise
0.617
0.487
0
1
Instrumental
Distance to extension office (km)c
In kilometer
13.334
7.418
2.500
30
Share of CSA adopted in the villagec
Percentage
64.712
33.694
0
100
Control variables
Total area (acre)c
In acre
2.872
2.223
0
8.750
Farming experience (yrs)c
In number of years
25.603
12.79
1
65
Age of HH (yrs)c
In number of years
50.719
11.687
18
82
Access to govt ext (yes/no)d
If accessed govt extension = 1, 0 otherwise
0.703
0.458
0
1
Education of HH (yrs)c
Years of schooling
7.788
5.37
0
17
HH sizec
Numbers of HH members
4.886
1.656
1
11
Credit from cooperatived
If availed credit from cooperative = 1, 0 otherwise
0.462
0.499
0
1
Access to subsidiesd
If availed subsidies = 1, 0 otherwise
0.281
0.450
0
1
Asset indexc1
Index of assets1
2.988
1.421
1
5
Mechanization indexc2
Index of mechanization2
− 4.01*10^(-9)
1.3779
− 1.163
5.628
Migration (yes/no)d
If migrates = 1, 0 otherwise
0.261
0.439
0
1
Kendrapara districtd
If resident of the district = 1, 0 otherwise
0.293
0.456
0
1
Balangir districtd
If resident of the district = 1, 0 otherwise
0.210
0.408
0
1
Mayurbhanj districtd
If resident of the district = 1, 0 otherwise
0.497
0.501
0
1
cContinuous variable; ddummy variable; 1own house, types of houses, types of roofs, electricity connection, drinking water, toilet facility, ownership of mobile, ownership of television, ownership of radio; 2tractor drawn equipment, rice transplanter, chaff cutter, spray machine, sprinkler, harvester, water pump, tractor
Crop rotation, practiced on a seasonal or annual basis, involves changing crops in a field from one season to another, enhancing soil health. This practice holds prominent significance within climate-smart agriculture (CSA) as it contributes to soil health preservation, pest and weed control, and the maintenance of soil organic matter (Abegunde et al. 2019). Farmers in the Balangir district adhere to yearly crop rotation systems like rice, vegetables, oilseeds, maize-pulses/oilseeds, and fiber-pulses. Similarly, in the Kendrapara district, crop sequences such as jute–rice–pulses and rice–green gram/black gram/groundnut are followed. The Balangir and Mayurbhanj districts practice the annual rice–mustard/linseed/Bengal-gram/safflower/black-gram/lentil/green-gram rotation system. Despite the prominence of paddy cultivation, its susceptibility to climate uncertainties such as erratic rainfall, drought, cyclones, and market fluctuations renders it a risky and potentially less profitable endeavor. Farmers in the study area have adopted integrated soil management (ISM) techniques in response to climate change risks.
The first instrumental variable is the mean distance from the surveyed village to the block agriculture extension office, measuring 13 km (ranging from 2.5 to 30 km). Farmers frequently visit this extension office to gather information about various government schemes and obtain seed subsidies and other benefits. Furthermore, village agricultural extension staff typically visit one to two times per week. The second instrumental variable incorporated into the model is the percentage of adopter farmers in a village. This instrument assumes the role of a peer effect, motivating many farmers to engage in adoption. Consequently, this encourages other farmers to embrace adoption practices as well.
Within the study area, approximately 70% of farmers reported having access to extension services. The average age of sampled farmers stood at 50.7 years, with an average family size of five members. The household heads possessed an average of 7 years of education. The mean size of cultivable land is 2.8 acres, with 19% being irrigated and the remaining 81% relying on rainfed agriculture. About 59% of sampled farmers noted that their land was fertile, characterized by black clay and loamy soil capable of retaining water and moisture, making it conducive to cotton cultivation. Farmers, on average, have 25 years of farming experience. To assess their wealth, we created an asset index. We use principal component analysis (PCA), dividing the sample into five categories ranging from 1 (poorest) to 5 (richest), as outlined by Filmer and Pritchett (2001). Both asset and mechanization indexes were constructed and categorized according to low-to-high rankings.
The detailed items of PCA have been provided in Appendix Tables 9 and 10. The sample composition includes 21% of farmers from the Balangir district, 29% from the Kendrapara district, and 49% from the Mayurbhanj district.
3.2 First-stage IV results
Table 2 shows the initial outcome of the instrumental variable (IV) analysis. This outcome demonstrates the pertinence of the instruments regarding farmers’ decisions to adopt climate-smart agriculture (CSA) practices. In our context, the jointly significant nature of the endogenous variables adopting CSA is observed in conjunction with the instrumental variables. The significance holds individually and when considered together, emphasizing the influential role of both instruments in CSA adoption decisions. The significant signs of both instruments also uphold the expected directional relationship. The correlation between CSA adoption and distance to the extension office proves significant and negative, while the linkage between the percentage of multiple adopters and CSA adoption displays positive significance. The coefficient’s strength is slightly diminished when the two instruments are combined. The Cragg-Donald Wald F-statistic underscores the overall significance and importance of the instruments. In our study, this F-statistic exceeds the predefined threshold of ten for specific models, which indicates the instruments’ relevance as established by Stock et al. (2002).
Table 2
First-stage OLS regression results
I-1
I-2
Combined
I-1
I-2
Combined
Crop rotation
− 0.032*** (0.000)
0.006*** (0.000)
− 0.022*** (0.000)
0.002*** (0.018)
Integrated soil management
− 0.027*** (0000)
0.006*** (0.000)
− 0.008* (0.098)
0.004*** (0.000)
CD Wald F stat
12.250
11.190
11.910
11.390
12.880
12.250
I-1 instrumental variable-1 (average distance from village to extension office in km), I-2 instrumental variable-2 (share of household with multiple CSA adopters)
*Significant at 10% level; **significant at 5% level; ***significant at 1% level
3.3 Second-stage IV results
In the second phase of the instrumental variable (IV) procedure, the forecasted values of the endogenous variables, obtained from the initial structural equation, are included as explanatory elements in the subsequent structural equation model. The central variable of interest is presented in Table 3, while the comprehensive table incorporating control variables can be found in Appendix Table 8, Table 9, and Table 10. The two-stage least squares (2SLS) results reveal a positive impact of climate-smart agriculture (CSA) practices on farmers’ total agricultural income, as substantiated by studies like Fentie and Beyene (2019) and Sardar et al. (2021). The robustness of the IV estimate, unburdened by measurement errors, enhances its reliability compared to the ordinary least squares (OLS) estimate.
Table 3
Second-stage 2SLS regression results
Total log farm income per year
OLS
IV1
IV2
Combined
Crop rotation
0.061 (0.082)
0.486*** (0.158)
0.466*** (0.168)
0.479*** (0.153)
Integrated soil management
0.035* (0.085)
0.584*** (0.187)
0.473*** (0.168)
0.501*** (0.165)
*Significant at 10% level; **significant at 5% level; ***significant at 1% level, standard error in parenthesis
Compared to the OLS estimate, the IV estimate indicates a more pronounced effect of CSA on income. The instrumental variable (IV) analysis focuses on estimating the local average treatment effect (ATE), providing insights into the specific localized influence of the treatment on the outcome (Oreopoulos 2006; Sardar et al. 2021). While the ordinary least squares (OLS) estimate suggests insignificance between exogenous and dependent variables, the instrumental variable (IV) model illuminates a significant and positive effect on the dependent variable. This underscores the instrumental role in driving positive outcomes through adoption, particularly highlighting its affirmative impact on farmers’ total agricultural income (Abid et al. 2016). The adoption of crop rotation exhibits a noteworthy positive impact at a significance level of 5%. Farmers who embrace changing crops and diversifying their selections within the same plot over a year can anticipate larger farm income compared to their hypothetical counterparts.
Farmers adopting “integrated soil management” practices witness a significant increase in farm income positively compared to those who abstain from smart soil practices, holding other factors fixed. The consistency of these estimates across the three specifications validates the robust predictive capacity of the instruments for forecasting CSA adoption practices. This substantiates their efficacy in successfully identifying the causal impact of CSA adoption on farm family income within a specific year. As anticipated, various control variables, such as land size, access to extension services, education, farm mechanization, and credit accessibility, influence farm income (see Appendix Table 8 and Table 9). Farmers integrating farm mechanization into their operations can anticipate a 4–6% increase in farm income relative to those who do not employ mechanization. In contrast, farmers who secure agricultural credit from banks and cooperative societies are expected to experience a 10–12% reduction in their farm income compared to those without access to credit.
3.4 Results of PSM
The analysis of the average treatment effect on the treated (ATT) derived from the propensity score matching (PSM) model is detailed in Tables 5 and 6. The focal outcome variables are the total agricultural income (annually) and paddy yield (quintal/acre). PSM ensures a more balanced evaluation by mitigating bias between adopters and non-adopters in the sample selection process. The PSM results are presented in Tables 5 and 7, aligning with the study’s core objective to explore the influence of climate-smart agriculture (CSA) adoption on farmers’ primary agricultural income and productivity.
Balancing test outcomes can be found in Tables 7 and 8. These results indicate the absence of statistically significant mean disparities between CSA adopters and non-adopters across the PSM covariates. This is an advantageous characteristic indicating successful matching. The propensity score overlapping prior to and post-matching is depicted in Figs. 3 and 4. These plots offer insights into the degree of overlap in propensity scores between treated and control observations within each section of the adoption practices. Notably, the plots highlight the substantial enhancement in overlapping propensity scores after matching across all outcome indicators. This observed improvement in post-matching propensity score overlap is a highly desirable attribute of effective matching.
Fig. 3
Propensity score overlapping between treat and control groups, before and after matching for crop rotation
Fig. 4
Propensity score overlapping between treatment and control groups, before and after matching for soil treatment
×
×
3.4.1 Impact of crop rotation on yield and income
The average treatment effect (ATT) on total agricultural income is positive and statistically significant, indicating that the CSA adopters have higher agricultural income than the non-adopter group. This is also exhibited by the mean difference in total agricultural income, which amounts to 0.424 (radius), 0.455 (kernel), and 0.427 (5-nearest neighbor), as presented in Table 4. Similarly, Table 4 reveals that the adoption of crop rotation practices has a positive and significant effect on yield per acre. On average, there is an observed difference of 1.98 to 2.54 quintals in yield per acre when a farmer adopts crop rotation CSA practice.
Table 4
Average treatment effect on yield and total agriculture income
Outcome variables
Crop rotation
Matching methods
ATT
Standard error
Yield per acre
Radius
1.390***
0.632
Kernel
1.460***
0.652
5-Nearest neighbor
2.080***
0.910
Total agricultural income
Radius
0.424***
0.079
Kernel
0.455***
0.822
5-Nearest neighbor
0.427***
0.116
The results of three matching methods (radius, kernel, and 5-nearest neighbor method) for two outcomes variable yield and income are furnished. *Significant at 10% level; **significant at 5% level; ***significant at 1% level
The balancing test presented in Table 5 underscores the appropriate equilibrium achieved in covariates after matching. This indicates a successful alignment between adopters and non-adopters. The overlapping graph illustrated in Fig. 4 further confirms this observed balance, reinforcing the appropriateness of matching outcomes between the two groups.
Table 5
Statistical significance (p-value) of explanatory variables used in PSM before and after matching
Crop rotation
Kernel
Radius
Nearest neighbor
Un-matched
Matched
Un-matched
Matched
Un-matched
Matched
Total land area (acre)
0.816
0.194
0.816
0.373
0.816
0.029***
FarmingExperience (yrs)
0.369
0.960
0.369
0.862
0.369
0.244
Age of household head (yrs)
0.760
0.574
0.760
0.804
0.760
0.156
Access to government_extension
0.000***
0.877
0.000***
0.652
0.000***
0.699
Subsidies
0.830
0.316
0.830
0.582
0.830
0.080
Education of household head
0.147
0.512
0.147
0.769
0.147
0.191
Household size
0.613
0.710
0.613
0.772
0.613
0.410
Mechchanization Index
0.001***
0.914
0.001***
0.799
0.001***
0.127
Asset index
0.187
0.488
0.187
0.646
0.187
0.029***
Credit from cooperatives
0.004***
0.853
0.004***
0.707
0.004***
0.312
Migration
0.188
0.745
0.188
0.958
0.188
0.505
Mayurbhanj district
0.000***
0.777
0.000***
0.687
0.000***
0.098
Balangir district
0.005***
0.975
0.005***
0.772
0.005***
0.123
Overall indicators of covariate balancing
Before
After
Pseudo-R2
0.093
0.006
LR chi2
61.870
4.620
p-value of log-likelihood
0.000
0.990
Median absolute bias
12.700
3.600
Total % bias reduction (%)
75.100
18.100
Balancing test among explanatory variables. ***Significance at the 1% level, **significance at the 5% level, and *significance at the 10% level
3.4.2 Impact of integrated soil management on yield and income
Maintaining soil health through soil treatment constitutes a pivotal component of climate-smart agriculture (CSA). Excessive fertilizer, pesticide application, drought, and other adversities can transform fertile soil into one with hardened conditions. This alteration leads to soil acidity and rigidity. Counteracting this, farmland treatment involving gypsum, papermill sludges, and additional nutrients contributes to improved soil nutrition and heightened farm productivity. As shown in Table 7, integrated soil management practices significantly and positively impact both yield and total agricultural income. The average treatment effect (ATT) demonstrates a mean difference in total agricultural income between adopters and non-adopters, ranging from 27 to 34%. The balancing test presented in Table 6 underscores the appropriate equilibrium achieved in covariates after matching. This indicates a successful alignment between adopters and non-adopters.
Table 6
Statistical significance (p-value) of explanatory variables used in PSM before and after matching
Integrated soil management
Kernel
Radius
Nearest neighbor
Un-matched
Matched
Un-matched
Matched
Un-matched
Matched
Total land area (acre)
0.970
0.782
0.970
0.883
0.970
0.650
FarmingExperience (yrs)
0.046
0.592
0.046
0.426
0.046
0.315
Age of the household head (yrs)
0.280
0.504
0.280
0.302
0.280
0.153
Access to govtextension
0.000***
0.746
0.000***
0.976
0.000***
0.771
Subsidies
0.344
0.545
0.344
0.620
0.344
0.293
Education of the household head
0.467
0.772
0.467
0.701
0.467
0.694
Household size
0.333
0.744
0.333
0.761
0.333
0.226
Mechanization Index
0.022***
0.612
0.022***
0.784
0.022***
0.614
Asset Index
0.005
0.551
0.005
0.938
0.005
0.105
Credit from Cooperatives
0.044
0.804
0.044
0.699
0.044
0.044
Migration
0.662
0.297
0.662
0.342
0.662
0.033
Mayurbhanj district
0.000***
0.945
0.000***
0.630
0.000***
0.932
Balangir district
0.163
0.541
0.163
0.992
0.163
0.390
Overall indicators of covariate balancing
Before
After
Pseudo-R2
LR chi2
p-value of log-likelihood
Median absolute bias
Total % bias reduction (%)
0.116
0.022
75.830
4.200
0.000
0.994
21.300
3.900
85.600
17.400
Balancing test among explanatory variables: ***significance at the 1% level, **significance at the 5% level, and *significance at the 10% level
Table 7
Average treatment effect for yield and total agriculture income
Outcome variables
Matching methods
Integrated soil management
ATT
SE
Total agricultural income
Radius
0.348***
0.747
Kernel
0.323***
0.083
5-Nearest neighbor
0.477***
0.094
Yield in quintal per acre
Radius
1.750***
0.873
Kernel
0.707
0.675
5-Nearest neighbor
0.716
0.657
The results of three matching methods (radius, kernel, and 5-nearest neighbor method) for two outcomes variable yield and income furnished
*Significant at 10% level; **significant at 5% level; ***significant at 1% level
The overlapping graph illustrated in Fig. 4 further confirms this observed balance, reinforcing the appropriateness of matching outcomes between the two groups. Furthermore, the application of integrated soil management practices has the potential to increase yield per acre by an average of 2.5 to 3 quintals, as evidenced by the findings.
4 Discussion
Climate-smart agriculture (CSA) practices have undergone substantial innovation at the farm level, aimed at adapting to and mitigating the repercussions of climate change and other challenges linked to agricultural production. The uptake of these practices effectively diminishes the detrimental impacts of climate change and increases food output and farmers’ earnings (Eitzinger et al. 2014; Andati et al. 2023; Li et al. 2024; Singh et al. 2024). Mizik (2021) and Ishtiaque et al. (2024) revealed that the adoption rates of CSA practices remain low in developing nations. Among the reasons cited by these studies, the following factors stand out: a lack of awareness and training in these practices, weak organizational capacities, CSA technologies inadequately incentivized, limited monitoring and follow-up, lack of information dissemination, and financial incentives stemming from adopting them (Siedenburg et al. 2012). In this context, the current study holds immense significance. Our research reveals that farm households in Odisha are actively adopting various climate-smart agriculture practices, and these practices, in turn, have implications for their potential income and crop yields in the prevailing climate change conditions. As observed in the study areas, integrated soil management and crop rotation practices are popular practices within the CSA basket. These practices significantly impact economic aspects such as crop productivity and farm incomes, serving as primary drivers for resource-constrained farmers. We used standardized and rigorous impact evaluation techniques such as 2SLS and PSM to bolster the credibility of these findings. This conclusion is reaffirmed by both propensity score matching (PSM) and instrumental variable (IV) regression analyses performed in the study. Both the methods triangulated in the study have a synchronized result: they show the positive impact of CSA practices on income and yield.
The study region features a soil composition characterized by salinity and acidity. Coastal agricultural areas contend with salinity challenges due to their proximity to shorelines, while inland crop regions grapple with soil acidity. Integrated soil health management practices are designed to counter these issues by incorporating suitable anti-saline and anti-acidic agents like gypsum and lime. This approach aids in soil treatment and enhances productivity. The farmers of the study regions use a balanced fertilizer combination of bio- and chemical fertilizers. Consequently, adopters of integrated soil management practices enjoy relatively elevated yields. Both PSM and 2SLS models show that adopters of ISM have 32–58% higher income than non-adopters. A notable past study by Bhattacharyya et al. (2016) showed that the mean annual income per family increased by 43% by adopting soil and water conservation practices in India. Tiwari et al. (2010) also identified a positive impact of soil and crop management on farmers’ income and yields in Nepal. The study conducted by Bravo-Ureta et al. (2006) suggested that soil conservation practices demonstrate a positive and statistically significant association with farm income. In the drought-prone areas of India, households that adopted CSA practices had higher incomes of INR 54,717 versus the non-adopters during a drought year (Vatsa et al. 2023; Samuel et al. 2024).
Our results show that the adoption of ISM has a positive impact of 70 kg to 1.75 quintals per acre. Adolwa et al. (2019) demonstrated that adopting integrated soil fertility management (ISFM) boosted maize yields by 16–27% in Ghana. Similar findings resonate in studies conducted by Khatri-Chhetri et al. (2016) in the Indian Indo-Gangetic plains, as well as by Sardar et al. (2021) in Pakistan. Gathala et al. (2022) demonstrate a 10% increase in crop production through adopting conservation agriculture-based sustainable intensification (CASI). Our result is also in line with Khan et al. (2007) and Zhao et al. (2020) that ISM has a positive impact on crop yields in the rice–wheat cropping system. Soil nutrient management, crop improvement practices, seed management, and crop protection techniques have been enhanced by 41%, 40%, and 39% of crop yield, respectively (Andati et al. 2023). Moreover, households implementing CSA practices experienced a 20–30% increase in average annual farm income per hectare compared to those who did not adopt such practices (Belay et al. 2023). The adoption of CSA practices helped smallholder vegetable farmers increase their crop yields, net farm returns, and per capita consumption expenditures by 21%, 15%, and 13%, respectively (Torsu et al. 2024). Conservation agriculture-based sustainable intensification improves technical efficiency by 8% and 9% in productivity and technical efficiency, respectively (Paz et al. 2024). The uptake of climate-smart agricultural practices (CAPs) has a significant and positive impact on household income, net farm income, and income diversity (Sang et al. 2024).
Similarly, our findings highlight the significant benefit reaped by adopters of crop rotation practices, with a 42–45% higher agricultural income and 1.39 to 2.08 quintals of more rice yield per acre. Given the region’s vulnerability to climatic shifts, this outcome holds particular importance regarding extreme weather adaptation. Frequent droughts in the inland areas and floods along the coast make crop rotation a vital strategy. By alternating crops season-to-season, farmers can selectively cultivate crops suited to extreme climatic conditions, thus meaningfully enhancing their climate resilience and livelihood. In Africa, Kuntashula et al. (2014) reveal that crop rotation led to a maize productivity improvement of approximately 21–24% in Zambia. Wang et al. (2019) provide evidence that rice-wheat rotation with reduced fertilizer application results in significantly higher crop yield and economic benefits in People’s Republic of China. Our findings align with He et al. (2021) and Jena et al. (2023), which show that enhancing agricultural diversity via crop rotations has led to substantial improvements in the social, economic, and ecological advantages of rice production. Our findings are in line with the results of Zhao et al. (2020) that crop yields experienced a mean augmentation of 20% through the implementation of crop rotation, in contrast to the persistent monoculture methodology.
However, considerable heterogeneity within farming communities impedes the deep penetration of CSA practices, thereby limiting potential benefits. Adopting technology necessitates an awareness of its potential benefits, available technical extension support, and adequate financial backing, either through formal credit institutions or government subsidies (Tanti and Jena 2023). The Indian Government has sought to reinforce its farm subsidy policy through various input subsidy programs, encompassing provisions like stress-tolerant seeds, mechanized tool subsidies, biofertilizers, and subsidized soil testing facilities under the flagship “soil health program.”2 Therefore, to enhance scalability in climate-smart agriculture (CSA), we advocate for tailored government programs; capacity building through comprehensive training; financial incentives; practical demonstrations; and community involvement through a bottom-up, region-specific, and collaborative approach.
From focus group discussions (FGDs), we gleaned that around half of the farmers who were approached displayed little interest in soil tests. Widespread lack of primary education in rural parts of India often results in farmers’ aversion to new practices. Moreover, rough geographical terrains hinder extension agents’ access to these hinterlands, impeding the effectiveness of extension services. Challenges persist during subsidy disbursement drives, where funds are often unavailable when farmers require them. These incidents foster mistrust toward officials and government schemes, leading to adverse selection and moral hazard issues. Insights from expert group interviews with extension officials reveal that interested and dedicated farmers often withdraw from government subsidy schemes due to a lack of timely subsidy availability for seeds, fertilizer, and machinery. Conversely, less-committed farmers eventually obtain subsidies once they become available, often channeling them into non-agricultural activities. Such occurrences are prevalent across various parts of the country.
5 Conclusion, policy implications, and limitations
5.1 Conclusion
This study systematically examined the impact of adopting climate-smart agriculture (CSA) on both farmers’ income and yield per acre. The primary dataset utilized in this analysis was collected through a comprehensive survey conducted during 2019–2020 involving 494 farm households in three of Odisha’s climate-vulnerable districts. Our investigation specifically focused on evaluating the impact of two widely embraced CSA techniques—crop rotation and integrated soil management practices—on these farm households’ productivity and income levels. The results demonstrated that adopting CSA practices increases agricultural income and paddy yield. Notably, the robustness of the findings across various model specifications emphasizes the effectiveness of the instruments in providing accurate insights into the impact of CSA practice adoption. In conclusion, this study substantiates the positive influence of CSA adoption on farmers’ economic outcomes, shedding light on the potential benefits of incorporating climate-smart practices in agriculture.
Farmers are generally motivated by the potential for increased income even though they express a preference for environmental preservation. The crucial factor determining the adoption of CSA practices is the income-enhancing potential that can transform subsistence farming into a profoundly ingrained farming culture. Thus, it is essential to disseminate awareness about the benefits of CSA practices, particularly among small and marginal farmers who rely on continuous revenue generation. The findings strongly advocate for the upscaling of CSA adoption. Notably, the results highlight significant policy implications, emphasizing the influential role of economic gains and positive income effects in driving technology adoption.
5.2 Policy implications
Based on our key findings, we have outlined recommendations to promote adopting climate-smart agricultural (CSA) practices. By implementing these, policymakers can create an enabling environment that supports CSA adoption, enhances agricultural resilience, and contributes to sustainable development goals. Our suggestions are strategically designed to address the diverse challenges and opportunities associated with fostering CSA adoption. First, targeted extension services, which emphasize the importance of disseminating information about CSA practices effectively, should be promoted. Extension programs should prioritize outreach to subsistence and marginalized farmers with limited access to information and resources. Thus, by providing tailored training and support, extension services can enhance awareness and facilitate the adoption of CSA practices among these communities. Second, there is a critical need for investment in agricultural infrastructure to support CSA adoption. This includes improving access to irrigation systems, promoting sustainable land management practices, and enhancing storage and processing facilities. Policymakers can create an enabling environment that encourages farmers to adopt CSA practices and enhance agricultural productivity by investing in infrastructure upgrades. Third, financial incentives and support mechanisms to incentivize CSA adoption and scaling it up should be provided. This may include subsidies for CSA technologies, access to credit facilities, and insurance schemes to mitigate risks associated with climate variability. Thus, by providing financial support, policymakers can help alleviate the initial costs associated with adopting CSA practices and encourage widespread adoption among farmers. Fourth, the empowerment of farmers’ cooperatives, exemplified by India’s Farmer Producer Organization (FPO) system, serves as a linchpin in driving holistic climate change adaptation efforts. These cooperatives necessitate active involvement from farmers, supported by governmental financial and technical aid. Continuous financial backing is imperative to unlock the full potential of such initiatives. Fifth, capacity building and training programs are crucial in empowering farmers with the knowledge and skills needed to adopt CSA practices effectively. We recommend investing in training programs focusing on climate-smart farming techniques, sustainable land management practices, and risk mitigation strategies. By building farmers’ capacity, policymakers can equip them with the tools and resources to adapt to changing climatic conditions and improve agricultural resilience. Finally, we recommend the establishment of monitoring mechanisms. Despite the initiation of CSA practices by governments and non-governmental organizations, post-adoption monitoring and evaluation often fall by the wayside. A robust monitoring mechanism is critical to gauge the extent of CSA adoption and ensure sustained technology uptake. Stakeholders must be educated about the enduring benefits of CSA practices, recognizing that some practices require time to yield significant results. There is a need for a monitoring and evaluation mechanism to assess the continuation of CSA technology adoption. There is a need to educate farmers regarding the long-term benefits of CSA practices as some practices need a long time to get benefits.
5.3 Limitations and scope for future studies
This study provides valuable insights, but it is imperative to acknowledge its limitations. First, the scope of the research is constrained by the available resources and time frame, both in terms of finances and human resources, which restricted the study to only three districts and a sample size of just over 400 households. The dynamic nature of farmers’ adoption practices and their impact might not have been fully captured due to the lack of a panel dataset. The feasibility of repeated surveys is hindered by the time and resource constraints associated with collecting more primary data and also, farmers self-reported income data might affect estimations. The study also did not address the mitigation impact of CSA adoption and the environmental implications that should have been assessed. The agricultural production might have been shaped by factors other than climate change, such as climate variability and the rise in post-monsoon rainfall. Conducting a trend analysis and examining the impacts explicitly associated with climate change are imperative and not captured in this study. Furthermore, the geographical specificity of the primary data collection, limited to specific locations of rural settings, may need to be revised to allow the generalizability of findings to larger populations, rural-urban mixed setups, and community-based programs.
To address these limitations and enhance the study’s depth, a subsequent survey round could be conducted to assess the continuous and long-term impact of CSA adoption over the years. This approach would provide a more robust evaluation of the sustained effects of CSA practices. Furthermore, exploring the institutional dimension of CSA practices, particularly monitoring farmers’ relationships with various institutions, could offer valuable insights into how these relationships are constructed, maintained, strengthened, or dissolved over time. Future research endeavors could also look into the mitigation impact of CSA practices within the study area, offering a more comprehensive understanding of environmental implications, which would contribute to a more nuanced comprehension of the multifaceted aspects of CSA adoption.
Declarations
Conflict of interest
The authors declare no competing interests.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Description of the CSA practices considered by past literature
CSA practices
Description
Sources
Crop rotation
Crop rotation is an agricultural technique involving the sequential cultivation of various crops in a field over time, aiming to improve soil quality and reduce pest and disease problems
Abegunde et al. (2019); Khatri-Chhetri et al. (2017)
Integrated soil management
Integrated soil management (ISM) entails a holistic strategy for maintaining soil health and fertility through the synchronized implementation of diverse practices and technologies
Killham (2011); Bationo et al. (2007); Chen et al. (2011)
Instrumental variables
Distance of village from the Block Extension Office (in kilometer)
Total area (acre), years of farming experience, age of household head, access to government extension services, years of education of household head, numbers of members in the household, access to credit from cooperative societies, access to subsidies, migration, asset index, mechanization index
Tripathi and Mishra (2017); Abid et al. (2016); Bryan et al. (2009)
The table has been compiled using the literature survey
Table 9
Asset index components
Item no
Components
Description
1
Own house
If owns house = 1, otherwise 0
2
Types of houses
If house type pucca = 1, otherwise 0
3
If house type kutcha = 1, otherwise 0
4
Types of roofs
If the house has a concrete roof = 1, otherwise 0
5
If house has tiles = 1, otherwise 0
6
Electricity connection
If house has electricity = 1, otherwise 0
7
Drinking water
If house has tap water = 1, otherwise 0
8
Toilet facility
If house has toilet = 1, otherwise 0
9
Ownership of mobile
Own mobile phone = 1, otherwise 0
10
Ownership of television
Own television = 1, otherwise 0
11
Ownership of radio
Own radio = 1, otherwise 0
The asset index was constructed by considering all the above items in a table, and the index was developed using principal component analysis
Table 10
Mechanization index components
Item no
Component
Descriptions
1
Tractor drawn equipment
If own equipment = 1, otherwise 0
2
Rice transplanter
If own a rice transplanter = 1, otherwise 0
3
Chaff cutter
If own chaff cutter = 1, otherwise 0
4
Spray machine
If own spray machine = 1, otherwise 0
5
Sprinkler
If own sprinkler = 1, otherwise 0
6
Harvester
If own harvester = 1, otherwise 0
7
Water pump
If own water pump = 1, otherwise
8
Tractor
If own tractor = 1, otherwise
The mechanization index was constructed by considering all the above items in the table, and the index was developed using principal component analysis
Table 11
Second-stage full regression result
Total log income
Instrumental variable-1
Instrumental variable-2
Combined instruments
Crop rotation
0.487**
0.466**
0.479
(0.158)
(0.168)
(0.154)
Soil conservation
0.584**
0.473**
0.501**
(0.187)
(0.168)
(0.165)
Total land owned
0.0246
0.0238
0.0248
0.0249
0.0247
0.0246
(0.015)
(0.015)
(0.015)
(0.015)
(0.015)
(0.015)
Farming experience
0.0025
0.0031
0.002
0.003
0.002
0.002
(0.003)
(0.003)
(0.003)
(0.003)
(0.003)
(0.003)
Age
− 0.005
− 0.005
− 0.005
− 0.005
− 0.005
− 0.005
(0.003)
(0.003)
(0.004)
(0.003)
(0.004)
(0.003)
Govt extension
0.0614
0.0434
0.0652
0.0637
0.0627
0.0584
(0.072)
(0.074)
(0.073)
(0.072)
(0.072)
(0.072)
Subsidies
− 0.0401
− 0.0059
− 0.039
− 0.005
− 0.034
− 0.005
(0.073)
(0.071)
(0.073)
(0.070)
(0.073)
(0.070)
Years of schooling
− 0.001
− 0.003
− 0.001
− 0.003
− 0.001
− 0.003
(0.006)
(0.005)
(0.006)
(0.005)
(0.006)
(0.005)
HH size
0.023
0.014
0.022
0.016
0.023
0.015
(0.018)
(0.018)
(0.018)
(0.018)
(0.018)
(0.018)
Mech-Index
0.047
0.065**
0.048
0.063*
0.047
0.06*
(0.025)
(0.025)
(0.026)
(0.025)
(0.026)
(0.025)
Asset index
0.036
0.04
0.036
0.040*
0.036*
0.040
(0.024)
(0.024)
(0.024)
(0.024)
(0.024)
(0.024)
Cooperative credit
− 0.120*
− 0.104*
− 0.118*
− 0.101*
− 0.119*
− 0.102*
(0.063)
(0.061)
(0.063)
(0.061)
(0.062)
(0.061)
Migration
− 0.093
− 0.052
− 0.091
− 0.051
− 0.092
− 0.052
(0.070)
(0.068)
(0.070)
(0.067)
(0.070)
(0.067)
Kendrapara
0.368***
0.466***
0.364***
0.428***
0.367***
0.438***
(0.094)
(0.108)
(0.095)
(0.104)
(0.094)
(0.103)
Balangir
0.688***
0.695***
0.683***
0.673***
0.686***
0.678***
(0.110)
(0.109)
(0.111)
(0.107)
(0.110)
(0.107)
_cons
10.740***
10.640***
10.750***
10.710***
10.740***
10.690***
(0.196)
(0.206)
(0.197)
(0.200)
(0.195)
(0.199)
N
491
491
491
491
491
491
***Significance at the 1% level, **significance at the 5% level, and *significance at the 10% level
Over time, as the adoption of new technologies becomes routine, farmers become less dependent on government assistance. For instance, the government used high-yielding seeds and chemical fertilizers during the Green Revolution. Farmers adopted these practices with government support initially, and now, they independently utilize these inputs without relying on continuous government assistance. Similarly, the concept of CSA must be ingrained in the farming community. Once farmers accept and integrate CSA practices into their routines, they can carry them out independently. However, the state must provide support, especially for marginal farmers, who may need external assistance to adopt CSA practices.
