Additionality in Agricultural Conservation Programs and Extension Services

ABSTRACT

We present in this article estimates of the extent to which conservation cost-share programs and extension services are additional, using farm-level survey data from Louisiana. Farmers’ adoption decisions for 12 soil conservation practices on agricultural land are analyzed, using the propensity score matching approach. We find a varying impact of financial incentives and technical assistance across different soil management practices. Results reveal positive additionality for farm-specific conservation plans, conservation tillage, and zero-grade fields. Payment to avoid the burning of crop residue is nonadditional. These findings guide funding agencies in making conservation investments cost-effective while also attaining environmental goals.

Cost-share payments to farmers for adopting conservation practice(s) on working lands are crucial policy instruments to address soil health, water quality, and other natural resource concerns (Claassen and Ribaudo 2016; Wade et al. 2015). In the United States, the Farm Bill conservation programs are the primary source of federal funding for private land conservation (Burger et al. 2006; McGranahan et al. 2013). The 2018 Farm Bill reauthorized funding for major agricultural conservation programs to provide financial incentives and assist landowners in undertaking conservation practices on private agricultural lands.¹ Farmers receive financial and technical assistance as compensation for the opportunity costs of complying with environmental guidelines (Alix-Garcia et al. 2004; Karsenty et al. 2017). Agricultural conservation practices resemble impure public goods accruing private economic benefits and public environmental benefits (Canales et al. 2020; Horowitz and Just 2013; Woodward et al. 2016). Conservation spending constitutes a substantial portion of the Farm Bill allocation; therefore, cost-effectiveness has been in the spotlight alongside environmental protection (Karsenty et al. 2017). However, assessing the actual benefit of financial incentives and fostering cost effectiveness are challenging due to the inability to observe counterfactuals (Imbens and Woolridge 2009). In addition, self-selection bias arises from voluntary, nonrandom enrollment in cost-share programs, making beneficiary and nonbeneficiary groups nonequivalent, limiting the ability to plan for cost effectiveness by gauging program impact (Fleming 2017). In light of such challenges, a quasi-experimental approach such as matching methods (Rosenbaum and Rubin 1983) facilitates estimation of additionality. Additionality measures the average increase in the proportion of a given conservation practice among enrolled farmers compared with their counterfactual situation without financial incentives or technical assistance (Claassen et al. 2018; Mezzatesta et al. 2013).

Additionality is a useful criterion to promote the cost effectiveness of agricultural programs (Lawley 2019; Ribaudo et al. 2017). Without additionality, financial and technical assistance are both not cost-effective and thus are inefficient. Risks of nonadditionality in federally funded conservation programs abound (Howard 2020; Pannell and Claassen 2020). Therefore, it is essential to identify additional practices to enhance the cost effectiveness of existing conservation programs (Ribaudo et al. 2017). Given the importance of conservation practices for improving soil and water quality, we are interested in evaluating the additionality of cost-share programs and technical assistance on Louisiana's soil conservation practices using the propensity score matching (PSM) approach. This study quantifies the effect of enrollment in two federal cost-share programs – the Conservation Stewardship Program (CSP) and the Environmental Quality Incentives Program (EQIP) – and technical assistance from two extension service providers – the Louisiana State University Agricultural Center (LSU AgCenter) and the Natural Resources Conservation Service (NRCS) – on the outcome of interest, adoption of soil management practices. Results show a varying impact of financial incentives and technical assistance across different soil management practices. Conservation practices, including a farm-specific conservation plan, conservation tillage, and zero-grade fields, show positive additionality, while not burning crop residue is nonadditional. These additionality estimates could guide cost-effective conservation investments by funding agencies while also attaining environmental goals.

This article adds to the existing literature by Mezzatesta et al. (2013), Claassen et al. (2018), and Fleming et al. (2018) on the topic of additionality of agricultural conservation programs. It extends the analysis by estimating additionality from extension service providers such as the NRCS and the LSU AgCenter. The study supports the argument that technical assistance, extension, and outreach, along with financial incentives, strengthen conservation efforts (Burton et al. 2008; Reimer et al. 2012).

The following section is a literature review, followed by the data and the estimation method. The next section presents the empirical results and discussion, and the final section concludes.

Literature Review

CSP and EQIP remain the major working land conservation programs in the United States addressing environmental issues accruing from agricultural production, including non-point source pollution (USDA 2020). These programs are administered by the NRCS and rely on agricultural producers' voluntary participation to promote conservation efforts. The EQIP and CSP budgets amount to $36.6 billion for the 2019-2029 decade (Farm Bureau, 2019). Enhancing program sustainability will require that farmers use these funds efficiently to further practices with the potential for sustained improvement of the environment along with agricultural production.

Horowitz and Just (2013) mention that additionality holds when a particular environmental service is available with financial payments but absent without the payments. Knowledge of additionality is important to ascertain if cost-share programs incentivize farmers to adopt on-farm practices for meeting conservation goals. Measuring additionality can help identify any inefficient allocation of conservation funds by determining which practices farmers adopt with and without conservation incentives (Claassen et al. 2018). Thus, additionality is an essential measure of performance for voluntary conservation programs adopting the cost-share model (Mezzatesta et al. 2013; Pufahl and Weiss 2009). Recent research has measured the additionality of cost-share programs on specific conservation practices (Chabé-Ferret and Subervie 2013; Claassen et al. 2014; Fleming et al. 2018; Mezzatesta et al. 2013; Pufahl and Weiss 2009). These studies have partially addressed the pervading selection-bias problem in observational studies by exploiting econometric techniques such as matching methods. Matching is required because self-selection into treatment groups – that is, participants in conservation programs – makes it difficult to assess the treatment effect through the usual with- and without-program participation approach.

In general, practices with high initial costs or limited on-farm benefits or that provide benefits in the distant future typically have a high level of additionality in voluntary conservation payment programs (Claassen and Ribaudo 2016). Such practices include cover crops, filter strips, and riparian buffers (Claassen et al. 2018; Mezzatesta et al. 2013). On the other hand, management practices such as conservation tillage and nutrient management have shown low levels of additionality due to varying on-farm benefits and costs and the high risk of nonadditionality (Claassen et al. 2014). Interestingly, some practices are less additional but still cost-effective when the cost is low, or the public environmental benefits are high (Pannell and Claassen 2020). For instance, grass buffers used in limiting nutrient runoff possess a high benefit–cost ratio despite low additionality (Chabé-Ferret and Subervie 2013).

Pufahl and Weiss (2009) employed a semiparametric PSM method to estimate agri-environment programs' impact on Germany’s farm input and output. They reported a positive relationship between participation in agri-environment programs and the cultivation area. They found a negative relationship to farm chemicals purchase, leading to reduced nutrient load in surrounding water bodies. Similarly, Chabé-Ferret and Subervie (2013) applied difference-in-difference (DID) matching to estimate the additionality of agro-environmental schemes (AES) payment for cover crops. The authors found that cover crop acreage increased by 10 ha per farm but did not appear cost-effective. Mezzatesta et al. (2013) investigated the additionality of conservation programs for six practices in Ohio using the PSM method. Their results indicated positive additionality but with large variation between practices. Using data from the Agricultural Resource Management Survey, Claassen et al. (2018) estimated additionality for selected practices under conservation payment programs. Their results show greater than 95% additionality for filter strips and riparian buffers but less than 50% for conservation tillage. Fleming (2017) reports a substantial increase in cover crop and conservation tillage acreage because of cost-share programs. Similarly, Fleming et al. (2018) evaluated the effects of cover crop program participation in Maryland. They reported 97% additionality for program participants for cover crop use for nitrogen abatement.

The reviewed studies show that conservation cost-share payments result in additional adoption of specific practices while having no effect on other conservation practices, indicating the idiosyncratic nature of practices, farmers, and production practices. These studies suggest that different econometric strategies using state-level cross-section or panel data may influence additionality results. The reviewed studies also demonstrate the PSM methods' superiority when selection bias and information asymmetry are present in nonexperimental data and highlight the need to account for technical assistance in adoption decisions, which we include in the current analysis.

Conservation practices enhance ecosystem service delivery and generate a positive externality (Howard 2020). Farmers benefit from conservation practices; therefore, knowledge about factors influencing adoption of a conservation practice is essential (Canales et al. 2020; Mace et al. 2007). Both economic (e.g., financial incentives) and noneconomic (e.g., social image, climate, biophysical characteristics, environmental attitudes) factors influence farmers’ decisions to undertake conservation practices (Karali et al. 2014; Liu et al. 2018). Apart from financial incentives, education and outreach are equally crucial for farmers to adopt conservation practices and strengthen conservation stewardship (Prokopy and Genskow 2016; Ribaudo et al. 2017). According to McCann and Núñez (2005), farmers trusting and using public information sources are more likely to be aware of conservation programs and more likely to participate in conservation activities. Outreach offers a wide range of support to farmers, including one-on-one training, workshops, tours, demonstration projects, and peer-to-peer social diffusion (Jackson-Smith and McEvoy 2011; Lemke et al. 2010). The essentiality of outreach lies in streamlining conservation recommendations and receiving farmers' feedback (Jennings et al. 2012).

Furthermore, outreach provides a two-way stimulus to devise programs that meet farmers' needs and then assess the outcomes for further improvements. Lemke et al. (2010) studied the effect of outreach on conservation practices adoption in Illinois and reported a positive association of outreach with adoption. They further assert that outreach is primarily needed to generate awareness and provide impetus to implement best management practices (BMPs). Conservation agencies and land-grant universities have also been crucial to achieving conservation goals, and the results are apparent with the U.S. cooperative extension model (Ford and Babb 1989; Sweikert and Gigliotti 2019; Tucker and Napier 2002). Although studies focus on evaluating conservation practices supported by conservation payments, we can extend the same concept to evaluate conservation behavior for other forms of support such as technical assistance (Greiner et al. 2009; Claassen et al. 2018; Jackson-Smith et al. 2010; Claassen and Ribaudo 2016). Therefore, it is essential to account for technical assistance from agencies and land grant universities to measure the additionality of nonfinancial programs on conservation practices, which has not yet been evaluated in the literature.

Data

We designed a structured survey based on the tailored design method of Dillman et al. (2009) to determine on-farm soil management practices and technical support providers. We surveyed row crop and pasture farmers who actively managed their farms and had more than 250 acres of farmland.² We identified the farmers from the Louisiana Master Farmer Program database for this study.³ Among the 754 shortlisted farmers based on the above criteria, only 502 farmers in the database possessed complete contact information, constituting our study population. Using an online mode, we administered the survey to 502 farmers in 2016 and 2017. The survey questionnaire consisted of 28 closed-ended questions, including farmers’ sociodemographic characteristics, attitudes, information about the adoption of soil BMPs, enrollment in working lands programs, and technical service providers. Of the 502 farmers initially emailed, a total of 159 farmers responded. After removing partial and uncompliant responses, we proceed with 105 observations—a 21% response rate—for further analysis.

Due to the relatively small sample size, we used the variables farm size, age, and farming experience to evaluate respondents' representativeness in the study area. The latest 2017 Census of Agriculture data showed the Louisiana averages for farm size, age of principal farm operator, and years in farming as 292 acres, 58.4 years, and 16 years. In our sample, the above variables averaged 607 acres, 55 years, and 18 years, respectively. Overall, the survey-reported numbers are consistent with state averages except for farm size,⁴ indicating that our survey respondents' characteristics reflect state averages.

Variables

The dependent variables in the study include 12 soil management practices⁵ for assessing additionality and are presented in Table 1. The description of soil management practices under study follows: A farm-specific conservation plan reflects the current agricultural operation and includes BMPs to address erosion and sediment loss and protect the soil resource. Soil test refers to soil sample analysis to determine nutrient content and composition and is vital in the nutrient management planning process for crop fields. Conservation tillage is an approach to conserve soil by reducing erosion and increasing water availability. Grassed turn-rows around the fields are vegetated strips adjacent to the farmed field, which protect the surface water quality, reduce soil erosion, and improve habitat and biodiversity. Crop residue burning is a common, inexpensive, and effective method for providing ash fertilization, but not a recommended approach to remove excess residue, facilitate timely planting, and control pests and weeds (Dhammapala et al. 2006). Disking is a regular soil management practice among farmers to manage residues from the previous crop(s). Although disking crop residue helps promote the rapid decay of the plant materials, it can negatively affect the soil and disturb its structure. No-till in the fall is one of the key soil management practices that improve water quality by preventing erosion. Crop rotation discourages growing the same crop in the same spot for consecutive growing seasons to maintain soil health, combat pests and weeds, and slow erosion. Cover crops add organic matter to the soil, help smother weeds, control pests, and enhance soil moisture retention. Zero grade refers to fields that have been leveled to the point of no slope or grade, a new practice used in the rice production system, requiring significantly less water. The side inlet irrigation practice is accomplished by placing an aboveground irrigation pipe along the field with inlets, which provides greater water control. Biological pest control is an environmentally safe and acceptable farm practice, which relies on predation, parasitism, herbivory, or other natural mechanisms to control pests.

We include two treatment variables for the analysis (Table 1). The first treatment is enrollment in CSP/EQIP. The treatment group comprised of farmers who enrolled in CSP and EQIP for adopting soil management practice(s), while the control group included those farmers who did not enroll.⁶ The second treatment variable is the source of technical assistance, in which the treatment group included farmers obtaining technical assistance and production research support from NRCS and LSU AgCenter.⁷ We choose CSP and EQIP for evaluation in our analysis because these are the two largest working lands programs. Similarly, we chose the LSU AgCenter and NRCS because they remain the two most popular extension service providers in Louisiana. We present the adoption of different soil management practices by farmers under different scenarios in Tables 2 and 3.

For estimation of the likelihood of enrollment in the cost-share program and also the likelihood of obtaining technical assistance from extension service providers, we select independent variables, also known as covariates, based on existing studies on conservation adoption such as Adusumilli and Wang (2018), Baumgart-Getz et al. (2012), Carlisle (2016), Fleming (2017), Ma et al. (2012), Mezzatesta et al. (2013), Prokopy et al. (2008), and Reimer and Prokopy (2014). We describe the covariates in Table 1.

Estimation Method

Caliendo and Kopeinig (2008) pointed out that program applicants and nonapplicants are usually different due to unobserved heterogeneity; therefore, a simple comparison of the mean difference may not exhibit the real treatment effect of participating in a program or receiving technical assistance. There is no theoretical guidance for farmers to select the treatment group in nonexperimental studies. Therefore, selection bias may persist due to nonrandom enrollment and the heterogeneity of observed and unobserved characteristics of farmers, which affects the probability of receiving treatments in addition to outcome indicators.

Developed by Rosenbaum and Rubin (1983), PSM is a useful tool to control the selection bias during program evaluation with observational studies (Chagwiza et al. 2016). This method involves pairing participants and non-participants with similar characteristics (Getnet and Anullo 2012; Heckman et al. 1999). Rejesus et al. (2011) pointed out that the PSM controls for selection bias from observable covariates given the nonrandom nature of enrollment to federal conservation programs. Several researchers applied this technique to evaluate the additionality of conservation programs among farmers (Mezzatesta et al. 2013; Rejesus et al. 2011; Claassen et al. 2018; Adusumilli et al. 2019). The average treatment effect on the treated (ATT) gives the treatment effect in the subpopulation of program-enrolled farmers and provides a quantitative estimate of additionality. Simply, ATT can be expressed (Mezzatesta et al. 2013) as

where D is the treatment denoting enrollment in federal cost-share programs (CSP and/or EQIP) or obtaining technical assistance (LSU AgCenter and/or NRCS). Y₁ and Y₀ denote the outcome variable under the treated and control conditions, respectively. Equation (1) fails to control for different factors affecting both treatment and outcome, including self-selection, thereby leading to upwardly biased ATT estimates. Thus, we measure ATT following Rosenbaum and Rubin (1983) as

where Z includes a set of covariates satisfying the conditional independence assumption, that is, treatment and outcome assignment are independent. Thus, the ATT is defined as the expected effect of enrollment in the treatment group on adopting a particular soil management practice (outcome), conditional on the treatment(s): cost-share or technical assistance.

Participants are paired with nonparticipants with similar propensity scores through the matching process, thereby forming a treatment and control group. If the treatment group and the control group are balanced in terms of explanatory variables (covariates), then the difference between their outcomes is the ATT. The resulting ATT is the estimate of the additionality of soil conservation practices adopted by participants due to financial or technical incentives. According to Claassen et al. (2018), the matching estimator in Equation (2) can be written as

where ω_i = 1/n₁, n₁ represents the farmers that participated in a program (treatment), and i denotes treated observations from 1 to n. ω_ij denotes the matching weight assigned to the outcome from the matched jth control with the ith treated observations.

In this study, we employ the two commonly used techniques – nearest neighbor and kernel matching – for estimation of the additionality from matched samples. Under nearest neighbor matching, a treated observation is matched to the n₀ nearest control observations with weights ω_ij = 1/n₀ based on the closest propensity score value (Claassen et al. 2018; Dehejia and Wahba 2002). The ATT is calculated as the mean difference between the treated observations and the matching weighted counterfactual estimates.

Kernel matching is statistically more efficient as it uses all the control observations to construct the counterfactuals (Heckman et al. 1998) with weights as (Claassen et al. 2018)

where G(.) denotes a kernel density function, k represents the bandwidth parameter, and the denote the estimated propensity scores for the nonparticipants (j) and the participants (i). The bandwidth parameter k in ω_ij is allotted by the researcher and is a tradeoff between potential bias and efficiency for ATT estimates. This technique has a lower variance because more information is used. A disadvantage, however, is that it possibly includes observations that are bad matches (Caliendo and Kopeinig 2008). Using two matching methods can help to check for the disadvantages of each approach.

Results and Discussion

We present the summary statistics of different covariates used for matching in Table 4. Altogether, 45 out of 105 farmers enrolled in the CSP/EQIP. Similarly, 60 out of 105 farmers obtained technical assistance from LSU AgCenter/NRCS. About 43% of the farmers enrolled in the CSP/EQIP also received extension support. Similarly, CSP and EQIP are the primary funding sources for farmers to adopt crop rotation and conservation tillage, followed by soil-test and grassed turn-row around the fields (Table 2). Likewise, the extension services of LSU AgCenter and NRCS are the primary sources of technical assistance and production support for farmers to avoid burning crop residue and to adopt triennial soil testing followed by conservation tillage practices and crop rotation (Table 2).

We apply the logit model to estimate the propensity scores – that is, the probability of treatment assignment – for each farmer, based on the observable covariates presented in Tables 1 and 4. We present the estimated coefficients from the logit model for both program enrollment and technical assistance in Table 5. The goodness-of-fit tests suggest that the selected covariates provide reasonable estimates of the likelihood of program enrollment (χ² value = 31.46, p = 0.00, and pseudo R² = 0.22) and technical support (χ² value = 18.61, p = 0.04, and pseudo R² = 0.13).⁸ Five variables, including age, belief, experience, rented, and crop acres, are significant in explaining the enrollment of farmers in CSP/EQIP. Similarly, three variables, including experience, rented, and north, are significant in explaining the likelihood of obtaining technical assistance from LSU AgCenter/NRCS.

The positive and significant influence of belief indicates that the probability of participating in a federal cost-share program such as CSP/EQIP increases if a farmer perceives that farming practice affects water quality. This result is consistent with previous studies such as Baumgart-Getz et al. (2012), Haghjou et al. (2014), Lubell and Fulton (2008), and Nunez and McCann (2004) that report higher adoption of conservation practices among farmers aware of the association of water quality with different farming practices. Our results further reinforce the need for extension and outreach on BMPs that improve water quality.

The significantly positive influence of land ownership and total crop acres implies that a farmer is more likely to participate in federal cost-share programs if the proportion of rented land is more than 50% of farmed land and the total crop acres are less than 500. The result may be due to conservation practices' contribution to enhancing soil productivity, vital for sustained farm income. Lambert et al. (2006) reported that the greater the proportion of land owned to total acres farmed, the greater is the participation in conservation programs. In Louisiana, the number of acres planted by tenants and part owners is on an upward trend. Therefore, tenant farmers see conservation practices that can reduce input use and improve profits as favorable solutions. In contrast, farmers renting more than 50% of their land are less likely to seek technical assistance from state-level extension providers. This result may be because various conservation practices have positive net returns only in the long run (Soule et al. 2000). Tenant farmers' contract obligations can influence the outlook toward the prospects of such long-term practices, leading to reluctance to seek technical assistance.

The probability of enrollment in CSP/EQIP decreases with farming experience, and the effect is significant. Older farmers tend not to participate due to the complexities of program requirements, paperwork, and funding availability (Claassen and Duquette 2012). In our case, a farmer with more than 20 years of farming experience is less likely to participate in a federal cost-share program because of less time for gaining returns from conservation activities. However, Lambert et al. (2006) report that experienced farmers are more likely to integrate conservation practices into their current production practices. In our case, there is a negative association of experience with obtaining technical assistance and production support.

The variable north, which evaluates regional impact, has a positive influence on the likelihood of obtaining technical assistance and production support. Farmers with farmland located in the northwest and northeast geographic regions of the state are more likely to receive technical assistance and support from LSU AgCenter/NRCS because these two regions represent dominant row crop acres in the Red River and the Mississippi River basins. Farmers in this region rely on irrigation for crop water needs. Profit margins can vary substantially within the area if farmers overapply irrigation or crops become vulnerable to poor management activities. Hence, farmers tend to seek input from agencies for appropriate management actions. However, there is no regional difference in the likelihood of enrollment in CSP/EQIP.

As mentioned earlier, the propensity score denotes the probability that a farmer in the full sample receives the treatment conditional on a set of observable characteristics (Rosenbaum and Rubin 1983) that may affect program enrollment or receiving technical assistance. We used the predicted propensity score interval to check the region of common support, ensuring that any combination of characteristics observed in the treatment group is observed among the control group observations to find adequate matches. In our case, the predicted propensity scores for program enrollment and technical assistance range from 0.06 to 0.93 (mean = 0.44) and 0.33 to 0.93 (mean = 0.63), respectively. We disregarded all other observations outside this interval from the analysis. After assessing the propensity scores for both treatments, we checked for the common support condition. Figure 1 presents the common support region for CSP/EQIP enrollment and assistance from LSU AgCenter/NRCS. There is a good overlap of propensity score distribution between CSP/EQIP enrolled and non-enrolled groups. The common support is also satisfactorily extended for LSU AgCenter/NRCS technical assistance recipients and nonrecipients. Moreover, the balancing property is satisfied in both cases.

We then match each farmer with treatment to farmers in the control group based on propensity score using the nearest 10 neighbors and kernel matching methods. We checked whether two matched groups are the same after matching using a standardized t-test for each outcome group and found that farmers' observed characteristics are well-matched under both methods. Table 6 presents the ATT of CSP/EQIP enrollment and technical assistance using two matching approaches. Both the nearest 10 and kernel matching techniques yield similar results. The additionality of CSP/EQIP enrollment is positive and statistically significant on a farm-specific conservation plan, conservation tillage practices, and using zero-grade fields. However, enrollment in CSP/EQIP has a statistically significant nonadditional effect on not burning crop residues. Similarly, technical assistance from NRCS/LSU AgCenter has significant additionality on the on-farm conservation plan, not disc crop residue, cover crops, and biological pest control.

The farm-specific conservation plan adoption rate is 40–41% higher in a situation where farmers are enrolled in CSP/EQIP (p < 0.01). Similarly, it is about 17–19% higher when farmers obtained technical assistance from NRCS/LSU AgCenter. The farmwide conservation plan is a long-term plan that includes addressing multiple resource concerns on the farm and involves adjusting practices as farm conditions change. The positive and statistically significant ATTs suggest that enrollment in CSP/EQIP and obtaining technical assistance achieve a significantly positive level of additionality for this practice.

In the case of conservation tillage practices, enrolled farmers have a 29–30% higher adoption rate than nonenrolled counterparts (p < 0.05). Changes in tillage practices require specific modifications to tillage equipment, and such changes incur upfront costs to most farmers. Participation in conservation programs provides reimbursement for those costs. NRCS promotes cover crops along with conservation tillage as a suite of soil management practices. The adoption of cover crops is associated with an increase in no-till adoption and crop diversification (Lee and McCann 2019). Results show that farmers obtaining technical assistance from LSU AgCenter/NRCS have 11% higher adoption of cover crops (p < 0.10). The same practice did not have a significant additional effect from participating in EQIP/CSP. This result corroborates farmers' concerns regarding cover crop termination guidelines, challenges in achieving effective control of some cover species, seed unavailability, and others. This further implies that monetary incentives alone may not be sufficient to drive conservation efforts when other technicalities of given practice(s) are a major concern. Mezzatesta et al. (2013), Fleming (2017), and Fleming et al. (2018) reported additionality for cover crops as 0.237, 0.269, and 0.289, respectively. The reported additionality for cover crops from our study is slightly lower than their estimates. However, our results are almost consistent with the ATT for cover crops (0.178) reported by Plastina et al. (2018) and Sawadgo et al. (2019) from a study in Ohio. Similarly, our ATT estimate for conservation tillage (0.29) is slightly higher than that reported by Mezzatesta et al. (2013), 0.149, but similar to that reported by Fleming (2017), 0.258.

For zero-grade fields in rice production, we find a 16% higher adoption rate for CSP/EQIP enrolled farmers than for nonenrolled farmers (p < 0.05). Grading is an expensive practice with costs ranging up to a few hundred dollars per acre, depending on the amount of soil moved. As a result, farmers would consider participation in federal conservation programs to recoup some of the costs.

Furthermore, practicing not burning crop residue is 23–24% lower in a situation where farmers enrolled in CSP/EQIP and is statistically significant at the 1% level. Adequate residues on cropland reduce soil erosion and enhance water conservation (Unger et al. 2001; Wilhelm et al. 1986). Crop residue is burned mainly for pest management purposes. The practice promoted as an alternative to burning crop residue, often done in sugar cane, is sweeping. However, sweeping requires sugar cane farmers to change certain equipment or purchase new equipment capable of sweeping the residue. Some farmers indicated being closer to neighborhoods as their primary reason for the adoption of sweeping. Others pointed to added soil benefits when sweeping is combined with conservation tillage as their reason. However, burning residue has long been a common practice in agriculture. Conservation programs discourage burning of crop residue. Nonadditionality of this practice may be due to the flexibility concerning practice implementation without cost-share program support. For example, with the forecast of a good stretch of wind and temperature during residue cleaning, farmers prefer to burn residue as the quickest option. Farmers mostly agree with the benefits of sweeping the residue; however, the NRCS incentive of $11–13 per acre often does not fully cover the costs of equipment changes. In addition, burning is convenient for farmers to clear the fields for the following crop due to the narrow window of time. Therefore, the economic concern stands out as a significant hurdle to adopting the practice. Not disking crop residue is another residue management practice, and its adoption rate is 19-20% higher among those obtaining technical assistance with NRCS/LSU AgCenter. However, there is no significant additionality from programs paying to discourage the disking of crop residue.

Our results show that additionality does not hold for the soil management practices, including soil test, grassed turn-rows, no-till in the fall, crop rotation, and side inlet. Nonadditionality in several practices may be due to crowding out and the pay-for-nothing effect, as mentioned in Howard (2020). Our results reveal that technical assistance from extension service providers holds an additionality criterion for few conservation practices. Extension services, including technical assistance and production support, would enable farmers with the necessary tools to embrace various BMPs with synergistic benefits effectively (Canales et al. 2020). Therefore, the targeted extension service could aid in meeting conservation goals. Our results may serve as a reference for delineating and prioritizing BMPs. We derive our results with alternative specifications to ensure robustness; however, we acknowledge that these results stem from a relatively small sample and may require additional analysis with a larger sample to extract even more insightful information. Furthermore, we consider cost-share and extension programs as independent, but their actual relationship may be subject to further investigation.

Conclusions

Federal cost-share incentives and technical assistance are key policy tools for improving farmers’ adoption of conservation practices and addressing non-point source pollution concerns. However, evaluating these conservation programs' effects is complicated by selection bias and unobservable heterogeneity. Using data from an online survey of Louisiana farmers, we employ the propensity-score matching approach to assess additionality in two major cost-share conservation programs—CSP and EQIP—and two major extension service providers—LSU AgCenter and NRCS—on the 12 soil management practices.

This study reveals significant additionality of CSP/EQIP programs on a farm-specific conservation plan, conservation tillage practices, and zero-grade rice fields; however, not burning crop residue shows “subtractual” additionality. The results suggest that the cost-share programs and technical assistance encourage farmers to adopt certain soil management practices that many farmers would not undertake in the absence of such programs. We posit that more environmental benefits could be harnessed by promoting soil conservation practices with high additionality and carefully evaluating others that are nonadditional.

Funding allocation for conservation programs at the state level relies on priorities identified by local soil and water conservation districts. However, conservation funding has been on a downward trend over the years. As a result, the allocation of funds to improve adoption needs to measure its impact carefully. The framework and the study results can help evaluate the effect of both funding and technical assistance on conservation practices adoption. Program allocations can consider such frameworks to evaluate program funding and expand additional conservation practices on the ground. Estimating additionality coupled with environmental benefits and costs indicators on a national basis, and investigating the nature of the relationship—complements or substitutes—of cost-share and extension programs could further the understanding of conservation initiatives and is left for future research.

Notes

References

Author notes