Improving nitrogen management in US agriculture: Yield Reserve versus land retirement

Abstract

Pressure to protect and restore water quality continues to drive demand for greater reductions in nitrogen (N) loads from US agriculture. Researchers have been evaluating agricultural production choices along both the extensive and intensive margins to improve ecosystem services. This study uses the US Department of Agriculture's Regional Environment and Agriculture Programming (REAP) model, a partial equilibrium simulation model that integrates agricultural production, land use, and environmental outcomes, to evaluate the cost‐effectiveness of a Yield Reserve Program compared to an expansion of the Conservation Reserve Program (CRP) on revenues, costs, output, and potential reductions in N loads from the production of 10 major crops, both nationally and regionally. The findings indicate that the Yield Reserve Program outperforms the CRP in terms of achieving N reduction under equivalent government budget expenditures. However, the N reduction under the Yield Reserve Program is partially offset by the “rebound effect” on corn (Zea mays L.) acreage whereby corn acreage increases with the subsidized N reduction. The CRP expansion demonstrates a strong “slippage effect,” where the expansion of CRP acreage simply brings marginal land into crop production resulting in a smaller‐than‐expected N reduction. Sensitivity analysis shows that higher percentage of Yield Reserve in terms of the amount of subsidized N reduction tends to be more cost‐effective, and more inelastic land supply tends to reduce the “slippage” of CRP expansion.

Plain Language Summary

Our study evaluates two strategies for reducing nitrogen runoff from US agriculture: the Yield Reserve Program and the Conservation Reserve Program (CRP). Nitrogen runoff from farms harms water quality, so finding effective management practices is essential. Using an agricultural economic simulation model (Regional Environment and Agriculture Programming model), we compare the cost‐effectiveness of both programs in reducing nitrogen pollution from major crops nationwide. We find the Yield Reserve, which compensates farmers for lowering nitrogen use, achieves greater reductions for the same cost compared to the CRP, which pays farmers to retire farmland. However, both programs face challenges. The Yield Reserve can lead to increased crop acreage, partially offsetting nitrogen savings, while the CRP experiences reduced effectiveness due to new land entering production elsewhere. Considering broader environmental benefits in future studies would provide a more complete assessment.

Abbreviations

CRP

Conservation Reserve Program

EPA

Environmental Protection Agency

ERS

Economic Research Service

FPR

farm production region

HEL

highly erodible land

HUC

hydrologic unit code

LRR

land resource region

NIFA

National Institute of Food and Agriculture

REAP

Regional Environment and Agriculture Programming

USDA

United States Department of Agriculture

1. INTRODUCTION

Water pollution caused by anthropogenic nutrient sources, particularly nitrogen (N), remains a critical environmental issue in the United States. Agriculture, the primary contributor to nutrient runoff, is estimated to account for approximately 60% of N loadings into the Gulf of Mexico (Alexander et al., 2008; Marshall et al., 2018). For the Chesapeake Bay watershed, agricultural activities produced 116.37 million pounds of N load, accounting for 46% of total simulated N load to the Bay (Chesapeake Progress, 2024). Row‐crop agriculture, particularly corn, is a major contributor to nutrient pollution (Alexander et al., 2008). Despite agronomic recommendations, farmers often apply more fertilizers than necessary, leading to nutrient runoff (Sheriff, 2005). The US average corn yields rose from 64 bushels per acre in 1964 to over 176 bushels per acre in 2018, while the US average N fertilizer applications to corn increased from 58 pounds per acre in 1964 to 149 pounds per acre in 2018 (United States Department of Agriculture [USDA] Economic Research Service [ERS], 2019). On average, 34 pounds of N per acre per year are lost from crop fields to waterways (USDA Natural Resources Conservation Service [NRCS], 2017). Long‐term fertilization practices have resulted in a significant accumulation of nutrients in soils, lakes, riverbeds, and groundwater (Van Meter et al., 2017).

Federal and state governments have taken actions to protect and restore water quality, but progress is not satisfactory. The United States Geological Survey calculated that the total N load from the Mississippi/Atchafalaya River Basin to the Gulf in Water Year 2017 was approximately 3320 million pounds and that the action plan requires a cut in N load to the Gulf by 45% or more (US Environmental Protection Agency [EPA], 2022). Recognizing the enormity of the task of reducing N loads on a subcontinental scale, the US EPA extended the time for reaching its original goal of reducing the areal extent of the hypoxic zone in the Gulf of Mexico from 2015 to 2035 (US EPA, 2022). The total simulated N load to Chesapeake Bay in 2024 was 252 million pounds, while the interim target with Changing Environmental Conditions is 220 million pounds, which means 32 million pounds (12.6%) of “excess N” needs to be removed annually (Chesapeake Progress, 2024). Achieving this target requires a huge reduction in N loads from agriculture, which could have significant economic impacts on producers and consumers.

Researchers have been evaluating agricultural production choices along both the extensive and intensive margins to improve ecosystem services while considering the cost‐effectiveness of these choices (Bravard et al., 2022; Hellerstein et al., 2015; Hodge et al., 2024; Kirwan et al., 2005; Marshall et al., 2018; Metcalfe et al., 2007; Miao et al., 2016; Rabotyagov et al., 2010). The extensive margin refers to changes in crop production and associated environmental impacts resulting from adjustments in the total area of land under cultivation, such as expanding or retiring cropland. In contrast, the intensive margin captures how production levels and environmental outcomes respond to changes in input use (e.g., fertilizer, labor, or irrigation) on existing cropland (Babcock, 2015; Earnhart & Hendricks, 2023; Hodge et al., 2024; Sela et al., 2017). Understanding these two margins is critical for evaluating how farmers respond to policy incentives and for identifying which interventions, whether targeting land use or input intensity, are more cost‐effective in achieving nitrogen management improvements. One widely discussed intensive margin approach is the Yield Reserve Program (Chesapeake Bay Commission, 2004; Henry & Wallace Center for Agricultural and Environmental Policy at Winrock International, 2001; Metcalfe et al., 2007; US Congress, Senate, 2002), a nitrogen management policy that compensates producers for reducing fertilizer application rates while continuing crop production. Rather than removing land from cultivation, Yield Reserve Program aims to reduce excess N loading by incentivizing suboptimal N application rates and allowing farmers to partially offset yield losses through adjustments in other inputs or practices. One widely discussed and implemented extensive margin approach is land retirement, such as the Conservation Reserve Program (CRP) (Fleming, 2014; Hansen, 2007; Morefield et al., 2016; Wu, 2000, 2005), established in 1985 and administered by the USDA. The CRP incentivizes landowners to voluntarily retire environmentally sensitive cropland from production for a period of 10–15 years in exchange for annual rental payments and cost‐share support for conservation practices. The CRP primarily reduces nutrient loads through land retirement, targeting reductions in erosion, N runoff, and other environmental harms by taking high‐risk acres out of agricultural use entirely. To prioritize enrollments, the CRP uses the Environmental Benefits Index, which ranks offers based on multiple criteria, including expected improvements in water quality, erosion control, and wildlife habitat, ensuring that co‐benefits such as biodiversity and ecosystem services are incorporated into program selection. Wu (2000) discussed the slippage effect of the CRP, whereby retiring cropland may bring non‐cropland into crop production. Wu found that for each 100 acres of cropland retired under the CRP in the central United States, 20 acres of non‐cropland were converted to cropland, offsetting the CRP water and wind erosion reduction benefits, respectively. Fleming (2014) modeled county‐level slippage empirically using satellite imagery in the Midwest states, and the results also suggested the existence of CRP slippage, although at a rate lower than the 20%reported by Wu (2000). However, most analyses of these approaches are conducted at the regional level, few if any studies have compared these two approaches in terms of their impacts on the whole US agriculture sector as well as their potential ecosystem benefits.

We examine the national and regional effects of a Yield Reserve Program and an expansion of a land retirement program (CRP), assuming equal government spending, on revenues, costs, output, and potential reductions in N loads from the production of 10 major crops in the United States. We also test the sensitivity of the Yield Reserve Program through different levels of Yield Reserve, extra subsidies, and enrollment rates, and the sensitivity of CRP expansion through different levels of marginal land supply elasticities, which measure how responsive the amount of land allocated to a particular use is to changes in its price. We find that the Yield Reserve Program outperforms the CRP in terms of achieving N reduction under equivalent government budget expenditures. Sensitivity analysis indicates that implementing a higher percentage of Yield Reserve, such as 2% instead of 1%, tends to be more cost‐effective in N reduction and that a more inelastic land supply tends to reduce the “slippage effect” of CRP expansion.

Core ideas

• Yield Reserve reduces nitrogen loads while increasing overall crop acreage.

• Yield Reserve optimizes nitrogen use, but expanded corn acreage partly offsets lower per‐acre N applications.

• Conservation Reserve Program (CRP) expansion shows limited nitrogen reduction due to slippage, undermining environmental benefits.

• Yield Reserve produces greater environmental benefits than the CRP in terms of nitrogen load reduction.

2. MATERIALS AND METHODS

2.1. REAP model

This study uses the Regional Environment and Agriculture Programming (REAP) model, a partial equilibrium model implemented using nonlinear mathematical programming. The REAP model was developed by USDA's Economic Research Service to analyze the intersection of agriculture and the environment for policy applications (Johansson et al., 2007). Recent issues analyzed using the REAP model include policies to reduce hypoxia in the Gulf of Mexico (Marshall et al., 2018), climate change adaptation (Marshall et al., 2015), and environmental implications of biofuels (Marshall, 2011; Sands et al., 2017).

The REAP model maximizes total net welfare (producer plus consumer surplus) from production of crops, livestock, and processed products subject to land supply constraints and production balance requirements (Sands et al., 2017). The REAP is a price‐endogenous model with prices determined by the intersection of demand and supply curves. As a partial equilibrium model, the REAP does not have endogenous markets for factor inputs. The REAP simulates a baseline equilibrium as well as alternative scenarios in which prices, policies, or other parameters to the model are changed. Effects of the scenarios can be evaluated for model outcomes including US and regional values for land use, crop and livestock production, prices, farm income, and other indicators including environmental indicators such as erosion, wildlife habitat, and nutrient and pesticide loadings (Sands et al., 2017).

The REAP simulates the regional allocation of land and other inputs by production agriculture in the United States and how those allocations change in response to changes in policy, market, or bio‐physical conditions. The model divides the United States into 10 farm production regions (FPRs). Livestock production (dairy, layers, broilers, turkey, hogs, grazed and feedlot cattle, stocker calves) is assumed to be homogeneous within each FPR. The United States is further subdivided into 48 regions that are formed by the intersection of FPRs with land resource regions (LRRs) defined by the Natural Resources Conservation Service of the USDA (Sands et al., 2017). These regions are further divided into 273 LRR‐HUC (hydrologic unit codes) units, which are intersections with watershed HUC boundaries. Many but not all LRR‐HUC units are further subdivided into highly erodible land (HEL) and non‐highly erodible land. In total, there are 456 combinations of LRR‐HUCs with HEL or non‐HEL land. Crop production is assumed to be homogeneous within each LRR‐HUC‐HEL or non‐HEL unit.

The basic decision unit in the REAP is the crop rotation made up of some combination of 10 field crops (barley [Hordeum vulgare L.], corn [Zea mays L.], cotton [Gossypium hirsutum L.], hay [commonly Medicago sativa L. or Festuca arundinacea Schreb.], oats [Avena sativa L.], rice [Oryza sativa L.], silage [typically corn silage, Zea mays L.], sorghum [Sorghum bicolor (L.) Moench], soybeans [Glycine max (L.) Merr.], and wheat [Triticum aestivum L.]). The model is initialized with the observed rotations from the National Resources Inventory. The initialized point is adapted to be consistent with other REAP input data such as observed crop acreage by FPR as reported by the National Agricultural Statistics Service. The model is calibrated to a baseline of year 2030 mainly based on USDA Agricultural Projections to 2030 (USDA, 2021).

For this application, we simplify the REAP model to a crop‐only model. Appendices provide further information on objective function (Supporting Information Appendix) and data input to the model including commodity demand elasticities (Table S1), crop acreage and production (Table S2), cropland supply (Table S3), and major crop yields and production costs (Tables S4 and S5).

The 2022 USDA Budget includes $4.6 billion for the Farm Bill Conservation Programs including $2.3 billion for the CRP. In this study, we simulate three scenarios of increasing subsidy payments of approximately $500 million, $750 million, and $1 billion to farmers to achieve N load reductions from all US croplands. We model the three scenarios under two program implementations, the CRP or Yield Reserve, as described in Section 2.2 and 2.3 that would reduce the N load from cropland while maximizing net welfare (producer plus consumer surplus) of 10 major crops.

2.1.1. Land supply function

The cropland supply function in the REAP model was originally formulated as a kinked linear function. This specification allocates the initial X% of cropland (fixed land) at a constant price, with any additional cropland (variable land) supplied at a higher price along an upward‐sloping linear curve. While this approach facilitates model convergence during the calibration of a baseline scenario, it fails to adequately capture realistic land market dynamics under external shocks. To address this limitation, we revised the specification of the variable land supply function to an exponential form (Figure S1). This modification better reflects the increasing marginal cost of land as cropland expands, thereby improving the model's ability to simulate realistic market responses to exogenous changes.

2.2. Yield Reserve Program in REAP

A Yield Reserve subsidy that represents an income transfer from taxpayers to agricultural producers results in a very small increase in net welfare from the agricultural sector but also a substantial reduction in “excess N,” which represents an environmental benefit. The basic decision unit of crop production in the REAP is the crop rotation made up of some combination of 10 field crops (barley, corn, cotton, hay, oats, rice, silage, sorghum, soybeans, and wheat). We model a Yield Reserve Program that would compensate farmers for reducing N applications in corn production given that corn typically has the highest N fertilizer applications among the 10 field crops. Specifically, the REAP simulates farmers’ profit‐maximizing decisions by reallocating land and adjusting input use in response to changes in net returns across all major crops. When payments are offered to reduce N applications, the model endogenously adjusts crop rotation, fertilizer use, and land allocation to reflect how farmers would likely respond to altered economic conditions. This allows us to account for both direct and indirect effects on planting decisions and nitrogen use. Assuming that farmers currently follow Extension recommendations in applying N, subsidies are designed to compensate farmers for the opportunity cost of foregone net revenues from the N that is not applied. Foregone net revenues, $\mathrm{\Delta }NR,$ are calculated from the loss of yield net of the savings in N‐ and yield‐related costs relative to those that would have been obtained prior to the Yield Reserve Program:

$$\mathrm{\Delta }NR=P_c\left(Y_{\mathrm{base}}-Y_{\mathrm{year}}\right)-P_{\mathrm{N}}\left({\mathrm{N}}_{\mathrm{base}}-{\mathrm{N}}_{\mathrm{year}}\right),$$ (1)

where $P_c$ is the price of crop, $P_{\mathrm{N}}$ is the price of N fertilizer, $Y_{\mathrm{base}}$ is the crop yield under the baseline condition, $Y_{\mathrm{year}}$ is the yield obtained with the N application mandated by the Yield Reserve Program, ${\mathrm{N}}_{\mathrm{base}}$ is the amount of N applied under the baseline condition, and ${\mathrm{N}}_{\mathrm{year}}$ is the amount of N applied under the Yield Reserve Program.

Yield loss in response to reduced N applications is estimated using empirical studies of the response of crop yield to N applications. Studies for the Corn Belt (Huang & LeBlanc, 1994; Vanotti & Bundy, 1994) show a consistent pattern in which yields rise steeply at low N rates and then plateau as applications approach economically optimal levels. We adjusted the general corn yield function estimated for the Corn Belt (Huang & LeBlanc, 1994; Vanotti & Bundy, 1994) to serve as a simplified illustration of the general functional form used in the REAP:

$$\mathrm{Yield}\left(\mathrm{corn}\right)={\mathrm{\alpha }}_{\mathrm{reap}}+0.73N-0.001127N^2$$ (2)

where ${\mathrm{\alpha }}_{\mathrm{reap}}$ is the intercept of yield response function with y‐axis in each REAP region. Regions with lower baseline productivity will have correspondingly lower ${\mathrm{\alpha }}_{\mathrm{reap}}$ values, resulting in lower predicted yields for the same nitrogen application rate. The application of 196.5 pounds of N fertilizer per acre would yield 196.5 bushels of corn grain per acre in this corn yield response function when the ${\mathrm{\alpha }}_{\mathrm{reap}}$ is equal to the national average of 96.57.

The reduction in N load is estimated as the reduction in excess N from corn production. Excess N is defined as the amount of N available to runoff to watersheds, calculated as the total amount of N fertilizer applied to corn minus the amount of N removed by corn (Ribaudo et al., 2017). The estimated N removal by corn varies by state. In Virginia and Michigan, the estimated N removed by corn is approximately 0.90 pound per bushel (Michigan State University Extension, 2017; Virginia Cooperative Extension Service, 2000). In California, the estimated N removed by corn is approximately 0.81 pound per bushel (University of California, Davis, 2009). We take the average of the three states and use a rate of 0.87 pound per bushel to count the N removed by corn for all regions, acknowledging that localized variation may exist and that these states are not among the largest corn‐producing regions. This constant removal coefficient is a simplifying assumption that may be refined as more region‐specific data become available. The total reduction in excess N from corn production under Yield Reserve is estimated as follows:

$$\begin{array}{ccc}\hfill \mathrm{\Delta }\mathrm{Excess}\mathrm{N}& =& ({\mathrm{N}}_{\mathrm{base}}-0.87\times {\mathrm{Y}}_{\mathrm{base}})\times {\mathrm{A}}_{\mathrm{base}}\hfill \\ & & -({\mathrm{N}}_{\mathrm{year}}-0.87\times {\mathrm{Y}}_{\mathrm{year}})\times A_{\mathrm{year}},\hfill \end{array}$$ (3)

where $A_{\mathrm{base}}$ is the corn acreage under the baseline condition and $A_{\mathrm{year}}$ is the corn acreage obtained under the Yield Reserve Program. The excess N per acre from corn production at baseline is estimated to be approximately 25.55 pounds according to the baseline corn yield of 196.5 bushels per acre and N application of 196.5 pounds per acre.

Scenarios of the Yield Reserve Program were introduced in the REAP as follows: First, we estimate foregone net revenues as represented in Equation (1) equal to the three program funding scenarios of approximately $500 million, $750 million, and $1 billion. Second, the subsidies of estimated foregone net revenues are added back to the corresponding rotations containing corn production. The estimated yield losses and foregone net revenues depend on the total subsidy amount available for the program as described below.

To simulate the scale of N reduction needed for the Yield Reserve Program to result in estimated subsidy payments of approximately $500 million, $750 million, and $1 billion, we begin by estimating foregone net revenue at the national level. We assume that the national average price of corn remains unchanged despite the production change induced by the Yield Reserve Program. This simplifying assumption is consistent with how REAP is typically employed in policy evaluation scenarios and allows us to focus on the physical and economic trade‐offs in nitrogen management. However, it also limits our ability to capture feedback effects between yield reductions and market prices. In reality, a reduction in corn yields from lower nitrogen application could exert upward pressure on market prices, partially offsetting revenue losses for producers and influencing subsequent production decisions. Consequently, our fixed‐price assumption may either overstate or understate the true economic cost of nitrogen reductions, depending on how market prices would respond in a fully dynamic setting. Based on 2030 baseline data, we use $3.55 per bushel as the average corn price in the United States, which, when multiplied by predicted yields from the response function (Equation 2), gives the estimated gross revenue per acre (see Figure 1).

FIGURE 1.

Average total revenue and cost of corn production per acre at baseline and first Yield Reserve scenario of $501 million.

Total production costs per acre are estimated by summing three components: (1) the average land cost ($163.41per acre), (2) the average variable production cost excluding N fertilizer ($276.56 per acre), and (3) the cost of N fertilizer at $0.39 per pound (Schnitkey et al., 2022; USDA ERS, 2023). We estimate that nitrogen fertilizer accounts for approximately 22% of the total variable cost based on USDA estimates.

For example, under the first scenario targeting a total subsidy of $500 million, a 3.41% reduction in N application leads to a 1% yield loss relative to the baseline scenario. This N reduction also results in a 0.75% decrease in variable costs due to lower fertilizer expenditures. Using the net revenue loss function (Equation 1), the net loss per acre, after accounting for yield reduction and variable cost savings, is approximately $4.37. With 99 million acres of corn planted nationwide, this corresponds to a total estimated foregone revenue of $4.37 × 99 million = $433 million.

Additionally, the reduction in N application lowers production costs and slightly increases net returns per acre, creating an incentive for farmers to expand corn acreage. In our simulation, this “rebound effect” adds 4.35 million corn acres. Because the Yield Reserve subsidy is paid on a per‐acre basis, these additional acres also receive payments, increasing the total program cost to approximately $501 million, even though the per‐acre subsidy rate is unchanged. This cost increase occurs as a result of behavioral responses to the subsidy, not because the program is designed to achieve a fixed N‐reduction target.

For the second scenario ($750 million subsidy), a 5.04% reduction in N application leads to a 1.5% yield loss and a 1.11% reduction in variable costs of corn production. This results in a foregone net revenue of approximately $6.58 per acre, and with a rebound effect expanding corn acreage by 6.72 million acres, the total subsidy is approximately $772 million.

For the third scenario ($1 billion subsidy), a 6.62% reduction in N application results in a 2% yield loss and a 1.46% reduction in variable costs, with a foregone net revenue of about $8.80 per acre. Factoring in the rebound effect and acreage expansion of 9.19 million acres, the total subsidy is approximately $1058 million.

2.3. CRP in REAP

Changes of CRP acreage enrolled act as the land retirement shocks. We evaluated three scenarios of uniform percentage expansion of CRP acreage across all REAP regions that would cost additional payments of approximately $500 million, $750 million, and $1 billion to farmers. The total acreage enrolled in the CRP is projected to be 26.9 million acres in the baseline year 2030 (USDA, 2021), which is almost 10 million acres less than the modern peak of 36.8 million acres in 2007. The total payment to the CRP land is estimated to be $2057 million in the baseline year; therefore, a 37% uniform expansion of CRP land in the baseline year would cost approximately an additional $750 million and restore the acreage of CRP land to its peak. We found that 25% and 50% uniform expansions of CRP land in the baseline year would cost approximately an additional $500 million and $1 billion, respectively. Therefore, to be budget equivalent to the three scenarios in the Yield Reserve Program, three scenarios of uniform expansion of the CRP land by 25%, 37%, and 50% across all REAP regions were introduced in REAP with corresponding additional CRP payments of approximately $515 million, $762 million, and $1029 million added into the total welfare, respectively.

3. RESULTS AND DISCUSSION

We organize the results in two main parts. Section 3.1 reports the outcomes of the Yield Reserve Program, focusing first on the national‐level changes in N application, corn yields, and total welfare, followed by the effects on crop output, prices, and regional distribution of acreage and N reductions. Section 3.2 presents the corresponding results for the CRP, using the same structure to allow direct comparison. For each program, we also provide sensitivity analyses to assess how results vary under alternative parameter assumptions. Tables 1 and 2 summarize the main quantitative findings, while figures illustrate key patterns and spatial variations.

TABLE 1.

Simulated results under Yield Reserve scenarios.

	Corn Yield Reserve scenario
	Baseline	1%	1.5%	2%
N fertilizer application on corn (pound per acre)	196.5	189.8	186.6	183.5
% change		−3.41	−5.04	−6.62
% change of corn variable cost		−0.75	−1.11	−1.46
Corn yield (bushel per acre)	196.5	194.5	193.5	192.6
Forgone net revenue (dollar per acre)		4.37	6.58	8.80
Simulated Yield Reserve subsidy (million dollars)	0	501	772	1058
Simulated excess N (million pounds)	2388	2019	1823	1618
Simulated excess N reduction (million pounds)		369	565	771
Total welfare of 10 crops (billion dollars)	414.6	419.5	422.0	424.6
% change		1.18	1.78	2.41

TABLE 2.

Simulated results of the Conservation Reserve Program (CRP) and total crop acreage under CRP expansion scenarios.

	CRP expansion
	Baseline	25%	37%	50%
CRP land (million acres)	25.92	32.40	35.51	38.88
CRP payment (million dollars)	2057	2572	2819	3086
Total crop acreage (million acres)	325.0	323.6	322.8	321.8
Slippage a		78.5%	77.2%	75.7%
Total corn acreage (million acres)	99.2	98.5	98.2	97.8
Simulated N reduction from corn production (million pounds)		7.6	11.6	16.1
Total welfare (billion dollars)	414.6	413.9	413.5	413.9
% change		−0.17	−0.26	−0.36

^a Slippage = 100% − (ΔCrop acreage/ΔCRP acreage). For example, the slippage of 25% CRP expansion scenario is 100% − (325.0 − 323.6)/(32.40 − 25.92) = 78.5%.

3.1. Yield Reserve Program

The simulated results in Table 1 present two performance measures for the Yield Reserve Program: (1) the reduction in excess N from corn production and (2) the total welfare from the 10 crops. These measures are shown for corn Yield Reserves of 1%, 1.5%, and 2%. Compared with the baseline, N fertilizer applications on corn fall by 3.41%, 5.04%, and 6.62%, translating into excess N reductions of 369, 565, and 771 million pounds, respectively. Over the same range, total welfare rises by 1.18% to 2.41%, showing that larger Yield Reserves deliver both greater environmental gains and higher net benefits.

3.1.1. Crop output and prices under Yield Reserve Program

With N fertilizer reductions of 3.41%, 5.04%, and 6.62% under Yield Reserve of 1%, 1.5%, and 2%, correspondingly, the production of seven major crops in the United States declines, while that of corn, soybeans, and sorghum increases (Figure 2, Table S6). The increase in corn production reflects the rebound effect of Yield Reserve subsidies, in which the savings from lower N application per acre alongside the incentive of receiving subsidies bring more cropland into corn production. Initially, farmers are overapplying N fertilizer and losing profits, and a reduction in N application moves the production of corn closer to the point of maximum net revenue. A subsidy to compensate for foregone net revenue, which “overshoots” the estimated losses from reduced yield and results in a net profit, provides additional incentives for farmers to expand corn production, particularly in regions with relatively flat marginal cost curves or low opportunity costs of yield reduction. This overshooting phenomenon arises when incentives are not finely tuned to the heterogeneous economic and biophysical conditions faced by producers, causing some to over‐respond relative to what is socially optimal. This aligns with the literature on nonlinear policy response, which finds that the behavioral elasticity of input use tends to vary significantly across farm types, soil quality, and production systems (Kopper et al., 2020; Tittonell et al., 2008). When subsidies are not differentiated by these characteristics, the result can be inefficient over‐adjustment by some producers and under‐adjustment by others. The increase in production of soybeans and sorghum along with corn occurs mainly because the expansion of corn production happens in rotations of corn and soybeans and rotations of corn and sorghum, which have higher profitability than other rotations containing corn. The decline in production of the other seven crops reflects the effects of leftward supply shifts as a result of increased production of corn, soybeans, and sorghum combined with downward sloping demand. Reductions vary among crops with a larger percentage reduction for barley and oats, which are less profitable or are mainly used as feed grains.

FIGURE 2.

Crop commodity output change by Yield Reserve scenarios.

The change in crop prices is opposite to the change of crop production reflecting the downward sloping demand for crops (Figure 3, Table S7). Corn shows the largest percentage price decrease among the crops that has increased production. However, percentage price decreases are more than the percentage increase in the production of corn reflecting the downward sloping demand for corn with its low domestic demand elasticity of −0.23 (meaning that for every 1% increase in the price of corn, the quantity demanded decreases by 0.23%) For example, while corn production increases by 2.96, 4.47, and 6.01 percent across the three Yield RYeserve scenarios, corresponding corn price decreases are 6.15%, 9.31%, and 12.50%.

FIGURE 3.

Crop commodity price change by Yield Reserve scenarios.

3.1.2. Regional crop distribution under Yield Reserve Program

With corn Yield Reserves of 2%, total acres of 10 major crops increase by 9.02 million acres (Table S8), which indicates that the total expansion of corn, soybeans, and sorghum production offsets the decline in the other seven major crops (Figure S4). With corn Yield Reserves of 2%, total acres of corn increase by 9.19 million acres in the United States (Table S9) with the biggest expansion of 7.42 million acres in Northern Plains followed by Corn Belt and Lake States. The change of total acres of 10 major crops shows a similar pattern from the corn acreage change at farm production region level (Tables S8 and S9, Figures S5 and S6), which indicates the changes of total acres of 10 major crops are dominated by corn acreage changes.

3.1.3. Regional N reduction under Yield Reserve Program

With a corn Yield Reserve of 2%, the simulated N reduction across the United States totals approximately 771 million pounds. As illustrated in Figure 4, the most significant N reductions are concentrated in the Corn Belt, Lake States, and the southern portion of the Northern Plains, where the implementation of Yield Reserves has a substantial impact. In these regions, there is a notable decrease in excess N even with a slight rebound in corn acreage, demonstrating that the Yield Reserve approach effectively reduces N loads despite potential increases in corn production. However, the map also reveals areas in the northern portion of the Northern Plains where N levels have increased, as indicated by the dark shading. This suggests that in these regions, the expansion of corn acreage has partially or completely offset the N reduction benefits associated with the Yield Reserve. This pattern highlights the spatial variability in the effectiveness of Yield Reserves, emphasizing that while Yield Reserves can lead to significant reductions in certain areas, the outcome may differ where corn expansion is prevalent. This nuanced spatial response points to the need for region‐specific considerations when implementing N reduction strategies tied to Yield Reserve.

FIGURE 4.

N reduction by Regional Environment and Agriculture Programming (REAP) region under 2% Yield Reserve scenario.

3.1.4. Sensitivity of Yield Reserve Program

Evidence from the existing conservation programs suggests that the overall enrollment rate of Yield Reserve program is unlikely to be 100% due to incentives, qualifications, and restrictions (Barnes et al., 2020; Institute for Agriculture & Trade Policy, 2024). A past study estimated that Yield Reserve implementation would cost $2–$3 per acre to make sure that the program is appropriately placed and any losses are documented (Metcalfe, 2006).

We tested the sensitivity of the Yield Reserve program through different levels of Yield Reserve, extra subsidies, and enrollment rates. The heatmaps in Figure 5 show the results of varying extra subsidies (dollars per acre) and percent enrollment rates on program cost, N reduction, cost‐effectiveness (the cost rebound of nitrogen reduction achieved under each program), and corn acreage expansion for Yield Reserve 1% and 2% scenarios. Each heatmap represents different scenarios by adding extra subsidies and varying enrollment levels, where the highlighted black or red boxes represent the realistic scenarios within the budget constraint of $500 million to $1 billion. Assuming higher enrollment rates are associated with increased extra subsidies, the most realistic scenarios are most likely to be around the 45° line across each heatmap. In the N reduction heatmap, the black‐boxed scenarios show that the achievable N reduction levels within the given budget under 2% Yield Reserve are higher than 1% Yield Reserve.

FIGURE 5.

Sensitivity of Yield Reserve with different enrollment rates and extra subsidies.

Correspondingly, in the cost‐effectiveness heatmap, the red‐boxed scenarios show that the cost‐effectiveness within the given budget under 2% Yield Reserve is better than under 1% Yield Reserve. This is also reflected by the heatmap of corn acreage expansion, where the black‐boxed scenarios show that there is less corn acreage expansion under 2% Yield Reserve. Evidence suggests that farmers frequently overapply N fertilizers based on profit maximization (Del Rossi et al., 2023; Ribaudo et al., 2011). Implementing a 1% Yield Reserve could potentially align corn production more closely with its economic optimum compared to a 2% Yield Reserve, as reflected by the scenarios around the 45° line across the heatmaps of cost‐effectiveness and corn acreage expansion.

In summary, the highlighted black or red boxes provide different possible scenarios of farmers’ participation in the Yield Reserve program within a budget of $500 million to $1 billion. They point to combinations of Yield Reserve levels, subsidy levels, and enrollment rates that balance cost‐effectiveness with significant N reduction and minimal corn acreage expansion. The higher percentage of Yield Reserve (2% vs. 1%) tends to be more cost‐effective. These scenarios emphasize that strategic funding allocations can achieve meaningful environmental outcomes without compromising budgetary constraints, offering a pragmatic approach to N reduction through the Yield Reserve program.

3.2. Conservation Reserve Program

The results from the three scenarios of CRP expansion show how the 10 major crops in the United States react to a uniform expansion of CRP land by 25%, 37%, and 50% across all REAP regions (Table 2). The estimated slippage effect of the CRP is approximately 78.5%, 77.2%, and 75.7% across the three scenarios of CRP expansion, which means for each 100 acres of cropland retired under the CRP in the United States, approximately 76–79 acres of non‐cropland would be converted to cropland, partially offsetting the N reduction benefits of the CRP. For example, the total corn acreage decreases by 0.70%, 1.03%, and 1.41% across the three scenarios of CRP expansion (Table S13); however, the simulated excess N decreases only 0.32%, 0.48%, and 0.68%, respectively. The CRP expansion retires relatively low yield corn land and pushes the corn production to marginal cropland, which requires more N application. As a result, the N reduction from corn production under CRP expansion scenarios is approximately 2% of the simulated N reduction from corn production under Yield Reserve scenarios (Tables 1 and 2). Our evaluation is restricted to assessing N reductions and does not account for other ecosystem co‐benefits, such as wildlife habitat improvements, phosphorus sequestration, or additional nitrogen cycle impacts. Consequently, the reported results represent a conservative estimate of CRP's overall environmental performance. Incorporating these broader ecosystem services into future analyses would provide a more comprehensive assessment of the environmental benefits offered by the CRP.

3.2.1. Crop output and prices under the CRP

With the uniform expansion of CRP land by 25%, 37%, and 50% across all REAP regions, the production of crops declines except oats (Figure 6, Table S10). The decline in crop production reflects the effects of leftward supply shifts as a result of increased acreage of CRP land combined with downward sloping demand. Output reductions vary among crops with the largest in hay, followed by sorghum, wheat, and barley.

FIGURE 6.

Crop commodity output change by Conservation Reserve Program (CRP) scenarios.

The change in crop prices is opposite to the change in crop production reflecting the downward sloping demand for crops (Figure 7, Table S11). Barley shows the largest percentage increase in price, followed by hay, wheat, and sorghum.

FIGURE 7.

Crop commodity price change by Conservation Reserve Program (CRP) scenarios.

With the exception of hay and silage, the change in crop acreage shows a similar pattern to the change of crop production (Figure 6, Figure S8, Table S12). The acreage of hay and silage increases, but the output declines, which means the expansion of CRP land takes over the land originally planted to high yield hay and silage and pushes the production of hay and silage to lower yield land.

3.2.2. Regional crop distribution under the CRP

With the uniform expansion of CRP land by 50% across all REAP regions, total acres of 10 major crops decrease by 3.14 million acres (Table S12), which indicates the total decline in other seven major crops offsets the total expansion of oats, hay, and silage (Figure S6). With the uniform expansion of CRP land by 50% across all REAP regions, the biggest decline in crop acreage is in Pacific States, followed by Appalachia, Delta States, and Southeast (Table S12, Figure S7).

3.2.3. Sensitivity of the CRP

We tested the sensitivity of the CRP through different levels of land supply elasticities to understand how land supply elasticity influences slippage rates and the overall effectiveness of CRP in reducing N loads. Figure 8 illustrates an inverse relationship between slippage and elasticity, where the “beta land” is a measure of land supply elasticity derived from the ratio of baseline price, quantity, and elasticity parameters, confirming that as land supply becomes more inelastic, the slippage effect decreases and leads to more N reduction. The CRP's effectiveness in reducing N runoff is highly dependent on land supply elasticity.

FIGURE 8.

Slippage rate and N reduction under decreased land supply elasticity scenarios.

4. SUMMARY

This study finds that the Yield Reserve Program outperforms the CRP in terms of achieving N reduction under equivalent government budget expenditures. The CRP shows a strong “slippage effect,” where the expansion of CRP acreage simply brings marginal land into crop production resulting in small excess N reductions. These findings suggest that subsidizing adoption of conservation practices such as Yield Reserve on cropland may be more cost‐effective than land retirement in reducing N loadings. Further research is needed to compare the cost‐effectiveness of Yield Reserve with other working land conservation practices such as nutrient management, reduced tillage, and cropland buffers.

The N reduction under the Yield Reserve Program is partially offset by the expansion of corn acreage. The increase in corn acreage under Yield Reserve Program reflects the rebound effect where the Yield Reserve subsidies plus savings in N application costs outweigh the losses in revenue causing corn acreage to increase. The expansion of corn acreage under the Yield Reserve Program is mainly in the Northern Plains, followed by Corn Belt and Lake States, which are all regions with high corn acreage.

This study finds that the slippage effect of the CRP is approximately 77% across the three scenarios of CRP expansion, which offsets the N reduction benefits of the CRP. In corn production, we find that the CRP expansion retires relatively low yield corn land and pushes the corn production to marginal cropland which requires more N application. As a result, the N reduction from corn production under CRP expansion scenarios is almost negligible. Sensitivity analysis of the CRP through different marginal land supply elasticities shows that more inelastic land supply tends to reduce the “slippage” of CRP expansion. Further research on land supply elasticities is needed to provide more accurate estimates of the effects of land retirement programs on slippage and achievement of environmental goals.

Our evaluation focuses solely on N reductions, omitting other ecosystem benefits such as wildlife habitat and phosphorus sequestration. Thus, our results conservatively estimate the overall environmental effectiveness of the CRP. Incorporating these broader ecosystem services into future analyses would provide a more comprehensive assessment of the environmental benefits offered by the CRP.

AUTHOR CONTRIBUTIONS

Chenyang Hu: Conceptualization; data curation; formal analysis; methodology; validation; visualization; writing—original draft; writing—review and editing. Wei Zhang: Funding acquisition; methodology; project administration; resources; software; supervision; writing—review and editing. Darrell Bosch: Funding acquisition; methodology; project administration; resources; software; supervision; writing—review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Supplementary Material

Hu, C. , Zhang, W. , & Bosch, D. (2025). Improving nitrogen management in US agriculture: Yield Reserve versus land retirement. Journal of Environmental Quality, 54, 1307–1321. 10.1002/jeq2.70094