What drives landowners’ conservation decisions? Evidence from Iowa

Results and Discussion

As table 1 shows, no-till is used on 27% of Iowa farmland, making it the most prevalent of the four studied conservation practices. No-till is most concentrated in the southwest and west-central CRDs—used on 56% and 40% of farmland, respectively—possibly due to erosion common in the Loess Hill soils in western Iowa, as evidenced by the high percentages of HEL in that area. We test this idea by examining how conservation practice use differs by soil quality using counties’ HEL shares summary data from the Environmental Working Group. We find a positive relationship, suggesting that a 10 percentage-point increase in a county’s HEL corresponds to a 3.7 percentage-point increase (2.4 to 5.1, 95% confidence interval) in the county’s share of no-till (p-value < 0.001, R² = 0.248) (figure 1a).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

County shares of highly erodible land vs. share of (a) no-till and (b) cover crops. Data obtained via personal communication with Soren Rundquist, Director of Spatial Analysis, Environmental Working Group, based on data from USDA Farm Service Agency-Common Land Unit and the USDA National Agricultural Statistics Service, Cropland Data Layer.

Cover crops are used on 4% of Iowa’s farmland. Cover crops are used the most in the southeast CRD (12% of farmland), followed by the northeast and south-central CRDs. All three districts are high in beef or dairy cattle production, possibly due to spillovers across farming enterprises—Plastina et al. (2018) finds that grazing a cover crop or harvesting it for forage adds around US$49.42 ha^–1 (US$20 ac^–1) in cost savings on animal feed. Figure 1b shows on average a 10 percentage-point increase in a county’s HEL share corresponds to a 0.59 percentage-point increase in cover crops (0.02 to 1.2, 95% confidence interval; p-value = 0.043; R² = 0.042). The presence of HEL appears to be a larger driving factor in the adoption of no-till than for cover crops.

The positive relationship between HEL and the use of these conservation practices can be explained by compliance provisions that require farmers to agree on a conservation plan for HEL with their local USDA Natural Resources Conservation Service (NRCS) office before participating in most Farm Service Agency or Risk Management Agency programs or receiving federal government crop insurance subsidies. The dispersion in these regressions could be due to variations in recommendations given and funding provided by the decentralized NRCS offices.

Buffer strips are used on 3% of land statewide and 6% of the land in the north-central and northeast CRDs. Ponds are used on 2% of land statewide and are most prevalent in the south-central CRD. Ponds are predominantly used in high livestock production areas, likely because they also provide water for cattle.

In the next two subsections, we analyze conservation practice use across four categories: (1) land tenure (whether the parcel is operated by the landowner or rented out), (2) operator status (whether the landowner farms), (3) farming experience, and (4) local versus absentee farmer. Figure 2 shows the distribution of Iowa farmland by land leasing arrangements and landowner’s farming experience and residency status. Operator-landowners include full-time farmer landowners (farm and have no off-farm job) as well as part-time farmer landowners (farm and have off-farm employment). NOLs include owners who have never farmed and those retired from farming. Owner-operated land is farmed by the surveyed landowner or a co-owner, whereas rented-out farmland is farmed by a tenant who is not one of the owners. Absentee landowners do not reside in Iowa and local landowners reside in Iowa at least part of the year. Figure 2 shows that 53% of Iowa farmland is rented out by the landowner—only 37% is operated by a landowner. The remaining 10% is custom farmed or in government programs, such as the CRP. Among the rented farmland, 45% belongs to NOLs, and landowners that currently farm own the remaining 8%. Individuals who have never farmed own 25% of the state’s farmland. Almost one-third of farmland owned by those who have never farmed belongs to absentee landowners.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Breakdown of Iowa farmland by landowner type. Note: “Co-owner of NOL operates land” represents land for which the surveyed landowner did not farm the land, but the respondent indicated that another owner did.

Land Tenure. We examine how no-till, cover crop, buffer strip, and pond use differ by land tenure, and test our hypothesis that conservation practices are used on a greater share of owner-operated farmland than rented-out farmland. No-till is a short-term conservation practice and may even be profitable in the short term (Ibendahl 2016); thus, we expect land rental arrangements not to hinder no-till adoption. Figure 3 shows the between-group t-test results when comparing the share of owner-operated and rented-out farmland under each conservation practice (1 versus 2 in figure 2). At the state level, buffer strips and ponds are more prevalent on owner-operated land (p-values = 0.071 and 0.083), and there is no statistically significant difference for cover crops (p-value = 0.220), which may be due to the duration and expense of those practices and that tenants are less likely to adopt long-term practices. Thirty percent of rented-out farmland and 20% of owner-operated farmland uses no-till (p-value = 0.006), which is in line with the idea of no-till generating short-term benefits (Ibendahl 2016). We cannot directly infer from figure 3 why no-till is more prevalent on rented land than owner-operated land—many factors affect conservation practice use, and other variables may confound the effect of land tenure. Thus, we explore the effect of land tenure on conservation practice use by region, landholdings, and farming status.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Iowa conservation practice farmland shares by land tenure and practice type. Bars reflect standard error of the mean.

Land tenure results could be driven by regional specificities, since regional-level soil and land characteristics may affect conservation practice use and the share of rented land varies throughout the state. Figure 4 shows the share of no-till and cover crops on owner-operated and rented land by CRD. The statewide relationship between land tenure and no-till is consistent across CRDs—eight of nine CRDs have higher rates of no-till on rented farmland. However, the differences are statistically significantly different from zero only for the central and south-central CRDs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Iowa share of farmland with (a) no-till and (b) cover crops by land tenure and crop-reporting district (Northwest [NW], North Central [NC], Northeast [NE], West Central [WC], Central [C], East Central [EC], Southwest [SW], South Central [SC], and Southeast [SE]). Bars reflect standard error of the mean.

The share of farmland with cover crops is higher on rented land in five CRDs, but only one is statistically significant at a 95% confidence level. Figure 4, therefore, confirms that our results are not driven by an anomaly in any one particular district.

We examine whether farming experience affects the share of owner-operated and rented farmland using no-till and cover crops. We use full- versus part-time farming as a proxy for farming operation scale, as full-time farmers operate a greater area of land than part-time farmers, in general. We do not consider retired landowners or those who have never farmed because they do not have owner-operated land. For our robustness checks, we compare the rates of no-till and cover crops among groups 1.A versus 2.A and 1.B versus 2.B from figure 2. Landowners who farm full time use no-till on about the same proportion of their operated and rented-out land (29% versus 31%, p-value = 0.865) (figure 5a). However, part-time farmer landowners use no-till on a significantly lower share of their owner-operated land (13%) than on the land they lease out to others (39%) (p-value = 0.002) (figure 5a), suggesting that lower no-till adoption on owner-operated land is largely due to low adoption by part-time farmers. We believe this is due to part-time farmers typically operating less land than full-time farmers. Several studies document that conservation-tillage adoption is positively correlated with area of farmland operated (Lee and Stewart 1983; Rahm and Huffman 1984; Epplin and Tice 1986; Gould et al. 1989; Sheikh et al. 2003; Davey and Furtan 2008; Vitale et al. 2011; Wade and Claassen 2017; Canales et al. 2018). Additionally, part-time farmers may not have as much time to engage with other farmers to learn about no-till, and may be less likely to adopt due to the learning curve associated with using no-till. Full- and part-time owner-operators have cover crops on a greater proportion of their owner-operated than rented-out land (7% versus 3% and 3% versus 1%, respectively), but the difference is only statistically significant for landowners who farm full time (p-values = 0.005 and 0.101, respectively) (figure 5b).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Share of owner-operated vs. rented out Iowa farmland that has (a) no-till and (b) cover crops by landowner farming status. Bars reflect standard error of the mean.

Operator Status, Farming Experience, and Iowa Residency. We analyze how operator status, farming experience, and residency affect conservation practice use (table 2) and find that operator-landowners have all four conservation practices on higher proportions of their farmland than do NOLs, which is significant because NOLs own 57% of the state’s farmland (Zhang et al. 2018). We expect landowners with farming experience to be more likely to use conservation practices because they likely possess more knowledge of them. We observe this expectation across all four conservation practices when we compare full-time farmer landowners with those who have never farmed, but not for part-time or retired farmers. This is especially concerning because landowners who have never farmed own 34% of the state’s farmland (Zhang et al. 2018).

All four conservation practices are implemented on a lower share of absentee landowner land when compared to local landowners (table 2c). Novel approaches may be needed to increase the effectiveness of outreach to NOLs, especially absentee NOLs, per Petrzelka and Armstrong (2015).

Financial Characteristics. Table 3 shows how conservation use differs by owner’s landholdings, percentage of agriculture-based income, and percentage of land that has been paid for. On average, landowners with more land tend to use no-till at a higher rate. Landowners with more than 809 ha (2,000 ac) use no-till on 36% of their land compared to just 20% for those who own 0 to 40 ha (0 to 99 ac). There was no obvious pattern for cover crops, buffer strips, or ponds.

Landowners with higher percentages of agriculture-based income have higher farmland shares of no-till and cover crops, with the exception of landowners with entirely agriculture-based income. Among landowners with entirely agriculture-based income, 80% are operator landowners and 60% state that current income is the most important reason for owning farmland. There is not a clear relationship between percentage of land that is paid for and conservation practice use.

Landowner Demographics. Tables 4a, 4b, and 4c show Iowa’s conservation practice use by landowners’ age, gender, and education, respectively. Table 4a shows there is no consistent relationship between landowner age and use of conservation practices. No-till is least prevalent on land owned by someone less than 55 years old, which contrasts with prior literature that suggests older farmers are less likely to adopt conservation practices because they may have less time to obtain the benefits (Prokopy et al. 2008). Landowners younger than 55 use cover crops and ponds at the highest rates.

Gender does not have an effect on conservation practice adoption. Each practice is used on about the same proportion of farmland owned by males and females, which fails to support women landowners being more likely to adopt the conservation practices studied, despite findings that female landowners may be more conservation oriented than male landowners (Eells and Soulis 2013; Druschke and Secchi 2014). Druschke and Secchi (2014) find that although women are favorable toward conservation, they have lower knowledge levels about conservation practices, and Carolan (2005) finds female landowners may feel alienated and less comfortable making recommendations to male tenants. Women own 47% of Iowa farmland (Zhang et al. 2018); thus, increasing outreach efforts to female landowners could have a sizable impact on conservation use.

The relationship between education and conservation practice prevalence is unclear. There is a direct relationship for ponds—landowners with higher levels of education have ponds on a greater share of their land. However, the opposite is observed for no-till—high-school educated landowners have no-till on 34% of their land, compared to 21% for landowners with a graduate degree. This contrasts with past studies that find a positive relationship between education and conservation practice adoption (Prokopy et al. 2008).

Landowner Perspectives and Future Intentions. We look at how conservation practice use is expected to evolve in the near future. Tables 5a and 5b show that landowners are open to having more cover crops on their land—in the next five years, 18% are likely to use them and an additional 34% might use them. These two groups of landowners own 19% and 38% of Iowa farmland, respectively. It is unlikely that these farmers will adopt cover crops on all of their land; however, it would be a substantial increase from the 4% of farmland that is currently cover cropped. Only 10% of landowners state that they are likely to use no-till, 4% are likely to use buffer strips, and 2% are likely to use ponds over the next five years.

Table 6 shows which policies could be effective at inducing landowners to adopt conservation practices. Tax credits or deductions in exchange for implementation of conservation practices would be most effective—45% of landowners state that they would be likely or very likely to adopt more conservation practices under such a policy. Thirty-six percent of landowners are likely or very likely to adopt more conservation practices if tax-free cost sharing is available, and 22% if land enrolled in conservation programs is excluded from the value of the estate for tax purposes. Enthusiasm among landowners is understandable; however, their future adoption will be contingent on whether and how these new policies are delineated.

We also investigate landowners’ willingness to encourage their tenants to use cover crops. Barriers to conservation practices on rented land exist on both the landowner and tenant side (Carolan 2005); thus, we disentangle these factors by examining whether and in what ways landowners are willing to encourage their tenants to use more cover crops. As table 7 shows, about one-third of landowners would pay for a portion of cover crop planting costs, or increase the length of a tenant’s lease if they adopted or increased the area under cover crops, which is important because prior literature cites high costs (Roesch-McNally et al. 2018) and short leases (Ranjan et al. 2019; Carolan 2005) as barriers to tenants adopting conservation practices.

Table 7 shows landowners’ willingness to help tenants use cover crops based on whether the landowner has any no-till farmland, which helps us determine if having a conservation practice on their land makes them more willing to help tenants adopt a different conservation practice on rented land. We find landowners’ willingness to help tenants is higher for no-till users than for those who don’t use no-till—42% (41%) of landowners using no-till are willing to give tenants a longer lease (pay for a portion of planting costs) in exchange for planting more cover crops, compared to just 12% (30%) of those without any no-till acres. This suggests there may be links between the adoption of no-till and cover crops.

Table 8 shows landowners’ responses to an open-ended question about their main reasons for not using each conservation practice. For NOLs, the primary reason for not using no-till is that they deem it not suitable for the land (46%), while the primary reason for not using cover crops is that the decision is up to the tenant (36%).

For operator landowners, the primary reasons for not using no-till is that it hurts crop yield (17%), it is the tenant’s decision (15%), and that they tried it but did not like it (13%); and, the main reasons for not using cover crops are that it is the tenant’s decision (19%), the cost of terminating the cover crop is too high (19%), and they do not have time to plant them in the fall (16%). A relatively low proportion of NOLs indicate that no-till and cover crop use are decisions made solely by the tenant (6% and 36%, respectively), which suggests that landowners and tenants typically make joint conservation decisions on rented land. This confirms Arbuckle’s (2019) survey in which 38% of farmers stated that the tenant should solely be responsible for conservation decisions.

The overwhelming reason for not using buffer strips (84%) or ponds (88%) is that the landowner deems they are not needed on the land. These values are similar across owner-operated and rented-out land. We also note that previous research such as Zhang et al. (2016) and Arbuckle and Roesch-McNally (2015) show that farmers’ conservationist identity and perceived efficacy of a conservation practice in reducing erosion or runoff is critical for higher adoption.

Perceptions related to economics and factors of production drive many landowners’ reasons for not using cover crops, whereas land attributes are the main reason for not using buffer strips and ponds (and somewhat no-till). As research results emerge to address preconceived ideas, perceptions may evolve. For instance, while some landowners’ reasons for not using cover crops fall in line with previous literature, others are inconsistent with what has been observed. Roesch-McNally et al. (2018) find that barriers to adopting cover crops include costs and lack of time to plant them in the fall, which is similar to what the landowners stated. However, the sources of costs differ from what the landowners mentioned. Plastina et al. (2018) find cover crops’ greatest expense is planting costs—costs of terminating the cover crop are minor, as on average, farmers do not use additional inputs or machinery passes (e.g., extra tillage) than they use in absence of cover crops. NOLs cite termination costs as a reason for not using cover crops at a greater rate than operator landowners, suggesting a gap in perceptions.