1 Factors influencing land rental market participation: a case study 1 in Northern Ireland 2 Adewale H. Adenugaa and Claire Jackb 3 aGibson Institute for Land, Food and Environment, School of Biological Sciences, Queen’s 4 University, Belfast, United Kingdom 5 bEconomics Research Branch, Agri-food and Biosciences Institute, 18a Newforge Lane, 6 Belfast, BT9 5PX, UK 7 Corresponding author: a.adenuga@qub.ac.uk 8 This article has been accepted for publication and undergone full peer review but has not been 9 through the copyediting, typesetting, pagination and proofreading process, which may lead to 10 differences between this version and the Version of Record. 11 Please cite this article as: 12 Adenuga A.H., and Jack C. (2025). Factors influencing land rental market participation: a 13 case study in Northern Ireland, Just Accepted. DOI: 10.36253/bae-16400 14 Abstract 15 Agricultural land mobility through an efficient land rental market has been shown to contribute 16 to the productive and sustainable utilisation of land, by facilitating the transfer of land from 17 less productive farmers to more productive farmers. However, this is not the case in Northern 18 Ireland where the sale of agricultural land is limited with a constrained tenanted sector. The 19 objective of this study is to analyse the factors influencing participation in the land rental 20 market in Northern Ireland. To achieve our objective, data from 1466 farmland owners was 21 analysed using principal component analysis (PCA) and multinomial logistic regression model. 22 The results show that land rental market participation is impacted by motivational and 23 socioeconomic factors. The study recommends the development of schemes that support the 24 early and comfortable retirement of older farmers to increase the access of young farmers to 25 land and improve the land rental market. 26 27 Keywords: land; rental market; sustainability; conacre; multinomial logit model 28 29 JEL Classification code : Q15 30 31 1. INTRODUCTION 32 Agricultural land mobility has been shown to contribute to the productive and effective 33 utilisation of land as a resource by facilitating the transfer of land from less productive farmers 34 to more productive farmers (Bradfield, Butler, Dillon, & Hennessy, 2020; Deininger & Jin, 35 2008; Li, Ma, Mishra, & Gao, 2020; Tesfay, 2023). This can be achieved through an efficient 36 land rental market, playing an important role in shaping farmers’ land-use decisions and 37 2 supporting sustainable agricultural production (Min, Waibel, & Huang, 2017; Udimal, Peng, 38 & Guillaume, 2021). A previous study in Poland by Marks-Bielska (2021) has shown that with 39 stable long-term agreements, farming on leased land is comparable to farming on owned land 40 as long as the rights of the lessor and a lessee are protected. 41 Land mobility is a significant issue in Northern Ireland (NI) as historically there have 42 always been strong sentimental and cultural ties to land (Adenuga, Jack, & McCarry, 2023; 43 Bradfield et al., 2023a). The majority of farming in the region is undertaken on owned land and 44 the transfer of land through sale is limited due to high land sales prices. For example, the price of 45 agricultural land in NI ranges between £11500 and £20,000 per acre and less than one percent of the 46 total agricultural land area is offered for sale each year (Harris, 2022). This is reflected in the high 47 proportion of farming undertaken on owned land compared to what occurs in other countries. Owned 48 land as a percentage of farmed area in NI is 72% which in relative terms is high compared to 49 other countries such as France and Germany where the proportion is 38% and 39% respectively 50 and at the EU level, which is 48% (Adenuga, Jack, & McCarry, 2021; DAERA, 2023a; 51 Eurostat, 2022). This makes the purchase of land for agricultural production in NI less optimal 52 (Adenuga, Jack, & McCarry, 2021). This is because land ownership requires significant capital 53 investment, and further purchases may not be financially optimal when alternative 54 arrangements, such as renting, provide operational flexibility without the long-term financial 55 burden of land acquisition. Additionally, owning a large proportion of land may indicate that a 56 farmer’s operational needs are already met, reducing the necessity for further expansion 57 through purchase. In addition, the average agricultural land area in the region can be regarded 58 as small relative to the other regions of the UK and has become highly fragmented with 59 increased concentration and competition for use among active and intending farmers (Adenuga, 60 Jack, McCarry, & Caskie, 2023; Milne et al., 2022). The low level of land mobility, and high 61 fragmentation have a consequential effect on the overall competitiveness and productivity of 62 3 the NI agri-food sector. This potentially constrains opportunities for new entrant farmers to 63 access land (Milne et al., 2022). 64 Access to land through the rental market provides an avenue for farmers and aspiring 65 farmers to access land and increase their competitiveness since it requires less capital outlay 66 (Adenuga, Jack, & McCarry, 2021; Bradfield et al., 2020; Jin & Deininger, 2009). With solid 67 legal regulation, the land rental market can be regarded as a rational land management strategy 68 (Marks-Bielska, 2013). Compared to the sale of land, it offers greater flexibility, with an 69 opportunity to design contractual terms to suit both the lessor and the lessee (Zhang et al., 2018) 70 It allows farmers to alter farm size, exploit economies of scale, increase operation as well as 71 technical efficiency, and capture technological advances to achieve an optimal level of 72 production (Bradfield et al., 2021; Bradfield et al., 2020; Geoghegan et al., 2021; Li et al., 73 2020; Zou & Luo, 2018). In addition, the land rental market makes it possible for rural 74 households to generate additional income from their land (Lan Zhang et al., 2018; Zou, Mishra, 75 & Luo, 2020). An efficient and fully functional land rental market that supports optimal 76 allocation and transfer of land is therefore necessary to bring land to its most productive use 77 and provide opportunities to transform the rural economies and improve the welfare of rural 78 households (Bradfield et al., 2020; De Janvry, Gordillo, Sadoulet, & Platteau, 2001; Huy, Lyne, 79 Ratna, & Nuthall, 2016; Jin & Deininger, 2009). For example, a study by Bradfield et al. (2020) 80 for dairy farmers in Ireland, showed that farms that rent land generated a higher net margin 81 than farms with no rented land. The study also showed that increased access to land through 82 the land rental market enhanced farming households’ succession plans. 83 Given the high cost and limited access to the sale of land in NI, an efficient land rental 84 market provides an avenue for farmers in the region to increase their farmland area. This is 85 particularly important given the shift towards sustainable farming with agricultural support 86 policies now targeted at environment-based payments (Adenuga, Jack, McCarry, et al., 2023; 87 4 Little, Lyon, & Tsouvalis, 2023). In addition, it can also provide pathways for 88 accessing agricultural land for those who may otherwise have very limited access, for example, 89 young or new-entrant farmers (Abay et al., 2021). This will be vital for farmers to meet their 90 environmental commitments. This paper sought to examine the determinants of participation 91 in land rental markets in NI as a mechanism to improve access to land. To achieve our 92 objective, we empirically analysed the motivational and socioeconomic factors influencing 93 land rental market participation in NI. To the best of our knowledge, this study provides the 94 first insights into the complex motivations and behavioural factors underlying farmers’ 95 decisions to participate in the land rental market. While these motivations may also apply to 96 farming in general, understanding, how they influence farmers’ land rental decisions is 97 essential to improving the land market. A previous study in Ireland by Geoghegan et al. (2021) 98 show that attitudinal factors are a significant predictor of openness to land mobility, both on 99 the supply and demand side of the market This study employs the multinomial logistic 100 regression model which allows us to consider not just farmers who rent-in and rent-out land 101 but also those who neither rent-out nor rent-in land. A previous study by Udimal et al. (2021) 102 also employed a similar approach. However, they did not consider motivational factors that 103 could influence participation in the land rental market. It remains unclear how the different 104 farmers’ motivations and behavioural factors influence their decision to participate in the land 105 rental market. Results from this study will be useful in providing the requisite evidence base to 106 inform the formulation of policies aimed at encouraging farmers and landowners to engage in 107 the land rental market. 108 2. RESEARCH METHODOLOGY 109 In this study, we employed the multinomial logit (MNL) model to analyse the 110 motivational and socioeconomic factors influencing land rental market participation in NI. 111 Farms in NI are typically family-owned with a small, tenanted sector compared to other regions 112 5 of the United Kingdom. By making use of a couple of attitudinal statements, we derived 113 different farming motivations of farmers using principal component analysis (PCA). 114 2.1 Data collection 115 116 The dataset used in this study was obtained from a cross-sectional survey of farmers in 117 NI. The sampling frame was the farm census data for NI which consisted of 12 747 farmers for 118 the year 2020 from which we selected 4029 farmers using a stratified random sampling 119 technique. The farmers were classified into six strata based on farmland ownership and rental status: 120 those farming solely on owned land, those farming both on owned and rented land, those farming on 121 owned and rented land while also leasing out land, those farming only on owned but also leasing out 122 land, those farming exclusively on rented land and landowners who lease out all their land. Due to the 123 large number of farmers in the “owned land only” and “owned and rented land” groups, we randomly 124 selected 20% of farmers from these two groups. A well-structured questionnaire was developed 125 after a comprehensive literature review and key informant interviews (Adenuga, Jack, & 126 McCarry, 2021). The questionnaire was organised around some themes. This includes land 127 ownership and rental status and duration, socioeconomic as well as farmers’ attitudinal 128 characteristics. The questionnaire was developed in a hybrid format, making it possible to be 129 completed online and on paper. The Snap survey software was used to design the online version 130 of the questionnaire with a QR code that was generated and placed on the front page of the 131 paper version of the questionnaire. Respondents either completed the paper questionnaire 132 directly or scanned the QR code with their phone to complete it online. The survey took place 133 between December 2021 and February 2022. We sent two reminders over this period with the 134 QR code included in the letters and the respondents could request a paper copy of the 135 questionnaire if they require a new one. No personally identifiable information was collected, 136 and farmers were assured of their anonymity in reports or publications resulting from the 137 project. To encourage the farmers to complete the questionnaire, each completed and returned 138 6 questionnaire was entered into a prize draw for 1 of 10 £100 e-vouchers. This was for farmers 139 who had indicated their intention to participate in the draw. Out of the 4029 questionnaires 140 administered, 1228 paper questionnaires were returned in the pre-paid envelopes sent alongside 141 the questionnaire while 499 questionnaires were completed online. In total, we received 1727 142 responses. This number was reduced to 1466 following the dropping of 91 farmers who both 143 rented out and rented in land, and farmers with no information relating on land ownership or 144 rentals. Table 1 gives a summary of the socioeconomic characteristics of the farmers. Some of 145 the variables do not add up to 1,466 because some of the farmers did not fully complete the 146 questionnaire, omitting some questions. 147 2.2 Empirical model 148 149 The land rental market participation is modeled based on the random utility theory 150 (Udimal et al., 2021). The theory assumes that each farmer 𝑖 has different options (𝑘 =151 1,2,3…𝐾) available to them and they chose a particular option (y) that offers the maximum 152 utility by considering the economic and environmental risk associated with the various options 153 (Diriye et al., 2022; Grant et al., 2019; McFadden, 1972). To analyse the factors influencing 154 land rental market participation in this study, we employed the multinomial logit model 155 (MNL)(Daly, 1987; Gensch & Recker, 1979). The MNL model has been used extensively in 156 the literature to analyse farmers’ behaviour and choices in decision-making (Diriye et al., 2022; 157 Duressa, 2021; Otieno, 2022; Ouattara et al., 2022). The model, unlike the bivariate Tobit 158 model employed in previous studies, for example, the study by Rahman (2010), can 159 accommodate the non-binary, multivariate nature of the dependent variable (Dang & Pham, 160 2022; Osanya et al., 2020; Yasmin et al., 2022). Previous studies have also shown that the MNL 161 model performs better than the multinomial probit model (MNP) (Dow & Endersby, 2004). 162 Besides, the MNL model possesses fewer computational problems compared to MNP because 163 the probit likelihood function is often flat near its optimum and it generally requires numerical 164 7 approximation for the multivariate integrals(Dow & Endersby, 2004). It is also not prone to 165 optimization errors. Although it should be acknowledged that MNL model relies on the 166 Independence of Irrelevant Alternatives (IIA) assumption, which requires that the relative 167 probabilities of choosing between any two alternatives remain unchanged when other 168 alternatives are introduced or removed. In this study, certain alternatives, such as the sale and 169 purchase of agricultural land, were excluded due to the thin market for land transactions, as 170 discussed in the introduction. While these alternatives may theoretically exist, their practical 171 relevance is limited in this context. Nevertheless, we assess the validity of the IIA assumption 172 and the results indicate that the exclusion of these alternatives does not significantly influence 173 our findings. 174 Generally, our choice of a multivariate dependent variable is based on the need to 175 capture the multiple dimensions of farmers’ land rental behaviour within a single model 176 framework which allows us to identify distinct factors influencing each dimension 177 simultaneously. It should also be acknowledged however that our use of the MNL model 178 means we are not modelling the actual amount of land rented-in or rented-out which could 179 provide additional insight. The choice of our model is in line with the objective of this paper 180 which was mainly to identify factors influencing farmers’ land rental behaviour. In this study, 181 it is assumed that a farmer can choose from any of three alternatives. That is rent-out land; rent-182 in land and neither rent-in nor rent-out land. The observed outcome, 𝑌𝑖𝑘 as a function of the 183 variables 𝑋𝑖𝑘 is presented in equation (1). 184 𝑌𝑖𝑘 = 𝑋𝑖𝑘𝛽𝑘 + 𝜀𝑖𝑘 (1) 185 Where 𝑋 is a vector of explanatory variables which include motivational and farmer-specific 186 socioeconomic characteristics that are hypothesised to influence land rental market 187 participation, 𝛽 represents the parameters to be estimated and 𝜀 is the error term or random 188 8 component of the model and it is assumed to be independently and identically distributed 189 (Diriye et al., 2022). The choice probability of the 𝑖𝑡ℎ farmer in choosing option k from the list 190 of K options following Mellon-Bedi et al. (2020) is presented in equation 2. 191 (𝑌𝑖 = 𝑘|𝑋) = exp⁡(𝛽𝑘𝑋𝑖) 1 + ∑ exp⁡(𝛽𝑗𝑋𝑖) 𝐾 𝑗=2 , 𝑓𝑜𝑟⁡𝑘 > 1⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡(2) 192 In estimating the model, one of the categories was normalised by equating it to one 193 (equation 3). In our analysis, we used neither renting-out nor renting-in land category as the 194 baseline such that only two equations are estimated. The probability of renting out and renting 195 in land is compared to the probability of neither renting in nor renting out land. 196 𝑃(𝑌𝑖 = 1) = 1 1 + ∑ exp⁡(𝛽𝑗𝑋𝑖) 𝐾 𝑗=2 , 𝑓𝑜𝑟⁡𝑘 = 1⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡(3) 197 Based on this parameterisation, the jth element of the vector 𝛽𝑘 is interpreted as the 198 increase in the log-odds ratio of the kth category relative to the reference category holding the 199 other explanatory variables constant to their mean values (Carpita et al., 2014). 𝑌𝑖 is the 200 dependent variable representing farmers that rent out and rent in land in the two models 201 respectively. The variables hypothesised to influence participation in the land rental market 202 include the age of the farmer measured in years, membership of a business development group 203 (BDG) measured as a dummy variable, having a diversification enterprise (this refers to owning 204 a diversification enterprise on the farm) measured as a dummy variable, off-farm employment 205 (defined as income earned outside of the farm by the primary decision-maker and may include 206 spousal income, which could influence household labour availability and financial 207 stability) measured as a dummy variable, identification of a successor measured as a dummy 208 variable, the enterprise type of the farmer (dairy, beef sheep and others which include arable 209 and horticulture enterprises) , area of farmland owned measured in hectares, level of 210 agricultural qualification measured as a dummy variable, level of formal education, time 211 9 commitment to farming (full-time or part-time) and land type (land classification based on their 212 agricultural conditions as lowland, disadvantage or severely disadvantaged). These variables 213 were obtained from the literature and are considered important factors likely to influence land 214 market participation (Adenuga, Jack, & McCarry, 2021, 2023; Bradfield et al., 2020; Che, 215 2016; Zou et al., 2020). The descriptive statistics of the variables are presented in Table 1. 216 To incorporate the motivational factors into our analysis, principal component analysis 217 (PCA) was employed. The PCA is a statistical technique that examines the pattern of 218 correlations amongst the explanatory variables and creates a smaller set of uncorrelated linear 219 combinations of the original variables (Lapple & Kelley, 2010; O’Kane et al., 2017). The 220 higher a respondent’s score on each of the factor variables, the higher their associated level of 221 agreement with the specific attitudinal component (Howley, 2015; O’Kane et al., 2017). The 222 analytical technique was used to reduce sixteen attitudinal statements which represent varying 223 motivations of farmers into four components. The statements included in our principal 224 component analysis were obtained from a comprehensive review of related literature on 225 farmers’ behaviour (Adenuga, Jack, & McCarry, 2023; Howley, 2015; Howley et al., 2015). 226 We used the promax rotation to facilitate the interpretation of the components. As is the usual 227 practice, components with an Eigenvalue of at least one were retained (O’Kane et al., 2017). 228 We retained statements with loadings greater or equal to 0.3 on their target factor. Statements 229 that did not load greater or equal to 0.3 on any component were dropped. The four motivational 230 components obtained from the PCA include Principal Components (PC) 1 which shows high 231 loadings for items relating to efficient farm management and technology adoption and was 232 termed “progressive construct”, PC2 loads highly on statements that do not support pro-233 environmental behaviour such as “I am not that concerned about environmental issues” and 234 was termed “environmental apathy” construct., PC3 which loads on the statements that 235 prioritise the protection of the environment was termed “pro-environmental construct ” and 236 10 PC4 which loads highly on statements that relate to being risk averse and was termed “risk 237 averse” construct. Two statements that did not align with any of the four components were 238 dropped. The internal consistency of each component was assessed using Cronbach’s alpha 239 coefficient which ranged between 0.60 and 0.68, showing good reliability. With a Kaiser–240 Meyer–Olkin measure of sampling adequacy of 0.75, the components represent a significant 241 proportion of the variance in the data (Läpple & Kelley, 2013). The results of the PCA analysis 242 are presented in Table A1 in the appendix. The STATA spost13 post-estimation command was 243 employed to allow the coefficients of the MNL model to be interpreted in terms of the 244 percentage change in odds for a unit change in the explanatory variable (Long & Freese, 2005). 245 3. RESULTS AND DISCUSSION 246 3.1 Descriptive and socioeconomic characteristics 247 Table 1 gives an overview of the socioeconomic and farm characteristics of the 248 respondents. Similar to the general farming population in Northern Ireland, about 90% of the 249 farmers in our sample are livestock farmers (DAERA, 2023b). Our analysis showed that more 250 than half (55%) of the farmers undertake their farming activities in land categorised as either 251 disadvantaged (DA) or severely disadvantaged areas (SDA). The SDA or DA (less favoured 252 areas (LFA)) land types refers to land located in parts of the country which, because of their 253 relatively poor agricultural conditions, have been so designated under EU legislation(Caskie et 254 al., 2001; DAERA, 2023b). The majority of respondents are male (91.5%) and 75% of 255 respondents are married. Thirty-two percent of the farmers stated that they had no general 256 education qualifications. Based on land rental participation status, about 45% of those who rent 257 out land have a diploma or degree level education while it is 35%, of those who rent in land. 258 In terms of agricultural qualification, 33% of the farmers stated that they have formal 259 agricultural qualifications. The value is higher for those that rent in land with 40% of them 260 stating that they have a formal agricultural qualification. Only 30% of those who rent out land 261 11 stated that they have formal agricultural qualifications, and it is 33% for those who neither rent 262 in nor rent out land. For 37.9% of respondents, farming was on a full-time basis (these are 263 farmers who work, at least on average, 38.0 hours per week on the farm). Twenty-six percent 264 of sheep farmers surveyed reported that they were farming full-time. Ninety-eight percent of 265 the farms are family farms held either as sole ownership or partnership. The average years of 266 farming experience is 35 years and 55% of the farmers stated that they have off-farm 267 employment. The proportion of farmers with off-farm employment that rent in land was 61% 268 while it was 47% for those that rent out land and 55% for those that neither rent in nor rent out 269 land. Only 13% of the farmers are members of the Business Development Groups (BDGs). 270 Nineteen percent of those who rent in land were in the BDG group compared to just 9% for 271 those who rent out land and those who neither rent in nor rent out land. The BDGs is a 272 knowledge transfer scheme developed by the Northern Ireland College of Agriculture, Food 273 and Rural Enterprise (CAFRE) in March 2016. The scheme employs a group approach aimed 274 at improving the performance of farm businesses through facilitated ‘peer-to-peer’ learning to 275 encourage the fostering of knowledge capital and knowledge exchange between actors (Jack et 276 al., 2020). Forty percent of the farmers stated that a successor to the farm has been identified. 277 This was 34% among those who rent out land, 41% among those who rent in land, and 42% 278 among those who neither rent out nor rent in land. The modal age group was 55 to 64 years 279 and 68% of the farmers are aged over 55 years. Based on land rental market participation, as 280 much as 83% of those who rent out land are older than 55 years, while it is 44% of those who 281 rent in land and 66% of those who neither rent in nor rent out land. This is an indication that 282 those who rent out land are older and more likely to include retired farmers. The average 283 amount of cultivated land owned is 36 hectares. The value is 43 hectares for those that rent in 284 land, 37 hectares for those that rent out land, and 28 hectares for those that neither rent in nor 285 rent out land. 286 12 287 288 289 290 291 Table 1: Socioeconomic characteristics of respondents 292 Variables Frequency Percentage- (%) Time commitment (n= 1354) Full-time 513 37.9 Part-time 841 62.1 Land types (n = 1351) Lowland 608 45.0 Disadvantaged 428 31.7 Severely Disadvantaged 315 23.3 Diversification activities (n=1466) Yes 258 17.6 No 1,208 82.4 Identification of successor (n=1347) A successor has been identified 536 39.8 Successor not yet identified 811 60.2 BDG membership (n = 1466) Yes 192 13.1 No 1,274 86.9 Age of the farmer (n = 1466) 13 Less than 30 49 3.3 30-40 118 8.05 41-54 303 20.7 55-64 400 27.3 65-74 358 24.4 75 or older 238 16.2 Education of the farmer (n=1466) No formal qualification 471 32.1 Less than 5 GCSEs 136 9.3 5 GCSEs or equivalent 219 14.9 A level or equivalent 97 6.6 Higher education - diploma or equivalent 242 16.5 Degree level or higher 301 20.5 Agricultural qualifications (n=1447) No formal agricultural qualification 963 66.6 National Diploma NVQ Level 2 or equivalent 171 11.8 National Diploma NVQ Level 3 or equivalent 101 6.9 HND Level or equivalent (just below degree level) 74 5.1 Degree level or equivalent 68 4.7 Others 70 4.8 Off farm employment (n=1466) Yes 808 55.1 No 658 44.8 Source: own elaboration 293 294 3.2 Land market participation 295 296 14 The results of our analysis in terms of land rental characteristics showed a relatively 297 even distribution among the three categories of farmers. We found that 31% of the farmers 298 farmed on owned land only (i.e., neither rent in nor rent out land) while 37% and 31% of the 299 farmers rented in and rented out land respectively. Table 2 provides an overview of how land 300 is rented in and rented out in conacre and on long-term lease. The conacre is the predominant 301 form of land rental system in the region. It is a traditional short-term land rental system unique 302 to the island of Ireland in which land is let to a farmer nominally for 11 months or 364 days 303 without the need for either party to enter a long-term commitment. Specifically, 91% and 78% 304 of the land that is rented out and rented in respectively are in conacre (Adenuga, Jack, & 305 McCarry, 2023). While the duration for renting land in conacre is 11 months or 364 days, the 306 arrangement between the landlord and the tenant is such that it can be rolled over for several 307 years with no formal contract signed. For example, as shown in Table 2, among landlords 308 renting out and farmers renting in land, it can be observed that 48% and 67% have been rented 309 out and rented in respectively for more than 10 years. The long-term lease refers to a land rental 310 arrangement in which a formal contract has been signed between the landlord and the tenant. 311 This proportion of long-term leases among landowners and farmers in NI is still relatively small 312 with a higher proportion on 5-year leases as shown in Table 2. 313 Table 2: Land rental system and duration of rental 314 Land Rental Characteristics Renting out land Renting in land Frequency Percentage (%) Frequency Percentage (%) Ways by which land is rented Conacre 422 91.1 426 78.0 A long-term lease 22 4.8 66 12.1 A combination of conacre and long- term lease 19 4.1 54 9.9 Duration of Conacre (years) Less than 3 40 9.1 34 7.1 15 3 413 9.3 17 3.5 4 25 5.7 11 2.3 5 34 7.7 28 5.8 More than 5 but less than 10 88 20.0 68 14.2 10 or more years 212 48.2 322 67.1 Duration of long-term lease Less than 3 2 4.9 9 7.5 3 2 4.9 8 6.7 4 1 2.4 0 0 5 20 48.8 42 35.0 More than 5 but less than 10 5 12.2 15 12.5 10 or more years 11 26.8 46 38.3 315 As presented in Table 3, we also analysed the land market participation characteristics 316 of the respondents based on enterprise types. The results show that renting in land was more 317 common among dairy farmers compared to other enterprise types with 79% of the dairy farmers 318 renting in land, in addition to farming on owned land. Only 5.5% rented out land. This may 319 reflect the intensive nature of dairy farming in Northern Ireland (Adenuga et al., 2020). It also 320 may have been driven by the relatively higher incomes of dairy farms and their ability to pay 321 higher rents. This is evidenced by the high percentage (84%) of dairy farmers who undertake 322 farming on a full-time basis. Although the majority (84%) of the land rented in was in conacre 323 rather than long-term leases. 324 Table 3: Land rental characteristics by enterprise type 325 Enterprises Frequency percentage Rent out land (%) Rent in land (%) Neither rent out nor rent in land (%) Proportion that are full-time farmers (%) Beef suckler 409 32.2 18.3 43.3 38.4 36.3 Beef finishing 316 24.9 25.0 39.9 35.1 35.4 16 Source: own elaboration 326 3.3 Results of econometric analysis 327 328 The results of the parameter estimates resulting from the MNL model are presented in 329 Table 3. The likelihood ratio chi-square of 366.61 with a p-value < 0.0001 indicates that the 330 model fits well. A test for The Independence of Irrelevant Alternatives (IIA)showed that it was 331 not violated. Like other categorical response models, the interpretation of the MNL model is 332 complex because of its nonlinearity. To interpret our MNL model, we have employed the Long 333 and Freese (2006) SPost13 command, listcoef which can provide a single table of the estimates 334 for all the comparisons of outcome categories for each variable included in the model. The 335 coefficients are interpreted in terms of the standardized or percentage change in odds for a unit 336 change in the explanatory variable holding all other variables constant (Howley et al., 2015; 337 Long & Freese, 2006). 338 The results show that motivational and socioeconomic factors influence the likelihood 339 of land rental market participation. In total, fourteen explanatory variables were statistically 340 significant for factors influencing the decision to rent out, land. For the factors influencing the 341 decision to rent in, land, we found seven explanatory variables to be statistically significant 342 compared to the baseline (the decision to neither rent in nor rent out land). We found the 343 progressive construct variable to be a statistically significant factor (p < 0.05) for the decision 344 to rent out and rent in land. In specific terms, one standard deviation increase in the progressive 345 Sheep 306 24.1 29.1 35.3 35.6 25.8 Dairy 110 8.7 5.5 79.1 15.5 83.6 Arable 81 6.4 49.4 27.2 23.5 43.2 Poultry 30 2.4 33.3 23.3 43.3 66.7 Horticulture 13 1.0 61.5 23.1 15.4 38.5 Pig 5 0.3 20.0 60.0 20.0 80.0 17 orientation factor decreases the odds of renting out, land for the average farmer by 15.9%. On 346 the other hand, a standard deviation increase in the progressive orientation factor increases the 347 odds of renting in, land by 18.6%. This implies that farmers, with a progressive and positive 348 mindset and the motivation to maximise profit, have a lesser tendency to rent out land compared 349 to the baseline and are more likely to rent in land. This is understandable as more land is often 350 required to take advantage of economies of scale, adopt new technology, and increase farm 351 incomes (Huy et al., 2016; Geoghegan et al., 2021). The pro-environment factor was found to 352 have a positive relationship with renting out land, but it was not statistically significant. It was 353 however statistically significant and negatively related to renting-in land. A standard deviation 354 increase in the pro-environment factor reduces the odds of renting in, land by 13.5%. The 355 negative relationship between pro-environmental behaviour and renting in, land may be linked 356 to the predominantly short-term conacre land rental system predominant in NI. This is because 357 environmentally friendly agricultural practices often require long-term investments in soil 358 health and this is difficult to achieve on short-term land rentals without adequate tenure 359 security. Besides, farmers with pro-environmental behaviour may avoid renting additional land 360 to retain effective control of their land in line with their values of ensuring proper stewardship 361 of the land. 362 In terms of the socioeconomic characteristics, we found the enterprise type and the time 363 commitment to farming to be statistically significant factors in both the decision to rent out and 364 rent in land but with opposite signs. Specifically, farms that are classified as dairy, beef or 365 sheep enterprises are less likely to rent out land and more likely to rent in relative to the baseline 366 enterprises of arable, horticulture, poultry, and pig farms which were categorised as “others”. 367 Having a dairy, beef or sheep enterprise reduces the odds of renting out land by 75%, 63% and 368 46% respectively and increases the odds of renting in land by 378%, 68% and 28% 369 respectively. This reflects the intensive nature of land usage in livestock production in NI 370 18 relative to other enterprises with about 79% of the total farmed area used for livestock 371 production (DAERA, 2023a).On the other hand, the age of the farmer, land type, identification 372 of a successor, the education of the farmer, the area of land owned, and land type were found 373 to be a statistically significant factors influencing the decision to rent out land. Membership of 374 the BDG and access to off-farm income were found to be statistically significant factor 375 influencing the decision to rent in land. 376 The identification of a successor reduces the odds of renting out land by 37% relative 377 to the baseline. With a successor already identified, farmers would rather keep their land to 378 themselves to keep the successor on the farm rather than renting out land. A previous study by 379 Bradfield et al. (2020) has also shown that with a successor identified, farmers are more likely 380 to keep their land and instead rent in land to provide the successor with immediate employment 381 and skill development. Similarly, study by Daniele (2024) has shown that the identification of 382 a successor in a farming household significantly influences farmers’ strategic choices. Farmers 383 with higher education are more likely to rent out land. A standard deviation increase in degree 384 level or higher education increases the odds of renting out land relative to the baseline by 34%. 385 This may imply that farmers who are highly educated may be spending less time in direct 386 farming activities. This is because higher education tends to increase the opportunity cost of 387 farming due to the increased potential of securing higher-paying jobs outside agriculture. 388 Consequently, these farmers may choose to rent out their land instead of farming it themselves. 389 This connection between education and land rental decisions reflects broader economic trends 390 where higher education leads to more diverse career options, thereby influencing land use 391 choices. This result corresponds to that obtained by Rahman (2010), Bizimana (2011), and 392 Zhang et al. (2022) in which they found that farmers with higher levels of education are more 393 likely to rent out land. The result is however in contrast to that obtained by Tesfay (2023) in 394 which they found that the less educated farmers are more likely to rent out their land. 395 19 Our result also showed that farmers older than 55 years are more likely to rent out land. 396 These are farmers who may be approaching retirement and probably do not have a successor. 397 Such farmers will be better off renting out part of their land for additional income. This result 398 corresponds to that obtained by Tesfay (2023) and Min et al. (2017) in which they found older 399 farmers to be more likely to rent out land. The result is also in line with that obtained by Mellon-400 Bedi et al. (2020) in their study of smallholder participation in the land rental market in China 401 in which they found that households with a higher share of older people were more likely to 402 participate in the land rental market, It is however in contrast to that obtained by Zhang et al. 403 (2022) who found that aged farmers participate less in renting out land. Our result also shows 404 that farmers who own larger farm areas are more likely to rent out land. One standard deviation 405 increase in owned land area increases the odds of renting out land by 34% relative to the 406 baseline. A similar result was obtained by Vranken and Swinnen (2006) in which they found 407 farming households who own more land to be more likely to rent out land. Farmers that farm 408 in lands categorised as disadvantaged or severely disadvantaged are less likely to rent out land 409 compared to farming on low land relative to the baseline. This may be because most of the 410 farmers that farm on disadvantaged and severely disadvantaged lands are small beef and sheep 411 farmers who undertake farming usually on a part-time bases. For this group, farming is mostly 412 to keep the family enterprise with a greater attachment to the land. As a result, they are less 413 likely to rent out their land. 414 Table 4: Determinants of land market participation (N= 1,121) 415 Rent-out land Rent-in land Variables Coef. Std. Err. % %StdX Coef. Std. Err. % %StdX Environmental- apathy 0.029 0.052 2.9 4.7 -0.035 0.047 -3.5 -5.6 Risk-averse 0.117 0.080 12.4 13.5 0.051 0.068 5.2 5.6 Pro-environment 0.079 0.078 8.2 9.3 -0.128* 0.071 -12.0 -13.5 Progressives -0.104** 0.052 -9.9 -15.9 0.103** 0.049 10.8 18.6 BDG membership 0.187 0.267 20.5 7.0 0.611*** 0.223 84.1 24.6 Off-farm income -0.149 0.195 -13.8 -7.1 0.484*** 0.185 62.3 27.3 Successor -0.466*** 0.175 -37.3 -20.5 -0.095 0.159 -9.1 -4.6 Dairy enterprise -1.376*** 0.534 -74.7 -32.9 1.566*** 0.375 378.8 57.5 20 Beef enterprise -0.989*** 0.231 -62.8 -38.9 0.520** 0.264 68.2 29.6 Sheep enterprise -0.609** 0.259 -45.6 -22.1 0.250 0.296 28.4 10.8 Owned land area (ha) 0.006*** 0.002 0.6 33.5 0.003 0.002 0.3 17.7 Agricultural qualification 0.251 0.210 28.5 12.7 0.021 0.195 2.2 1.0 Lower than 5 GCSEs -0.590* 0.348 -44.6 -15.5 -0.266 0.276 -23.3 -7.3 5 GCSEs or equivalent 0.070 0.273 7.3 2.6 -0.206 0.238 -18.7 -7.2 A level or equivalent 0.758** 0.352 113.5 21.0 0.251 0.345 28.5 6.5 Higher education—diploma or equivalent 0.540* 0.283 71.6 22.7 0.126 0.263 13.4 4.9 Degree level or higher 0.724*** 0.253 106.3 33.6 -0.260 0.258 -22.9 -9.9 Full-time -0.402* 0.206 -33.1 -17.9 0.804*** 0.184 123.4 48.4 Age (greater than 55) 1.142*** 0.225 213.2 70.6 0.003 0.191 0.3 0.1 Disadvantaged -0.656*** 0.189 -48.1 -26.4 -0.191 0.180 -17.4 -8.5 Severely Disadvantaged -0.984*** 0.231 -62.6 -33.8 -0.159 0.194 -14.7 -6.5 Source: own elaboration 416 Note: % is the percent change in odds for unit increase in our explanatory variable; %StdX is the percent 417 change in odds for a standard deviation change in our explanatory variable; single, double, and triple 418 asterisks (*, **, ***) indicate significance at the 10%, 5%, and 1% level, respectively 419 The result for the factors influencing the decision to rent in land showed that farmers 420 with off-farm income have a positive and statistically significant (p < 0.01) effect on the 421 decision to rent in land. This may imply that the farming households with off-farm income 422 have enough money to invest in and expand their farming activities compared to households 423 with no off-farm income. A previous study by Zou et al. (2020) has also found that farmers 424 with part-time, off-farm employment have a greater likelihood of renting in, land. Similar result 425 was obtained by Geoghegan et al. (2021). However, the studies by Vranken and Swinnen 426 (2006) and Kung (2002) found a contrasting result as they showed that greater availability of 427 off-farm income reduces the probability of renting in land. One possible explanation for the 428 contrasting result is that off-farm income, as defined in this study, may include spousal income. 429 If this is the case, the variable may also serve as a proxy for household labour availability, as 430 married farmers may have additional family members contributing to farm work. This 431 increased labour capacity could, in turn, make land expansion through rental agreements more 432 viable. Farmers who are members of the BDGs are also more likely to rent in, land. A previous 433 study by Adenuga, Jack, Ashfield, et al. (2021) has shown that farmers who are members of 434 the BDG operate a more profitable enterprise. Their higher profitability and access to 435 21 information through membership of the BDG may contribute to the decision to rent in, more 436 land and earn more income from farming. Similarly, farmers who undertake farming on a full-437 time basis are more likely to rent in, land compared to the baseline. This may be linked to the 438 assumption that farmers who undertake farming on a full-time basis are more likely to increase 439 their farm size by renting more land. 440 4. CONCLUSIONS 441 In this paper, the effect of motivational and socioeconomic factors on farmers’ 442 participation in the land rental market has been analysed using a multinomial MNL model. The 443 use of the MNL model allows for greater flexibility as it is able to incorporate not just the 444 demand side but also the supply side of participation and non-participation in the land rental 445 market simultaneously. An efficient land rental market is essential to increase the efficiency 446 and sustainability of agricultural production through greater access of young and productive 447 farmers to land. Our results showed that participation in the land rental market is not only 448 influenced by socioeconomic factors but also motivational factors. Most land rentals in NI are 449 still on short-term leases called conacre with a majority of the farmers undertaking farming on 450 owned land. From a policy perspective, it implies that the development of appropriate strategies 451 to encourage land market participation can contribute to the transformation of the rural 452 economy through efficient land use. The enterprise type of the farmer, the identification of a 453 successor and the amount of time devoted to farming are particularly significant factors in the 454 decision to participate in the land rental market. While the result shows that farmers with 455 successors are less likely to rent out land, only 40% of the farmers in our sample already have 456 a successor identified. Policies that support succession planning and the transfer of sustainable 457 practices could help maintain the viability of farms in disadvantage areas while programmes 458 that facilitate the transfer of land from farmers without successors to younger and more 459 productive farmers will strengthen the rental market. Our results showed that older farmers are 460 22 also more likely to rent out their land compared to younger farmers.. The study recommends 461 the development of policies that encourage the younger generation to engage in farming on a 462 full-time basis through schemes that allow for early and comfortable retirement of the older 463 farmers who are happy to make their land available for rent. An example is the NI land mobility 464 scheme which although has now been replaced with a new scheme (farming for future 465 generations) gave young farmers the opportunity to partner with retiree farmers. This allowed 466 young farmers to learn their trade from those hoping to retire and prove themselves with the 467 eventual aim that they will be able to take over and lease the land in a few years. This is essential 468 to promote generational renewal and the modernisation of agriculture for environmental 469 improvements. 470 Another important result of this study is the negative relationship between pro-471 environmental behaviour and renting in, land. While this may be linked to the fact that 472 sustainable agricultural practices often require long-term investment in land, an alternative 473 explanation is that more intensive farming systems, which tend to generate higher profits, 474 create a greater demand for rented land. This is supported by our data in Table 3, which shows 475 that dairy farmers, who typically engage in more intensive production, rent the most land. 476 These factors highlight the complex relationship between environmental practices and land 477 rental decisions. This supports the need for policymakers to develop measures aimed at 478 encouraging the adoption of long-term land leasing which provides the tenure security needed 479 to invest in sustainable practices. This could incentivise farmers with pro-environmental 480 behaviour to rent in, land, improving land quality and protecting the environment. Financial 481 incentives could also be provided to landowners to rent to less intensive tenants. A previous 482 study by Adenuga, Jack, McCarry, et al. (2023) has shown that incentives such as income tax 483 incentives to landlords and tenants for sustainable management of agricultural land could 484 encourage long-term land leasing and increase the likelihood of farmers with pro-485 23 environmental behaviour renting in more land. This will enhance long-term productivity and 486 improve the efficiency of land use by allowing more diverse and sustainable practices. 487 Our finding that progressive farmers are less likely to rent out land and more likely to 488 rent in land also has important policy implications. By developing a policy framework that 489 encourages the renting of more land by progressive farmers, land will be put to more effective 490 and productive use, leading to a more functional land market. This can be achieved if the 491 government enacts policies that promote long-term flexible arrangements and encourage 492 landowners to lease out their land to more productive farmers if they are not being used 493 efficiently. The creation of an enabling environment that allows progressive farmers to rent in 494 more land will contribute to a more dynamic and competitive land market 495 The fact the majority of the land in the region is located in disadvantaged areas has 496 implications for the rental market with most farmland being fragmented. This supports the need 497 to encourage pro-environmental practices that ensure long-term stewardship of the land. There 498 is also the need to provide targeted financial incentives and infrastructure to make the renting 499 of the land more attractive. A previous study by Onofri et al. (2023) also supports the need to 500 provide incentives (tax incentives and subsidies), to drive decision-making of both tenants and 501 landowners. Income tax relief provided for renting out land on long-term lease in the Republic of 502 Ireland has resulted in increased land rental (Bradfield et al., 2023b). Recently, the NI government 503 has embarked on a soil health nutrient scheme (SHNS) aimed at testing all soils in NI between 504 2022 and 2026 to improve sustainability and efficiency in the farming sector. This is a good 505 starting point as it provides up-to-date data on the conditions of the land and how it can be 506 improved. This may encourage a more balanced land rental market in which disadvantage areas 507 are not left behind in terms of productivity gains and land market participation. A limitation of 508 our study is that our methodology does not consider the area of land (as a proportion of total 509 land area) rented in or rented out. This should be considered in future research as it has the 510 24 potential to influence landowners and farmers’ motivations to rent out or rent in land. Future 511 research should also consider possible interaction between the explanatory variables as this 512 could have an effect on the results. 513 REFERENCES 514 Abay, K. A., Chamberlin, J., & Berhane, G. (2021). Are land rental markets responding to rising 515 population pressures and land scarcity in sub-Saharan Africa? Land use policy, 101, 105139. 516 https://doi.org/https://doi.org/10.1016/j.landusepol.2020.105139 517 Adenuga, A. H., Davis, J., Hutchinson, G., Patton, M., & Donnellan, T. (2020). Modelling environmental 518 technical efficiency and phosphorus pollution abatement cost in dairy farms. Science of The 519 Total Environment, 714, 136690. 520 https://doi.org/https://doi.org/10.1016/j.scitotenv.2020.136690 521 Adenuga, A. H., Jack, C., Ashfield, A., & Wallace, M. (2021). Assessing the Impact of Participatory 522 Extension Programme Membership on Farm Business Performance in Northern Ireland. 523 Agriculture, 11(10), 949. https://www.mdpi.com/2077-0472/11/10/949 524 Adenuga, A. H., Jack, C., & McCarry, R. (2021). The Case for Long-Term Land Leasing: A Review of the 525 Empirical Literature. Land, 10(3), 238. https://www.mdpi.com/2073-445X/10/3/238 526 Adenuga, A. H., Jack, C., & McCarry, R. (2023). Investigating the Factors Influencing the Intention to 527 Adopt Long-Term Land Leasing in Northern Ireland. Land, 12(3), 649. 528 https://www.mdpi.com/2073-445X/12/3/649 529 Adenuga, A. H., Jack, C., McCarry, R., & Caskie, P. (2023). Barriers and Enablers of Long-term Land 530 Leasing: a Case Study of Northern Ireland. EuroChoices, 22(2), 20-27. 531 https://doi.org/https://doi.org/10.1111/1746-692X.12404 532 Bizimana, C. (2011). Determinants of land rental markets: Theory and econometric evidence from rural 533 Rwanda. Journal of Development and Agricultural Economics, 3(4), 183-189. 534 Bradfield, T., Butler, R., Dillon, E., Hennessy, T., & Kilgarriff, P. (2021). The Effect of Land Fragmentation 535 on the Technical Inefficiency of Dairy Farms. Journal of Agricultural Economics, 72(2), 486-499. 536 https://doi.org/https://doi.org/10.1111/1477-9552.12413 537 Bradfield, T., Butler, R., Dillon, E. J., & Hennessy, T. (2020). The factors influencing the profitability of 538 leased land on dairy farms in Ireland. Land use policy, 95, 104649. 539 Bradfield, T., Butler, R., Dillon, E. J., Hennessy, T., & Loughrey, J. (2023). Attachment to land and its 540 downfalls: Can policy encourage land mobility? Journal of Rural Studies, 97, 192-201. 541 https://doi.org/https://doi.org/10.1016/j.jrurstud.2022.12.014 542 Bradfield, T., Butler, R., Dillon, E. J., Hennessy, T., & Loughrey, J. (2023). The impact of long-term land 543 leases on farm investment: Evidence from the Irish dairy sector. Land Use Policy, 126, 106553 544 Carpita, M., Sandri, M., Simonetto, A., & Zuccolotto, P. (2014). Chapter 14 - Football Mining with R. In 545 Y. Zhao & Y. Cen (Eds.), Data Mining Applications with R (pp. 397-433). Academic Press. 546 https://doi.org/https://doi.org/10.1016/B978-0-12-411511-8.00015-3 547 Caskie, P., Davis, J., & Wallace, M. (2001). Targeting disadvantage in agriculture. Journal of Rural 548 Studies, 17(4), 471-479. https://doi.org/https://doi.org/10.1016/S0743-0167(01)00016-X 549 Che, Y. (2016). Off-farm employments and land rental behavior: evidence from rural China. China 550 Agricultural Economic Review, 8(1), 37-54. https://doi.org/10.1108/caer-09-2014-0086 551 DAERA. (2023a). Statistical review of Northern Ireland agriculture 2022. Policy, Economics and 552 Statistics Division, Department of Agriculture, Environment and Rural Affairs. 553 https://www.daera-554 ni.gov.uk/sites/default/files/publications/daera/Stats%20Review%20Internal%20for%20202555 2%20ONLINE.pdf 556 https://doi.org/https:/doi.org/10.1016/j.jrurstud.2022.12.014 25 557 DAERA. (2023b). The Agricultural Census in Northern Ireland. Belfast: Department of Agriculture 558 Environment and Rural Affairs Retrieved from https://www.daera-559 ni.gov.uk/sites/default/files/publications/daera/Agricultural%20Census%202023%20Publicat560 ion.pdf 561 Daly, A. (1987). Estimating “tree” logit models. Transportation Research Part B: Methodological, 21(4), 562 251-267. https://doi.org/https://doi.org/10.1016/0191-2615(87)90026-9 563 Dang, H. D., & Pham, T. T. (2022). Predicting Contract Participation in the Mekong Delta, Vietnam: A 564 Comparison Between the Artificial Neural Network and the Multinomial Logit Model. Journal 565 of Agricultural & Food Industrial Organization, 20(2), 135-147. 566 https://doi.org/doi:10.1515/jafio-2020-0023 567 Daniele, B.-C. (2024). The farm succession effect on farmers’ management choices. Land use policy, 568 137, 107014. https://doi.org/10.1016/j.landusepol.2023.107014 569 570 Deininger, K., & Jin, S. (2008). Land sales and rental markets in transition: Evidence from rural Vietnam. 571 Oxford bulletin of Economics and Statistics, 70(1), 67-101. 572 573 Diriye, A. W., Jama, O. M., Diriye, J. W., & Abdi, A. M. (2022). Public preference for sustainable land 574 use policies – Empirical results from multinomial logit model analysis. Land use policy, 114, 575 105975. https://doi.org/https://doi.org/10.1016/j.landusepol.2022.105975 576 Dow, J. K., & Endersby, J. W. (2004). Multinomial probit and multinomial logit: a comparison of choice 577 models for voting research. Electoral studies, 23(1), 107-122. 578 Duressa, D. S. (2021). Multinomial Logistic Regression Analysis of Livelihood Diversification Strategies 579 of Rural Farm Households: A Case of Limmu District, East Wollega Zone of Oromiya Regional 580 State, Ethiopia. 581 Eurostat. (2022). Farm tenure. European Commission. Retrieved 582 from https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Farm_tenure 583 Gensch, D. H., & Recker, W. W. (1979). The Multinomial, Multiattribute Logit Choice Model. Journal of 584 Marketing Research, 16(1), 124-132. https://doi.org/10.1177/002224377901600117 585 586 Geoghegan, C., Kinsella, A., & O’Donoghue, C. (2021). The effect of farmer attitudes on openness to 587 land transactions: evidence for Ireland. Bio-based and Applied Economics, 10(2), 153-168. 588 589 Grant, S. W., Hickey, G. L., & Head, S. J. (2019). Statistical primer: multivariable regression 590 considerations and pitfalls. European Journal of Cardio-Thoracic Surgery, 55(2), 179-185. 591 Harris, L. (2022). Mixed fortune for Wales and NI land despite high demand. Farmers Weekly. 592 https://www.fwi.co.uk/business/markets-and-trends/land-markets/mixed-fortune-for-593 wales-and-ni-land-despite-high-594 demand#:~:text=Usually%2C%20less%20than%201%25%20of,case%20in%202022%2C%20h595 e%20said. 596 Howley, P. (2015). The happy farmer: The effect of nonpecuniary benefits on behavior. American 597 Journal of Agricultural Economics, 97(4), 1072-1086. 598 Howley, P., Buckley, C., Donoghue, C. O., & Ryan, M. (2015). Explaining the economic ‘irrationality’of 599 farmers' land use behaviour: the role of productivist attitudes and non-pecuniary benefits. 600 Ecological Economics, 109, 186-193. 601 Huy, H. T., Lyne, M., Ratna, N., & Nuthall, P. (2016). Drivers of transaction costs affecting participation 602 in the rental market for cropland in Vietnam. Australian Journal of Agricultural and Resource 603 Economics, 60(3), 476-492. https://doi.org/https://doi.org/10.1111/1467-8489.12149 604 Jack, C., Adenuga, A. H., Ashfield, A., & Wallace, M. (2020). Investigating the Drivers of Farmers’ 605 Engagement in a Participatory Extension Programme: The Case of Northern Ireland Business 606 Development Groups. Sustainability, 12(11), 4510. 607 https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Farm_tenure https://doi.org/10.1177/002224377901600117 26 Kung, J. K. S. (2002). Off-farm labor markets and the emergence of land rental markets in rural China 608 [Article; Proceedings Paper]. Journal of Comparative Economics, 30(2), 395-414. 609 https://doi.org/10.1006/jcec.2002.1780 610 Lapple, D., & Kelley, H. (2010). Understanding farmers’ uptake of organic farming: An application of 611 the theory of planned behaviour. 612 Läpple, D., & Kelley, H. (2013). Understanding the uptake of organic farming: Accounting for 613 heterogeneities among Irish farmers. Ecological Economics, 88, 11-19. 614 https://doi.org/https://doi.org/10.1016/j.ecolecon.2012.12.025 615 Li, C., Ma, W., Mishra, A. K., & Gao, L. (2020). Access to credit and farmland rental market participation: 616 Evidence from rural China. China Economic Review, 63, 101523. 617 https://doi.org/https://doi.org/10.1016/j.chieco.2020.101523 618 Little, R., Lyon, J., & Tsouvalis, J. (2023). The co-design of post-Brexit Agri-environmental policy—619 Focusing on environmental land management in England. Dr Ludivine Petetin and Dr Mary 620 Dobbs, Authors of, 54. 621 Long, J. S., & Freese, J. (2005). Regression Models for Categorical Outcomes Using Stata, 2nd Edn 622 College Station. TX: Stata Press.[Google Scholar]. 623 Long, J. S., & Freese, J. (2006). Regression models for categorical dependent variables using Stata (Vol. 624 7). Stata press. 625 Marks-Bielska, R. (2013). Factors shaping the agricultural land market in Poland. Land use policy, 30(1), 626 791-799. 627 Marks-Bielska, R. (2021). Conditions underlying agricultural land lease in Poland, in the context of the 628 agency theory. Land use policy, 102, 105251. 629 https://doi.org/https://doi.org/10.1016/j.landusepol.2020.105251 630 McFadden, D. (1972). Conditional logit analysis of qualitative choice behavior. 631 Mellon-Bedi, S., Descheemaeker, K., Hundie-Kotu, B., Frimpong, S., & Groot, J. C. J. (2020). 632 Motivational factors influencing farming practices in northern Ghana. NJAS - Wageningen 633 Journal of Life Sciences, 92, 100326. 634 https://doi.org/https://doi.org/10.1016/j.njas.2020.100326 635 Min, S., Waibel, H., & Huang, J. (2017). Smallholder participation in the land rental market in a 636 mountainous region of Southern China: Impact of population aging, land tenure security and 637 ethnicity. Land use policy, 68, 625-637. 638 https://doi.org/https://doi.org/10.1016/j.landusepol.2017.08.033 639 O’Kane, H., Ferguson, E., Kaler, J., & Green, L. (2017). Associations between sheep farmer attitudes, 640 beliefs, emotions and personality, and their barriers to uptake of best practice: The example 641 of footrot. Preventive veterinary medicine, 139, 123-133. 642 Onofri, L., Trestini, S., & Mamine, F. (2023). Understanding agricultural land leasing in Ireland: a 643 transaction cost approach. Agric Econ, 11. https://doi.org/10.1186/s40100-023-00254-x 644 Osanya, J., Adam, R. I., Otieno, D. J., Nyikal, R., & Jaleta, M. (2020). An analysis of the respective 645 contributions of husband and wife in farming households in Kenya to decisions regarding the 646 use of income: A multinomial logit approach. Women's Studies International Forum, 83, 647 102419. https://doi.org/https://doi.org/10.1016/j.wsif.2020.102419 648 Otieno, D. J. (2022). Determinants and Effects of Rural Households’ Participation in Land Markets on 649 Agricultural Output and Food Security in Siaya County, Kenya. 650 Ouattara, N. B., Xiong, X., Bakayoko, M., Bi, T. B. A. Y., Sedebo, D. A., & Ballo, Z. (2022). What Influences 651 Rice Farmers’ Choices of Credit Sources in Côte d’Ivoire? An Econometric Analysis using the 652 Multinomial Conditional Logit Model. Progress in Development Studies, 22(2), 149-173. 653 https://doi.org/10.1177/14649934211066453 654 Rahman, S. (2010). Determinants of agricultural land rental market transactions in Bangladesh 655 [Article]. Land use policy, 27(3), 957-964. https://doi.org/10.1016/j.landusepol.2009.12.009 656 https://doi.org/https:/doi.org/10.1016/j.chieco.2020.101523 27 Tesfay, M. (2023). Factors affecting renting in and renting out of land in a semi-arid economy of Tigrai, 657 northern Ethiopia: a generalized random effect order probit model. Cogent Economics & 658 Finance, 11(1), 2132649. https://doi.org/10.1080/23322039.2022.2132649 659 Udimal, T. B., Peng, Z., & Guillaume, N. (2021). Farmland Lease Options in the Rural China: Key 660 Determinants and Policy Implications. Asia-Pacific Journal of Rural Development, 31(2), 218-661 233. https://doi.org/10.1177/10185291211065228 662 Vranken, L., & Swinnen, J. (2006). Land rental markets in transition: Theory and evidence from 663 Hungary. World Development, 34(3), 481-500. 664 Yasmin, I., Akram, W., Adeel, S., & Chandio, A. A. (2022). Non-adoption decision of biogas in rural 665 Pakistan: use of multinomial logit model. Environmental Science and Pollution Research, 666 29(35), 53884-53905. https://doi.org/10.1007/s11356-022-19539-7 667 Zhang, L., Cao, Y., & Bai, Y. (2022). The impact of the land certificated program on the farmland rental 668 market in rural China. Journal of Rural Studies, 93, 165-175. 669 https://doi.org/https://doi.org/10.1016/j.jrurstud.2019.03.007 670 Zhang, L., Feng, S., Heerink, N., Qu, F., & Kuyvenhoven, A. (2018). How do land rental markets affect 671 household income? Evidence from rural Jiangsu, PR China. Land use policy, 74, 151-165. 672 Zou, B., & Luo, B. L. (2018). Why the Uncertain Term Occurs in the Farmland Lease Market: Evidence 673 from Rural China [Article]. Sustainability, 10(8), 15, Article 2813. 674 https://doi.org/10.3390/su10082813 675 Zou, B. L., Mishra, A. K., & Luo, B. L. (2020). Grain subsidy, off-farm labor supply and farmland leasing: 676 Evidence from China. China Economic Review, 62, Article 101293. 677 https://doi.org/10.1016/j.chieco.2019.04.001 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 https://doi.org/10.1016/j.chieco.2019.04.001 28 Appendix 697 A. Principal component analysis 698 This study employed the principal component analysis (PCA) to reduce 14 attitudinal 699 statements to four main motivational constructs which were hypothesized to influence 700 farmers’ participation in the land rental market. The statements and the constructs are 701 included in Table A1. 702 Table A1: Principal components (component loadings) for farming motivations (values > .3 are 703 highlighted in bold) (N= 1,121). 704 Variables PC1 PC2 PC3 PC4 Progressive (α = 0.68) Environmental apathy (α = 0.60) Pro- environment (α = 0.63) Risk Averse (α = 0.66) I am generally keen to adopt new technologies 0.5360 -0.0368 0.0201 -0.2164 I try to find new ways of increasing profit on the farm 0.4888 -0.0139 -0.0174 -0.0547 I find farming rewarding from a quality-of-life perspective 0.3374 -0.1124 -0.0162 0.1251 I think good record keeping is very important in managing a farm business 0.4264 -0.1116 0.0329 0.0461 It is more important to maximize profits than protect the environment -0.0833 0.4775 0.0411 -0.0721 I believe society places too much emphasis on environmental issues -0.0387 0.5274 0.0356 0.0335 I am not that concerned about environmental issues -0.1169 0.4402 -0.0541 -0.0787 I think the media exaggerate the negative impact of agricultural activities on the environment 0.1206 0.3663 0.0557 0.0120 I take some actions to protect the environment when managing my farm because I feel it is the right thing to do 0.0919 0.0829 0.4373 0.0621 | Farmers should receive subsidies for protecting the environment and not for the total amount of land farmed -0.0818 0.0928 0.5357 0.0333 In terms of what I produce on my farm, I think it is important to take the environment into consideration, even if it lowers profit 0.0119 -0.0593 0.4794 -0.0226 I am concerned about the loss of biodiversity in our farmed environment 0.0293 -0.0479 0.5210 -0.0420 I try to avoid taking risky farm business decisions -0.1034 -0.0046 -0.0144 0.6889 I try to keep debt levels as low as possible -0.0568 -0.0381 0.0267 0.6628 Initial eigenvalues 2.92 2.70 1.38 1.17 705 706 707 708 709