Forest Policy and Economics 39 (2014) 13–20

Contents lists available at ScienceDirect

Forest Policy and Economics

j ourna l homepage: www.e lsev ie r .com/ locate / fo rpo l

Factors influencing Irish farmers' afforestation intention

Stefanie Duesberg ⁎, Vincent Upton, Deirdre O'Connor, Áine Ní Dhubháina School of Agriculture, Food Science & Veterinary Medicine, University College Dublin, Ireland

⁎ Corresponding author. Tel.: +353 86 3244856, +353E-mail addresses: stefanie.duesberg@ucd.ie, steffidom

1389-9341/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.forpol.2013.11.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 19 June 2013Received in revised form 13 November 2013Accepted 14 November 2013Available online 4 December 2013

Keywords:AgricultureForestryDecision-makingLogistic regression

The natural conditions in Ireland have a positive influence on tree growth as themean annual increment is twice ashigh as that in mainland Europe. However, due to centuries of resource exploitation and the expansion of agricul-tural land the island has the second lowest forest cover in the EU. An increased forest cover would encourage theestablishment of a range of processing industries and thus support necessary economic development in ruralareas. Furthermore through farm afforestation farmers are given the opportunity to diversify their businesses, asmarket output of the majority of cattle and sheep farms in Ireland often does not cover the production cost. To in-crease forest cover, the Government in 1989 introduced a scheme supporting farm afforestation, which is encour-aged through premium payments that are high enough to make forestry more profitable than the majority ofdrystock farming. Afforestation targets, however, have not beenmet and previous studies have failed to offer a con-sistent explanation for the shortfall in planting rates. Thus, the objective of this work was to identify the factorsinfluencing farmers' afforestation decision. More specifically the study aimed at identifying the combined effectof structural, socio-demographic and attitudinal factors on the probability to plant. Based on previous findingsfrom in-depth interviews with Irish farmers' about their goals and values regarding farming and afforestation, apostal surveywas conducted in Spring 2012 including question on farm structure and socio-demographic variablesas well as questions on reasons for planting/not planting. The data was analysed using logistic regression. The de-veloped logit model showed that while profit goals did not significantly influence the decision-makingwith regardto farm afforestation, structural as well as attitudinal factors played a vital role in this process. Thiswas identified asone reason as to why the current incentive scheme failed to deliver the outlined afforestation targets. Other policytools are needed in addition to the incentives to further encourage afforestation.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

1.1. Policy background

Due to its temperate north-Atlantic climate, the natural conditionsfor tree growth in Ireland are very favourable. The mean annual incre-ment is almost double than the European average (Kearney andO'Connor, 1993). Forest cover however is only about 12% and it is theGovernment's target to increase it to at least 17% by the year 2030(DAFF, 1996). To achieve this target, planting levels of 25,000 ha perannum to the year 2000, and 20,000 ha per annum from 2000 to2030, have been set in the Government's Forestry Strategy ‘Growingfor the future’ (DAFF, 1996). The majority of this afforestation is to beundertaken by private landowners, more specifically farmers. For thispurpose, an afforestation schemewas launched in 1989 and continuallyimproved over the years in order to encourage Irish farmers to afforest(see Fig. 1 for premium and planting rates).

Currently the scheme covers all planting and establishment costsand pays an annual premium for the duration of 20 years to offsetthe loss of income from the time of planting until the first revenues

1 716 7755.bret@gmx.de (S. Duesberg).

ghts reserved.

from timber harvesting. The rationale behind this strategy is two-fold: first, the achievement of the planting targets will lead to a crit-ical mass of timber output that will facilitate the development of arange of processing industries. Second, by offering grants andpremiums to farmers they are encouraged to diversify their busi-nesses and create alternative income streams. Such alternatives arenecessary as most farms in Ireland are not economically viable with-out the EU subsidies. In particular, the market returns from sheepand non-dairy cattle farming do not cover all production costs(Hennessy et al., 2011); these farm types make up 76% of all farmsin Ireland (CSO, 2012). Carbon sequestration as another objectiveof the afforestation scheme has become increasingly important inrecent years in order to meet the Government's internationallyagreed climate change targets.

Initially, the interest in afforestation by farmers was high withplanting rates reaching a peak of 17,000 ha planted in 1995 (ForestService, 2009) (Fig. 1). However, since this time planting rates havebeen consistently and significantly below the target. In the periodfrom 1996 to 2009, only 48% of the targeted area of farmland wasplanted with trees (Forest Service, 2009). Despite continuous improve-ments in funding, planting rates have remained below the target. Thus,the Department of Agriculture, Fisheries and Food stated in its RuralDevelopment Programme for the period from 2007 to 2013 that ‘the

Fig. 1. Private afforestation rates (ha/year) and rate of annually paid farm afforestation premiums (Euros/ha) in Ireland 1990–2012.Source: N.N. (1990); Irish Farmers Association (1991–1996); Irish Timber Growers Association (1997–2010); (Forest Service, 2010, 2012).

14 S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

major difficulty with the [afforestation] programme at the moment is thelow rate of take-up’ (DAFF, 2010).

The first objective of the study was to quantify the importance ofthe previously identified factors influencing Irish farmers' afforesta-tion decision-making for the wider farming community in Irelandand to develop a model that would describe the likelihood that afarmer will afforest based on these factors. The second objectivewas to establish for what proportion of farmers a lack of detailedinformation about the afforestation scheme's benefits is a barrier toplanting and to identify which group of farmers should be addressedwith such information in order to address that potential barrier.Finally, the results will be discussed as to their implications forpolicy-making to further encourage afforestation.

The paper will first review the literature looking at factorsinfluencing farmers' afforestation decision. Second, data collectionand the analytical tools are explained. Third we present the resultsin form of the two logit models developed describing A) the proba-bility of a farmer to afforest and B) the factors influencing a farmerto change mind in favour of planting after being given detailed infor-mation on the scheme. Finally the results are discussed and conclu-sion is drawn with regard to policy recommendations.

1.2. Factors influencing farmers' afforestation decisions

A number of studies have been conducted to explain the shortfall inplanting rates, mainly looking at the influence of economic and socio-demographic factors. Few studies included attitudinal factors such asfarmers' values and their attitudes towards forestry.

The majority of studies tried to explain the shortfall in plantingrates by comparing the economic returns of afforested land tothose of the displaced agricultural use. They were based on the as-sumption that farmers' decisions to afforest are influenced by profitmaximisation goals. The results of these studies were mixed. Forexample Wiemers and Behan (2004) employed a real optionsmodel to calculate forestry returns that would trigger afforestationon various land-use types. According to that study, Irish farmers inthe past made economically optimal decisions with regard to afforesta-tion. However Collier et al. (2002), Behan (2002 cited in Wiemers andBehan, 2004), Duesberg (2008) and more recently Breen et al. (2010)showed that forestry returns would exceed those from drystock beefand sheep farming and that afforestation should have taken place to agreater extent if all farmers were acting as profit maximisers. In 2005,farm afforestation was made even more financially attractive given thatfarmers who planted continued to receive agricultural direct paymentson the afforested land. According to calculations done by Wiemers andBehan (2004) and Bacon (2004), this reform should have had a positiveeffect on farm afforestation. In reality however, planting declined fromaround 10,000 ha in 2005 to 6000 ha in 2008.

Other studies looked at the relationship between farmers' afforesta-tion intentions and farm structure as well as socio-demographicvariables such as farm size, enterprise type, off-farm employment, edu-cation level, age, marital status, successor situation and region (Collieret al., 2002; Farrelly, 2006; Frawley and Leavy, 2001; Hannan andCommins, 1993; Ní Dhubháin and Gardiner, 1994). The only variablethat consistently emerged as having an influence on farm afforestationin Ireland as well as in the UK was farm size: farmers with larger thanaverage farms were more likely to plant (Frawley, 1998; Frawley andLeavy, 2001; Ilbery, 1992; Mather, 1998; Ní Dhubháin and Gardiner,1994; Watkins et al., 1996).

Another research focus to explain Irish farmers' decision-makingwith regard to afforestation has been attitudinal factors or the goalsand values of farmers. Collier et al. (2002) and similarly Frawley andLeavy (2001) found that farmers in general recognize the need for agreater forest cover in Ireland; however they do not want forests ontheir own land or in close proximity. As Fléchard et al. (2006) observed,some rural dwellers associated forestry with bringing isolation and de-population to their areas. This might be due to a lack of integration ofthese plantations into the existing landscape, as Nijnik and Mather(2008) and Nijnik et al. (2008) found in studies on the public prefer-ences regarding woodland development in Scotland that woodlandsare to play an important role in the integration of aesthetic, ecologicaland socio-economic components in landscape management. Nijnikand Mather (2008) furthermore found that the public in Scotlandholds a diversity of views with regard to preferences and normative ex-pectations as to the tree-planting in Scotland. In the authors' previouswork on farm afforestation decision-making, farmers' most importantreasons for not planting or planting were influenced by non-monetaryreasons rather than by profit goals (Duesberg et al., 2013). For that pre-vious research, 62 in-depth interviews with farmers were conducted. Inthese interviews the importance of producing food, land-use flexibilityand the enjoyment of the work tasks related to farming were identifiedas the most prominent reasons for not planting (Duesberg et al., 2013).Similarly McDonagh et al. (2010) discovered that the main barriers toplanting for Irish farmers were the inflexibility resulting from afforesta-tion and their assertion that they needed all their land for agriculture. Anumber of earlier studies similarly found that the majority of farmersonly considered afforesting land that could not be used agriculturallyor that was ‘good for nothing else’ (Collier et al., 2002; Frawley, 1998;Frawley and Leavy, 2001; Hannan and Commins, 1993; Kearney,2001; McCarthy et al., 2003; Ní Dhubháin and Gardiner, 1994; NíDhubháin and Kavanagh, 2003). This finding is underpinned by thefact that private forests in Ireland are mainly growing on land consid-ered marginal for agriculture such as peat (30%), poorly drained gleysoils (30%) or podzols (10%) (Farrelly, 2006). Similar findings weremade in England, Spain, Finland, Scotland and Northern Ireland,where farmers were also more willing to afforest marginal land such

15S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

as fallows, unimproved bog or rough grazing ground (Clark and Johnson,1993; Edwards and Guyer, 1992; Marey-Perez and Rodriguez-Vicente,2009; Selby and Petäjistö, 1995; Watkins et al., 1996). Furthermore, themajority of farmers afforesting in the UK indicated to have multiple rea-sons for afforesting, the most important of which was to enhance thelandscape, while timber production only ranked sixth (Nijnik andMather, 2008).

Few studies have been conducted to explore farmers' attitudinal bar-riers to afforestation of farmland. Burton (1998) studied the influence offarmers' self-identity on their participation in a community woodlandscheme in England. He found that farmers gain little satisfaction fromthe management of woodland and thus are disinclined to establishone. In our own previous research mentioned above, we explored thevalues and goals underlying a farmer's afforestation decision and cameto the conclusion that the majority of farmers make this decisionbased on intrinsic, expressive and social values about farming ratherthan on profit maximisation (Duesberg et al., 2013). According to NíDhubháin and Wall (1999), the negative attitude of Irish farmers to-wards forests arises, in part, from the historical association of treeswith land-owning gentry. Additionally, the extensive area of bogs thatare found in many parts of the country resulted in peat being used asthe primary fuel source rather than wood. This further contributed tothe lack of interest in establishing trees and the development of afarm forestry tradition (Ní Dhubháin and Wall, 1999). Similarly, Nijnikand Mather (2008) contend that there is a lack of forest culture inScotland.

In the context of understanding the decision-making process withrespect to Irish farm afforestation, structural, socio-demographic andattitudinal factors were examined. However, to date, no attempt hasbeen made to combine explanatory factors from different areas todevelop a holistic model explaining farmers' afforestation deci-sions. One sociological theory that attempts to overcome the di-chotomy of sociological research focusing either on actors orstructure, on the macro- or micro-level is Anthony Giddens' theoryof structuration (Giddens, 1984). He argues that the social sciencesshould focus their analysis more on social practices rather than onindividual experience or social structure only. According toGiddens' theory of structuration social practices such as land-useand land-use change are influenced by structure as well as by indi-vidual agents' actions (Giddens, 1984). He defines structure as the‘rules’ (e.g. agricultural policy) and ‘resources’ (e.g. farm structure)being a condition to social practices, but also being the outcome ofagents' actions (‘duality of structure’). Agent factors that influencesocial practices for example are socio-demographics and attitudes.As social practices such as land-use change are influenced by bothstructure and agency factors there is scope to develop a model de-scribing the combined effect of such factors on land-use change ormore specifically on farmers' decision-making to change land-use,e.g. to forestry.

Looking at the more specific literature on the decision-making offarmers, Giddens' theory is paralleled by concepts of Battershill andGilg (1997), Edwards-Jones (2006) and Burton and Wilson (2006).These authors conceptualize farmers' behaviour and decision-makingwith regard to land-use change as being influenced by structural (gov-ernment policies, financial situation, physical geography), socio-demographic (age, family structure, education), and individual farmer(agent) factors such as attitudes, goals and values.

2. Data

2.1. Data collection and survey design

The study sets out to identify the factors influencing a farmer's affor-estation decision. More specifically the study aimed at describing thecombined effect of structural, socio-demographic and attitudinal factorson the probability to plant. For this purpose survey was distributed by

mail in Spring 2012 to a random sample of 4000 farmers in Ireland.The random samplewas drawn froma list of 136,000 Irish farmers in re-ceipt of direct payments, which represents approximately 97% of theIrish farming population. The mailing was administered with the sup-port of the Department of Agriculture, Food and the Marine. Of thetotal number of survey forms administered, 1529 forms were sentback resulting in a relatively high response rate of 38%. Havingdiscarded forms with missing values, a sample of 1077 responses wasused for data analysis. The survey form consisted of four pages compris-ing questions about farm structure and socio-demographic variables, aswell as questions regarding issues such as profit goals and farmingvalues. Including goals and values into the questionnaire facilitatedthe analysis of the importance of structural and socio-demographic aswell as attitudinal factors in a farmer's decision to afforest. The attitudi-nal questions were designed based on the previously conducted 62 in-depth interviews on the goals and values of farmers with regard to af-forestation (Duesberg et al., 2013). In that study, three different profitgoalswere identified among Irish farmers–profitmaximisation, satisfy-ing profit, making no profit/hobby farmers – and a number of intrinsic,expressive and social values that play a role in farmers' decision-makingfor farming in general and with regard to afforestation. The three profitgoals as well as the most important intrinsic, expressive and socialvalues were included in the questionnaire. Participants were asked tochoose from the three profit goals the one they would agree mostwith. Furthermore they were asked how strongly they would agreewith statements representing the following intrinsic, expressive andsocial values using a Likert-type scale:

• Enjoyment of farming activities and lifestyle (LFST)• Importance of food production (FOOD)• Independence (INDI)• Taking on new challenges (CHAL)• Family tradition (TRAD).

The phrasing of the profit goal- and the farming-value-statementswere based on typical representative quotes made by farmers duringthe previously conducted in-depth interviews.

Additionally, to establish whether a lack of detailed informationabout the afforestation scheme is a barrier to further planting, the ques-tionnaire provided participantswho indicated that theywould not plantwith detailed information about the benefits of the scheme. Havingbeen presented with this information, participants were then askedagain if they would be interested in planting to see whether receipt ofthe information had changed their choice.

2.2. Data analysis

The assumption is that farmer decision-making with regard to af-forestation is a ‘social practice’ that is influenced by structural andindividual agents' factors. Thus, the study sets out to examinewhich farm structure, socio-demographic and attitudinal variablesinfluence the probability of Irish farmers considering the afforesta-tion under the State's support scheme. In addition, the characteris-tics of those farmers who changed their mind about planting oncethey were provided with detailed information concerning the affor-estation scheme's benefits were also explored. In both situations,the variable of interest takes a binary form, considering planting ornot, hence logit models were used. Logit models have been widelyused to describe farmers' behaviour, first from the late 1950s inadoption–diffusion research and more recently in research onfarmers' uptake of multifunctional farming or agri-environmentalmeasures (Crabtree et al., 1998; Finger and El Benni, 2013;Jongeneel et al., 2008; Mettepenningen et al., 2013; Poppenborgand Koellner, 2013; Rodrìguez-Vicente and Marey-Pèrez, 2009;Sheikh et al., 2003; Wauters et al., 2010; Yiridoe et al., 2010).

Table 1Overview of participants' characteristics: enterprises.

Percentage of surveyparticipants

Nationalaverage

Cattle (specialist beef) 55% 55%a

Mixed cattle and sheep 11% 11%a

Sheep 11% 10%a

Dairy 11% 11%a

Mixed tillage and grazing livestock 5% 2%a

Tillage 4% 3%a

Mixed field cropsb 7%a

Other 3% 1%a

Total 100% 100%a

Average size of land owned 39 ha 33 haa,c

Average size of land owned— farms with forest 56 ha –

Forest ownership 14% 12%a

Average size of forest 14 ha 10 had

Average age 54 years 55 yearse

a Census of Agriculture 2010 (CSO, 2012).b This category was only surveyed in the Census of Agriculture.c The difference between the sample and the national average farm size is most likely

due to differences in the way this figure is defined by farmers and the Census.d Forest Service (2010b).e National Farm Survey 2007 (Connolly et al., 2008).

16 S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

Under a logit specification the probability of a binary outcome isidentified as:

Pi ¼eβxi

1þ eβxi

where Pi is the probability of outcome i, xi represents the independentvariables or characteristics related to outcome i, including a constant,and β represents the model coefficients. The model can be estimatedusing maximum likelihood estimation. Given the nature of the model,the coefficients are not directly interpretable. Thus, in this study, mar-ginal effects are also reported, which identify the change in the proba-bility of choice at the sample means given a unit increase in thevariable. For dummy variables, the reported marginal effects describethe change in probability due to the inclusion of the variable versus itsomission. Results from the qualitative interviews and statements fromthe survey can be considered as reporting about cause–effect relationsas perceived by the interviewees.

3. Results

Two logit models were created from the collected data. The first de-scribes farmers' probability to afforest depending on a number of

Table 2Summary of variables in the logit model describing farmers' probability to afforest.

Variable Mean St. dev. Unit Missing

Considering 0.097 – Dummy (1 if considering,0 if not)

204

Past planting 0.142 – Dummy (1 if previouslyplanted forestry, 0 if not)

159

Area owned 39 31.838 Hectare 65Full-time farmer 0.576 – Dummy (1 if full-time,

0 if part-time, retired, other)49

Dairy enterprise 0.114 – Dummy (1 if dairy farmer,0 if cattle, sheep. Tillage,mixed enterprise, other)

134

Forest cover ofcounty

0.448 – Dummy (1 if forest coverof county is greater than11%, 0 if less)

0

LFST 4.114 0.739 Likert scale (1–5) 24CHAL 3.824 0.889 Likert scale (1–5) 33TRAD 4.345 0.729 Likert scale (1–5) 23

structural and attitudinal variables. The second describes the character-istics of farmerswho changed theirmind in favour of planting on receiptof detailed information about the afforestation scheme's benefits.Table 1 gives an overview of respondents' characteristics.

3.1. Probability to afforest

For each logit model, a number of independent variables were en-tered into the data analysis. Appendix 1 gives an overview of all vari-ables surveyed. In the first model describing farmers' probability toafforest, eight variables turned out to be significant (Table 3). Table 2gives an overview of the dependent and independent variables in thefinal logit model describing farmers' probability to afforest. Of theeight significant independent variables in themodel, five were of struc-tural and three of attitudinal nature (Tables 2 & 3).

3.1.1. Structural variables

3.1.1.1. Past afforestation and farm size. The variable ‘Past planting’was positively correlated to respondents' intention to plant. Farmerswho already had planted some forest in the past were 12%more like-ly to plant in the future than those who hadn't (Table 3). Farm sizewas another significant structural variable in the logit model toexplain farmers' probability to afforest (Table 3). Farmers with largerfarmsweremore likely to afforest. Additionally the average farm sizeof those who had planted in the past was with 56 ha above thenational average of 33 ha (CSO, 2012). This confirms findings of pre-vious studies that had already shown the dominance of relativelylarger farms among those where afforestation takes place (Frawley,1998; Frawley and Leavy, 2001; Ilbery and Kidd, 1992; Mather,1998; Ní Dhubháin and Gardiner, 1994).

3.1.1.2. Occupation and enterprises. Of the occupation variables enteredinto the logit analysis, only full-time farming was shown to be correlat-ed to the afforestation decision: full-time farmers were less likely to de-cide in favour of afforestation (Table 3). Farming enterprises typicallyoperated in full-time are dairy and tillage or mixed tillage farms. Fromall the enterprise variables entered into the analysis, only dairy farmingturned out to be a variable of significance in the model. Dairy farmerswere less likely to join the afforestation scheme and plant trees(Table 3).

3.1.1.3. Average forest cover. Farmers living in counties with above-average forest cover were more likely to consider afforesting their land(Table 3). The average forest cover per county ranges from 22% in countyWicklow to 3% in county Meath. While county Wicklow is characterizedby hilly terrain, which limits agricultural land-use, county Meath is amore or less flat midland county with fertile soils suitable for a widerange of agricultural land-uses. Forest cover is likely to reflect local soil

Table 3Logit model on factors influencing Irish farmers' probability to consider afforestation.

Variable Coefficient S.E. Sig. P-value Marginal effect

Past planting 1.210 0.245 *** 0.000 0.119CHAL 0.439 0.156 *** 0.005 0.029Area owned (ha) 0.008 0.003 ** 0.014 0.001TRAD −0.383 0.159 ** 0.016 −0.026Full-time farmer −0.554 0.239 ** 0.021 −0.039Dairy enterprise −0.946 0.490 * 0.054 −0.048County with above-averageforest cover

0.413 0.221 * 0.061 0.029

LFST −0.282 0.161 * 0.080 −0.019Constant −1.602 0.749 ** 0.032 −0.108Number of observations 1077Log likelihood −299.987

Significance level: * 10%, **5%, and ***1%.

Table 4Variables of the logit-model explaining farmers changing their mind in favour of planting.

Variable Mean St. dev. Unit Missing

Changed mind 0.158 – Dummy (1 if considering, 0 if not) 133Dairy enterprise 0.114 – Dummy (1 if dairy farmer, 0 if cattle, sheep. Tillage, mixed enterprise, other) 134Forest cover of county 0.448 – Dummy (1 if forest cover of county is greater than 11%, 0 if less) 0Knowledge of scheme 5.276 2.669 Score (0–8) 213Married with children 0.64 – Dummy (1 if married with children, 0 if married without children, living with partner, single, widowed, divorced/separated) 95Aged 45 to 64 0.523 – Dummy (1 if aged between 45 and 64, 0 if less than 45 or more than 64) 42Past planted 0.142 – Dummy (1 if previously planted forestry, 0 if not) 159

17S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

types and climate and, consequently, the range and profitability of poten-tial land-uses. Thus the fact that farmers living in counties with above-average forest cover are more likely to plant is probably correlated tothese geographic parameters.

3.1.1.4. Attitudinal variables. The survey included two questionsconcerning attitudinal variables: Profit goals and general farmingvalues. Respondents were asked which of the three profit goals theywere presented with (maximum/satisfying/none) they would agreemost with. None of these profit goals were a variable of significance inthe logit model — the likelihood of planting did not significantly in-crease or decrease depending on the profit goals. However, three ofthe five non-monetary farming value variables entered into the analysisturned out to have a significant influence on farmers' afforestation deci-sion (Table 3).

The non-monetary farming value variable with the highest signifi-cance was the one representing the expressive value of taking on newchallenges (CHAL) (Table 3). In the questionnaire this optionwas repre-sented by the following statement: “I like taking on new challenges and Ihave a lot of ambition for my farm and many plans about how I want tomanage it in the future”. Farmers who agreed with this statement weremore likely to afforest. From the in-depth interviews we know thatfarmers who are inclined to taking on new challenges were also morewilling to take risks and in general exhibited a more business-oriented, entrepreneurial thinking (Duesberg et al., 2013).

The two other attitudinal variables, which were significant in themodel, were the ‘Tradition’ (TRAD) and the ‘Lifestyle’ (LFST) variables.Both were negatively correlated to the intention to afforest. The ‘Tradi-tion’-variable was represented in the questionnaire by the followingstatement: “I regard the farm as a family asset that I'm keeping in a goodcondition to pass on to my successors one day.” Farmers who agreedwith this statement were less likely to afforest. The ‘Lifestyle’-optionwas represented in the questionnaire by the following statement: “Ienjoy the activities, work tasks and lifestyle related to farming”. Thosefarmers did not want to see the farm business replaced by a forest be-cause it would deprive them of an important source of satisfaction intheir life. We also know from our previous study that for farmers whodo not plant for lifestyle reasonsmaking a profit from farming in generalwas less important.

Table 5Logit-model on factors influencing Irish farmers changing their mind in favour of planting.

Variable Coefficient

Past planting 1.238Knowledge of scheme (0–8) −0.108Aged 45–64 (reference is aged b 45 and aged N 65) 0.466Dairy enterprise −0.711Married with children 0.430County with above-average forest cover −0.320Constant −1.434N = 943Log likelihood −423.864

Significance level: * 10%, **5%, and ***1%.

3.2. Intention to plant after provision of detailed information

The second logit model developed from the data concerned farmerswho changed their mind in favour of planting on receipt of more de-tailed information about the afforestation scheme's benefits. Over 87%of the respondents in general were aware of the availability of thescheme and this was not influenced by farmer characteristics. Respon-dents who had no intention of plantingwere provided with detailed in-formation concerning the benefits of the afforestation scheme andwerethen asked again whether they would consider planting. In total, thenumber of those interested in planting rose from 10% to 26%. Thosewho changed their mind in favour of planting were analysed againusing a logit model. Table 4 gives an overview of the dependent and in-dependent variables in that logit model. The analysis showed that thosewho had planted in the past, were aged between 45 and 64 and weremarried with children were more likely to change their mind(Table 5). Dairy farmers and farmers living in counties with above-average forest cover were less likely to change their mind after beinggiven more information (Table 5). Also the more respondents alreadyknew about the scheme the less likely they were to change their mind.

4. Discussion and conclusions

The study sets out tomodel the probability that a farmerwill afforestbased on structural, socio-demographic and attitudinal variables. Thesecond objective was to establish whether addressing a lack of detailedinformation about the afforestation scheme's benefits would get morefarmers interested in planting and, if so, who those farmers were. Thechosen methodological approach proved useful as it allowed a moregeneral assessment of the afforestation scheme than for example a strictapplication of the theory of planned behaviour (TPB). The TPB has beenwidely used in researching farmer behaviour; however it has beencriticised for not being capable to produce a broad enough picture offarmer motivation (Burton, 2004).

4.1. Farmers considering afforestation

As to the first objective, the data analysis showed that five struc-tural and three attitudinal variables have a high probability to affectfarmers' decision-making with regard to afforestation. This proves

S.E. Sig. P-value Marginal effect

0.231 *** 0.000 0.2270.038 *** 0.004 −0.0150.184 ** 0.011 0.0640.313 ** 0.023 −0.0830.197 ** 0.029 0.0570.179 * 0.074 −0.0440.269 *** 0.000 −0.200

1 The average farm size in Ireland ranges from 22 ha in CountyMayo to 44 ha in CountyKildare.

18 S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

the importance of individual farmer factors, such as farming values,in this specific decision-making situation. Farmers who liked takingon new challenges were more likely to plant, while farmers forwhom farming lifestyle and family tradition were important wereless likely to consider afforestation. To encourage more farmers toplant, those values need to be taken into account in policy develop-ment. For example, to get ‘lifestyle farmers’ interested in plantingthey would need to be shown how farmers can get involved in in-teresting work tasks around establishing and managing a forest.Addressing those farmers for whom family tradition is importantcould focus on the future value of a forest for their successors.From the results, we also know that profit goals did not have a sig-nificant influence on the decision to afforest, demonstrating that itis not primarily related to considerations about the comparativereturns from farming and forestry.

There were five structural variables that turned out to play a sig-nificant role in the afforestation decision. Past planting, local forestcover and farm size had a positive effect; while dairying and fulltimefarming had a negative effect on the probability to afforest. SimilarlyIlbery and Kidd (1992) and Crabtree et al. (1998) in studies conduct-ed in the UK concluded that farmers who have planted in the pastwere more likely to join an afforestation scheme. Farmers who hadplanted in the past not only were positively inclined to consider af-forestation again, but they were also more likely to change theirmind in favour of planting (again) after being givenmore detailed in-formation about the scheme. This indicates that the experience frompast afforestation has been positive. This group could be easily iden-tified and addressed through a simple information campaign in orderto increase afforestation rates. Another advantage of encouragingpast planters to afforest more land would be that larger forestsmight be created when planting fields adjacent to the previouslyplanted areas. Our own previous research, as well as other studies,had shown that farmers would only afforest ‘bad land’ (Collieret al., 2002; Frawley, 1998; Frawley and Leavy, 2001; Hannan andCommins, 1993; Kearney, 2001; McCarthy et al., 2003; Ní Dhubháinand Gardiner, 1994; Ní Dhubháin and Kavanagh, 2003). Furtherresearch could reveal whether past planters intend to afforestremaining patches of ‘bad land’ or, if due to a positive afforestationexperience, they would consider planting even better quality land,which would indicate an improvement in the attitude towardsforestry as a farm enterprise.

The positive experience from planting could be passed on byword-of-mouth to neighbouring farmers, which might explainwhy farmers living in counties with above-average forest coverwere more likely to afforest. Another reason for this phenomenoncould be that farmers living in counties with high forest cover ingeneral have a more positive attitude towards forestry (Frawleyand Leavy, 2001).

From a rural development perspective, one of the afforestationscheme's objective is to offer income support to those farmers whostruggle to make a living from farming, which typically are small-scaledrystock farmers (Hennessy et al., 2011). The study showed thatdrystock farmers are neither significantly inclined nor disinclined toplanting. However, targeting small farms could be difficult, as the resultsshowed that larger farms were more likely to be planted. A scheme ini-tiating and supporting group plantings of small neighbouring fieldscould enable small-scale (or below average farm size) farmers toplant. This would also have the advantage of increasing the averagefarm forest's size, improving their value for forestry, nature conserva-tion and recreation aswell as the bargaining power of the forest ownersonce it comes to thinning and harvesting operations. In the Netherlandsenvironmental cooperatives proved successful in motivating farmers tojoin agri-environmental and rural development activities (Renting andVan Der Ploeg, 2001).

Past studies in Ireland and the UK had already shown the dominanceof relatively larger farms among those where afforestation takes place

(Frawley, 1998; Frawley and Leavy, 2001; Ilbery and Kidd, 1992;Mather, 1998; Ní Dhubháin and Gardiner, 1994). There is however aninteresting difference between Irish and UK farmers as to the farmsize they consider big enough for planting. While Irish farmers inthis study on average planted forests if their farm size was at least56 ha, farmers in a study undertaken in West-Nottinghamshireconsidered planting from a farm size of at least 100 ha (Watkinset al., 1996). The overall average farm size in that area howeverwas with 197 ha much bigger when compared to the overall Irishaverage farm size of 33 ha (CSO, 2012; Watkins et al., 1996). Itseems that the farm sizes deemed big enough for planting are re-gionally flexible. An average farm size could be assessed as ‘bigenough’ for planting if it is above the local average. As there is con-siderable difference in farm sizes within Ireland, a farm size ‘bigenough’ for planting could change between counties.1 The factthat there is regional flexibility in the farm sizes deemed bigenough for planting raises the question if there also is a temporaland sectoral flexibility. Average farm sizes have continually in-creased in the past; in Ireland average farm size grew from 22 hain the 1980s (the decade where the first afforestation programmeswere launched) to 33 ha in 2010. Thus the average farm size reck-oned big enough for planting could have risen over the years, too.Also farm sizes differ between enterprises with tillage farms aver-aging 56 ha and specialist beef farms averaging 28 ha (CSO,2012); thus farm sizes big enough for planting could also be varyingbetween enterprise types. As the farm size plays a pivotal role in thedecision-making with regard to farm afforestation the regional,temporal and sectoral flexibilities in average farm sizes mighthave to be considered when developing strategies to encouragemore farmers to plant. Further research would be needed to con-firm these conclusions.

Another result of the study was that full-time and dairy farmerswere less inclined towards planting. The latter is noteworthy as theaverage farm size of Irish dairy farms is with 55 ha above the overallaverage of 33 ha (CSO, 2012) and farm size had a significantly posi-tive influence on the probability to afforest. One reason for this effectcould be the comparatively high profitability of dairy farms. Dairyfarming in the past has been the most profitable farm enterprise inIreland (Hennessy et al., 2011). It is a highly specialised businessthat needs a high level of investment in machinery and technicalequipment, which is typically financed by loans (Hennessy et al.,2011). Such sunk costs determine the course of the farm businessfor many years into the future, also termed as ‘path dependency’ byeconomists. Another explanation for dairy farmers being less likelyto join the afforestation scheme could be that they typically operateon fertile or ‘good’ agricultural land (see above). As our previous re-search and other studies have shown, farmers in general are reluc-tant to plant such land. While dairy farmers might be less likely toplant, it is questionable whether such a group should be targetedwhen designing policy tools to encourage farm afforestation andwhether it makes sense from a rural development perspective tooffer alternative income streams to viable farm business such asdairy farms.

On the other hand, tillage farmers were not significantly disin-clined to plant, despite them also typically, being viable businessesand operating on fertile land (CSO, 2012; Hennessy et al., 2011).One reason might be that fewer tillage farms run their businesseswith loans compared to dairy farming (Hennessy et al., 2011). Alsoprofit margins on tillage farms have been decreasing in the pastdue to the continuous increase in fertilizer and fuel prices. Anotherpossible explanation could be a number of unusually wet summersand cold winters in Ireland. From personal communication with for-esters in Ireland, we know that the interest in planting by tillage

Characteristic Attribute

Type of farm DairyDry stock (beef cattle and sheep)Tillage and mixed tillage and dry stockOther

Farm size HectaresForest size (ifapplicable)

Hectares

Occupation Full-time farmersPart-time farmers (part-time off-farm job)Full-time off-farm jobRetiredOther

Education PrimarySecondaryAgricultural trainingTrade based qualificationThird level or above

Awareness aboutscheme

YesNo

Detailed informationabout scheme

Did you know (y/n) that the scheme …

Covers all planting cost?Pays an annual premium?The annual premium is paid for 20 years?The minimum annual premium is 369 Euros?Pays premiums that are tax-free?Allow farmers to keep the Single Farmpayment on the landplanted?Offers additional compensatory payments to REPS farmersof 200 Euros per hectare?Farmers retain ownership of the land after planting?

Age groups b44 years45–64 yearsN65 years

Gender FemaleMale

Marital status MarriedLiving with partnerSingleWidow/-erDivorced/separated

Family stage No childrenChildren b 18 yearsChildren N 18 years

19S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

farmers rises after extreme weather situations. This is confirmed byfindings of Sutherland et al. (2012) according to which farmers aremore likely tomakemajor changes in farmmanagement after triggerevents. While other farm enterprises, too, suffer from bad weatherthe effect can be more devastating to tillage farmers, as crops canbe irreplaceably destroyed by a single extreme weather event. An-other reason for tillage farmers being less opposed to forestrymight be that growing trees is closer to their understanding of an ag-ricultural product than is the case for dairy farmers. As the averagesize of a tillage farm is 56 ha, which is significantly greater than thenational average of 33 ha (CSO, 2012), and larger farms are morelikely to be planted, targeting tillage farmers with afforestation cam-paigns could prove successful, especially after trigger events. Tillagefarms typically operate on fertile soils, which would make them par-ticularly interesting as sites for establishing forests of high naturevalue. Ireland offers a specific support scheme to create such forests.This scheme should be promoted when encouraging tillage farmersto afforest. As concluded above, however, the farm size big enoughfor planting could be flexible between enterprises. Thus a tillagefarmer could consider 56 ha as being not big enough for planting.

4.2. Farmers changing their mind

The second objective of the study was to establish whether a lack ofdetailed information about the afforestation scheme's benefits was abarrier to more planting, and if so, to whom specifically. In total, 16%of farmers changed their mind in favour of planting following the provi-sion of such information as part of the survey. Encouraging 16% offarmers to join the afforestation scheme and plant could significantly in-crease the number of hectares planted in Ireland. Furthermore, such anencouragement could be achievedwith comparatively simple tools suchas mailings. Again, past planters were more likely to change their mind,whichwas somewhat surprising as onemight assume that past plantersalready knew about the scheme's details. However, new benefits wereadded to the scheme over the years and past planting might havebeen undertaken some time ago. This again leads to the conclusionthat past plantersmight be easily convinced to plant somemore forestrythrough a simple information campaign specifically addressed to them.Employing such information campaigns after trigger events negativelyaffecting the course of the farm business could improve their efficiency(Sutherland et al., 2012).

In addition, married farmers aged between 45 and 64 with chil-dren were more likely to change their mind, which was interesting,as in the first logit model regarding intention to plant, no socio-demographic variable emerged as significant. From our previousstudy, we know that the most important reason for planting, forthose who already had planted, was generating an asset for theirsuccessors. Providing information about the benefits of the affores-tation scheme might have demonstrated to farmers with childrenthe value of a forest to them and their successors. Dairy farmerswere less likely to change their mind in favour of planting,confirming the first logit model's results, which showed that theyin general are less likely to plant. Farmers living in counties withabove-average forest cover were less likely to change their mind,which is different from the first logit model in which they weremore inclined to plant. One explanation could be that informationabout afforestation might be more easily accessible in thosecounties, for example through neighbours who have planted. Also,having seen neighbours planting, farmers might have seriouslyconsidered afforesting their own land and have already exploredthe conditions. Thus, farmers in counties with above-average forestcover might have rejected the afforestation option based on an in-formed decision. The detailed information presented was not newto them and thus did not change their mind. The same explanationmight apply to the fact that the more farmers knew about thescheme the less likely they were to plant.

To recommend more specific policy actions based on the study'sfindings (e.g. who exactly to address and how) the afforestationpolicy would need to specify more detailed goals. The Irish state'safforestation policy neither indicates regional focuses for furtherafforestation nor does it outline which farmers specifically shouldbe encouraged to plant. From a rural development, but also from aforestry perspective, it would be necessary to outline regions andfarm enterprises that future afforestation policies should focus on.Such planning could ensure that resources are concentrated onareas where the natural conditions would be most suitable forforestry and where local economies would benefit most from astrong forest sector.

Acknowledgements

This research was funded by COFORD, Department of Agriculture,Food and the Marine under the National Development Plan. Theauthors would also like to thank the Department of Agriculture,Food and the Marine for facilitating participant sampling and theadministration of the survey mailing.

Furthermore, we would like to thank the two anonymous reviewersfor their valuable comments.

Appendix 1. Overview of all variables entered into the modellingprocess

20 S. Duesberg et al. / Forest Policy and Economics 39 (2014) 13–20

References

Bacon, P., 2004. A Review and Appraisal of Ireland's Forestry Development Strategy. FinalReport, Killinick, Ireland.

Battershill, M.R.J., Gilg, A.W., 1997. Socio-economic constraints and environmentallyfriendly farming in the Southwest of England. J. Rural. Stud. 13, 213–228.

Behan, J., 2002. Returns from farm forestry vs other farm enterprises. IFA Farm ForestryConference, Limerick.

Breen, J., Clancy, D., Ryan, M., Wallace, M., 2010. Can't See the Wood for the Trees: TheReturns to Farm Forestry in Ireland. Working Paper Series.The Rural Economy Re-search Center, Teagasc, Athenry.

Burton, R.J.F., 1998. The Role of Farmer Self-identity in Agricultural DecisionMaking in theMarston Vale Community Forest. De Montfort University, Leicester 301.

Burton, R.J.F., 2004. Reconceptualising the ‘behavioural approach’ in agricultural studies: asocio-psychological perspective. J. Rural. Stud. 20, 359–371.

Burton, R.J.F., Wilson, G.A., 2006. Injecting social psychology theory into conceptualisationsof agricultural agency: towards a post-productivist farmer self-identity? J. Rural. Stud.22, 95–115.

Clark, G.M., Johnson, J.A., 1993. Farm woodlands in the Central Belt of Scotland: a socio-economic critique. Scott. For. 47, 15–24.

Collier, P., Dorgan, J., Bell, P., 2002. Factors Influencing Farmer Participation in Forestry.COFORD, Dublin.

Connolly, L., Kinsella, A., Quinlan, G., Moran, B., 2008. National Farm Survey Report Report2007. Teagasc, Athenry.

Crabtree, B., Chalmers, N., Barron, N.-J., 1998. Information for policy design: modellingparticipation in a farm woodland incentive scheme. J. Agric. Econ. 49, 306–320.

CSO, 2012. Census of Agriculture 2010. Central Statistics Office, Dublin.DAFF, 1996. Growing for the future. In: Department of Agriculture Fisheries and Food

(Ed.), A Strategic Plan for the Development of the Forestry Sector in Ireland (Dublin).DAFF, 2010. Rural Development Programme Ireland 2007–2013. In: Department of

Agriculture Fisheries and Food (Ed.), Rural Development Programme Ireland 2007–2013 (Dublin).

Duesberg, S., 2008. Trees, beef or sheep? A comparison of the net present values of bareland of forestry and two agricultural land-use systems in Ireland. Unpublished report.University College Dublin.

Duesberg, S., O'Connor, D., Ní Dhubháin, Á., 2013. To plant or not to plant—Irish farmers'goals and values with regard to afforestation. Land Use Policy 32, 155–164.

Edwards, C., Guyer, C., 1992. Farmwoodland policy: an assessment of the response to thefarm woodland scheme in Northern Ireland. J. Environ. Manag. 34, 197–209.

Edwards-Jones, G., 2006. Modelling farmer decision-making: concepts, progress andchallenges. Anim. Sci. 82, 783–790.

Farrelly, N., 2006. A Review of Afforestation and Potential Volume Output from PrivateForests in Ireland. Teagasc, Athenry.

Finger, R., El Benni, N., 2013. Farmers' adoption of extensive wheat production —

determinants and implications. Land Use Policy 30, 206–213.Fléchard, M.-C., Ní Dhubháin, A.i, Carroll, M., Cohn, P., 2006. Forestry and the Local Com-

munity in Ireland: A Case Study in the Arigna Region. Small-scale Forestry and RuralDevelopment: The Intersection of Ecosystems, Economics and Society, Galway,Ireland.

Forest Service, 2009. Afforestation Statistics 2009. Government of Ireland, Johnstown.Forest Service, 2010. Afforestation Statistics 2010. Government of Ireland, Johnstown.Forest Service, 2012. Afforestation Statistics 2012. Government of Ireland, Johnstown.Frawley, J.P., 1998. Farmers' attitudes towards forestry as a farm enterprise in Ireland. In:

Wiersum, F.K. (Ed.), Forestry in the Context of Rural Development, Proceedings of theFinal Scientific Conference COST Action E3. Vila Real, Portugal.

Frawley, J.P., Leavy, A., 2001. Farm Forestry: Land Availability, Take-up Rates and Econom-ics, Rural Economy Research Series. The Rural Economy Research Centre, Teagasc.

Giddens, A., 1984. The Constitution of Society. Polity Press, Cambridge.Hannan, D.E., Commins, P., 1993. Factors Affecting Land Availability for Forestry. Econom-

ic and Social Research Institute, Dublin.Hennessy, T., Kinsella, A., Moran, B., Quinlan, G., 2011. National Farm Survey 2011.

Teagasc, Athenry.

Ilbery, B., 1992. State-assisted farm diversification in the United Kingdom. In: Bowler, I.R.,Bryant, C.D., Nellis, M.D. (Eds.), Contemporary Rural Systems in Transition. CAB Inter-national, Wallingford, pp. 100–118.

Ilbery, B., Kidd, J., 1992. Adoption of the farm woodland scheme in England. Geography77, 363–377.

Irish Farmers Association, 1991–1996. Irish Farmers Handbook (Dublin).Irish Timber Growers Association, 1997–2010. Forestry and Timber Yearbook (Dublin).Jongeneel, R.A., Polman, N.B.P., Slangen, L.H.G., 2008. Why are Dutch farmers going mul-

tifunctional? Land Use Policy 25, 81–94.Kearney, B., 2001. A Review of Relevant Studies Concerning Farm Forestry Trends and

Farmers Attitudes to Forestry. COFORD, Dublin.Kearney, B., O'Connor, R., 1993. Economic issues in Irish forestry. J. Stat. Soc. Inq. Soc. Irel.

XXVI (Pt.5), 179–209.Marey-Perez, M.F., Rodriguez-Vicente, V., 2009. Forest transition in Northern Spain: local re-

sponses on large-scale programmes of field-afforestation. Land Use Policy 26, 139–156.Mather, A., 1998. The changing role of forests. In: Ilbery, B. (Ed.), The Geography of Rural

Change. Longman, Harlow.McCarthy, S., Matthews, A., Riordan, B., 2003. Economic determinants of private afforesta-

tion in the Republic of Ireland. Land Use Policy 20, 51–59.McDonagh, J., Farrell, M., Mahon, M., Ryan, M., 2010. New opportunities and cautionary

steps? Farmers, forestry and rural development in Ireland. Eur. Countryside 2,236–251.

Mettepenningen, E., Vandermeulen, V., Delaet, K., Van Huylenbroeck, G., Wailes, E.J., 2013.Investigating the influence of the institutional organisation of agri-environmentalschemes on scheme adoption. Land Use Policy 33, 20–30.

N.N., 1990. Farm Afforestation Grants and Premiums, Irish Farmers' Journal, Dublin.Ní Dhubháin, Á., Gardiner, J.J., 1994. Farmers' attitudes to forestry. Ir. For. 51, 21–26.Ní Dhubháin, Á., Kavanagh, T., 2003. Joint ventures in private forestry in Ireland. Small

Scale For. 2, 9–19.Ní Dhubháin, Á., Wall, S., 1999. The new owners of small private forests in Ireland. J. For.

28–33.Nijnik, M., Mather, A., 2008. Analyzing public preferences concerning woodland develop-

ment in rural landscapes in Scotland. Landsc. Urban Plan. 86, 267–275.Nijnik, M., Zahvoyska, L., Nijnik, A., Ode, A., 2008. Public evaluation of landscape content

and change: several examples from Europe. Land Use Policy 26, 77–86.Poppenborg, P., Koellner, T., 2013. Do attitudes toward ecosystem services determine

agricultural land use practices? An analysis of farmers' decision-making in a SouthKorean watershed. Land Use Policy 31, 422–429.

Renting, H., Van Der Ploeg, J.D., 2001. Reconnecting nature, farming and society: environ-mental cooperatives in the Netherlands as institutional arrangements for creating co-herence. J. Environ. Policy Plan. 3, 85–101.

Rodrìguez-Vicente, V., Marey-Pèrez, M.F., 2009. Land-use and land-base patterns in non-industrial private forests: factors affecting forest management in Northern Spain. For.Policy Econ. 11, 475–490.

Selby, A., Petäjistö, L., 1995. Attitudinal aspects of the resistance to field afforestation inFinland. Sociol. Rural. 35, 67–92.

Sheikh, A.D., Rehman, T., Yates, C.M., 2003. Logit models for identifying the factors that in-fluence the uptake of new tillage-technologies by farmers in the rice–wheat and thecotton–wheat farming systems of Pakistan's Punjab. Agric. Syst. 75, 79–95.

Sutherland, L.-A., Burton, R.J.F., Ingram, J., Blackstock, K., Slee, B., Gotts, N., 2012. Triggeringchange: towards a conceptualisation of major change processes in farm decision-making. J. Environ. Manag. 104, 142–151.

Watkins, C., Williams, D., Lloyd, T., 1996. Constraints on farm woodland planting in En-gland: a study of Nottinghamshire farmers. Forestry 69, 167–176.

Wauters, E., Bielders, C., Poesen, J., Govers, G., Mathijs, E., 2010. Adoption of soil conserva-tion practices in Belgium: an examination of the theory of planned behaviour in theagri-environmental domain. Land Use Policy 27, 86–94.

Wiemers, E., Behan, J., 2004. Farm forestry investment in Ireland under uncertainty. Econ.Soc. Rev. 35, 305–320.

Yiridoe, E.K., Atari, D.O.A., Gordon, R., Smale, S., 2010. Factors influencing participa-tion in the Nova Scotia Environmental Farm Plan Program. Land Use Policy 27,1097–1106.