A neural network model of Irish farmers’
perceptions of land mobility
Marija Banovic1, Alan Renwick1, Mark T. Keane2&
Pat Bogue3
1
School of Agriculture & Food Science, University
College Dublin, Belfield, Dublin 4, Ireland
2
School of Computer Science & Informatics,
University College Dublin, Belfield, Dublin 4, Ireland
3
Broadmore Research, Ennis, Ireland
Introduction
Land has always been one of the most important and
controversial assets in Ireland and land mobility
continues to be a critical issue to the future success of
the Irish agricultural sector (FH2020). The Irish
agricultural sector is still portrayed by a low level of
land mobility and late transfer pattern with small farms
and an older farming population. Policies and schemes
applied to the agricultural sector to improve land
mobility situation appear to be failing to have the
desired effect. The overall objective of this study is to
assess the present situation and identify potential
solutions that could improve land mobility and smooth
land transfer in the Irish agricultural sector as perceived
by the Irish farmer.
Materials, Methods & Model
The data used in this study comes from a Macra na
Feírme survey conducted in 2012 on a random sample
of 421 Irish farmers aged over 50 years to determine
the future plans related to land transfer, farming and
land ownership (Bogue, 2012). A subset of 201 farmers
was used to better understand potential measures for
land transfer and land mobility. Collected data was
analysed by using the Interactive Activation and
Competition (IAC) neural network
analysis
(McClelland, 2014; McClelland & Rumelhart, 1988).
IAC models are useful for showing the supporting and
competing constraints between different factors in a
problem domain and can reveal generalisations over
data sets describing individuals. Here the model was
applied to farmers’ perceptions, their properties (e.g.,
location, age, sex), and land transfer/mobility
measures. An IAC network consists of a collection of
nodes representing features of interest (e.g., farmer age,
sex, location, farmer perception that a mobility will not
reduce tax) and excitatory links between these nodes
indicating that these features are related in a particular
case (i.e., the properties/responses of a particular
farmer in the survey). Nodes are also organized into
pools, indicating that these feature-nodes are mutually
exclusive with inhibitory links between them (e.g., the
sex pool has two mutually exclusive nodes, male vs
female) As such, the network as a whole represents the
properties and response choices of the farmers in the
study and the co-dependencies and constraints between
these factors. If one sets the activation of one featurenode (e.g., a land transfer option) to a high level (i.e.,
clamp it) and propagates activation through the
network, a generalisation of the overall dependencies
and constraints linked to this feature can be found by
reading the activation levels in other nodes, when the
network settles.
Results
From running the model, the results show that land
transfer is mainly related to the farmer’s personal and
family characteristics with a complex interaction of
factors affecting the land transfer decision. The
traditional and intense relationship with the male heir
stands out as a key factor. The model also shows that,
on average, land mobility and transfer measures are
connected to the farmer’s internal processes; his
perceptions of different land mobility solutions, as well
as personal characteristics (age group) and economic
resources (size of the farm, enterprise type). Figure 1
shows a network of land mobility solutions associated
to early retirement and young farmer’s incentives as
perceived by farmers in the survey.
Fig.1. Network of farmers’ perceptions of potential
solutions to encouraging land mobility.
Conclusion
In this study we have addressed land mobility issues as
perceived by the Irish farmers that could help nurture
change in land transfer patterns. The benefits of
applying some of these measures could result in better
social and economic conditions for encouragement of
young farmers and security for elderly farmers who
wish to retire. But, implementation of these measures,
such as early retirement and young farmers’ incentive
schemes, should be a part of a larger network and
policy change. A more dynamic and coherent
programme including farmers’ visions for land transfer
is needed to be able to encourage land mobility and
early land transfer.
Acknowledgments
This study arises from the DAFM funded Stimulus
project AgLandMarket – “Analysis of the functioning
of Irish agricultural land markets”. We also
acknowledge Macra na Feírme for allowing us use of
the data for this study.
References
Bogue, P. (2012). Land Mobility and Succession in
Ireland, Research report – Macra na Feírme.
McClelland, J.L. & Rumelhart, D.E. (1988).
Explorations in Parallel Distributed Processing, The
MIT Press, USA.
McClelland, J.L. (2014). Explorations in Parallel
Distributed
Processing,
Second
edition
(https://web.stanford.edu/group/pdplab/pdphandbook/).