Contemporary and 'messy' rural in-migration processes: Comparing counterurban and lateral rural migration.

This paper questions the ongoing dominant coverage given to counterurbanisation in the rural population literature. It is argued that this provides only a partial account of the true diversity of contemporary migration processes operating in rural areas and has the potential to fuse together different in-migration processes. Speci ﬁ cally, lateral rural migration has been under-researched to date. Using empirical data from a survey of 260 migrant households to 3 UK case study areas (in Scotland, Wales, and Northern Ireland), the signi ﬁ cance of lateral rural migration is revealed and compared with counterurban migration and migrants. The last change of address shows that 59% relocated from an urban area (participating in a counterurban ﬂ ow) whilst 41% moved from another rural location (lateral rural ﬂ ow). The boundary between migration processes can, however, be blurred: Some moves are an example of both counterurbanisation and lateral rural ﬂ ows. Incorporating lifetime migration histories data demonstrates the contemporary complexity and messiness of rural in-migration processes. For example, 26% of these migrant households only ever undertook a lateral rural move during their lifetime. For others, the direction of migration has changed numerous times and intertwined with each move are aspects of life course, return, and inter-regional migration. Comparing the survey characteristics and motivations of counterurban and lateral rural migrants, alongside interview material, highlights important similarities and differences. The paper concludes by calling on rural population geographers to more fully engage with the complexity, totality, and indeed messiness of contemporary rural in-migration processes. © 2015 The Authors. Population, Space and Place . Published by John Wiley & Sons, Ltd.

INTRODUCTION P apers on internal migration are a prominent feature of population geography journals. Arguably, one migration process stands out as receiving considerable academic attention, that of counterurbanisation. Whilst various aspects of rural population growth and rural in-migration have been studied (e.g. rural gentrification (Smith & Phillips, 2001;Phillips, 2005;Stockdale, 2010), back to the land movement (Halfacree, 2006), return migration (Ni Laoire, 2007)) much has been framed in terms of counterurbanisation processes, and accordingly, urban-to-rural migration flows have been assumed to dominate. Indeed, the literature is awash with counterurbanisation studies. However, have other rural migration flows been ignored as a consequence, such as lateral rural migration, including movement up (from a remote rural location into a small rural settlement) and down (from a small rural town or village to the open countryside) the rural settlement hierarchy? In this paper, it is argued that the dominant research narrative on urban-rural migration offers only a partial account of both rural migration processes per se and in-migration processes specifically. In addition, such a focus has the potential to overemphasise the assumed dominance and importance of counterurbanisation and consequently fails to acknowledge the heterogeneous nature of contemporary rural in-migration processes. Contrary then to the singular onedirectional flow of counterurbanisation commonly portrayed in the rural population literature, rural in-migration (and migration generally) is a messy process comprising a multitude of separate flows. Milbourne (2007, p. 385) said as much: 'Rural population change, though, is much more complicated… Being composed of movements into, out of, within and through rural places; journeys of a few hundred yards as well as those of many hundreds of miles; linear flows between particular locations and more complex spatial patterns of movement'. Nevertheless, this complexity or messiness is only occasionally acknowledged (Stockdale, 2006;, and all too rarely are these other migration flows studied either in their own right or directly compared with counterurbanisation. As such, there is a dearth of empirical evidence to support Milbourne (2007). This paper is intended to address this empirical void and seeks to challenge what has become the established narrative of counterurbanisation. It is shown that lateral rural migration is an important but neglected component of contemporary rural migration processes. The catalyst for this paper came from the attainment of unexpected findings in a recent research project: namely, the scale of lateral rural migration evident. This forced the author to reflect on two important questions. First, has the dominant counterurbanisation discourse blinkered us to the fact that other rural in-migration processes might also be evident and important? Second, with the exception of the obvious origin differences between migrants, what are the similarities and differences between counterurban and lateral rural migration flows?
The diversity of rural in-migration processes, including lateral rural migration or movement from one rural area to another, has generally gone unreported with the exception of a few recent and tentative studies (Grimsrud 2011;Bijker et al. 2013). Because of this, we know little about the nature of such flows, participating migrants, or the consequences for origin and destination rural areas. How significant is lateral rural migration? Is it comprised of long-distance movement or flows between essentially neighbouring rural areas? Are lateral rural migrants attracted to similar destination areas as counterurban migrants? What are the migrants' characteristics? What are their motivations for moving? In what ways do lateral rural migrants contribute to their destination areas? How do these compare with counterurbanisation flows and counterurbanite migrants? Whilst it is impossible to answer all of these questions in one short paper, the intention is to shed some light on many of these questions and in doing so present a case for increased academic attention to be given to lateral rural migration. This paper is, therefore, intended to stimulate further reflection and debate within the academy and by doing so encourage future research to move beyond what is now traditional counterurbanisation studies and explore other, contemporary, rural in-migration flows and processes.
The remainder of the paper is organised into five sections. First, the rural population and migration literature is reviewed to highlight the dominance of counterurbanisation studies and the associated stereotyping of rural in-migration processes and migrants. Second, the current study is introduced, including a description of its methodology and case studies. Third, using last change of address and migration histories data, a multitude of rural migration flows is identified and the relative importance of lateral rural migration in particular assessed before, fourth, progressing to examine the characteristics and motivations associated with such mobility. Throughout these empirical sections of the paper direct comparisons are made with counterurbanisation mobility. Finally, the paper concludes by highlighting not only the contemporary importance of lateral rural migration but also other forms of rural in-migration and calls for greater prominence to be given to them in future population and rural research. be explained sufficiently by the prevailing counterurbanisation models alone'. Similar conclusions have been reached by others (Milbourne, 2007;Halfacree, 2008;Grimsrud, 2011;. Nevertheless, it is the counterbanisation story that has largely prevailed in the rural population literature for near on 40 years in the UK (Perry et al., 1986;Champion, 1987;Bolton & Chalkley, 1990;Halliday & Coombes, 1995;Stockdale et al., 2000;Phillips, 2005;Stockdale and McLeod, 2013), Europe, and further afield (Berry, 1976;Fielding, 1982;Vertiainen, 1989;Kontuly, 1998;Lindgren, 2003;Herslund, 2012;Simon, 2014). Milbourne (2007), in particular, is critical of such studies and alleges that the evidence from different countries is 'that the dominant focus is on unidirectional flows of people to rural areas' (p. 384) and '[r]ural researchers have been pre-occupied with longer distance movements and with migrations from urban to rural places' (p. 385).
Much of the counterurbanisation literature has given only a cursory acknowledgement of the existence of other migration flows. For example, Champion (1998, p. 21) 'points to the existence of rural-to-rural migration that can lead to depopulation of certain rural localities while others are repopulating', and Bell and Osti (2010, p. 200) report that '[t]here is also much rural-rural movement within countries'. Mitchell (2004) too acknowledges that other migration flows may operate simultaneously alongside counterurbanisation, one of which is lateral rural migration; however, she appears to dismiss this type of relocation, noting 'it has less impact on the overall pattern of population distribution at the settlement scale' (p. 24).
This 'love affair' with counterurbanisation by rural population researchers is not surprising: but why has it endured for several decades? Not only that, but these decades (since the 1970s) have arguably witnessed a dramatic growth in and changes to mobility (Sheller & Urry, 2007), with Findlay and Wahba (2013) reporting that both the context and demographic regimes within which migration occurs have fundamentally changed. In simple terms, nothing short of a 'mobility revolution' has occurred with, for example, advancements to transport and telecommunications facilitating mobility over greater distances, multiple migrations, and indeed increasingly diverse and complex migration flows (Smith & King, 2012). However, rural population researchers have seemingly remained blinkered to, or dismissive of (Mitchell, 2004), the increasing complexity of migration and possible emergence and importance of other rural inmigration flows to accompany the expected urban origin dominance.
First identified in the US in the mid-1970s (Beale, 1975), rural or non-metropolitan population growth soon attracted considerable academic attention with the phenomenon having spread from the US (Berry, 1976;Vining & Strauss, 1977) to include other countries (Vining & Kontuly, 1978), including most of Western Europe (Fielding, 1982), Australia (Hugo & Smailes, 1985), and more recently parts of Africa (Potts, 2005) and post-Socialist states (Simon, 2014). Rural and population researchers worldwide embraced this new population phenomenoncounterurbanisation (see Champion, 1989, andMitchell, 2004, for a comprehensive overview of the concept)and so embarked on a plethora of Census and rural-based migration research focusing on the magnitude and spatial coverage of the trends, migrants' characteristics, motivations, and destination consequences, which has continued to the present day.
Early research tended to interrogate national quantitative data sets whilst more recent studies have adopted a largely qualitative approachso that the study of rural migration (and specifically counterurbanisation) mirrors what Smith (2007) refers to as the abandonment of quantitative approaches to population change in rural studies. Similarly, Grimsrud (2011, p. 643) claims: 'a factor fuelling belief in the counterurbanisation trend is the recent upswing in qualitative approaches to rural research'. An additional, if related, factor has been the growth of case study research. Such research, rather than investigating rural migration flows within these localities per se, has often concentrated on one migration process, such as urban-to-rural migration, with these findings frequently then generalised and assumed to be representative of all rural in-migration patterns and processes. By assuming that counterurbanisation dominates, researchers have stopped looking for or at other forms of rural in-migration. Moreover, with reference to counterurbanisation specifically, several authors (including Halfacree, 2008) acknowledge that counterurbanisation may take on different 601 Research on Lateral Rural Migration guises and warn that future research 'should be careful not to use too simplified understandings of counterurbanisation' (Bijker et al., 2013, p. 591). Thus, not only is all rural in-migration too frequently framed in terms of counterurbanisation, but also counterurbanisation itself may be too broadly defined and as a result 'fuses together' what are very different rural in-migration processes.
Rural in-migration has become stereotyped along the lines of 'mainstream counterurbanisation' (Halfacree, 2008). According to Milbourne (2007, p. 382): '[m]eta-narratives of rural population have become dominant…, based on lifestyle-led voluntary movements of middleclass groups to rural areas and associated sociocultural and housing-related conflicts and problems'. Equally, Grimsrud (2011, p. 642) alleges that 'the stereotypical rural in-migrant is routinely portrayed as someone who escapes the harried city for a more "down-to-earth" way of life', and Halfacree (2008, p. 479) claims that '[t] his almost taken-for-granted presentation of wealthier people moving to rural areas is the dominant image today'. In the same paper, Halfacree bluntly remarks: 'counterurbanisationwhat more is there for researchers to say about it in the twenty-first century? It is now seemingly an exhausted… research topic, so thoroughly inscribed in our texts and our background knowledge… that there is little new to add' (Halfacree, 2008, p. 480).
Halfacree argues that 'there is more to say' (p. 481) and in doing so presents a threedimensional counterurbanisation model. This model is a welcome addition to the counterurbanisation literature in that it highlights motivational differences for rural in-migration, culturally imagined dimensions, and the growing role of international labour migration. However, has Halfacree, like others, inadvertently dismissed many other rural migration flows or simply 'fused them together' with counterurbanisation (e.g. international labour migration to rural areas)?
The presence of other rural in-migratory flows has recently been acknowledged. Stockdale (2006) identifies a diversity of migration processes operating in rural areas, whilst Andersen (2011) and Bijker et al. (2013) observe a diverse group of rural in-migrant types. 'Return to the rural' migrants , birthplace returnees (Lundholm, 2012), and rural-to-rural migration Gkartzios & Scott, 2010;Grimsrud, 2011;Stockdale & McLeod, 2013) have all been observed. Previous research has also observed different motivations and characteristics amongst lateral rural and counterurban migrants as well as motivational differences by type of rural area moved to. These studies demonstrate the inappropriateness of assuming that rural in-migration and counterurbanisation are one and the same or that counterurbanisation is the dominant rural inmigration flow. Indeed, some who have captured an element of lateral rural migration in their work fail to explore it any furtherfalling as it often does outside the specific remit of their paper (Gkartzios & Scott, 2010;Stockdale & McLeod, 2013). Too often, lateral rural migration is reported as an interesting aside. The current paper seeks to address this by moving lateral rural migration 'centre stage' from its more customary and at best 'in the wings' position. It is intended to fill a gap in our knowledge by examining this under-researched migration flow and by doing so argues, first, that there is more than counterurbanisation going on in rural areas and, second, that there are important similarities and differences between counterurban and lateral rural flows, processes, and migrants. To continue to neglect the study of lateral rural migration in favour of counterurbanisation research is to oversimplify, stereotype, and provide only a partial understanding of contemporary migratory processes. Furthermore, in contrast to several of the studies reported earlier (Grimsrud, 2011, and the various papers by Bijker and colleagues) that compare migration patterns in different types of rural areas, this paper compares counterurban and lateral rural migrants moving to the same rural areas. In addition, and again distinct from previous research, data are analysed relating to the last change of address and the lifetime migration histories of migrant households. Lifetime migration data provide greater insights into not only the messiness of contemporary rural inmigration processes but also demonstrate that counterurbanisation as a process is far from the 'tidy' migration flow commonly portrayed in the literature. Finally, in contrast to the dominant 602 A. Stockdale qualitative approach that now characterises much rural migration research, the present study adopts a mixed methods approach.

METHODOLOGY
This paper utilises data from a research project in three remote and peripheral UK locations: the Powys region of Mid-Wales, the Isle of Bute in Scotland, and the Glens of Antrim in Northern Ireland. All three areas have a strong tradition of youth out-migration, fragile economic structures, and low wage economies, outstanding scenic environments, and proximity to small rural market towns. The three case studies, nevertheless, have different demographic profiles and have experienced different overall population trends. For example, Powys has experienced notable population growth in recent decades, whereas the populations of the other two study areas have declined. The Glens of Antrim possesses the youngest population profile, while the Isle of Bute has the oldest population profile with approximately one-third of its population aged 60 years or older. It is important to say from the outset that the research project was not designed with lateral rural migration specifically in mind. Instead, the project investigated middle-aged and retirement-led rural in-migration. As such, the case studies were selected on the basis that they had experienced high rates of in-migration by those aged 45 years or older (Sample of Anonymised Records and Special Migration Statistics, 2001 Census of Population). In addition, at least subconsciously, during the design stage it was assumed that counterurbanisation (urban-rural moves) would be the dominant in-migration process.
A 'drop off and collect' household survey was carried out using a random list of target addresses generated from the most recent edited electoral register for each study area. Market and other small towns were excluded so that the survey targeted village and open countryside residents and therefore was concentrated on the most rural parts of each study area. The household survey was supplemented by a postal survey that targeted specifically (given the pre-retirement age focus of the original project) migrants aged 50-64 years at the time of their move into the area. This purposive sample was comprised of persons living in the study areas who had first registered with a local GP (since 2000) when aged 50 years or older. This postal survey was administered on the author's behalf by national agencies responsible for patient registration data.
Collectively, both surveys generated data for 644 households of which 260 were identified as rural in-migrant households (as distinct from non-moving and local mover households). This paper analyses data for the in-migrant households only. In-migrant households were defined as households whose last change of address since 2000 involved a move of at least 15 km to their current address. This arbitrary 15 km cut-off served the purpose of disaggregating local 'within study area' changes of address from inmigration. In reality, many migrant households moved considerable distances: 26% moved 40-79 km, 21% 80-159 km, and 31% relocated over a distance of at least 160 km.
The survey focused on the migration histories and motivations of households and individual occupants. Specifically, data were collected relating to each household's last change of address (to its current place of residence), its previous change of address (i.e. the move to the origin of the last change of address), and lifetime migration by the householder (self-defined) and their spouse/partner. In addition, the survey obtained data relating to the demographic and socioeconomic characteristics of all current household occupants. This detailed dataset permitted the identification of different rural in-migration flows and direct comparisons to be made between counterurban (urban-to-rural) and lateral rural (rural-to-rural) migration and migrants. The research also included 30 in-depth interviews with migrant households (10 in each area). These interviews adopted a life history approach by chronologically identifying moves related to specific life events (such as childhood, employment, marriage, family formation, empty-nest, retirement, and widowhood) and sought to more fully understand the households' migration histories at different stages of the life course (including the move to their current address). This qualitative data is more fully analysed elsewhere (Stockdale & McLeod, 2013;Stockdale, 2014), with interview excerpts incorporated in the discussion that follows to illustrate and shed deeper understanding on the multiple migration processes at work.

MORE THAN COUNTERURBANISATION
The analysis of household migration data is disaggregated: First, the focus is on the household's last change of address and identifies multiple inmigration processes on the basis of the origin of the household. Origin is examined in terms of the type of environment moved from (namely, city, large town, rural market or small town, village, and open countryside) and the UK region of origin. Second, the analysis incorporates lifetime migration history data for the migrant householder and their spouse/partner. Here, the 'messiness' or complexity of past household migration leading to a move to the three rural study areas is revealed. Table 1 displays the type of environment origins of the last change of address by migrant households and hence the types of in-migration experienced in these study areas. Several distinct in-migration flows are evident. First, there is a classical counterurbanisation pattern that involves a move from cities/large towns to these rural locations and accounts for 59% of the last change of address undertaken by migrant households in the sample. Such urban-rural migration took place over varying distances. The modal distance moved (37%) was in excess of 160 km; however, 1 in 10 relocated no more than 39 km from their previous place of residence. Second, there is lateral rural migration accounting for 41% of the last change of address undertaken by migrant households (with their modal distance moved (31%) found to be 16-39 km). However, this category of migrant can be disaggregated further. Movement down the rural settlement hierarchy (from rural market and small towns to villages and the open countryside) accounts for 14% of all migrant households captured by the survey (or 35% of all households undertaking a lateral rural change of address). Such moves conform to a counterurbanisation model: namely, Champion's (2005) 'counterurbanisation cascade'. Even though the move is between or within rural areas, it is taking place in a typical counterurban direction. In addition, there is migration between similar rural environments: namely, rural villages/open countryside. This represents a more distinctive lateral rural migration flow and accounts for 27% of all migrant households (or 65% of all lateral rural changes of address).

Last Change of Address
The migration trends evident in these study areas would seem to justify the dominant counterurbanisation research focus displayed by rural population geographers. However, lateral rural migration accounts for a not insignificant share of the total migrant households captured by this survey: 41% (106 households) of all migrant households or 27% (69 households) if restricted to movement within/between villages/open countryside. The fact that such a sizeable rural in-migration flow has not been extensively researched in the literature confirms the partial nature of current rural migration research. If the results from this study are in any way indicative, as many as two out of every five moves into rural areas go unreported and unresearched. Moreover, movement down the rural settlement hierarchy can be both an example of lateral rural migration and counterurbanisation. This raises questions about how migration processes are defined. For example, it could be argued that much of the counterurbanisation literature has focused on migration from the largest settlements (cities and large metropolitan centres) and accordingly defines counterurbanisation in very narrow terms. Relocations further down the settlement hierarchy, and migration within and down the rural settlement hierarchy specifically, have been ignored. In other words, research has tended to concentrate on the 'larger' steps/moves down the settlement hierarchy (moves directly from a city into the countryside) and ignored 'shorter' step moves (from small rural towns into villages or between similar size rural settlement types).
A further component of rural in-migration in these case study areas is uncovered if the last change of address considers the region of origin. Bearing in mind that the study areas were in Wales, Scotland, and Northern Ireland, 85 (56%) and 36 (34%) counterurban and lateral rural migrant households, respectively, moved directly from England. Counterurban flows into these areas, in particular, are therefore characterised by sizeable flows directly from England. This is less apparent amongst lateral rural flows but nevertheless still accounts for one-third of all lateral rural mobility captured by the survey. In numerical and proportional terms, English inmigration seems therefore to be as significant as counterurban migration and more significant than lateral rural migration for these study areas. How can we make sense of these different social and geographical forms of in-migration? This English inflow is explored further in the succeeding sections.
With only 37 households participating in a movement down the rural settlement hierarchy specifically (Table 1), this category is combined with moves between similar rural environments to give an overarching lateral rural migration category for quantitative analysis in the remainder of this paper. Inevitably, this results in some loss of detail and can be criticised for using by default a narrow definition of counterurban migration (moves from cities and large towns only).

Lifetime Migration Histories
Migration history data for the household (and head of household and spouse/partner) demonstrate the complex nature of rural in-migration. It is far from a tidy unidirectional process. Figure 1 depicts the key migration history characteristics associated with counterurban and lateral rural last changes of address. A number of migration processes are apparent.
First, the dominance of counterurbanisation is once again evident (Figure 1 (column B)). Close to half (49%) of the 106 lateral rural (last change of address) migrant households had at some time previously also participated in urban-rural (counterurban) migration (i.e. a move that originated in a city or other large urban centre). As a consequence, at least three quarters (74%) of the households in the entire migrant sample had, at some time in their residential history, undertaken a counterurban move. Counterurbanisation is therefore the dominant rural in-migration trend and lends further support (in numerical terms at least) for its dominant coverage in the literature. What this data also confirms, however, is that having made a counterurban move, many households subsequently move within the rural environment (and move further down the rural

605
Research on Lateral Rural Migration settlement hierarchy specifically). This is rarely reported or studied.
Second, there is evidence of return migration and different forms of return migration. Amongst those households whose last change of address involved a counterurban relocation, 18% had previously participated in a ruralurban migration flow (Figure 1, column A). They had, therefore, at some time in the past taken part in rural out-migration. As such, a particular strand of counterurbanisation comprised a return to the rural by households who had previously lived in a rural area. Arguably, this return to the rural in-migration has been neglected in the literature, although Feijten et al. (2008), Farrell et al. (2012, and  do acknowledge its significance. Return migration to a specific place of former residence is also observed. A total of 18% and 28% respectively, of households who undertook a counterurban and lateral rural last change of address had previously lived in the general destination area. Birthplace return migration is also noted. Taking into account lifetime migration (measured in terms of the place of upbringing for the head of each migrant household and their spouse/partner), it is noted that 22% of urban (and 26% of rural) origin migrant households contained at least one returnee to their area of upbringing. Return migration then, and particularly birthplace return migration, is an important component of both counterurban and lateral rural in-migration, yet it is rarely acknowledged in the academic literature. Extending the birthplace analysis further to include the type of area brought up in (defined by the head of household themselves) illustrates once again that rural inmigration possesses strong urban origins. A total of 70% of heads of households participating in an urban-rural last change of address reported that they had been brought up in an urban environment (either a city or large town), and at least one in every two households undertaking a lateral rural move included a head who had been brought up in an urban environment.
Third, English-born in-migration is observed. Supporting the high number of households reported earlier whose last change of address involved a move directly from England, it is found that 48% and 51% of urban and rural origin (last change of address) migrant households, respectively, possessed a head of household who had been brought up in England. Rural inmigration to the study areas is not only, therefore, characterised by moves directly from England (by, e.g., previous out-migrants from Scotland, Wales, and Northern Ireland) but importantly also includes significant English-born inmigration. English in-migration to non-English rural areas of the UK has received limited, and in many cases now dated, attention in the literature. Studies by Jedrej and Nuttall (1996) and Short and Stockdale (1999) observed English inmigration to rural areas of Scotland during the 1990s. More recently, Cloke et al. (1998), Day et al. (2008), Day (2011), Milbourne (2011, and Stockdale (2014) report on English migrants to rural Wales. The results from this study demonstrate not only the continued importance of rural inmigration from England but also the rural inmigration of English-born migrants. In terms of the consequences for the rural destination areas, the urban origin of in-migrants has frequently been reported as destabilising established modes of rural living because of social and cultural differences between them and long-term residents. Intertwine a significant English component to rural in-migration in these study areas and the potential impact may be significant.
Overall, the identification of migration processes affecting these three remote and peripheral rural areas serves to highlight several important points that remain largely absent from recent rural population and migration research. Whilst numerically counterurban migration dominates the moves undertaken by households in this sample, the proportion of households whose last change of address (41%) involved or had only ever undertaken (26%) a lateral rural move is worthy of greater academic attention than received to date.
Rather than being the 'tidy' one-directional flow (often portrayed in the literature), rural inmigration is a considerably more 'messy' and complex set of migration processes. Household migration (eventually leading to a rural residence) frequently involves multiple migration events and changes of migration direction as summarised in Figure 1. This only becomes apparent when the households' and occupants' migration histories are studied. Research that incorporates only the last change of address provides a partial account of the migration processes at work. For example, a counterurban (urban-rural) last change of 606 A. Stockdale address may have been preceded by a ruralurban flow (and whilst beyond the scope of data collected here may have been followed by a return move to an urban environment). Equally, a household participating in a lateral rural (rural-rural) change of address may have participated in a previous rural-urban and/or urban-rural move, so that, in the case of the latter, a counterurban migration event is on occasion followed by a (lateral) move within the rural environment. Superimposing the migration histories of the householder and their spouse/partner adds a layer of migration complexity. Indeed, the direction of migration may have changed numerous times during the household's and occupants' lifetime: a fact that is not always acknowledged in migration research. Instead, focusing on the most recent change of address oversimplifies the diversity of migration flows and patterns evident in any given area and experienced by migrant households.
Other migration processes are at work including return migration (in different guises), English in-migration and, as shown later, retirement migration. Arguably, more contemporary forms of each have emerged, for example, return migration now includes a return to the rural (alongside a place-specific return) component. Specific to English rural in-migration, in numerical terms, this is as significant as counterurban (and more significant than lateral rural) migration in these study areas. English in-migration (Stockdale, 2014), return migration , and retirement migration (Stockdale & McLeod, 2013) are explored further elsewhere. Here, the attention focuses on a comparison between counterurban and lateral rural in-migrant households (as measured by last change of address).

WHO ARE THE COUNTERURBAN AND LATERAL RURAL MIGRANTS?
Given the criteria used to select case study areas (areas that in the Census had recorded high rates of in-migration by those aged 45 years and older) and the dominance of retirement as a motivation for the last change of address (see next section), not surprisingly approximately one in every five making an urban-to-rural or lateral rural last change of address did so when the head of household was at or above the UK State Pension eligibility age of 65 years, with a further 40% having moved when the head was aged 50-64 years ( Table 2). There are no statistically significant differences between the age of the head of household for counterurban and lateral rural migrants. Indeed, 48% and 42% of counterurban and lateral rural heads of migrant households, respectively, were retired immediately following the move to their current address. In contrast to a literature that almost exclusively associates rural in-migration with middle-aged or retirement-aged persons, a sizeable inflow of households headed by individuals aged younger than 50 years is also observed. Indeed, amongst lateral rural migration this age group accounts for 40% of the households moving into these study areas. The participation of younger age groups in rural in-migration is rarely acknowledged with the possible exception of Bijker and Haartsen (2012), Smith and Higley (2012), and Stockdale and Catney (2014).
Most moved as a married couple irrespective of the households' origins or head's age at the time of the move. Whilst any analysis of income data will be skewed by the large proportion in possession of a pension (and ignores the fact that at least two-thirds own their homes outright, i.e. without a mortgage), the modal household income is found to differ between counterurban and lateral rural migrant households (but not necessarily in the direction one would expect based on an academic literature reporting wealthy counterurbanites). The modal annual income amongst counterurban households is between £10-20,000 (approximately one-third) whilst amongst lateral rural households, it is between £20-30,000 (27%). Moreover, one in every four lateral rural movers possessed a pretax annual household income of £30-50,000

Research on Lateral Rural Migration
(compared with only 16% of counterurban households). In contrast then to the counterurbanisation literature, the evidence presented here is that lateral rural flows are associated with a greater share of 'wealthy' households. This can be explained by the greater number of younger and economically active households within the lateral rural migrant sample. It is because fewer lateral rural migrant households (when compared with counterurban migrant households) rely solely on a pension for their income that lateral rural migrants appear wealthier.
In terms of the householders' highest level of educational attainment and occupational classification there are few differences between the two migrant groups. Indeed, counterurban and lateral rural migrants are remarkably similar. Almost identical proportions possess a degree (and also no qualifications): 29% and 17%, respectively, of urban origin households and 25% and 19%, respectively, of those moving from another rural location. At least half of both migrant samples are headed by someone in the managerial and senior officials, and professional occupations groups and 17% and 19%, respectively, of counterurban and lateral rural migrant households contained at least one person who was either self-employed or ran their own business. A total of 56 businesses were owned by the combined migrant sample, the majority of which (78%) were associated with households headed by an individual aged younger than 50 years at the time of their move to the current address. These 'entrepreneurs' gave employment (39%) and quality of life (28%) reasons for moving to their current address. The latter has frequently been associated with lifestyle entrepreneurs in the rural literature (Bosworth & Willett, 2011).
However, notable differences are observed when we compare the types of businesses operated by counterurban and lateral rural migrants. Traditional rural businesses (e.g. agriculture, fishing, and forestry) account for 35% of those owned by lateral rural migrant households and only 17% of the counterurban business ventures. Other businesses commonly run by lateral rural migrants relate to accommodation and food and professional services. By comparison, a more diverse range of business types characterises counterurban migrant households, including: arts and crafts (20%), retail (13%), and accommodation and food (13%).
A total of 113 persons (including the migrant owner) were employed in these businesses with the number employed per business greatest amongst those owned by lateral rural movers: 2.2 per business vis-a-vis 1.8 per business owned by a counterurban mover. This job creation differential reflects that the more agri-type and accommodation and food businesses (associated with lateral rural households) are capable of employing more staff than say the arts and crafts ventures more favoured by counterurban movers (Akgun et al., 2011;Stockdale & McLeod, 2013). Lateral rural migration, by virtue of the nature of the businesses run, offers greater potential for rural job creation in these study areas than counterurban in-migration.
At interview, more information was obtained about the entrepreneurial activity of migrants. It ranges from sizeable businesses to one-man/ self-employed ventures, includes production, manufacturing, and services, was on occasions a financial necessity following an inadequate pension or an opportunity to fulfil a long-held ambition to undertake a particular venture, and involves those with considerable business experience as well as the first-time entrepreneur. Huw (living in Powys) reflected on migrants generally coming into the area: '…people coming to this areathey come in and they give a lot of things a new lease of life. Fresh blood, so to speak… They start businesses, [one] they make box files and their products you see in WH Smiths. I can think of the odd painter and decorator, … there are a lot of one man bands doing crafts and things.' Likewise, Craig who moved to the island of Bute initially to take up employment and has since become self-employed in the professional services, recounted some of the businesses that had been established by migrants: he's set up this preserving pickle thing. There's one couple, they are older I suppose, they've opened up a gallery. Some do a little farming [and contribute to] "Eat Bute" doing more of the local produce, and pushing that. Then there's a furniture manufacturer out at [specified location].'

608
A. Stockdale Margaret, who moved to Bute with her late husband from Glasgow, 'for a different life and… my sonhe'd met a girl and was now resident on the island here', established a mushroom business: 'I must be a crazy woman -I'd always wanted to grow mushrooms and my brother said to me when we moved, it [Bute] would be a great place to grow mushrooms. …I attended an open weekend at Inverness, …found out about these mushrooms and decided I would have a go at growing them.' Rebecca and Simon moved to Powys from a city when Simon retired and soon set up a small business in a nearby market town. Retirement was not what Simon expected: 'while we thought "this is retirement" yay! … At sixty years, I can retirewonderful, but it wasn't to be. [The pension] was poor but with our savingswith that money we've been able to set up the book shop.' Rosemary and her husband (reported in the following section) ran businesses all their working lives. When they returned to the Glens of Antrim from Scotland they quickly established two businesses.
Returning briefly to the small sub-group of counterurbanites who subsequently relocated within the rural environment, they are found to consist of a younger group of high income professionals when compared with the counterurban and lateral rural migrant samples. Three quarters of these household heads had been aged younger than 50 years when they undertook the counteruban move (compared with 42% of all counterurban migrants). Furthermore, given the importance of employment-led motivations amongst this sub-group (observed below), one might have hypothesised that such migrants were more likely to be associated with traditional rural occupations or to be entrepreneurial. This was not found to be the case: Only 8% ran their own business and most were connected to professional occupations, including accountancy, directorship of a public sector partnership, chartered surveying, medical practice, teaching (including further education), and so onoccupations that are not generally confined to rural areas. Whilst employment was frequently given as a reason for moving, this was not necessarily to work in a rural area: The majority instead commuted to a nearby urban settlement. Equally, and no doubt as a consequence of their largely professional and 'still in employment' status, one-third possessed a pre-tax annual income in excess of £30,000 (compared with 22% of all counterurban migrant households) so that counterurban migrants who then relocate within the rural represent the 'wealthiest' migrant cohort.

MOTIVATIONS FOR MOVING BY COUNTERURBAN AND LATERAL RURAL MIGRANTS
The main motivations for moving to the current address are presented in Table 3 and compared for counterurban and lateral rural migrant households. These household survey data are supplemented by interview quotes to provide further motivational insights into the range of migration flows evident. Survey respondents

Research on Lateral Rural Migration
were asked to identify their main reason for moving. It is acknowledged that migration decision-making will involve multiple considerations with the move unlikely to be reducible to any one factor Stockdale, 2014). Indeed, the interview quotes reported in the succeeding texts emphasise several motivations and/or considerations at work. Nevertheless, asking respondents to select only one motivational factor sheds considerable insights into the perceived dominating role played by a number of potential influences.
The motivations expressed by counterurban and lateral rural migrant households in the survey are found to be statistically different at the 99% confidence level. However, the three most frequently stated motivations (retirement, quality of life, and employment) are shared by both migrant groups, with a common modal reason identified: retirement accounted for 33% of urban-rural and 28% of rural-rural last changes of address. The importance of retirement as a motivational factor is clearly influenced by the selection of study areas. The case study areas were chosen on the basis that they had recorded high levels of in-migration by those aged 45 years and older. Nevertheless, the fact that retirement-led in-migration to these areas characterises both counterurban and lateral rural in-migration is significant. Even though life course migration has been receiving increasing attention (Feijten et al., 2008;Geist & McManus, 2008;Kley & Mulder, 2010), rural in-migration at the retirement life course stage is commonly reported in terms of urban-rural flows only. Here, we also have strong evidence that some rural residents move within the rural environment on retirement. A move at retirement is not, therefore, only the preserve of city/town dwellers.
Eva and her husband moved from Belfast to the Glens of Antrim on their retirement. Eva, originally from the Glens and therefore a return migrant at retirement, explained: 'Maybe my roots were asserting themselves, I don't know. … In my mind I was absolutely sure that I wanted to come and live here.' In the case of Brian and Wendy, who moved to Bute from a rural setting in the south of England on retirement, their holiday home, association with the island for many years, was influential: '[The holiday home] was just successful way beyond our wildest dreams. We discovered that we were making lots and lots of friends up here and going out and doing things we just didn't do in [England]…so eventually we decided that we would retire here.' It is the relative importance, however, of employment and quality of life reasons that help to explain the motivational differences between the two samples expressed in the survey. Quality of life motivations are significantly more important for counterurban mobility (26% vis-a-vis 15%) and are supportive of the wider literature on counterurbanisation that commonly views such migration for lifestyle and/or quality of life reasons (Benson & O'Reilly, 2009;van Dam et al., 2002). For example, James and his wife moved to Wales from England at retirement and sought a different and more active lifestyle. In the survey, they had explained their move as being primarily for quality of life reasons. They bought a smallholding and as James remarked: 'It gets to the point where you think, I've just got to do something [after retirement] because otherwise you just sit and do nothing and find yourself watching television in the middle of the afternoon and having naps… It [was] a waste of the time I've got left. …We moved here, bought some piglets and we enjoy it… we have two sheep and we're in the process of breeding them the third time. [We also have] turkeys and ducks… I think it's the experience, I get pleasure with the animals [rather] than wasting my time.' Younger couples too moved for quality of life reasons. Susan moved from the south of England to the Isle of Bute with her husband and two small children. Her husband continues to commute weekly to work in London (leaving Bute on Monday morning and returning Thursday evening). In the survey, Susan gave quality of life as the reason for moving and elaborated at interview: 'We wanted the lifestyle -I don't like that word; everybody's lifestyle here is different. But we wanted the atmosphere… We wanted to live somewhere where we didn't have to lock our car when we went to pay for petrol.

610
A. Stockdale Gosh, in Bute somebody comes and pumps it for youyou don't have to get out. That's what we like about it.' Similarly, lateral rural moves including movement further down the rural settlement hierarchy also displayed strong quality of life motivations. Emma and Alfie relocated from a small market town to the Welsh countryside, and accordingly, their move is an example of both lateral rural and counterurban migration. Speaking about their former residence, Emma explained: '[At one time] you used to walk down the main street and knew everybody. You could park your car in the main street and never lock it. You could open the back door and throw the shopping in and nobody would steal anything out of it. Not anymore. ....Moving to the country …it was nice, the children were small and there was a lot of freedom, they could run around in the fields and we don't need to worry about traffic or that sort of thing. We got a horse… I think we saw it as a good step as a family really.' In comparison, employment-motivated changes of address (as expressed in the survey) are relatively more important amongst lateral rural migrants (22% vis-a-vis 11% of those of urban origin). This accords with Grimsrud's (2011) findings in Norway. Neil and Yvonne (who moved from mainland Scotland to Bute) are typical of many employment-motivated migrants. Neil explains: 'I was working in [central belt] and then for one reason or another things didn't go so well. So I was looking for a job and I got the offer of a job over here [on Bute]. I thought a spell on the island wouldn't do the career any harm. I came here without a view to being long term at all, but [I] became a director.' Similarly, Rosemary (a native of the Glens of Antrim) and her husband returned to the area after selling their business in the Scottish Borders. In the Borders, they lived outside a small rural town. Rosemary outlined business changes as influencing their decision to move.
'A large company [a national chain retailer] came and asked could they buy our business.
And we said yeah because the price was right it was a great price, we were shocked. …I then did art and I loved it…and I took over a wee shop here in the Glens and turned it into an art gallery. My husband rented the business across the road.' Two further motivations are notable from Table 3. First is the relative importance of housing factors amongst lateral rural movers (accounting for 10% of all such moves). Housing-led reasons have traditionally been associated with short distance relocations, a feature confirmed during some interviews. Mr and Mrs Murphy moved between villages in the Antrim Glens. They reside in public sector housing and referring to their previous village of residence, reported: 'We weren't content living there…not because of the neighbours, it was just that village. [Where we lived] our house was straight up the back.
[There was] a big square behind us and it was full of children morning, noon and night. …we just didn't belong there: I just didn't feel comfortable living there. There was a girl living in this housewho came from that areaso the two of us swapped [public sector] houses.' Others, as they grew older, realised they had to move from a more remote setting to be closer to services and facilities and/or downsize their home in relation to their physical capabilities. This is typified by Kevin and Claire who relocated within Powys (following their original move from an English city). Kevin and Claire are an example of counterurban movers who then participated in a lateral rural move (but towards an urban centre): ' [We] did know that we would have to move into a town when we got older…the reality is we had to move closer to Builth [small market town]. … Coming to Wales we've left our children back in England.... They can't pop mum down to the grocery shop or the doctors. We also knew we would have to downsize. We were realistic with the size of garden we'd got. …It was physically hard work to keep it under control.' Second, a further 10% of all counterurban and lateral rural moves to these study areas were 611 Research on Lateral Rural Migration motivated by a desire to be nearer parents or adult children. One can speculate that this may involve a move to be closer to a now elderly parent(s) or for an elderly parent to move closer to familial support. Such moves are likely to be consistent with the various aspects of return migration noted.
Geoff, for example, along with his wife returned to the Glens of Antrim (from outside London) shortly after his mother died.
'It wasn't that he [my father] wasn't wellhe was just getting elderly and I was never here when my mum, I was never here to look after Mum… So I thought I'm not going to leave Dad on his own, so that was the reason. …It wasn't planned, it was more a spontaneous thing.' Equally, it may involve an elderly parent(s) moving closer to adult children perhaps following widowhood or the onset of declining health. Agnes, aged in her 80s at the time of interview, explained how she had her Powys home up for sale following the recent death of her husband and her own ongoing illness. She and her husband had moved to Powys from Birmingham: It was noted earlier that some households (38 in total) who had last undertaken a lateral rural move had participated in a counterurban relocation previously. In other words, the original counterurban change of address was followed by relocation within the rural environment. The survey asked respondents to identify their main motivation for each move and consequently permitted this sequence of moves to be explored further in terms of the motivational influences. Kevin and Claire (reported above) are typical of such relocations by ageing migrants. In addition, and in contrast, to the retirement and quality of life motivations that characterised a large share of counterurban last changes of address (reported earlier), amongst this group of subsequent movers several initially moved from an urban-to-rural area for employment (36%) and personal/familial (33%) reasons with these same reasons reported again to explain the subsequent relocation within the rural (personal/familial by 32% and employment by 24%). For example, Phil and Grace initially moved from the West Midlands to Devon (counterurban move). Whilst they reported this move in the survey as being primarily for employment reasons, at interview Phil's explanation of why they moved includes reference to future retirement and quality of life/lifestyle considerations. In addition, he introduces different reasons for the decision(s) to move and the choice of destination(s): 'I always fancied moving away…in preparation for retirement. Working in a city was brilliant, it was wonderful…but do you want to live with that after you've retired? I negotiated a move [within the firm] to Devon. [Devon] was perceived as a better lifestyle and away from the rough and tumble of the city and all the rest of it.' However, Devon did not live up to their expectations, and the couple subsequently relocated (participating in a lateral rural move) to rural Wales (following retirement) for personal reasons it was claimed in the survey: 'We were right in the middle of a village [in Devon] and [the other residents] were less welcoming in the village than I've ever experienced. Within a week or two of getting there we were being described as incomers buying up property that their young people couldn't buy. …[We] had a problem with a neighbour over boundaries and rights of way. That caused a problem, stress for Grace [who was at home] while I was working away all day… She got me to promise that we would move because she couldn't live there with them.' But commenting on the move to Wales, Phil introduces quality of life and community motivations: Importantly, as a group initial counterurban movers who subsequently relocated between rural areas (such as Phil and Grace and Kevin and Claire) have chosen to stay within a rural environment (if a different rural area from that first moved to) as opposed to abandoning the rural experience altogether. This is the group of migrants for which Halfacree and Rivera (2012, p. 92) ask: '[h]ow is it they become "the ones that stay"?' CONCLUSION This paper took as its starting point that undue attention in the rural population literature has been given to counterurbanisation at the expense of other rural in-migration flows and processes. It argues that the importance of lateral rural migration, in particular, has been neglected and that counterurbanisation itself as a process has been too broadly framed: fusing together very different migration processes. Other researchers (Milbourne, 2007;Halfacree, 2008) too have questioned the dominance of counterurbanisation studies; however, their arguments have been weakened by the absence of supporting empirical evidence. The contribution of this paper to advance these arguments is it analyses empirical data to make direct comparisons between counterurban and lateral rural migration flows and migrants. Utilising migration histories data, in addition to the more commonly used 'last change of address', has enabled a greater unravelling of rural inmigration processes. The previous focus on counterurbanisation and using only last change of address data has oversimplified and 'fused together' different in-migration processes. In effect, it 'tidied up' what is an increasingly messy and evolving set of migration processes operating at any point in time or over the course of a household's or individual's lifetime. Migration research that concentrates largely on counterurbanisation at best then gives a partial account of contemporary rural in-migration and at worst fails to distinguish between separate in-migration processes.
Defining counterurban migration on the basis of the last move (i.e. urban origin) allowed (perhaps for the first time) direct comparisons to be made with lateral rural (i.e. rural origin) moves. Incorporating migration histories data permitted the separation of increasingly fused together rural in-migration processes, such as return and English in-migration. In addition, combining quantitative and qualitative data in this study has enabled greater insight than is possible from either alone. The inclusion of interview excerpts added considerable depth of understanding, especially in relation to the migrant motivations associated with different in-migration flows. Several important conclusions have emerged.
First, the diversity of migration processes affecting contemporary rural areas. Not only is counterurban migration evident in the areas studied here but lateral rural, return (to rural, birthplace, or a specific previous place of residence), English, and retirement migration also feature prominently. Equally significant, and perhaps for the first time, empirical data are presented to demonstrate that an initial counterurban move may be followed by a relocation within the rural environment. In relation to counterurbanisation itself, the evidence presented here support claims elsewhere (Halfacree, 2008;Bijker et al., 2013) that it comes in different guises. It is found to include not only 'large steps' down the urban settlement hierarchy (from a city into the countryside) but also 'shorter steps' between settlement types (e.g. from a market town into the countryside). These shorter steps are rarely studied. Moreover, lateral rural migration is frequently bound up with urban-rural flows (many had previously undertaken a counterurban move before relocating 'within the rural') and on occasions included an element of urbanisation (a move up the rural settlement hierarchy), for example, a move from the countryside into or nearer to a village or small town. The inclusion of migration histories data in particular has helped to demonstrate that rural in-migration is far from the unidirectional flow commonly reported. To the contrary, multiple and messy migration processes are operating in any given rural area and multiple and diverse acts of migration take place across a household or individual's lifetime.
Second, in numerical terms, counterurbanisation is the dominant in-migration process affecting these study areas. However, given the sizeable share of 613 Research on Lateral Rural Migration migrant households that had only ever undertaken (26%) or whose last change of address (41%) involved a lateral rural move it would seem appropriate for lateral rural in-migration to be given a higher place on rural and population research agendas. Moreover, English in-migration (both in terms of origin and/or birthplace) and retirement migration are as significant in these study areas as counterurbanisation (and more significant than lateral rural migration) in numerical terms. The prominence of retirement migration is, however, a consequence of the case study selection criteria. In sociocultural terms, the English in-migration identified here may be more significant than counterurbanisation. Future research should seek then to determine if English in-migration is a component of a broader process of counterurbanisation or is something more distinctive. Similarly, is return migration a component of, or distinct from, counterurbanisation? In other words, research should attempt to disentangle the various inmigration strands rather than fuse them together.
Third, there are notable differences between counterurban and lateral rural migration in terms of migrant characteristics and motivations. These differences provide further support for increased attention to be given to lateral rural flows in migration studies. Accompanying the middle-aged and retired migrants frequently reported in the literature, a sizeable group of younger in-migrant households has also been identified amongst both counterurban and lateral rural flows. Rural in-migration by younger households is rarely studied. Contrary to the literature that purports middle-class and wealthy counterurban migrants, lateral rural migrants in this study are associated with higher annual pre-tax incomes. Whilst other socio-economic variables display remarkable similarities for both groups, the data on businesses owned by migrants show that lateral rural migrants are more likely to operate traditional rural enterprises (agriculture, fishing, and forestry) and create more jobs. In comparison, counterurban migrants are more likely to be associated with arts and crafts and retail enterprises. In terms of migrant motivations for their last change of address, notable differences are observed. Quality of life reasons, in common with previous studies, are significantly more important amongst counterurban migrants, whilst amongst lateral rural changes of address, one in every four was employment motivated. This latter finding further supports increased attention being given to lateral rural migration as distinct from counterurbanisation studies. Nevertheless, whilst migrant motivations can be assigned to broad categories (quality of life, retirement, housing, employment, etc.,) interview excerpts demonstrate that the decision to move is a highly personalised set of circumstances and considerations. Researchers should be careful not to overgeneralise the motivations of in-migrants.
In calling for more research into the totality of rural in-migration, and for greater attention to be given to lateral rural (and English and different forms of return) migration flows in particular, the current paper is not without its weaknesses. The study areas were selected on the basis of high levels of middle-aged or older in-migration so that the migration processes identified here may be very different from those experienced in other rural areas. Indeed, inmigration by younger age groups may be even higher in other rural areas than recorded here. The original study was designed with the expectation that counterurbanisation would dominate. Undoubtedly, a design phase that acknowledged different and varied migration flows would have resulted in a more comprehensive quantitative and qualitative dataset. Nevertheless, this paper is intended to stimulate debate and help deliver a research agenda that encompasses the messiness of contemporary rural migration. Rather than continuing with a dominant narrative that oversimplifies rural in-migration into a unidirectional counterurban flow, it is time that rural research sought to disentangle the individual migration strands and become reacquainted with the complexity and totality of rural migration flows and processes.