Applied Geography (1982). 2. 22 I-230

0 1982 Butterworths

221

Migration planning: the case for Tunisia

Allan Findlay

Department of Geography, The Unioersity, Glasgow GI2 8QQ, Scotland

Abstract

Many government policies indirectly influence migration patterns, ye! relatively few governments have sought to relate these diverse influences in such a way as to form an integrated migration strategy. The situation is illustrated from the Tunisian experience. It is agreed that knowledge gained from spatial analysis of migration systems can be directly applied to increasing the efficiency of government investments designed to temporarily modify the orientation and volume of migration flows. From a review of the possible measures available for altering migration patterns it is concluded that longerterm changes in patterns of population redistribution can be achieved only by migration strategies which either (i) seek to modify the urban bias in the aspirations of migrant populations. or (ii) plan to fundamentally alter the structure of national settlement systems.

Introduction

Policies for planning internal migration tend to reflect the personal predilections of policy makers. This is not surprising in view of the absence of any recognized structure for migration policy formulation and in view of the almost intractable nature of the problems which migration policies seek to solve. No single policy can be defined as a blueprint solution which will be applicable to all situations, but there are certain characteristics common to most migration patterns which migration planners should consider if they wish their policies to be relevant and effective.

Migration planning can be defined as an attempt to bring about a change in population distributions, through measures influencing population mobility. Clearly, policy makers may wish to modify population patterns in a number of different ways, with both the objectives and policy measures being governed by the ideological basis of the decision-making process. Policy objectives may vary as a result, from strategies which attempt to restrict or redirect migration flows, to those policies designed to stimulate selective new migration streams to specific destinations. Measures for implementing migration policies have ranged from fiscal measures to encourage worker or family migration, as in the case of the British assisted mobility policy (Beaumont 1976) through to forceful directives arising from attempts to modify political space rapidly, such as the population dispersal policy of Israel (Romann 198 1). While these examples illustrate positive migration strategies, government influence on migration in most countries. and most notably in the Third World, has come largely through the indirect effect of measures not designed to implement migration policies but to achieve some other goal. The majority of government measures do, however, have some indirect effect on patterns of population mobility through their differential influence on the regional or local welfare of specific populations.

In the Maghreb a number of different policies have been adopted with regard to

222 Migration planning: the casefor Tunisia

international migration, but policies directly impinging on internal migration have scarcely been considered and those measures which have been introduced to influence the volume and direction of population movement have been naive in their conception and implementation.

This paper examines the specific case of Tunisia, not merely to demonstrate the inefficacy of forceful intervention to modify migration patterns, and the contradictions arising from an incoherent approach to migration planning, but to illustrate some broader principles. Comprehension of the basic patterns of the Tunisian migration system is shown to give rise to specific suggestions for policy formulation, and to the application of measures which are sensitive to the mechanisms underpinning the spatial dimensions of the migration process.

Trends in Tunisian migration

Temporal and spatial specificity is an essential ingredient in migration policy formulation. Four characteristics of the Tunisian migration system will be discussed to illustrate this point.

The structure of Tunisian migration has changed considerably with the passage of time. This is evident from analysis of the two most recent national censuses held in Tunisia in 1966 and 1975. Both censuses report on patterns of migration from birth-place, as well as giving information on movements from place of last residence (1966 census) and place of residence in 1969 (1975 census). Table 1 offers a comparison of migration movements as classified by urban and rural origins and destinations. In the absence of any satisfactory functional definition of Tunisian urban areas, localities (dklbgations) have rather arbitrarily been defined as urban where more than 50 per cent of their population lived in urban communities.

While 35.7 per cent of the migrants recorded in the 1966 census had moved from localities of primarily rural character to urban zones, between 1969 and 1975 this category of migrants had shrunk to only 18.7 per cent of the total. Contemporaneously inter-urban

Table 1. Migration from rural and urban localities (N)

Previous place of residence (date unspecified)

Urban Rural Total

Residence in 1966 Urban 26.7 35.1 62.4 Rural 15.9 21.7 37.6 Total 42.6 57.4 100.0

Place of residence in 1969

Urban Rural Total

Residence in I975

Urban 50.3 18.7 6Y.O

Rural 16.3 14.7 31.0

Total 66.6 33.4 100.0

Source: Calculated from the results of the 1966 and 1975 census reports

A. Findlay 223

migration accelerated extremely rapidly to become the dominant form of population movement. The implications for policy making are evident: no longer will measures designed to stem a rural exodus be the most appropriate instrument for policy makers wishing to halt further migration into the largest cities.

A survey of migrant job seekers in Tunis in 1977 revealed that 88 per cent came from other urban centres (Findlay 1980b; OTTEEFP 1978). Most of these were from the smaller and medium-sized towns of northwestern Tunisia rather than from rural communities. In summary, a transition in the character of Tunisian population mobility has occurred during the last decade, the volume of rural-urban movements being exceeded by urban-urban migration.

A second noteworthy characteristic of Tunisian migration patterns is the very great regional variations which exist in the significance of population movements. The force of out-migration has been much greater from the mountainous northwestern districts and from the far south of the country than it has been from the pastoral steppelands of central Tunisia. Making reference once more to census statistics on migration, it has been shown by researchers that patterns of in-migration rates decline with increasing distance from the country’s largest city, Tunis. This is true both for the results of the 1966 census (Findlay 1980a) and for the recently published 1975 census (Findlay 1982). By contrast, only very low rates of in-migration have been recorded in the peripheral regions of the country.

A third characteristic of internal migration in Tunisia of significance to planners is the relatively low ‘efficiency’ of migration movements. Efficiency is a term which was originally proposed by Shryock (1959) to describe the ratio of net migration to gross migration. Although having a number of shortcomings [such as those recently put forward by Sternstein (1979)1 it remains a useful measure for exploratory data analysis. In Tunisia between 1969 and 1975, migration efficiency was only 24.3 per cent. That is to say, only one-quarter of migration movements resulted in a net redistribution of population. The vast majority of migration flows were closely balanced by counterflows in the opposite direction. Net migration flows of greater than 500 persons are mapped in Fig. 1, and emphasize the spatially specific nature of population redistribution from the northwestern gouvernoruts of Tunisia towards the capital. By comparison with this trend redistributive movements between other regions were relatively minor during the same period. It can be seen that in the Tunisian migration system out-migration has not been experienced from all the rural areas of western and southern Tunisia, nor has in-migration been significant in all the urban settlements of the eastern littoral. Net migration has accentuated the spatial concentration of population in the Tunis region. Net flows to other areas with large settlements, such as Sousse and Sfax, have been relatively small. To generalize the Tunisian migration system only in terms of rural-urban drift would therefore be to misconstrue the true nature of population redistribution.

Fourthly, it is worth commenting on the age-sex characteristics of Tunisian migrants. As in most cases of population movement, Tunisian migration has been highly age and sex selective, mainly affecting men in the age cohort 20-29 years. Between 1969 and 1975. 11.6 per cent of all men between 25 and 29 years of age and 7.1 per cent of women of that age group, changed their gouvernorut of residence. By contrast, only 6.5 per cent of men and 3.6 per cent of women between 35 and 39 years of age were migrants.

The four characteristics of Tunisian migration which have been discussed highlight the need for careful spatio-temporal analysis of migration processes prior to policy formulation. The indiscriminate application of migration policies is inappropriate in view of the intrinsically selective spatial and demographic character of migration. Simplistic ‘blanket policies’ fail to recognize the ever-increasing proportion of migration from small and medium-sized towns.

It is unnecessary to repaint the already familiar picture of poverty and congestion

224 Migration plannirrg: the case for Tunisiu

ALGERIA I 1,

/ ,___._-__t i

,

i ,j

,’ ,’

“- Net Population Migration

* 500-999

.-b 1000-1999

-b 2000-3999

4000 -7999

--- Gouvernorat Boundaries

Figure 1. Net migration ofTunisians (1969%1975).

A. Findlal 225

associated with persistent in-migration to the capital city. Like many of its counterparts in the Third World, Tunis is encircled with ill-planned and poorly serviced ‘gourbivilles’ (squatter settlements). Large numbers of migrants initially fail to find work in the regulated sector of the economy and are forced to subsist in the already overcrowded small-scale or so-called ‘informal’ sector. Surveys of migrant job seekers in Tunis have clearly demonstrated that spatial mobility has not been accompanied by occupational and social mobility. On the contrary, it has often led to migrants living in poorer housing conditions than in their areas of origin, and to extended periods of casual employment and underemployment.

Many of the regions of net population loss through internal migration have also experienced heavy emigration to foreign labour markets. and migrant remittances have become the life-blood of many communities. The disintegration of the social fabric and economic structures of these regions of departure has been widely observed (e.g. Koelstra and Tieleman 1977). As a result the government and many social commentators view the persistence of present migration trends as undesirable, and yet all conscious attempts at altering migration patterns have thus far been in vain. A national settlement strategy is urgently required which will either reduce or redirect the volume of net migration or else plan to cope with the problems resulting from continued population redistribution.

Government policies on migration

The Tunisian government has long been aware of the urban problems arising from in-migration to Tunis. During the colonial era the French army was known to establish two defence rings round the city to prevent displaced nomads from the steppelands from entering the capital and attempting to settle there (Clarke 1956). More recently Centres d’Accueif et d’orientation (Centres for the redirection of migrants) were established whose explicit function was to redirect poorly integrated migrants to their regions of origin. Centres were set up in other cities, such as the one in Kairouan, which handled between one-fifth and one-quarter of the rural migrants entering the holy city between 1968 and 1973 (Trabelsi 1976).

Further forceful intervention in the Tunisian migration system followed the riots of January 1978. Politicians openly admitted that unemployed persons who had moved to the cities from rural areas would be sent back home. The creation of an army of compulsory labourers, to be employed temporarily in regional development projects, was seen as a ‘means of removing the jobless from the city streets, thus depriving militants of an opportunity to stir further unrest’ (Hill 1978).

In spite of these examples, it should be noted that Tunisia has had very few restrictive policies concerning migration in comparison with Tanzania (Riddell 1978) or the countries of Eastern Europe (Fuchs and Demko 1978). All existing evidence suggests that forceful intervention in the internal migration process is extremely difficult and costly to implement. It is impractical to enforce such policies and interference in patterns of personal mobility may also have undesirable political repercussions.

Unplanned intervention

It is valuable to consider the incidental impact on migration of other Tunisian legislation and planning policies. Strategies such as the encouragement of tourist complexes along the eastern littoral have not been instituted with reference to specific migration goals, yet their impact on migration is considerable.

Despite the extensive aggregate impact of government policies on migration trends, the result is dysfunctional since the effect on migration is uncoordinated and does not lead to

226 Migration planning: the casefor Tunisia

an improved organization of human resources. The probable effects on Tunisian migration of a number of government programmes are proposed in Table 2. Since the government bodies and institutions listed in the table have emerged from legislation which was not designed to achieve specific migration goals, it should be no revelation to discover that the impact on migration of these different initiatives has frequently been contradictory.

One example of this is the contrasting influence on migration of industrial and agricultural investment policies. The availability of industrial investment incentives under the government’s ‘open door’ development policy has stimulated considerable foreign investment in the most profitable locations of northeastern Tunisia, and left the remainder of the country almost bereft of new employment opportunities in manufacturing industry. With 56.9 per cent of the new jobs created between 1973 and 1975 by the open door policy being located in the northeastern gouvernorats, it is hardly surprising that a strong desire exists in other parts of the country to migrate to this region (INS 1975).

The migration impact of industrial investment policies conflicts with that of projects for rural reconstruction. Attempts to stabilize rural populations have been less than successful because of the high aspirations of young workers in peripheral areas and their desire to share in the opportunities available to urban youth.

It is impossible to assess the precise quantitative impact on migration of all government projects and policies. On balance it appears that the influence of current Tunisian policies

Table 2. Policies and programmes indirectly influencing migration

Policy/programme Effects

I. Improvement of urban fabric (Sociktt: Immobilitre de Tunisie: Fonds National d’Am8nagement de I’Habitat)

2. Improvement of urban services (Office National d’Assainissement, and others)

3. Improvement of national infrastructure (Min. de Travaux Publics)

4. ‘Open door’ industrial investment (Agence de Promotion des Investissements)

5. Tourist developments (Office National de Tourisme)

6. Manpower training and employment agencies (Office des Travailleurs Tunisiens B VEtranger. de I’Emploi et de la Formation Professionnelle)

7. Centralization of decision-making 8. Placement of government employees 9. Industrial decentralization (Fonds de

Promotion et de D&centralisation Industrielle)

10. Reafforestation

11. Extension of irrigation 12. Rural development programmes

(Programme de Dkveloppement Rural)

Encourages migration to the largest cities

Encourages migration to the largest cities

Encourages migration mainly towards northeastern Tunisia

Encourages migration mainly towards northeastern Tunisia

Encourages migration to the eastern littoral

Encourages migration to foreign labour markets, and facilitates migration towards the northeast

Encourages migration to Tunis Encourages migration to peripheral regions Attempts to discourage out-migration

Encourages short-distance moves to new settlements

Discourages out-migration Attempts to discourage out-migration

A. Findla_v 227

has been to accentuate rather than to moderate trends towards further population concentration in the northeast of the country. A strategy which could resolve at least some of the conflicting influences on migration of government legislation would be a major step towards migration planning and might in some instances obviate the need for further state intervention in the migration system.

Towards a spatial policy for migration planning

The planning policies employed by the Tunisian government at the present time are inadequate and inappropriate to deal with the problems created by migration. They are inconsistent in their effect on migration and do not have the correct spatial bias, increasing rather than reducing the problems of spatial imbalance. It becomes pertinent to demand what type of policy would be appropriate to the Tunisian case and whether this type of approach would contain more general applications. It is suggested that a spatial policy for migration planning should be employed.

Three types of mobility are involved: spatial mobility, social mobility, and occupational mobility. Any attempt to influence spatial mobility must equally consider the other two aspects of mobility. Two approaches to changing patterns of spatial mobility may be considered. Firstly there are short-term policies aimed at reducing the hardships caused by population redistribution. Ameliorative policies treat the symptoms arising from migration rather than adopting measures to encourage new and more satisfactory migration patterns. Secondly there are policies which may take a longer time to implement, but which would by their innovative nature have a more lasting effect.

Spatially differentiated taxation systems, subsidies on rural incomes, rural industrial- ization and measures to improve infrastructure and local services have been proposed as means of reducing the pressure to migrate from rural areas (Weiner 1975; US Bureau of Census 1977). Similar measures need also to be considered for the urban communities of the less favoured areas since these appear to have become critical nodes of out-migration. There is a danger, however, that intervention will facilitate rather than discourage out-movement. For example, the establishment of two small factories in the steppeland town of Hajeb el Ayoun did not initially halt out-migration, but on the contrary it stimulated local aspirations to achieve still greater occupational and social mobility through out-migration (Trabelsi 1975).

One method of controlling the rate of migration to the metropolis would be the regulation of job recruitment in the large-scale ‘modern’ sector of the economy (Gugler 1976). At present a ‘queuing system’ appears to operate whereby migrants move to the largest cities, and work in the so-called ‘informal’ or small-scale sector, while looking for more permanent employment in the regulated sector. Migration to the city has become the first step in the job search procedure. In short, spatial mobility has become the key to occupational and social mobility in the eyes of many migrants. Hay’s survey of migrants in Tunis revealed that 17 per cent did not find any employment in the first month, while 50 per cent had to accept employment initially in the small-scale sector (Hay 1974). A policy which made jobs available to persons still resident in the towns and villages of northwestern Tunisia would discriminate against speculative migration and might reduce the number of unsuccessful migrants forced to subsist for long periods in the small-scale sector. The policy of allocating city jobs to potential migrants still living in zones of out-migration seems practical in the Tunisian context because of the well-established network of government employment bureaux across the country. It represents an interesting form of intervention capable of modifying some aspects of the Tunisian migration system.

Innovative planning intended to achieve long-term changes in migration patterns

228 Migratiorz piarznirlg: the cme,fbr Tutrisk

requires much more far-reaching intervention in the social and economic structures influencing migration than has been proposed thus far. A recent national survey of school-leavers established that only 3 per cent wished to enter agricultural employment although 43 per cent of the present workforce is employed in this sector. This indicates the pressure to find urban employment which exists amongst young job seekers (Clemenceau and Hadjadj 1976). The scope of migrant aspirations extends beyond the desire for secondary and tertiary employment and encompasses the ambition to live and work in a modern environment and to be absorbed in a cosmopolitan milieu. For example. in Nefza, a village in northwestern Tunisia, 75 per cent of local job seekers claimed that they would migrate to another part of their gowernoraf or further afield if the opportunity arose. despite prospects of future employment in a local shoe factory. Amongst the job seekers 24 per cent had already worked for a short time in Tunis and it seemed likely that many of them would return. In the long term the government might hope to modify migrant aspirations, through a re-emphasis in the education system of the significance and value of rural life-styles, but it seems probable that in the foreseeable future most young job seekers will continue to hope for modern-sector jobs in the nation’s largest cities.

Imbalance in the spatial distribution of large centres is one of the fundamental causes of interregional population redistribution in Tunisia. It is proposed that the policies most likely to achieve success are those which accept existing trends towards increased population concentration in the higher levels of the Tunisian settlement hierarchy. and which attempt the very major task of changing the hierarchy itself. rather than merely modifying the migration system in isolation from other factors. Effective intervention in the settlement system could occur at two levels in the hierarchy. Firstly, the largest regional centres in the northwest, the steppelands and the south could be strengthened. By increasing the intervening opportunities accessible to potential migrants, some out- migration to Tunis from these regions might be redirected to urban nodes at a lower level in the hierarchy. Secondly, a policy might be pursued to modify the nature of growth of the primate settlement of the system.

The development of provincial towns in northwestern and central Tunisia has been stunted, largely because their potential functions as industrial nodes and centres ofdecision making have never had the opportunity to grow. Under colonial rule and even following independence the only major role given to these towns was as market centres. For example, Beja, which had a population of 41 000 in 1975, is the largest town in the northwest, yet its only significant manufacturing employment is provided by a small sugar factory. As a consequence of the very limited interest taken in the potential of regional centres as nodes of non-agricultural employment and as centres offering high order services to surrounding villages, provincial towns have become a springboard to regional out-migration, instead of acting as catalysts to regional development.

Out-migration from rural areas and from small and medium-sized towns could be channelled towards preselected regional cities. Within each region of out-migration. the growth of no more than one or two cities should be encouraged in order that sufficient resources may be invested in these nodes to raise their urban status to the high levels aspired to by most migrants. As has already been demonstrated from analysis of census material. it is increasingly the residents of small towns who are departing to live in Tunis. It is these persons, as well as rural migrants, who must be attracted to settle in regional cities rather than migrating to Tunis.

Selective intervention in the Tunisian settlement system is most urgently required in the northwestern region. By raising the status of one of the towns of the region to that of a city, through the concentration of new employment opportunities and service functions. the scale of population movements might be modified from inter-regional to intra-regional migration. Multi-terminal network analysis of the spatial character of out-migration from

A . Findlay 229

the northwest has suggested that Beja would be the most appropriate node at which to intervene in the regional migration system (Findlay 1980b). Both rural-urban and urban-urban migration would persist at the intra-regional level towards such a centre. but regional imbalances in the distribution of human resources would be reduced and replaced by shorter-distance moves at the locality level.

An alternative strategy which might be adopted would be to encourage linear urbanization within the existing zone of attraction in the northeast (Abu-Lughod 1973). This policy would attempt to reduce the congestion of both physical and human resources caused by continued in-migration to the Tunis metropolis. Linear urbanization on the axis between Tunis and Bizerta would serve as an intervening opportunity to migrants from the northwest originally destined for Tunis or the settlements of the eastern littoral.

The absence of a coordinated migration policy in Tunisia is a policy in itself, and one which in the future may be shown to have had very unfortunate results. Intervention in the spatial organization of the Tunisian settlement system at either or both levels proposed would constitute a positive long-term migration strategy and would help to reduce the wide regional inequalities which currently exist in the country.

Conclusions

A major shortcoming of Tunisian national planning is the failure to specify a spatial policy for population distribution. Not surprisingly, social and economic policies have had a diverse and conflicting influence on patterns of internal migration. A number of ameliorative measures have been discussed which might reduce the volume of population redistribution. Failure of such measures in other countries has often occurred because they have been applied unselectively, not being designed specifically to influence the persons most prone to migration. The fundamental difficulty in slowing out-migration from towns and villages in peripheral areas remains the aspirations of many of the populace to live and work in large cities and to participate in the urban milieu. It is proposed that no minor fiscal incentives or regulatory mechanisms can make a lasting effect on the value systems of the population of potential migrants, and that the pressure to migrate towards the cosmopolitan environments of large cities will continue to escalate. Whether governments in developing countries consider the onerous task of restructuring their settlement systems to be a problem worthy of their attention is a decision which only they can take. Measures of lesser magnitude cannot be expected to have a serious long-term influence on the spatial organization of a nation’s ultimate resource-its population. Only measures capable of modifying either migrant aspirations or the spatial availability of opportunities for occupational and social mobility can hope to have any lasting impact on migration trends.

References

Abu-Lughod, J. (1973) Problems and policy implications of Middle Eastern urbanization. In Studies it? developmerlt problems it1 selected cowltries of the Middle East. pp. 42-62. New York: UNESOB.

Beaumont. P. (1976) The problem of return migration under a policy of assisted labour mobility. British Journal qf Industrial Relations 14. 82-88.

Clarke, J. (1956) A geographical stud,* qf nomadic migration iu Tunisia. Unpublished PhD thesis. University of Aberdeen.

Clemenceau. P. and Hadjadj. B. ( 1976) Contribution ri I’&ude des problPmes de l’enrploi et de la formation en Tunisie. Tunis: Ministire du Plan. Projet de Planification de I’Emploi.

Findlay, A. (1980a) Migration in space, immobility in society. In The changing Midddle Eastern city, (R. Lawless and G. Blake, eds). pp. 54-76. London: Croom Helm.

230 Migration planning: the case for Tunisia

Findlay. A. (1980b) Patterns and processes of Tunisian migration. Unpublished PhD thesis. University of Durham.

Findlay. A. (1982) Tendances recentes de la migration vers la ville de Tunis. Cahiers Gt!ographiques de I’Ouest 8 (in press).

Fuchs. R. and Demko, G. (1978) The post-war mobility transition in Eastern Europe. Geographical Reciew68 (2). 171-182.

Gugler. J. (1976) Migrating to urban centres of unemployment in tropical Africa. In Internatiorza( migration (A. Richmond et al.. eds) pp. 184-204. London: Sage.

Hay. M. (1974) An economic analysis of rural-urban migration in Tunisia. Unpublished PhD thesis, University of Minnesota.

Hill. P. (I 978) Tunisia to conscript unemployed. The Times. 3 March, 10. lnstitut National de la Statistique (1975) Recensement des Actkit& Industrielles, no. 7. Tunis: INS. Koelstra, R. and Tieleman, H. (1977) Dtkeloppernent ou Migration? The Hague: Remplod. Office des Travailleurs Tunisiens a I’Etranger de I’Emploi et de la Formation Professionnelle

(OTTEEFP) (1978) Le demande d’emploi d travers les bureaux de placement, am& 1977. Tunis: Minis&e des Affaires Sociales, OTTEEFP.

Riddell. J. B. (1978) The migration to cities of West Africa: some policy considerations. Jourrlal of Modern African Studies 16 (2). 24 l-26 1.

Romann. M. (198 I) Population redistribution policy in Israel: the special case of Jerusalem. In Policies qfpopulation redistribution (J. Webb et al.. eds). pp. 18 ! - 188. Oulu: IGU.

Shryock, H. S. (1959) The efficiency of internal migration in the United States. In international Union,for the Scientljic Stud\* qf Population. Vienna 1959. pp. 685-694. Vienna: IUSSP.

Sternstein. L. (1979) Migration streams. Paper presented to the Institute of British Geographers’ Conference on Population and Development, Cambridge. 19-20 September.

Trabelsi. M. (1975) L’industrie et son irnpac/ sur le de’reloppemerlt des :ones rurales. Tunis: CERES (Roneo).

Trabelsi. M. (1976) L’exode rural et son impact sur le diveloppement des villes regionales. Revue Tunisienne de Sciences Sociales 44. 147- I7 I.

US Bureau of Census (1977) Plaming.for internal migration. Washington DC: US Bureau of Census.

Weiner, M. (1975) Internal migration policies. In Policy sciences andpopulation (W. llchman et al.. eds). pp. 65-93. Massachusetts: Lexington.

(Revised manuscript receil)ed I5 February, 1982)