Illeris, Sv.: Recent Development of the Settlement Systems of
Advanced Market Economy Countries. Geografisk Tidskrift 78:
49-56. Copenhagen, June l, 1979.

In many countries of Western Europe and Anglo-America as well as in Japan, a shift in the distribution of population and economic activities from the higher to the lower levels of the settlement systems has taken place since 1970. Data on this shift åre presented, and possible explanations discussed.

Sven Illeris, Mag.scient. Research Institute on Local Government
and Planning. Nytorv 11, DK 1450 Copenhagen K.

The purpose of this paper is, firstly, to present evidence on the recent development of the settlement systems of a number of countries in Western Europe and Anglo as well as of Japan, where growth since 1970 mainly has taken place ind the medium and lower levels. Secondly, as this break with earlier tendencies of concentration into metropolitan areas had not been anticipated and is difficult to explain through conventional theories on the distribution of population and economic activities, it is attempted to discuss possible explanations of the observed development.

DISTRIBUTION OF POPULATION

Data on the distribution of population åre easily obtainable, frequently updated, and show important characteristics of settlement systems. A first attempt to describe recent developments must be basedon them.

Vining & Kontuly (1978) have collected regional migration statistics from a large number of European and Asiatic countries. These data show a shift in almost all advanced market-economy countries before or about 1970. Earlier the direction of net migration was to the metropolitan regions from other regions. Since 1970 migration is much more balanced, or there is even net emigration from (even over-bounded) metropolitan regions. The detailed picture varies from country to country and from years of economic boom to years of recession, but nowhere anything like a return to the tendencies of the 60's has been observed.

For the United States, Vining & Strauss (1977) have
shown a similar reversal in the devolopment of the distribution

of population. They used the Hoover index of
concentration, defined as

p. is the fraction of a nation's population
in subarea i in year t,
a. is the fraction of a nation's land area
in subarea l,
k is the number of subareas.

This means that if the population is perfectly uniformly distributed, H? will equal 0. If all of the population is located in one subarea, and if the subareas åre of negligible size relative to the total area of the nation, then H/ will approach 100.

Figure l shows the indexes calculated for 1900-1974 (and projected for 1980) for five different systems of areal subdivisions of the United States, reaching from 3000 odd counties to 9 large »geographical divisions«. It is clear that on the »geographical division« level, concentration has diminished throughout the century. On the county level, concentration increased up to 1970, but has declined since then.

For the study of settlement systems, regional data åre not sufficient, but up-to-date population data on individual settlements or meaningfully defined sizeclasses åre not so often available. Table l and 2, offering data from the United States and Denmark, have been selected because they illustrate the development of small towns both with and without commuting to major cities. Table l (Morrisen 1978) clearly shows a shift from metropolitan to non-metropolitan growth, mainly due to a change in migration patterns. Among metropolitan areas, data in Berry (1976) show that smaller ones now grow more than bigger ones, the latter having net outmigration. Among non-metropolitan counties, the fastest growth is experienced in counties with heavy outcommuting (extended suburbanization). But also entirely rural counties far from major cities show high growthrates.

In Denmark, where some decentralization of population could be observed already in the 60's, the present tendencies resemble those of the United States: Stagnation in the Copenhagen region (since the middle 70's even decrease); and accelerated growth in medium and small settlements (down to 500 inhabitants) both close to and far from major cities.

Wårneryd & Persson (1979), Dalmasso (1979), Cori (1979) and Drewett (1979) have presented material from Sweden, France, Italy, and Canada which show the same tendencies as tables l and 2. Figure 2 shows that in France, all settlement classes with over 1,000 inhabitants (except the Paris agglomeration) now show considerable growth.

In spite of many differences in the details from country to country, it may safely be concluded that since 1970, most aavanced market economy countries have experienced a shift in the population distribution from metropolitan areas to medium and small settlements, due to a changed pattern of net migration. Small settlements have expanded both within commuting distances from major cities (intra-regional decentralization) and far from major cities (inter-regional decentralization). This reversed development has been observed both in years of economic boom and in years of recession and unemployment, and cannot be regarded as a minor or unsignificant fluctuation.

DISTRIBUTION OF ECONOMIC ACTIVITIES

Another important aspect of the settlement system is the composition of economic activities in the different parts of the system. However, up-to-date statistics on the location of economic activities åre not always available. In this paper, only Danish data will be used. But som corroboration has been found in other countries, e.g. concerning manufacturing in the United States (Summers et al. 1976,

Sternlieb & Hughes 1977, Drewett 1979) and concerning
activities in the Nordic countries (Ekestam 1979).

Employment in agriculture has long been decreasing in advanced countries. However, in several countries the rate of decline has levelled off in the 70's: we seem to approach the bottom. Compared to earlier decades, this is a factor of stabilization for the small settlements and their tertiary activities.

In most advanced market economy countries, total employment in manufacturing has been stagnating or slightly declining in recent years. In Denmark, the 50's and 60's already witnessed some decentralization of manufacturing employment. From table 3 is seen that the shift in the 70's has been massive. It was observed both in the growth period up to 1973 and in the recession years after 1974. Small industries åre now widespread in all settlement classes down to 500-1,000 inhabitants.

The decentralization in the 60's mainly took place in low-wage branches. As illustrated by table 4, in the 70's the shift towards smaller settlements is still fastest in the low-wage branches (textiles, apparel, wood, and furniture). But in contrast to the 60's, high-wage branches (paper, printing, and chemicals) now also seem to decentralize. Numerous examples of flourishing, technologically advanced firms in small settlements åre found.

Data on private services åre sadly deficient. It is well known, however, that the most expansive parts of this sector åre to be found among high-level functions such as business services. Furthermore, within important branches such as food retailing, a concentration into fewer, larger establishments is going on. It may therefore be assumed that employment as a whole is shifting towards the medium and upper settlement classes.

The only sector with vigorously increasing employment in recent years is the public sector. In the 60's, the growth of this sector in Denmark mainly took place in the metropolitan areas and medium-sized towns, while public services in villages were reduced. This development was partly due to the faet that the most expansive institutions (e.g. universities and specialized hospital departments) were high-level services. And partly to the tendencies of concentration into fewer and bigger units which took place within many sub-sectors (such as primary schools and local authority administration).

The employment development in the 70's is shown in table 5. It appears that employment growth in the public sector has been fastest in small settlements. To some degree, this may be explained by the changed distribution of the population base of the services. However, the vigorous total growth of the sector makes the autonomous developments within the sector more decisive. As table 5 shows, growth is now much slower in the state (i.e. high-level) services than in the communal (i.e. local) services, such as child-care. Furthermore, most of the intra-sectoral concentration tendencies observed in the 60's have now more or less stopped.

To summarize: In the 70's, many different manufacturing activities seem to have shifted from the upper to the lower levels of the settlement system, and public services to have expanded more in the lower than in the upper levels. At the same time, the levelling off of agricultural decrease has tended to stabilize the smallest settlements. The only economic activities which possibly still concentrate towards bigger settlements åre private services. It may be assumed that the tendencies observed in Denmark more or less hold true in other advanced market economies.

Settlement systems cannot be characterized only by the population and the economic activities located in the individual settlements and in size-classes of settlements. Among other things, the functional linkages between the settlements should be considered. However, data on such interactions åre scarce. It has not been possible in this paper to consider recent developments of the linkages.

SEARCHING FOR EXPLANATIONS

The »trend-shift« had not been foreseen and can hardly be explained by existing theories on the distribution of population and economic activites. The observed facts åre a challenge to alle diciplines studying the spatial structure and development of society. New explanations have been discussed in recent years, especially in America. In Europe, it seems that the questions mainly have been taken up in international cooperation, such as the Nord (Ekestam 1979), the EEC (Eriplan 1978), the Vienna Centre for Social Studies (Drewett 1979), and the lIÅSA (Hansen 1978).

Before referring the hypotheses that have been suggested, it should be mentioned that we omit »explanations« which assume that the »trend-shift« mainly is a short-term consequence of business cycle variations. As already stated, the reversed distribution has been observed during so many business cycle stages in so many countries that this assumption clearly can be rejected (although business cycles certainly may cause minor fluctuations in the development of settlement systems).

The reasons that have been suggested as explanations for the shift in tendencies åre listed in figure 3. Usually the location of economic activities (left side of the figure) åre supposed to determine the location of households (the right side of the figure); influences in the opposite direction have also been suggested, however. It should be noticed that the figure as a whole applies to the interregional development. If only the intra-regional development is considered — where the location of households can change more or less irrespective of the location of jobs — the right half of the figure is the most relevant one. It should also be noticed that it has not been attempted to list all factors influencing the distribution of households and economic activities. We have only been interested in factors which may contribute to explain the changes in the development. But it should be kept in mind that some of these changes consist in removal of restrictions, e.g. mobility restrictions. In this way, factors which latently have existed before, have now been allowed to play an increased role. An example may be household preferences for living in/staying in small settlements.

On the job side we have first listed the »basic« activities.
Agriculture, mining and tourism åre, of course, tied to

the natura! resources they exploit, outside the metropolitan areas. The stabilization of agricultural employment — in some countries also of mining — and the increase in tourism all tend to improve the job development in small settlements in the 70's, compared to the 60's.

The shift in the location of manufacturing from bigger
to smaller settlements may be due to a variety of factors:

First of all, it may be a consequence of government intervention (regional policies). However, the decentralization of manufacturing has also taken place in countries with little or no government policies, so the latter cannot be more than partial explanations. In this connection, also increased local efforts to attrået firms should be mentioned.

Next, changes in the internal economies or diseconomies of scale have been suggested — connected with the faet that big plants åre over-represented in big settlements and small plants in small settlements. In many countries it has been observed that small physical units recently have performed better than big ones (irrespective of the organizational mergers that take place). The reasons åre not clear, but it has been suggested that small units might be more flexible in critical periods, and that modem organizational structures (»profit centres«) favour physical decentralization.

Third, the importance of externalities may have changed. The declining importance of the transport of goods in the total cost structure has already been observed some time ago. In contrast, the need of advanced firms for frequent face-to-face contacts with business partners, research, government authorities, and other organizations has been stressed much — perhaps even exaggerated. At any rate, the consequences of improved means of transport (increase of aviation) and telecommunication (data-transmission) probably åre that restrictions on the choice of location åre reduced, and that

advantages of agglomeration decrease. It has even been suggested that the nearness to qualified and specialized organizations in metropolitan areas now may be outweighed by the better personal understanding of problems and the less bureaucratic attitudes of public authorities and business connections in small settlements.

Finally, the availability of different kinds of labour — i._e. the location of households — plays a role. In the 60's, when the supply of unskilled labour was insufficient in metropolitan areas, workers leaving agriculture undoubtedly contributed to the decentralization of low-wage branches. Even under the present conditions, where the demand for labour tends to be smaller than the supply, a similar mechanism may operate — and contribute to the reduction of wage differences and reduction of unemployment differences which have been observed. Today, however, the advantage of small settlements may consist of a labour-force with more cooperative attitudes, less absenteeism, and smaller labour turn-over. Furthermore, qualified staff-members may be more easily available in small settlements than earlier — for reasons listed at the household side.

Now turning to the »non-basic« or service activites, traditional theory tells us that the location pattern of this sector is a rather passive consequence of the location of its customers. However, it must be recognized that also autonomous forces influence the location of the ever-increasing number of tertiary jobs.

Within private services, jobs possibly concentrate in the medium and upper levels of the settlement system — contrary to the overall tendency. Therefore, it may be more interesting to study the public services — keeping in mind that this sector generally is broader in Western Europe than in the United States.

Decentralized growth in public services, which now seems to be more important than in the 60's, may be the direct result of policies aiming at equal accessibilities — i._e. policies whose goals åre external to the sector in question.

More often than not, sector policies åre determined by considerations of internal benefits and costs. In this regard, there åre examples that earlier goals of specialized services (e.g. in schools and hospitals) now have given way to goals of improved human relationships — which may mean that existing buildings åre maintained instead of new, large institutions being constructed.

Finally, the growth-rates of the different sub-sectors åre very important. As already mentioned, it seems that the fastest growth rates have shifted from high-level public services to local public services.

On the household side it is necessary to distinguish between inter-regional changes and intra-regional changes. The latter may be determined by the factors discussed below. The former must generally be accompanied by changed job-locations. As mentioned, it is usually supposed that »the man must come (near) to the job« (except services which must locate near their customers), but clearly there åre also cases where »the job comes to the man«.

Let us start with the relationship between household and job location and notice that for an increasing share of all households, namely retired people, they do not exist. This means that other factors become decisive — e.g. the attraction of mild climates has become important in the United States.

Paradoxically, the tendency for married women to become economically active, and thus for households to have two bread-winners, also means that the importance of living near one vvorking place decreases. The household can choose its residence more freely within commuting distance from both jobs — and if one bread-winner becomes unemployed, the propensity to move to other regions is modest.

Of course, the enormous increase in private car-ownership in advanced market economies also opens up for bigger distances between place of work and place of residence. But this factor does not itself cause changes of location. It only removes a restriction on the choice of location and makes it possible for other factors — old and new — to operate more freely. Increased mobility is a necessary, but not a sufficient condition for decentralization.

Next, public intervention should be mentioned. However, the goals and tools vary so much from one country to another, and the administrative systems åre so different that public intervention hardly can be regarded as a decisive cause of the general tendencies observed in all countries. Intra-regional planning nowhere aims at extreme dispersal of house-building, nor at extreme concentration of population. But different intermediate solutions åre foreseen in different countries, including the new town policies of some countries. Differences in taxation may in some countries lead to — unintended — differences in the attractiveness of central cities and surrounding areas. Since 1970, many rural and suburban communes in Denmark have actively purchased land and offered it for house-building — whereas some urban communes have experienced bottlenecks in the road- and sewage construction etc. necessary for development. However, as mentioned below, the development of land prices indicate that not only land supply, but also land demand factors operate.

It has often been suggested that the differences in land prices between big and small settlements, combined with the increased demand for one-family houses, might be the decisive factor behind the intra-regional decentralization of households. Undoubtedly, this is -an important element in the explanation. But land price differences actually tend to become relatively smaller, whereas household decentralization if anything seems to accelerate.

If land supply at different prices were the only
cause, a deceleration of household decentralization
should have been expected.

Therefore we have to consider the demand side — household preferences — more closely. It has been suggested that what may broadly be called environmental factors play an increased and changed role. Increased: because easy access to natural amenities, better social contacts in small communities, in some countries a good climate, etc. may receive heavier weights in the preference structure than before. And changed: because telephones, cars, TVs, and — except in small villages — improved services have broken the former isolation of small settlements — whereas pollution, heavy traffic, crime and terrorism, and in some countries racial problems mean that metropolitan environments have deteriorated. Some of these social diseconomies of scale may also influence the tax bilis — but as mentioned, that depends on the different taxation systems.

To summarize: A considerable number of factors have been suggested as explanations for the changed distribution of population and economic activites. They may all contribute to Gibbs' (highly abstract) phase of decreasing concentration in advanced economies. The next research problem will be to study the relative importance of the different factors and their role in the general societal development. A few attempts have been made, using different approaches:

The traditional sociological method of interviewing people about their motives has been used i._a. by Mogensen, Mørkeberg & Sundbo (1979). They found that most respondents, having moved to small settlements, did so in order to purchase a one-family house cheaper than elsewhere. However, when the purpose is to study a change in the development, the method is difficult to apply: Interviews from before the change åre lacking. And so far as the change is due to people not leaving small settlements, as similar people would have done a decade ago, this method is simply unable to grasp the motives.

Another approach to the relative importance of different factors is through multiple regression analysis, which has been taken up by Beale (1977), McCarthy & Morrisen (1977), and Morrill (1978). They try out different independent variables to explain the distribution of population. They agree that the amount of manufacturing has lost in explanatory value, whereas retirement and recreation, development in tråde and manufacturing employment, and environmental indicators have gained in importance.

CONCLUSION

A number of factors have been suggested as responsible for the observed shifts in the distribution of population and economic activities. Only few and unsatisfactory attempts have so far been made to study the importance of

these factors in the general societal development. Since
our understanding still is so weak, it is not possible to
predict the future development of the settlement systems.

However, as among others Zelinsky (1977) and Berry (1978) have stressed, profound changes in the perception of economic development and in peoples' values and attitudes have occurred since 1970 in advanced marketeconomy countries. The shifts in the development of the settlement systems åre connected with these changes, and there is no reason to expect any return to earlier phases in societal development.

RESUME

I artiklen fremlægges data om den forskydning af vækstens placering fra store til mellemstore og små byer, der siden 1970 er sket i Vesteuropa, Angloamerika og Japan. Dette brud med den hidtidige koncentration var ikke forudset og kan vanskeligt forklares gennem de eksisterende teorier om befolknings- og erhvervsfordeling, det forsøges derfor at fremlægge tænkelige forklaringer.

I figur l og 2 samt tabel l og 2 fremlægges eksempler på forskydningerne i befolkningsf or delingen. Det konkluderes at forskydningerne, trods forskelle fra land til land, er veldokumenterede og beror på ændringer i vandringsmønsteret. Endvidere at små bebyggelser nu vokser både inden for pendlingsafstand fra større byer (intra-regional decentralisering) og langt fra større byer (inter-regional decentralisering). Og at forskydningerne er konstateret både i år med højkonjunktur og arbejdsløshedsår.

I tabel 3-5 belyses erhvervsfordelingen med tal fra Danmark. I 70'erne har industribeskæftigelsen — nu også i højlønsbrancher — forskudt sig fra store til små byer. Offentlig service har ekspanderet mere i små end i store byer. Tilbagegangen i landbrugsbeskæftigelsen er aftagende, hvilket medvirker til at stabilisere de små bebyggelser. Kun i privat service synes der stadig at foregå en koncentration.

En diskussion om årsagerne til denne konstaterede udvikling er blomstret op i Amerika og inden for internationale forskningsorganisationer. Figur 3 giver et overblik over de faktorer der kan forklare ændringerne.

Industriens forskydning kunne bl.a. skyldes det offentliges egnsudviklingspolitik, men ikke alle lande har ført en sådan. Endvidere at stordriftsfordelene i virksomhederne kunne være aftagende. Fly, data-transmission m.v. kan have formindsket kontaktafhængige virksomheders fordele ved at være lokaliseret i en storby. Endelig synes arbejdskraften i småbyer at være mere stabil end i storbyer, dvs. at husstandenes bosætningsvalg også får indflydelse på erhvervslokaliseringen.

Serviceudviklingen afhænger ikke blot af befolkningsunderlagets placering, men også af selvstændige faktorer. Den decentraliserede vækst inden for den offentlige sektor skyldes bl.a. bestræbelser på at gøre adgangen til service mere ligelig. Den koncentration inden for bl.a. folkeskoler og sygehuse, der skyldtes ønsket om bedre og mere specialiserede ydelser, er stilnet af, mens interessen for det interne sociale miljø er stigende. Endelig vokser den lokale service som helhed mere end den overordnede service.

Decentraliseringen af befolkningen antages i interregional
sammenhæng først og fremmest at skyldes arbejdspladsernes

decentralisering. Inden for de enkelte pendlingsoplande kom
mer yderligere faktorer til.

Det stigende antal pensionisthusstande er uafhængige a arbejdspladsernes lokalisering, og familier med to udearbej dende medlemmer kan også være tilbøjelige til at lægge mindn vægt herpå. Privatbilens udbredelse har også løsnet behovet foi at bo nær arbejdet. Fysisk planlægning, beskatningsforskelle jordpolitik kan påvirke bosætningsvalget, men her er stort forskelle fra land til land. Billigere grunde siges ofte at vær« afgørende for decentraliseringen, men grundprisforskellene ei aftagende. Der må derfor også ske forskydninger på efter sporgselssiden, og der kan her være tale om såvel præference ændringer som forbedrede levevilkår i småbyer og forringe miljø i storbyer.

Der er endnu kun foretaget få forsøg på at klargøre disse faktorers indbyrdes vægt og sammenhæng i den almindelige samfundsudvikling, og det er ikke muligt præcist at forudsige bymonstrets fremtidige udvikling. Nogen tilbagevenden ti tendenserne fra før 1970 er der dog ingen grund til at forvente

Reference?

Calvin L. Beale (1977): The Recent Shift of United States Popu
lation to Nonmetropolitan Areas, 1970-75. Internat. Reg.Sc
Rev. vol 2, no. 2, p. 113-22.

Brian J.L. Berry (1976): The counterurbanization Process Urban America Since 1970. P. 17-30 in B.J.L. Berry (ed.) Urbanization and Counterurbanization. Urb. Aff. Ann. Rev vol. 11. Sage Publ., Beverly Hills/London.

Brian J.L. Berry (1978): The Counterurbanization Process: How
General? P. 25-49 in N. Hansen (ed.): Human settlemenl
Systems. Ballinger, Cambridge Mass.

Berardo Cori (1979): Report on the Italian National Settlement systems. IGU Commission on National Settlemenl Systems & Polish Academy of Sciences: The National Settlement Systems. Warszawa.

Etienne Dalmasso (1979): Report on the French National Settlement System. IGU Commission on National Settlement Systems & Polish Academy of Sciences: The National Settlement Systems. Warszawa.

R. Drewett (1979): Urbanization Trends in Selected OECD
Countries. OECD Secretariat, (mimeo).

H. Ekestam (1979): Ortsystemets framtida utveckling. Nord
1.

European Communities' Concerted Research on the Growth of
Large Urban Concentrations. Final Report Interest Group 2,
Migration. Eriplan, Bruxelles 1978.

J. P. Gibbs (1963): The Evolution of Population Concentration.
Economic Geography.

Niles M. Hansen (1978): Preliminary Overview. P. 1-21 in N.
Hansen (ed.): Human Settlement Systems. Ballinger, Cambridge
Mass.

Sven Illeris (1979): Rapport om det fremtidige bymønster. Planstyrelsen,
København.

Richard, L. Morill (1978): Bases for Peripheral Growth. In: Pro
ceedings of the US-Hungarian Symposium on Urban Geogra
phy. Ann Arbor

Peter A. Morrison (1978): The Current Demographic Context o National Growth and Development. P. 473-79 in: L.S Bourne & J.W. Simmons (eds.): Systems of Cities. Oxforc University Press.

George Sternlieb & James W. Hughes (1977): New Reeional ane
Metropolitan Realities of America. Journ. Arner. Inst.Plan.
vol. 43 no. 3, July, p. 227-41.

G.F. Summers. S. D. Evans. F Clcmpntp FM Rcrk & l
Minkoff (1977): Industrial Invasion of Nonmetropolitar
America. A Quarter Century of Experience. Praeeer. N.Y.

Daniel R. Viningjr. & Thomas Kontulv (1978): Population Dis
persal from Major Metropolitan Regions: An Internationa
Comparison. Internat. Reg.Sc.Rev.. vol 3 no 1. P. 49-73.

D.R. Viningjr. & A. Strauss (1977): A Demonstration that the current deconcentration of population in the United States i; a clean break with the past. Environment and Planning A, vo 9 no. 7. P. 751-58.

Olof Warnervd & Per-Olof Persson (1979): Report on the Swedish National Settlement Systems. IGU Commission or national Settlement Systems & Polish Academy of Sciences The National Settlement Systems. Warszawa.

Wilbur Zelinskv (1977): Coping with the Migration Turn
around: The Theoretical Challenge. International Regiona
Science Review. Vol. 2. No. 2. pp. 175-178.

Kevin F. McCarthy & Peter A. Morrison (1977): The Changing Demographic and Economic Structure of Nonmetropolitan Areas in the United States. Internat. Reg. Se. Rev. Vol. 2 no. 2, p. 123-42.

G. Viby Mogensen, H. Mørkeberg & J. Sundbo (1979):Småbyer
og landdistrikter. Socialforskningsinstituttet, publ. nr. 86,
København.