WORKING PAPER 97/03 INTERNAL MIGRATION AND REGIONAL POPULATION DYNAMICS IN EUROPE: POLISH CASE STUDY

Transcription

1 WORKING PAPER 97/03 INTERNAL MIGRATION AND REGIONAL POPULATION DYNAMICS IN EUROPE: POLISH CASE STUDY Marek Kupiszewski Helen Durham Philip Rees November 1996 School of Geography University of Leeds Leeds LS2 9JT United Kingdom Report prepared for the Council of Europe (Directoriate of Social and Economic Affaires, Population and Migration Division) and for European Commission (Directoriate General V, Employment, Industrial Relations and Social Affaires, Unit E1, Analysis and Research on the Social Situation)

2 CONTENTS ABSTRACT...ii LIST OF TABLES...iii LIST OF FIGURES...iii ACKNOWLEDGEMENTS...iv 1. CONTEXT INTERNAL MIGRATION AND POPULATION CHANGE REVIEWED METHODS USED AND DATA EMPLOYED GEOGRAPHICAL SCALE AND GEOGRAPHICAL UNITS VARIABLES Population and population change data Migration Births and deaths Unemployment KEY INDICATORS Population density Unemployment Distance to the nearest urban centre MAPPING METHODS SPATIAL PATTERNS THE PATTERN OF POPULATION CHANGE: , AND THE PATTERN OF NET INTERNAL MIGRATION IN THE DEMOGRAPHIC SOURCES OF POPULATION CHANGE THE IMPACT OF MIGRATION ON THE DISTORTIONS OF DEMOGRAPHIC STRUCTURES RELATIONSHIP TO THE URBAN SYSTEM GAINING AND LOSING URBAN SYSTEMS DECONCENTRATION WITHIN URBAN REGIONS RELATION TO THE DISTANCE FROM URBAN CENTRES RELATION TO POPULATION DENSITY GEOGRAPHICAL PATTERNS OF POPULATION DENSITY RELATION BETWEEN POPULATION DENSITY, POPULATION GROWTH AND MIGRATION RELATION TO UNEMPLOYMENT GEOGRAPHICAL PATTERNS OF UNEMPLOYMENT RELATION BETWEEN UNEMPLOYMENT, POPULATION GROWTH AND MIGRATION SUMMARY AND CONCLUSIONS BIBLIOGRAPHY i

3 ABSTRACT The report analyses population migration and change in Poland over three periods , and The analysis is conducted for communes and municipalities, the finest administrative division for which population and migration data are available. The results of the investigation show substantial changes in population migration and dynamics over time. The most important changes between the eighties and the nineties are the reduction of outmigration from rural areas, the increase in the number of communes and towns suffering from structural depopulation and the decrease of the role of largest urban agglomerations as population growth poles. The suburbanization process, visible in already in the eighties, has been reinforced, but no trace of counterurbanisation can be detected. The pattern of the upward hierarchy migration (from rural areas to towns to cities; from areas with higher unemployment to areas with lower unemployment) is still visible, but not as evident as in the past. Migration patterns of males anf females differ, females being keener to move upward hierarchy, males being more selective. Some similarities between migration and population development in Poland now and in some West European countries in the past have been identified, leaving, however, a lot of room for Polish specificity. ii

4 LIST OF TABLES TABLE 1: TYPES OF POPULATION GROWTH IN POLAND IN TABLE 2: AGE STRUCTURE OF RURAL POPULATION IN SELECTED REGIONS, POLAND 1994 TABLE 3: POPULATION GROWTH OF TOWNS AND CITIES BY SIZE CLASS TABLE 4: DISTRIBUTION OF POPULATION GROWTH AND NET MIGRATION IN 1994 BY TYPE (RURAL VS. URBAN) AND URBAN SIZE CLASSES, POLAND TABLE 5: NET MIGRATION AND MIGRATION EFFECTIVENESS RATIOS IN 1994 FOR MALES AND FEMALES BY TYPE (RURAL VS. URBAN) AND URBAN SIZE CLASSES, POLAND. TABLE 6: NET MIGRATION AND MIGRATION EFFECTIVENESS RATIOS IN 1994 FOR MALES AND FEMALES FROM RURAL LOCATIONS BY DISTANCE FROM NEAREST TOWN OR CITY OVER INHABITANTS TABLE 7: NET MIGRATION AND MIGRATION EFFECTIVENESS RATIOS IN 1994 FOR MALES AND FEMALES FROM RURAL LOCATIONS AND TOWNS BELOW BY DISTANCE FROM NEAREST CITY OVER INHABITANTS TABLE 8: DISTRIBUTION OF POPULATION GROWTH , POPULATION CHANGE AND NET MIGRATION IN 1994 BY DENSITY CLASS, POLAND TABLE 9: NET MIGRATION AND MIGRATION EFFECTIVENESS RATIOS IN 1994 FOR MALES AND FEMALES BY POPULATION DENSITY CLASS, POLAND. TABLE 10: DISTRIBUTION OF POPULATION GROWTH , POPULATION CHANGE AND NET MIGRATION IN 1994 BY UNEMPLOYMENT CLASS, POLAND TABLE 11: NET MIGRATION AND MIGRATION EFFECTIVENESS RATIOS IN 1994 FOR MALES AND FEMALES BY UNEMPLOYMENT CLASS, POLAND. LIST OF FIGURES 1: DEPOPULATION REGIONS IN POLAND AFTER EBERHARDT (1989) 2: POPULATION CHANGE IN POLAND BY COMMUNES AND MUNICIPALITIES, : POPULATION CHANGE IN MAJOR URBAN AGGLOMERATIONS, : POPULATION CHANGE IN POLAND BY COMMUNES AND MUNICIPALITIES, : POPULATION CHANGE IN POLAND BY COMMUNES AND MUNICIPALITIES, : POPULATION CHANGE IN MAJOR URBAN AGGLOMERATIONS, : NET MIGRATION RATE PER 1000 POPULATION IN POLAND BY COMMUNES AND MUNICIPALITIES, : NET MIGRATION RATE PER 1000 POPULATION IN MAJOR URBAN AGLOMERATIONS, : WEBB CLASSIFICATION 10: WEBB CLASSIFICATION OF COMMUNES AND MUNICIPALITIES IN POLAND, : SEX STRUCTURE OF POPULATION BY COMMUNES AND MUNICIPALITIES IN POLAND, : POPULATION GROWTH OF POLISH TOWNS AND CITIES BY SIZE : POPULATION GROWTH OF POLISH TOWNS AND CITIES BY SIZE : POPULATION DENSITY IN POLAND BY COMMUNES AND MUNICIPALITIES, : UNEMPLOYMENT IN POLAND BY COMMUNES AND MUNICIPALITIES, 1992 DATA RECALCULATED TO 1994 BOUNDARIES iii

5 ACKNOWLEDGMENTS This case study report is part of a wider study of Internal Migration and Regional Population Dynamics, jointly being carried out by the Council of Europe (Population and Migration Division) and the European Commission (Directoriate General V, Employment, Industrial Relations and Social Affaires, Unit E1). The study is funded by awards from the Council of Europe and the European Commission. Our thanks are due to Mr. Franco Millich (Council of Europe) and Ms Isabelle de Pourbaix (European Commission) for their guidance and advice in the project. Central Statistical Office in Warsaw kindly provided data on population migration and stocks. Director Lucyna Nowak and Mrs. Elzbieta Mscichowska offered explanation, guidance and advice. The Computer Centres of Wojewodship Statistical Offices in Warsaw (Mrs. Anna Branicka) and in Olsztyn (Mr. Bohdan Nowak) have written necessary programs and extracted data. Alina Muziol-Weclawowicz, Piotr Kupiszewski, Krystyna and Dorota Kupiszewska, Christine Macdonald and Maureene Rosindale typed at various stages of the project the data which were not available in machine readable form. Marek Jagodzinski digitised and Dorota Kupiszewska checked maps used in the study. iv

6 1. CONTEXT The enormous political, social and economic changes in Central and Eastern Europe that we have witnessed in the last decade present a substantial challenge to the international community. For wealthy old Western democracies the challenge is to help to integrate weaker Central and East European democracies into the Western World. For the former communist states the challenge is to develop their economies and political systems to make them compatible with the Western style democracies. This process goes on on at many levels and in many different places. Project on Internal Migration and Regional Population Dynamics in Europe initiated by the Council of Europe and financed jointly with the European Commission is an example of the contribution to the process of mutual understanding, adjustment, and learning from each other. One of the important aspects of the project has been to collate a database on migration and population change on a subnational level for the member states of the Council of Europe. Existing statistical publications of the Eurostat provide valuable data for the European Union for NUTS (Nomenclature of Territorial Units for Statistics) level 1, 2 and 3 (Eurostat 1995), but they do not cover non-eu member states. The excellent demographic yearbook produced annually by the Council of Europe (see for example Council of Europe 1995) covers all Europe (both Council of Europe member and non-member countries) but on the national level. The assembling of the most of the data remained therefore in our hands. The results of our activities were twofold: the database itself and a network of collaborators and experts in academic and governmental institutions all over Europe. An important aspect of the data collection, not addressed by ourselves, is the comparability of definitions adopted and operational methods used for the data collection (Poulain 1994). Being well aware of this limitation we have to accept it as is. Bringing the demographic data in Europe to a common denominator by standardising definitions and collection procedures is a major task for intergovernmental and international cooperation for the next decade. The success of such a task is doubtful as in each country demographic data collection is to a large extent driven by 1

7 administrative requirements and the legal system. The existing differences do not invalidate our conclusions even when they make direct numerical comparisons difficult. 2

8 2. INTERNAL MIGRATION AND POPULATION CHANGE REVIEWED Poland witnessed a population explosion in the post war period. In the 1950 s the natural increase of population was over 450 thousand per year and in the period exceeded the half million mark. This phenomenon has been attributed to the postwar demographic compensation process (Rosset 1975). In the sixties the increase of population almost halved in comparison to the fifties and recovered in the seventies, oscillating around 340 thousand per year. From the mid eighties we have observed a rapid reduction in the natural increase: from 371 thousand (10 pro mille) in 1983 to 95 thousand (2.5 pro mille) in These changes were very different in urban and rural areas, which have reached below 10 pro mille natural increase threshold in 1962 and 1984 respectively, almost twenty years apart, with massive rural-urban migration acting as a balancing factor. The importance of the migration has reduced dramatically over the last twenty years: net gains of urban areas reached a quarter of a million in 1975 and shrunk in 1994 to 39 thousand, that is to slightly over 15% of the 1975 value. As a result of these processes the population of Poland has been growing over the entire post-war period. It stood at 23.8 million in 1946 and reached 38.5 million in However this growth was solely due to the growth of urban places. In fact rural population decreased from 15.6 million (66% of total population) in 1946 to 14.8 million in 1980 and 14.7 million (38% of total population) in The urban growth was due not only to migration and natural increase but also to changes of administrative boundaries. The population change has a distinctive geographical dimension which can be fully examined only using small spatial units. A fundamental analysis of this process over time was given by Eberhardt (1989). Starting from the result of the National Census and administrative division in 1978 he went back as far as to 1946 recalculating the results of all post-war National Censuses (i.e. in 1946, 1950, 1960 and 1970) to the administrative division in As a result he obtained a time series of populations for each commune and municipality in the unified spatial framework. This allowed for the classification 3

9 procedure aiming at a delimitation of persistently depopulating (Figure 1). An extension of Eberhardt's study for the period can be found in Stasiak (1990) and for the period in Kupiszewski (1992). The findings of both studies confirm the observation that a substantial part of rural Poland has been continuously losing population. In this study we will examine if this still happens and to what extent. There is a number of studies analysing the problem of depopulation at the regional scale (49 units), but they will not be discussed here as, in our view, this scale is inappropriate for any meaningful geographical examination of population change. So far attention has been paid to rural areas, particularly those losing population. Traditionally this population has migrated to towns, cities and industrial centres increasing population concentration there. This was in particular the case over the period from the end of the World War II till the late eighties when urbanisation driven by industralisation attracted rural population. As an excellent example of this process is the Katowice region, to which massive industrialisation attracted migrants from the most distant regions of Poland (Dziewonski and Korcelli 1981). Jerczynski and Gawryszewski (1984) described the changes in the urban settlement subsystem of Poland in the second half of the seventies. Extension of this analysis till mid eighties has been offered by Jerczynski (1989). According to their findings the significance of migration in the dynamics of urban growth has been decreasing over time. The role of small towns has diminished and the role of medium sized towns has grown. This was confirmed by Korcelli who noticed that highly urbanized regions with big cities (Korcelli 1990) as well as big cities themselves (Korcelli 1989) have been loosing their importance as concentration foci. This study will examine if there is a continuation of the process outlined above. The insight into the demographic characteristics of the process of population change has been provided by Rakowski and Rakowska (1990) who constructed Webb typology of population change in all rural communes in Poland in 1976, 1980, 1983, 1986, 1987 and over the periods and As the authors did not recalculate the data to one common set of administrative division their results are burdened with much less accuracy then Eberhardt's, Stasiak's and Kupiszewski s studies mentioned earlier as well as the present study. This refers in particular to the most interesting typologies encompassing longer periods of time rather than a single year. This study will examine the Webb typology for 1994 and see whether there are any substantial modifications of historical patterns. 4

10 Figure 1. Depopulation regions in Poland after Eberhardt (1989) kilometres 200 Regions according to EBERHARDT: 1 Zachodnio-Pomorski 5 Górna Narew 2 Warmi ńsko-mazurski 6 Podlasko-Nadburza ński 3 Suwalski 7 Janowsko-Hrubieszowski 4 Kurpiowsko-Biebrza ński 8 Roztocze 9 Nadwiślanski 10Napilicki 11 Środkowej Warty 12Sudecki Source: Adapted from Eberhardt (1989)

11 3. METHODS USED AND DATA EMPLOYED 3.1 Geographical scale and geographical units The spatial scale was in fact predetermined by the availability of the data. The investigation was conducted for communes and municipalities for the analysis of population change and sex structure, interplay between natural increase and migration, and migration inflow and outflow. The investigation was conducted for 3030 communes and municipalities, which is roughly speaking an equivalent of NUTS level 5. All data were recalculated to the administrative division as in Some references are made to regional level (49 units called wojewodztwa) which are comparable to NUTS level 2. Later on in this study wojewodztwa will be referred to as regions. 3.2 Variables Population and population change data Stocks of population for communes and municipalities in 1984 and 1990 were recalculated to the administrative division in 1994 and compared with the stock of population in The method of recalculation was based on the computation of areas of communes and municipalities resulting from the changes of administrative boundaries, and relocation of populations attributed to these areas. An assumption of homogenous distribution of population in space which was adopted in these operations is adequate for most rural areas. For urban areas, in particular when a new urban administrative unit is created by division of a rural unit into two units, one of which is urban and the other one is rural, the method may introduce errors. However, the number of divisions described above is very limited and the error introduced does not invalidate the results. The full description of the procedures used is given in Kupiszewski (1992). We identified that in some, very few, cases this approach gave unreliable results, generating either very high or very low population change. This is mainly due to the fact that the assumption of homogeneous distribution of population or other variable over the space of administrative unit does not hold. Sometimes population is 6

12 highly concentrated, what may lead to wrong results of the recalculation. These errors are few (for the recalculation of 1990 population data to 1994 administrative boundaries there were 10 units for which we thought the results were suspicious) and do not have any impact on the identification of the overall trends. The data employed in the study are based on the National Population Censuses of 1978 and 1988 respectively and adjusted with the information obtained from annual population account tables based on the registration of demographic events (deaths, births and migrations) in each spatial unit (communes and municipalities). Such tables are constructed every year by the Central Statistical Office. Official statistics overestimate the population in Poland, and in consequence population of regions, communes and municipalities, as they do not take into account migration losses due to unregistered international migration. It is difficult to assess the magnitude of these losses. Rather cautious estimates have been offered by Korcelli (1992). Kupiszewski (1993, Table 1) shows that the discrepancy between the outflow of Poles reported by the Central Statistical Office over the period and the aggregated inflow of Poles reported by selected countries over the same period differs by as much as 1.24 million. For the difference was 420 thousand. The results of the census of population held in December 1988 allowed for a slight correction of the count of resident population. The correction has never been widely publicised but may be traced indirectly by the comparison of half-annual increases of population. The population increased by 183 thousand in the first half of 1988, by 9 thousand in the second half, when the Census took place, and by 132 thousand in the first half of In the second half of 1988 the total increase of population calculated as a sum of natural increase and net international migration was in the region of thousand. That means that the Census corrected the population count by thousand. But still over the three years period well over 200 thousand has not been accounted for. 7

13 Table 1: Types of population growth in Poland in Description of type of growth Class name (see Error! Referenc e source not found.) Population change Number of units in each class Positive natural increase exceeds negative net migration A Positive 766 Positive net migration is smaller than positive natural increase B Positive 476 Positive net migration exceeds positive natural increase C Positive 362 Negative natural increase is smaller than positive net D Positive 119 migration Negative natural increase exceeds positive net migration E Negative 46 Negative natural increase exceeds negative net migration F Negative 63 Negative natural increase is smaller than negative net G Negative 235 migration Positive natural increase is smaller than negative net H Negative 816 migration Unknown\unclassified 145 Source: Webb 1963 and computation from the data provided by the Central Statistical Office. Note: to keep the description of classes short a simplification has been made. Whenever we refer to relations between variables with different signs we refer to relations between their absolute values. The Central Statistical Office must not be blamed for this difference. To blame is the political system which existed in Poland in the late eighties. It forced people to hide the fact of emigration. This was done by neither reporting long term absence of a member of family during the Census nor registering migration with the administration. Discussions of the reasons of this behaviour have been offered by Okolski (1991) and Kupiszewski (1993). 8

14 3.2.2 Migration On the commune and municipality level only data on inflow and outflow from/to each commune was by sex available. After aggregation of the data into unemployment bands, density bands, distance bands and population size bands, the full matrices of flows for males and females between the bands was estimated using biproportional fitting technique. This allowed for the construction of the matrices of net migration between bands and calculation of the effectiveness of migration Births and deaths Data on births and deaths in 1994 have been provided by the Central Statistical Office. for communes and municipalities. They were used to construct Webb typology of population change for each commune and municipality Unemployment Data on unemployment expressed as a percentage of unemployed in the total labour force by commune and municipality in 1992 has been taken from the publication of the Central Statistical Office (GUS 1993) and recalculated to the 1994 administrative division of the country. As in the case of population an assumption was made that unemployment characteristics are homogeneous in each of communes. 3.3 Key indicators In order to make findings for over 20 countries comparable it was necessary to use simple and easy to compute indicators which are meaningful virtually everywhere. The indicators used in this study are population density, unemployment and distance to the nearest urban centre Population density Population density indicates the intensity of human settlement. It was calculated in persons per square kilometer and constitutes an index which is probably the most comparable for all European countries. However, differences in spatial organisation 9

15 and above all differences in altitudes, land cover and morphology may have an impact on the comparability of population densities in various countries Unemployment Unemployment data by communes and municipalities in 1992 have been published by the Central Statistical Office (GUS 1993). They were recalculated to the administrative boundaries as in 1994 in the same way as the population was recalculated. The rate of unemployment is perceived as an indicator of the performance of local labour market. It may be even used as a crude indicator of the health of local economy. The direct comparability of unemployment rates between countries may easily lead to misunderstandings. Certainly unemployment in the very liberal, on European standards, United Kingdom economy means something different than unemployment in highly regulated economies of France or Poland which in turn will differ from unemployment in a planned communist-type economy of Bielarus Distance to the nearest urban centre The distance to the nearest urban centre may be interpreted as a crude measure of the accessibility to higher level services. The distances between communes and towns and cities were defined as the distance, measured in kilometers, between their centroids, that is their geometrical centres. 3.4 Mapping methods Mapping methods have been described in Rees, Durham and Kupiszewski (1996). The rules set out there are used in this study. 10

16 4. SPATIAL PATTERNS 4.1 The pattern of population change: , and The spatial patterns of the changes of population of Poland over the period and (see Figure 2 and Figure 3) are very similar to the pattern for the period described in Kupiszewski (1992) (Figure 4) and is very similar to the ones presented in publications mentioned in section 2. Therefore sections 4 and 5 will be based to some extent on Kupiszewski (1992). In one sentence one may say that in the last fifteen years the population redistribution in Poland consisted of two simultaneous phenomena: rural depopulation and urban concentration. However, both the process of urban growth and the process of rural decline vary depending on the size of town or commune and its geographical location. In certain cases we deal with urban decline rather than urban growth and rural growth rather than rural depopulation. A substantial part of rural Poland has been and still is growing - sometimes as an effect of the process of formation of urban agglomerations, sometimes because of high natural increase. There is also very clear change of the trend over time: in the mid eighties the rural urban migration was a quite profound phenomenon and the resulting depopulation was quite rapid. The mid nineties witnessed only a marginal outflow and in consequence the reduction of the rural depopulation. The geographical patterns of the change will be discussed below. Both in the period and the majority of the territory on which rural depopulation occurred might be quite liberally delimited with a line drawn from Frombork to Opole, then along upper Vistula and San rivers and the state boundary. Inside the territory delineated above the depopulation process has a patchy pattern. It may be easier to identify territories which have been growing. Almost all such patches are located around large cities i.e. ring of communes around Warsaw, extending to the South-East as far as to Kozienice and Deblin with a kind of an appendix around Lukow and to the South of it. Similar, but much smaller rings are around Kielce and Radom, to the North and North-West from Lublin as well as in the Northern Poland: crescents between Olsztyn and Ruciane-Nida or around Ostroleka. It is clear, that these clusters either form urban 11

17 agglomerations in various stages of maturity or at least grow in the vicinity of medium size towns, often new (since the administrative reform in 1975) region capitals. Both for the periods and the areas suffering the most from rural depopulation are located in Bialystok and Suwalki regions. Simultaneously, in these regions are located towns which experienced a very high growth (in some cases more than 30% over the decade ): Bialystok, Bielsk Podlaski, Siemiatycze and Suwalki. Only the latter two of these towns maintained this speed of growth till As Dziewonski (1990, Figure 3.3) has shown these towns have grown on the expense of its rural hinterland. The extent of rural depopulation on the North-Eastern and Eastern boundaries of Poland is much larger than these two regions. In the period all except 5 rural communes adjacent to the state boundary with the former USSR and located between Frombork on the shore of Zalew Wislany up to Wielkie Oczy in the South-East (Przemysl region) have lost population. Their number increased to 9 for the period The term "Eastern Wall" has been coined, denoting the stagnating and depopulating East of Poland. On the regional level all regions which constitute the Eastern Wall were characterised by a negative migration balance of their rural population and positive (with the exception of Bialystok region) natural increase. Only in south-eastern regions of Przemysl and Krosno were migration losses of rural areas compensated with natural increase. There is another significant cluster of communes loosing population in the Lodz region and in the ring of regions surrounding it. Lodz itself gained population over the period whereas in the period was a loser, mainly due to high natural losses and a migration balance close to zero. The Western part of Poland presents a mosaic of growing and declining communes. It is possible to identify depopulating areas in Gorzow Wielkopolski, Koszalin and northern part of Szczecin and Pila regions. Depopulation occurs also in the Lower Silesia in the Northern part of Legnica regions as well as to the South from Wroclaw and in the belt between Otmuchow and Kietrz. There are also rural areas with a prominent growth of population. The most important two are located in the North of Poland in Pomorze Gdanskie, in particular in Pojezierze Kaszubskie (Gdansk region), but also in Slupsk and Elblag regions and in the South in Carpathian Mountains and at its foothills (roughly speaking in Bielsko-Biala, Nowy Sacz, Krakow, Tarnow and part of Krosno regions). These regions combine high fertility with either migration gains or moderate losses. 12

18 Figure 2: Population change in Poland by communes and municipalities, Population Change (1984 = 1) 1.3 or more 1.2 to < to < to <1.1 1 to < to <1 0.9 to <0.95 Less than Kilometers 100

19 Figure 3: Population change in major urban agglomerations, Poznan and surrounding area Lodz and surrounding area Katowice, Krakow and surrounding areas Total population, to <5,000 5,000 to <10,000 10,000 to <25,000 25,000 to <50,000 50,000 to <100, ,000 to <250, ,000 to <500,000 More than 500,000 Population change (1984 = 1) 1.3 or more 1.2 to < to < to < to < to < to <0.95 Less than 0.9 Warsaw and surrounding area 0 10 Kilometers 20

20 Figure 4: Population change in Poland by communes and municipalities, Population Change (1980 = 1) 1.3 or more 1.2 to < to < to < to < to < to <0.95 Less than 0.9 Unknown 0 50 Kilometers 100

21 The differences between population change over two decades and suggest that in the period there were some important changes in the pattern of population redistribution. This hypothesis is supported by the marked changes in overall migration. It is also interesting to check whether there is any impact and if yes, what is it, of the political and economic changes started by Balcerowicz after the fall of communism. To examine it, maps of population change between 1990 and 1994 were prepared (Figure 5 and Figure 6). The comparison will be made with the changes over the decade , that is the period of a rapid degradation of the planned economy in Poland. The first observation is that the number of units losing population reduced substantially from 1281 over the period to 1061 over the period The most marked changes occurred in Olsztyn and Suwalki regions in the North-East and in Piotrkow region in Central Poland. The number of communes loosing population in these regions in reduced rapidly in comparison to The maps of intensity of changes (not shown in this repport; calculated as population change ratio per year) also differ. In the eighties the changes were faster than in the nineties, even if their geographic distribution was similar. The slow down of population redistribution may be attributed to the more difficult economic situation of households in the nineties and the shrinking labour market. Large cities, which in the eighties were all gaining population turned into losers in the first half of the nineties. The increase of medium towns slowed down. The changes in the population redistribution may suggest that Poland evolves in the direction which many developed Western countries went in the seventies and eighties: parallel processes of urban concentration and suburbanisation. So far no trace of the counterurbanisation may be identified. 16

22 Figure 5: Population change in Poland by communes and municipalities, Population Change (1990 = 1) 1.3 or more 1.2 to < to < to < to < to < to <0.95 Less than Kilometers 100

23 Figure 6: Population change in major urban agglomerations, Poznan and surrounding area Lodz and surrounding area Katowice, Krakow and surrounding areas Total population, to <5,000 5,000 to <10,000 10,000 to <25,000 25,000 to <50,000 50,000 to <100, ,000 to <250, ,000 to <500,000 More than 500,000 Population change (1990 = 1) 1.3 or more 1.2 to < to < to < to < to < to <0.95 Less than 0.9 Warsaw and surrounding area 0 10 Kilometers 20

24 4.2 The pattern of net internal migration in 1994 The interpretation of the net internal migration poses a lot of problems. For example positive net migration may be due to inflow of skilled and unskilled labour attracted by an investment project or due to inflow of opulent elderly attracted by mild climate and good amenities. The net migration resulted from both processes may be similar but their economic and social meaning very different. Classical economic equilibrium theories link migration gains or losses with the economic attractiveness of a region, often measured as the GDP per capita and unemployment rate. This interpretation, however weak, may be useful in the situation when mainly demographic information is available. Figure 7 and Figure 8 show the pattern of net migration rate in Poland in Such a measure is quite volatile due to a short period of observation and general instability of migration. Its accuracy is high as no estimation was involved in the calculations. One third of communes and towns in Poland suffered migration losses below five persons per thousand population. Another third suffered from losses larger than that and the remaining communes gained population. Negative net migration dominates rural Poland. This is a pattern prevailing everywhere except Kaszuby and rural communes surrounding urban agglomerations. In the latter case the classifications of these communes as rural is often disputable: many of them serve as areas of outmigration from core cities and enjoy the infrastructure and functionality of smaller towns. Large towns are all gainers. Medium size towns in most cases follow this pattern, with three classes of exceptions: new (after administrative reforms in 1975) regional capitals: Tarnow, Krosno, Piotrkow Trybunalski, Gorzow Wielkopolski, Radom and Koszalin; monofunctional industrial towns linked with the development of socialist investments, such as Pulawy, Kedzierzyn-Kozle (chemical industry), Stalowa Wola, Starachowice, Tychy (metal and automotive industry), Pabianice (textile industry); and towns linked with mining and heavy industries: Ruda Slaska, Swietochlowice, Jastrzebie Zdroj, Bytom, Piekary Siaskie, Raciborz, Dabrowa Gornicza in Upper Silesia as well as Walbrzych and Lubin. The latter two classes show that the governmental policy of subsidies randomly distributed to various gigantic socialist enterprises does not prevent population outflow. 19

25 Figure 7: Net migration rate per 1000 population in Poland by communes and municipalities, 1994 Net Migration Rate per 1000 population 50 or more 20 to <50 10 to <20 5 to <10 0 to <5-5 to <0-10 to <-5-20 to <-10 Less than -20 Unknown 0 50 Kilometers 100

26 Figure 8: Net migration rate per 1000 population in major urban agglomerations, 1994 Poznan and surrounding area Lodz and surrounding area Katowice, Krakow and surrounding areas Total population, to <5,000 5,000 to <10,000 10,000 to <25,000 25,000 to <50,000 50,000 to <100, ,000 to <250, ,000 to <500,000 More than 500,000 Net Migration Rate,1994 per 1000 population More than to <50 10 to <20 5 to <10 0 to <5-5 to <0-10 to <-5-20 to <-10 Less than -20 Warsaw and surrounding area Kilometers

27 If we accept the oversimplified interpretation that migration gains or losses demonstrates the economic well-being and quality of life of a places, we will have a picture of attractive Poland of large cities and suburban communes as well as of medium and small size towns and unattractive rural Poland and medium sized industrial towns. 4.3 The demographic sources of population change The spatial redistribution of population presents integrated effects of natural increase and mobility in each commune. The interplay of these factors will be examined in detail based on 1994 data. A long term historical perspective can be found in Rakowski and Rakowska (1990). A look at the rural population change in more global terms shows, that till the early 70's the migration losses had been compensated by natural increase. In 1970 for the first time the rural population reduced and the annual losses reached a peak in 1975 of 74.1 thousand but the negative migration balance stood in this year on thousand. This decrease was therefore solely due to huge rural emigration in the Gierek era. The total losses of rural areas in the eighties was about 111 thousand. In fact, only between 1982 and 1985 has the rural population been growing. An evident speed up of the process of decrease started in It was predominantly due to decrease in fertility, as mobility was relatively stable on the lowest level since early seventies. The trend continued till 1992 when unexpectedly the rural population increased by 158 thousand. The check of.the population accounts shows that in this year 167 thousand of inhabitants of rural areas were not accounted for by demographic events (fertility, mortality and migration) and occurred either due to changes of administrative boundaries or due to some undocumented corrections introduced by the Central Statistical Office. It is therefore justified to state that we have witnessed a decade long decrease in rural population. The pace of growth of urban population declined almost sixfold between 1980 and It was due to dramatic reduction of net migration and fertility and slight increase in mortality. In 1994 birth rate of urban population stood at 10.9 pro mille, death rate at 9.4 and internal migration gain at 1.6 pro mille totalling to a growth rate of a meagre 3.1 pro mille. After deduction of net international migration losses it was reduced to 2.4 pro mille. 22

28 An analysis of the interplay of natural growth and net migration on commune and municipality level has been conducted using Webb classification. This classification has been first proposed by John Webb (1963), an American geographer who invented it to study population change in England and Wales in the years The very simple classification is based on relations between natural increase and net migration (see Figure 9). As both net migration and natural increase may be either positive or negative and resulting total increase may be either negative or positive, it is possible to create eight classes described in Table 1. One of the problems with Webb classification is that it does not cater for cases where net migration or natural increase equal to zero nor for cases where the absolute values of the two are equal. Hence a large number of unclassified units (see Table 1 and Figure 10). The classification used allows for an immediate identification of the direction of population change, sign of net migration and natural growth and the leading force behind the population change. As Table 1 and Figure 10 demonstrate, majority of communes in Poland (60% of classified) grew in Of them 766 grew due to positive natural increase higher than negative net migration (type A). A vast majority of these communes is located in the North and North-West of Poland, inhabited by younger than average population, and in the South, which has traditionally high fertility. The communes which have positive both natural increase and net migration (types B and C), and therefore sound demographic growth are quite numerous (838 altogether). The former type of communes (B), dominated by natural growth are located in Kaszuby, Southern Poland, an what is interesting nearby large and medium size cities: they surround Kielce and form semi-circles East of Warsaw, East, South and West of Radom, South of Krakow, West of Bielsko-Biala and South-East of Poznan and East of Tarnow. This pattern may indicate a process of concentration of younger, procreative population around cities and may be linked with the economic activity and relative prosperity of this population allowing for both the investment in offspring and settlement in more favourable environmental conditions. This hypothesis should be tested in separate research. The latter type of communes (C), with the growth dominated by migration is characteristic for medium size cities, such as Gdansk, Gdynia, Torun, Bydgoszcz. Quite often they form crescents around large cities, such as Warsaw, Poznan, Lublin, Katowice. 23

29 Figure 9: Webb Classification NATURAL Natural gain exceeds negative net migration INCREASE Natural gain exceeds positive net migration I N C R E A S E NEGATIVE NET MIGRATION Negative net migration excceds natural gain Negative net migration exceeds natural loss H G A F B E C D Positive net migration exceeds natural gain Positive net migration exceeds natural loss POSITIVE NET MIGRATION D E C R E A S E Natural loss exceeds negative net migration NATURAL Natural loss exceeds positive net migration DECREASE Source: Adapted from Webb (1963)

30 Figure 10: Webb classification of communes and municipalities in Poland, 1994 Webb Classification Class Unclassified A - Natural gain > negative net migration B - Natural gain > positive net migration C - Positive net migration > natural gain D - Positive net migration > natural loss E - Natural loss > positive net migration F - Natural loss > negative net migration G - Negative net migration > natural loss H - Negative net migration > natural gain Unknown 0 50 Kilometers 100

31 If one looks at the combined pattern of the spatial distribution of these two classes the most striking is that the communes encircled towns and cities of various sizes and located all over Poland, with notable exception of South-Western Poland. Communes where positive net migration offset negative natural growth (type D) are located South from the line linking Poznan and Warsaw, and are often close to large cities as Warsaw, or are surrounding them, as in the case of Lodz. The most frequent reason for loosing population was the negative net migration exceeding positive natural increase (class H, 816 units). To this class belong mainly rural communes located all over Poland except Pomorze Gdanskie and infrequent in Southern Poland. Much more worrying are these communes where both components, net migration and natural growth are negative (type F and G). There are 428 communes belonging to either of these types. They are mainly rural, located in the former Russian partition (regions Bialostockie, Siedleckie, Piotrkowskie, Kieleckie, Czestochowskie). A cluster of towns in Upper Silesia (Swietochlowice, Siemianowice Slaskie, Czeladz, Wojciechowice, Dabrowa Gornicza, Sosnowiec, Slawkow, Ogrodzieniec, Pilica) fall into one of these classes, showing that environmental, economic and social problems translate directly into demographic problems. Class E, to which belong units where negative natural increase exceeds positive net migration are not very numerous. It is interesting to note, that large cities and towns such as Warsaw, Poznan, Lodz, Czestochowa and Zgierz fall into this category. The overall pattern of population change in 1994 shows that in comparison to the previous years the most alarming pattern - population decrease due to both negative natural increase and net migration is rapidly growing. There were 18 such communes in 1976, 103 in 1987 (Rakowski and Rakowska 1990) and since 1987 the number more than quadrupled. This rapid increase is not unexpected. It is a consequence of decades of rural depopulation and resulting from it, the distortion of age and sex structures: There is a deficit of females (see section 4.4) and an aged population (Gawryszewski 1989). A new phenomenon is the decrease of the population of large cities. It remains to be seen if this is a short lived fluctuation or a long term change of population distribution. 26

32 4.4 The impact of migration on the distortions of demographic structures The sex structure of population is mainly the result of selectivity of the mobility of population. In general, female population is more mobile than male. In particular, this is true when we examine the migration patterns of the rural population in Poland. There have been a long lasting, strong outflow of female population from rural areas (Gawryszewski 1989). It has reduced in the nineties. In 1994 the difference between outmigration of males and females reduced to bare 1 per thousand. Most rural areas experienced below national average (which stood at 51.3% in 1994) share of females (see Figure 11). Gawryszewski (1989, Table 3) demonstrates that depopulating rural regions have less females per 100 males than regions which do not suffer from depopulation. In Northern and North Eastern Poland this imbalance is particularly strong. The area suffering the most from the lack of women are in the Gdansk, Slupsk, Koszalin Szczecin, Suwalki, Bialystok, Lomza, Ostroleka, Ciechanow, Olsztyn and Elblag regions with the share of female under 49% in many rural communes. Quite severe distortion of the sex structure is observed in the South in mountainous communes of the Krosno and Nowy Sacz regions. Female population concentrates in towns and cities. In fact it is very unusual to have more males than females in a town and it does not happen in cities. Even areas dominated by mining and heavy industries are female-dominated, as are large agglomerations of Warsaw, Lodz, Poznan, Krakow and Wroclaw. As the migration process is highly age sensitive it is interesting to examine if there is any particular disproportion in sex structures in specific age groups. This must be done on regional level as data on the age structure of population on commune level is not available. Two clusters of regions which demonstrate particularly strong depopulation patterns have been selected: North -Eastern (regions: Bialystok, Ciechanow, Elblag, Lomza and Suwalki) and Central Poland (regions surrounding Lodz regions, Lodz region itself and Piotrkow, Plock, Sieradz and Skierniewice). Age groups and 65+ are examined (see Table 2). To avoid unusual differences in the proportion of males and females all values were standardised to the average value for Poland (right panel of the table). The female to male ratio for all ages, shown on Figure 11, does not reveals all truth. If we inspect sex structures in the age groups which are the most important from the point of view of marriages and procreation, i.e. between 20 and 34 years the picture would be much more serious. In all North - Eastern regions there are fewer that 85 women aged per

33 men at the same age. In Bialystok, Lomza and Suwalki regions this translates to less than 90 percent of value for rural Poland. Somewhat surprisingly this distortion of sex structure extends to the old age group. In the other cluster, centred on Lodz, it is difficult to identify any substantial distortion of sex structure. In Lodz and Plock regions the number of females per 100 males expressed as a percentage of the value for rural Poland exceeds 100% in the age group. In Piotrkow and Sieradz regions the deficit of females is clearly visible. Apparently in this area the process of depopulation does not have that strong impact on sex and age structures. This may be due to a different degree of ruralisation of rural areas, accessibility to facilities and cultural factors. Obviously the distortion of the balance of the sex structure has not only demographic, but much more important social and economic dimensions. It hampers the process of formation of families due to a lack of suitable female candidates for marriage. As the agriculture predominantly relies on family-run farms with traditional division of tasks between male and female members of family it also causes various fundamental problems in running farms. Leopold and Manteuffel (1982, p. 72) noted that on the average 17% farmers remain single, the value going in purely agricultural areas to as high as 30%. 28

34 Figure 11: Sex structure of population by communes and municipalities in Poland, 1994 Females per 100 Males Unknown 130 or more 120 to < to < to < to < to < to <95 Less than Kilometers 100

35 Table 2: Age structure of rural population in selected regions, Poland 1994 Region Females per 100 males Age Total Total 65+ North-East Bialystok Ciechanow Elblag Lomza Suwalki Central Lodz Piortkow Plock Sieradz Skierniewice Poland Females per 100 males in per cent of the value for Poland North-East Bialystok Ciechanow Elblag Lomza Suwalki Central Lodz Piortkow Plock Sieradz Skierniewice Poland Source: Computed from the data from GUS

36 Table 3: Population growth of towns and cities by size class Population growth over the period (%) Size class Under All cities and towns Source: Computed from the data provided by the Central Statistical Office. Note: for the period the growth rates were calculated based on sizes of urban units as in 1990, for other periods based on sizes of urban units as in RELATIONSHIP TO THE URBAN SYSTEM 5.1 Gaining and losing urban systems As was mentioned above the vast majority of the growth of population has occurred in urban territories. Increase in the urban population over the period was on average 310 thousand per year (calculated as a difference between end of year and start of year population to accommodate changes in administrative status of localities as well as changes of administrative boundaries). In 1975 it peaked at 543 thousand but from 1991 it stood below 150 thousand per year with the increase in 1994 by a meagre 60 thousand. This is partially due to a marked slow down in the increase of total population of Poland, partly due to marginalisation of the significance of migration gains in recent years. Jerczynski and Gawryszewski (1984) and Jerczynski (1990) noticed that this growth has not been distributed uniformly among all categories of cities and towns. 31