untitled Spatial turns of manufacturing since 1970 MARKKU TYKKYLÄINEN Tykkyläinen, Markku (2002). Spatial turns of manufacturing since 1970. Fen- nia 180: 1, pp. 213–226. Helsinki. ISSN 0015-0010. In this article, the geographical transformation of the Finnish manufacturing sector since the 1970s is analysed. Changes in manufacturing employment by municipalities are used as central indicators of the transformation. The urban– rural shift, deindustralisation, and emergence of the new economy have all occurred during the last three decades. A trend surface analysis indicates that manufacturing employment increased all over the country in the 1970s. The growth was higher the more northern and eastern the municipality’s location. This pattern was conformable to the goals of the Finnish regional policy of the 1970s. Deindustrialisation plagued the economy from the early 1980s to the mid-1990s, and its negative labour force impacts were focused especially on the southern mature industrial cit- ies. The industrial policy of the 1980s was without much result. Finally, the crisis of the early 1990s ruined a considerable part of industrial performance everywhere in the country. The era of telecommunications-driven economic growth in the 1990s led to the geographical pattern of growing employment in few urban areas and some localities adjacent to cities. The long, vigorous growth period (up to the year 2001), comparable with the growth of the late 1960s and early 1970s, diffused to rural municipalities, but the geographical pattern of this boom was oriented more towards the west and south than in the 1970s. Markku Tykkyläinen, Department of Geography, University of Joensuu, P. O. Box 111, FIN-80101 Joensuu, Finland. E-mail: markku.tykkylainen@joensuu.fi Background External factors, such as export demand, trade agreements, and wars, have strongly regulated the Finnish economy. Since 1917, Finland has been a sovereign actor in Europe. Until the mid-1990s, the only doctrine of the economic policy was to emphasise the nation-state as the prime regula- tive actor in the economy. The advantages of the economic integration for the specialised and ex- port-oriented economy led to a search for eco- nomic integration after the post-World War II reg- ulative period. As a corollary, the Finnish econ- omy has been integrated gradually into the core of the European economic market area: first as an associated member of the EFTA in 1961, then as a full member in 1986 and, finally, as a member of the EU in 1995. Nowadays Finland is part of the euro and the Schengen Treaty areas. The turbulent eras of Europe had radical im- pacts on the structures and geography of econom- ic activities in Finland (Hjerppe 1989). Forest- based production was the main source of export revenue for long after World War II. Finland was a sparsely populated country where the scattered settlement pattern was maintained by forestry (and agriculture). The logistics of the resource-based sector determined a great deal of the regional structures, especially outside a few industrial cen- tres (Tykkyläinen 1988: 343). Gradually, trans- formation from low-value products to high value- added wood and paper products and diversifica- tion to the metal industry began to take place. The nurturing of basic industries was part of the Finnish economic policy, leading to the growth of the manufacturing of metals, metal products and machinery, oil refining, and the production of chemicals over decades after World War II. The investors were largish companies, and many of them were in the state ownership, such as Kemira, Neste (now Fortum), Outokumpu, Rautaruukki, and Valmet (now Metso) (see www.kemira.fi; 214 FENNIA 180: 1–2 (2002)Markku Tykkyläinen www.fortum.com; www.outokumpu.com; www. rautaruukki.com; www.metso.com). The exports to the Soviet Union and other CMEA (Council for Mutual Economic Assistance) countries consisted of machines, ships and, later on, many consumer goods. The trade in consumer goods boosted the expansion of the light industry especially in the 1970s and 1980s. The industrial basis became more diversified, but the national economy was more controlled than it is today. For instance, in- dustrial companies were Finnish-owned with only a few exceptions until the early 1990s. Rational- isation led the spatial restructuring of industry to fewer units and, on the other hand, to the glob- alisation of Finnish industry. The current socio-economic landscape of Fin- land is the result of these economic transforma- tions. Finland has traditionally been divided into developed, urban southern and less-developed eastern and northern parts (Yli-Jokipii & Koski 1995). Furthermore, the western parts of the coun- try are considered more developed than the east- ern forestry-based hinterlands (Siirilä et al. 1990). During the period 2000–2006, the vast less-de- veloped areas of northern and eastern Finland constitute an EU Objective 1 area where GDP per capita is clearly less than the EU average (Hede- gaard 1999). The aim of this article is to depict the main trends in the sectoral transformation of the Finn- ish economy during the past decades. The study elaborates on the geographical manifestations of regional policy, the deindustrialisation of the 1980s, the economic depression of the early 1990s, and the growth of the Information and Communication Technology (ICT) cluster in the 1990s. The patterns of the geographical transfor- mation of the Finnish manufacturing employment are analysed by municipality from 1970 through 1998 in terms of the trend surface analysis. Employment data in all maps and analyses are extracted from Official Statistics of Finland com- piled by the Statistics Finland in Helsinki. Part of the data has been used earlier in studying changes in the manufacturing employment in Finland in the early 1980s (Tykkyläinen 1987). Employment data from 1970 through 1990 by municipality are based on the volumes of Industrial Statistics (Of- ficial Statistics… 1973–1992) and later data on electronic submissions from Altika and StatFin Service by Statistics Finland. All data are on file (Teollisuusaineisto 2001). Geographical impacts of regional policy up to 1980 Except for a few major industrial centres, the geo- graphical pattern of the Finnish industry was for a long time determined by the location of natural resources (Tykkyläinen 1988: 343), the impacts of colonisation (1922–1965) (Tykkyläinen 1996), and regional policy. The first wave of colonisation took place before World War II when land was allocated to landless people. The second wave followed after the war when refugees and veterans were resettled. The first regional policy law con- cerning underdeveloped regions went into effect in 1966 (Palomäki 1980: 197). When the first post-war generation entered the labour market and the post-war settlement policy came to an end in the 1960s, the industrialisa- tion of the rural areas and peripheral parts of the country was established as the goal of the indus- trial policy. The expansion of industry was pro- moted in regional policy explicitly from 1966 on- wards. Industrial location to less-developed are- as was thus supported by the state (Palomäki 1980). The net impact of regional policy on the employment increase in manufacturing in less- developed areas has been estimated to be 18,000 jobs in 1966–1980, or, taking into account the multiplicatory effects, 26,000 jobs (Tervo 1983: 170–174). Nationally, the growth period of the early 1970s was damaged when the first oil crisis led to the decline of the export demand of the Finnish industries; both production and industri- al employment declined in the manufacturing sec- tor in 1975 (Fig. 1). The Finnish economy recovered gradually from the first oil crisis and the growth rate peaked in 1980. The number of personnel in the manufac- turing sector reached its culmination, 568,000 people. This remained the record for the twenti- eth century. Proportionally, this number corre- sponds to 26.3 percent of the economically ac- tive population in Finland in 1980 (STV 1982: 46). The spatial pattern of the growth of the 1970s comprised the growth of employment in less-de- veloped areas. Industrial premises and estates were built in remote localities. The state support- ed this rural industrialisation with loans, grants, and tax relief. The highest support was directed to the most remote areas, Lapland and North Karelia (Palomäki 1980). The declining localities were industrial cities, such as Helsinki, Tampere, Turku, and Lahti. The winners of the 1970s were FENNIA 180: 1–2 (2002) 215Spatial turns of manufacturing since 1970 provincial centres, towns and rural communities (CD-Fig. 1). The explanations for this expansion of rural in- dustrialisation were the opening up of the Finn- ish economy, an abundant young labour force in the peripheries, and the implementation of re- gional policy. The pattern of rural industrialisation operated partly according to the globally-known production cycle model: advanced industry orig- inated in the core and late-stage, low-wage pro- duction was shifted to the periphery (e.g., Nor- ton & Rees 1979). Nevertheless, the Finnish data demonstrate that new industries grew in the pe- ripheries and that the shift of existing (and ma- ture-stage) activities from the south was in a mi- nor role (Jatila 2001: 149). The spill-over effect took place in the surroundings of most cities. More precisely, the growth of the new industries emerged in new relatively small localities outside the industrial cores. This expansion of the manu- facturing sector could not prevent the consider- able migration of people to cities and the neigh- Fig. 1. Growth of production and em- ployment in the manufacturing sector in Finland, 1960–1998 (Official sta- tistics… 1973–1992). 216 FENNIA 180: 1–2 (2002)Markku Tykkyläinen bouring Sweden. Migration to Sweden peaked in 1969–1970, when more than 80,000 people left Finland (Korkiasaari & Söderling 1994: 249). Since 1981 net immigration has been positive (STV 2000: 106). Rationalisation and deindustrialisation in the 1980s The years from 1981 to 1993 were years of de- cline in the manufacturing employment (Fig. 1). Industrial employment declined in almost all lo- calities. Only some suburban local districts were expansive, such as Espoo and Vantaa in the Great- er Helsinki area. In addition, some outlying com- munities succeeded in attracting new jobs, but the increase was small in absolute numbers. The vol- ume of production grew until 1990, however. The average annual growth of manufacturing output remained below the level of the growth period 1960–1974, but the relatively smaller growth rate does not fully explain the decline in the number of jobs. An additional factor, the growth of pro- ductivity, had a negative impact on employment as well. Rationalisation in traditional manufactur- ing industries took place in the 1980s. This ration- alisation, combined with moderate growth, main- tained the decline of labour in the manufacturing sector. The spatial pattern of the 1980s’ deindustriali- sation was not a withdrawal from the new remote emplacements that resulted from the booming growth period of the 1960s and early 1970s. The stronger growth of productivity than that of pro- duction reduced employment in major industrial cities and, to a lesser extent, in smaller localities and rural areas (CD-Fig. 1). Urban areas lost two- tenths of their manufacturing jobs and rural areas only one-tenth. The number of manufacturing jobs lost during the 1980s was 102,000, of which 89,000 (87%) were in cities and the rest in rural municipalities (for the urban–rural classification of the municipalities according to the type of mu- nicipality, see Municipalities… 1998: 22–23, 63– 68). The proportion of the manufacturing sector in GDP declined from 28.1 percent in 1980 to 23.3 percent in 1990 (Kansantalouden… 2000: 28–29). This proportion was relatively small compared with other industrialised countries. The competi- tiveness of the Finnish manufacturing sector was low, but credit expansion maintained the growth of domestic demand and import and the increase of asset prices. The opening up of the economy made it possible to increase foreign loans and ac- celerated speculation. The Finnish economy was showing more and more characteristics of Casino Economy: share prices tripled and housing prices doubled in the second half of the 1980s (Kiander & Vartia 1998: 69). The crisis of the early 1990s Share and housing prices peaked in 1989. Infla- tion dogged the economy. An increasing number of companies could not compete profitably in the market. The economic crisis led to a decline in both production and employment (Fig. 1). The proportion of the manufacturing sector in GDP dropped to 20.3 percent in 1991 (Kansanta- louden… 2000: 29). The number of closed enter- prises exceeded the openings by 2,250 in 1991 and by 1,548 in 1992 (STV 2000: 169). Bankrupt- cy proceedings tripled in 1992 compared with the numbers of the late 1980s. The end of barter trade with the Soviet Union at the end of 1990 exacer- bated the economic recession. In the late 1980s the proportion of the Finnish foreign trade to the Soviet Union had been about 15 percent, but sud- denly it decreased to around 5 percent in 1991. This brought about the loss of 30,000–40,000 jobs in Finland in 1991 (Rautava & Hukkinen 1992: 5). Multiplier effects have been taken into account in this calculation. The devaluation of the Finnish mark in 1991 and the decision in 1992 to float the mark im- proved the competitiveness of the economy, but the decline in the value of the Finnish currency led to problems in the financial markets. The for- eign debt in Finnish marks increased. The state fell deeper and deeper into debt during the econom- ic depression and even after it. Central govern- ment debt was only EUR 9 billion in 1986, but it increased to EUR 29 billion in 1992 and peaked to EUR 71 billion in 1998 (STV 2000: 304). In- debtedness increased for years after the econom- ic depression. Nevertheless, public debt is not high in Finland compared with other OECD-coun- tries (OECD… 2001: 262). Lowered values of assets worsened the eco- nomic situation in the early 1990s. The number of bankruptcies was at an all-time high, and em- ployment, real incomes, and demand declined. The economic decline was as deep as the decline FENNIA 180: 1–2 (2002) 217Spatial turns of manufacturing since 1970 in the countries in East Central Europe. The in- dustrial output decreased one-tenth, but started to recover from 1992 on. The decline of employ- ment in the manufacturing sector continued in the early years of the 1990s. The number of employed bottomed out to 357,000 in 1993 (in enterprises with a personnel of five or more) (STV 1997: 160). The drop was 24 percent from 1990 through 1993. The economic depression of the early 1990s, “the Great Depression II,” was even worse than that of the 1930s in Finland, but milder in relative terms than the economic depression of the 1930s in the USA (Kiander & Vartia 1998: 11). The export recovered in 1992 and manufacturing employment has grown since 1993, bringing about a long growth period until the beginning of 2001. The value of production has increased signifi- cantly in the long run in spite of the decline in the manufacturing employment. From 1980 to 1996, the output in the manufacturing sector in- creased by 60 percent. The recovery of the econ- omy took place in the high-tech sectors. The los- ers were traditional labour-intensive sectors, such as the manufacturing of wearing apparel, foot- wear, furniture, some metal products, and trans- port equipment, many of which exported goods to the CMEA countries in the 1970s and 1980s. The fastest-growing sector was the manufacture of electronic equipment, which produces commu- nications equipment, computers, and mobile phones and related equipment (Table 1). Structural adjustment in manufacturing reflects the loss of the Soviet market and the declining demand for some consumer goods and construc- tion materials. The industrial decline was especial- ly severe in labour-intensive sectors that employed low-skilled labour. On the other hand, the im- pacts of the growing Information Society are clear- ly visible in the figures. The geographical pattern of the crisis and the recovery of the early 1990s The spatial outcome of the crisis of the early 1990s was the decline in employment in most lo- calities (CD-Fig. 1). The recession hit cities and rural areas hard. The net change in manufactur- ing jobs was negative both in urban (–12%) and rural (–11%) municipalities. The crisis of the ear- ly 1990s spread geographically more evenly than the deindustrialisation of the 1980s. The total net decrease was 52,600 jobs, 41,000 of which in urban areas. The decrease in manu- facturing jobs in Helsinki, Vantaa, Lahti, Pori, Tampere, and Turku was to 21,000 from 1990 to 1995. Pori suffered most as 24.0 percent of jobs disappeared. The reductions in manufacturing jobs were 21.9 percent in Vantaa, 21.1 percent in Lahti, 16.7 percent in Helsinki, 11.4 percent in Tampere, and 11.3 percent in Turku. Unem- ployment grew rapidly and real estate prices plunged. Manufacturing employment grew only Table 1. The change of pro- duction in the Finnish manu- facturing industries in 1990– 1996 and their contribution to employment, total output, and exports in 1990 (STV 1997: 176–177; Yearbook… 1992: 74–93, 196–117; Tykkyläinen 1999). 218 FENNIA 180: 1–2 (2002)Markku Tykkyläinen in some localities, such as Salo (near Turku), by a net increase of 2,375 jobs, and Espoo (west of Helsinki), by 1,725. The employment growth was much due to the telecommunications-driven re- covery. The increase in research and development funding and the establishment of science parks and technopolies were responses to the transfor- mation, but there is no clear evidence of the visi- ble impacts of these inputs in the geographical pattern of industrial employment, as the map of 1990–1995 indicates (CD-Fig. 1). The loss of employment in the basic industry was bigger than the positive impacts of the dawn of the new economy that was based on the elec- tronics industry. The profitability of the basic in- dustry improved gradually and the Helsinki Stock Exchange recovered. Its annual HEX share index exceeded the level of 1990 in 1994 (STV 1995: 245). Towards the new economy The transformation of the Finnish industry con- tinued towards the end of the 1990s. The number of manufacturing jobs grew, as did the output (Fig. 1). The return to the growth rates of the 1960s, early 1970s, and the peak years of 1979–1980 was a fact (Fig. 1). The leading growth sectors were the manufacture of electronic equipment and the industries attached to it (Table 2). The scope of the growth industries is relatively narrow. The manufacturing of mobile phones and related industries were the growth engine of the economy in the latter half of the 1990s. This In- formation and Communication Technology (ICT) cluster consists of the Nokia Company and indus- tries supplying semi-finished products, comple- mentary production, and investment goods and services. According to the calculations of Ali- Yrkkö et al. (2000: 10), Nokia accounted for a lit- tle over 4 percent of the Finnish GDP and 23 per- cent of the total Finnish exports in 1999. These calculations do not take into account multiplier effects. Thus, the total impacts on GDP are even greater, because Nokia engages in a great deal of subcontracting. Nokia has specialised in electronics since the early 1980s, but found its success in the telecom- munications business as late as in the 1990s. The rapid expansion of Nokia lasted the entire dec- ade. In February of 1998, Nokia’s proportion of the total shares in the Helsinki Stock Exchange was 34 percent. On March 30, 2000, even after the backslide of ICT shares in global stock mar- kets, the proportion of Nokia was 62 percent. On August 15, 2001, the total value of Nokia shares was still 57 percent in Helsinki. Nokia has thus become the core asset of industrial capital in Fin- land. This development has a risky side, because the Finnish economy is now very dependent on the global demand for ICT products (Table 2). Table 2. Changes in industrial output (Bulletin… 2001: 18– 20) and the contribution of each industry to output and exports (Ennakkotietoja… 2001: 25–34). FENNIA 180: 1–2 (2002) 219Spatial turns of manufacturing since 1970 The success of Nokia has boosted attached in- dustries, such as the production of printed circuit boards, manufacturing of mobile phone enclo- sures, assembling industry and manufacturing services, and suppliers of production automation (see www.nokia.com; www.elcoteq.com; www.per- los.com; www.aspocomp.com; www.elektro- bit.com; www.eimo.com). The entire Information and Communication Technology cluster in Finland consists of Nokia (21,000 employees in 1998), other ICT equip- ment, network and related service producers (30,000 employees), components and contract manufacturers (5,000 employees), and telecom services (18,600 employees). Nokia’s contribution to the cluster’s sales was 50 percent in 1998 (Ali- Yrkkö et al. 2000: 27). The ICT cluster has been very profitable, but its contribution to investments negligible (Ali-Yrkkö et al. 2000: 13). The booming economy of the late 1990s in- creased industrial employment in most localities (CD-Fig. 1). Urban municipalities, such as Espoo, Salo, and Oulu, have been mentioned often as examples of the telecommunications-driven eco- nomic growth. The ICT cluster requires R&D in- puts. Spending on research and development per capita is thus very high in these growth areas (Tut- kimus- ja kehittämistoiminta… 1999: 7–9). Devel- opers of communications equipment, many sub- contractors, and companies providing automation technology and software have been located in advanced metropolitan environments and some localities adjacent to major cities in Finland, but having their manufacturing activities dispersed all over the world at the same time. The geographical pattern of the emergence of the new economy in Finland in the late 1990s in- dicates geographical clustering. Such a pattern is found in the high-tech sector where industries in their growth or innovative stages remain close to sources of skilled labour and specialised inputs as, for instance, Barkley’s 1988 study on the US American economy indicates. Moreover, the rea- son of spatial clustering can also be argued in a simple way: only a rather limited number of com- panies have expanded greatly, which results in an uneven spatial pattern. Both arguments can be supported empirically. The winners during the growth period 1995– 1998 in Finland were the city of Oulu (with a net increase of 3,896 jobs, or 48 %), the surrounding areas of Helsinki (including Espoo, Vantaa, Tuusu- la, Vihti, Nurmijärvi, and Järvenpää, with a net in- crease of 4,087 jobs, or 12%), the industrial com- munities of Central Finland (Jyväskylän maalaiskun- ta and Suolahti, 1,366/63%), Ylöjärvi–Tampere– Pirkkala conurbanisation (1,288/5%), Uusikaupun- ki (731/33%), and Salo (584/8%). The first four cas- es indicate that growth took place in the largest ur- ban areas, which comprise versatile learning envi- ronments (e.g., have universities and technical re- search) and largish travel-to-work areas. The popu- lation in these four locations ranges from 134,000 to 1.12 million inhabitants (STV 2000: 82). To gen- eralise, Finland has a new, 650-kilometre-long de- velopment axis from the south to the north, from Helsinki to Oulu via Tampere and Jyväskylä. The state has promoted development in these environ- ments, as the main “centres of excellence” related to ICT are located in these cities and their surround- ings (Osaamiskeskukset 2001). Uusikaupunki on the west coast is known for car manufacturing and Salo in the southwest for mobile phones. Some communities adjacent to the larger urban agglomerations were winners as well, such as Nurmo (480 jobs; adjacent to Seinäjoki), Hollola (422 jobs; adjacent to Lahti), Mustasaari (356 jobs; adjacent to Vaasa), and Raisio (387 jobs; adjacent to Turku). The industrial growth of Hol- lola is based on the expansion of the ICT cluster, but the localities have been boosted by other manufacturing sectors. These smaller growth com- munities are located in regions outside the tradi- tionally less-developed areas in Finland. The spatial expansion of the manufacturing sec- tor, and especially the ICT industry, has favoured urban and semi-urban areas in a selective way. This result can be understood in the light of many high-tech industry studies which make it known that the spatial organisation of the ICT industry is based on company-specific comparative advan- tage. This advantage is dependent on each prod- uct, production process, corporate purchasing policies, and type of final market. All these fac- tors are unique and thus produce specific loca- tional patterns (e.g., Glasmeier 1988). ICT pro- duction units were located near skilled labour and in areas close to cities and proper communica- tions networks. The vast less-developed Objective 1 area did not benefit much from the boom of the ICT cluster. The trickling-down effects of the ICT cluster are thus geographically limited and net- work-like. Growth takes place in suitable geo- graphical pockets (such as technopolies, industrial estates, and otherwise advantageous places) in multi-faceted environments. 220 FENNIA 180: 1–2 (2002)Markku Tykkyläinen Finland in comparison with other advanced countries Compared with other European countries and the OECD countries, Finland’s growth has been rap- id. The volume of industrial output increased by 41 percent from 1995 to May of 2000, whereas the increase in the USA was 26 percent, in the OECD countries 20 percent, and in the EU, 16 percent (Bulletin… 2000: 109). The Finnish econ- omy was booming. Growth combined with the high rate of auto- mation has not brought about labour-intensive production in manufacturing in Finland. The growth of productivity has been stronger in Fin- land than in other advanced countries (Table 3), which is largely due to the growth of productivity in telecommunications. Furthermore, investment in research and development increased rapidly in the 1990s, which contributed to the increase in productivity, but in the ICT industry in the main. The contribution of Nokia and other ICT com- panies to the increase in the productivity of the entire economy has been significant. The produc- tivity in the telecommunications equipment in- dustry grew 25 percent annually in the 1990s and, in the ICT industry as a whole, about 15 percent annually over the same period (Ali-Yrkkö et al. 2000: 14). This implies that in some industries growth of productivity remained close to zero and definitely below the long-term average growth. Aggregate figures are therefore somewhat mis- leading. Geographical pattern: industrial cities versus rural areas in 1970–1998 Economic development and restructuring reshape Finland’s economic landscape all the time. The industrialisation of the late nineteenth century took place in towns and industrial communities near the sources of raw materials and energy and along the rivers (Palmén J 1899; Moberg 1899; Palmén K 1899). Moreover, this industrialisation resulted in some largish industrial centres in the south, such as Helsinki, Tampere, and Turku. Job losses of the late twentieth-century manufactur- ing took place in these main manufacturing cen- tres (Fig. 2). Nevertheless, most of these industri- al cities are very viable centres attracting popula- tion from smaller places inasmuch as the service sector in these cities has grown. The manufactur- ing sector has many of its central administrative functions in the cities, and the growing post-in- dustrial economy and the positive migration bal- ance have been the engines of metropolitan growth. On the other hand, rural industrialisation, or the urban–rural shift, has taken place in Fin- land as in many other advanced countries (North 1998). The newest wave of industrialisation bene- fits some metropolitan areas and already devel- oped rural areas. During the examined three decades (from 1970 to 1998), the greatest winners in the competition for manufacturing jobs were Espoo (with a net in- crease of 11,497 jobs/291%), Vantaa (5,420/ 62%), Salo (4,304/110%), and Oulu (3,484/ Table 3. Productivity of the manufacturing sector, from 1990 to 1998 (STV 2000: 636). FENNIA 180: 1–2 (2002) 221Spatial turns of manufacturing since 1970 41%). The losers were Helsinki (with a net loss of 41,874 jobs/–57%), Tampere (14,693/–41%), Turku (14,024/–46%), Lahti (10,526/–53%), Pori (7,125/–46%), and Kotka (6,128/–55%). Unlike the employment data of 1970, the 1998 figures include the enterprises with less than five employ- ees, which represent 5.6 percent of the total manufacturing employment (STV 2000: 171). The growth is thus a little weaker and the decline deeper than the figures indicate. The restructur- ing of the manufacturing sector hit the industrial core of Finland, leading to a decline in manufac- turing employment in traditional industrial cities. The main boom of rural industrialisation took place in the very late stage of Fordism, in the 1970s. Especially during the last years of techno- logically-driven growth, new industrial spaces were created in dispersed metropolitan-type are- as: Espoo and Vantaa, in the technologically-ori- ented Oulu, and in Salo, which has a long tradi- Fig. 2. Changes in the number of em- ployees (including owners) in the Finnish manufacturing sector, in 1970–1998, by municipality. Note that the statistics for 1998 cover all enterprises in the manufacturing sec- tor, whereas the earlier statistics re- corded only the establishment of en- terprises with a personnel of five or more. These small enterprises repre- sented 5.6 percent of the total manu- facturing employment in 1998 (STV 2000: 171) (Official statistics… 1973– 1992; Teollisuusaineisto 2001). 222 FENNIA 180: 1–2 (2002)Markku Tykkyläinen tion of the electronics industry. Some smaller lo- calities succeeded, which proves that a long growth period impacts positively on employment in semi-remote rural areas. During the entire period 1970–1998, the diver- sification of the manufacturing sector formed a new economic landscape in Finland. This was first based on the new, light industries in the 1970s and, later, on the ICT industry. The geographical transformation consisted of various tendencies of industrial restructuring and growth, such as 1) the deindustrialisation of the major industrial cities in 1970–1998; 2) the industrialisation of rural municipalities, es- pecially in the 1970s and late 1990s; 3) the spill-over of manufacturing employment from industrial cities to the neighbouring com- munities in 1970–1990 and 1995–1998; 4) the industrial decline in the early 1990s and the growth of employment in a few localities; and 5) the growth impacts of the ICT industry on the competitive urban areas and localities adjacent to cities in the late 1990s. The mapping of the geographical transforma- tion indicates that the growth pattern of the manu- facturing sector became spatially selective in the 1990s. The results suggest that low-cost labour force as such is no longer attractive enough as a locational factor as it was in the 1970s. Instead, leading manufacturing branches look for skilled labour and competitive metropolitan environ- ments. How this phenomenon is constructed within the ICT industry, and whether it prevails outside the ICT industry, calls for further studies. Analysis of geographical growth patterns Finland is divided into developed, urban south- ern and less-developed eastern and northern parts, as stated earlier. Which parts of the coun- try were winners and which losers in the geo- graphical transformation of the manufacturing sector? Public opinion supposes that the spatial concentration of economic activities has shifted towards the south and its urban areas. It has al- ready been proven that the main industrial cities in Finland lost jobs in the manufacturing sector during the last three decades. But what about the north–south and east–west dimensions? In which areas have decline and growth taken place? Trend surface analysis can be used effectively to remove the local fluctuation that obscures the general spatial trend of the industrial transforma- tion. The x and y coordinates are used in the fol- lowing way: y indicates the distance from the equator and x indicates the east–west dimension in the way that the parallel of 27 degrees is set to be 3,500 kilometres. The coordinates of the sur- face are y (min) = 6,646 km and y (max) = 7,730 km and x (min) = 3,089 km and x (max) = 3,703 km. The dependent variable z is the absolute change in numbers of employees in the manufac- turing sector during a period. The periods are 1970–1980, 1980–1990, 1990–1995, 1980– 1995, 1995–1998, and 1970–1998. The growth decade of the 1970s favoured pe- ripheral locations. The manufacturing growth dif- fused to rural peripheries, as happened in many other advanced countries as well (e.g., Haynes & Machunda 1987). Equations 1 and 2 depict the geographical pattern of changes in Finland from 1970 through 1980: (1) z = – 545 + 0.056y + 0.097x (1970–1980) (2) z = – 36302 + 5.27y + 10.58x –0.0015xy (1970–1980) The first equation represents the simple planar surface which shows that the further north a place is located the higher was the increase in the number of jobs in the 1970s. Similarly, the more eastern the location is the higher was the growth in the manufacturing employment. The surface is steeper towards the east than the north, which in- dicates that an eastern location was even more beneficial than a northern one. This simple anal- ysis shows that regional development took place in the desired direction. All the z-values of the surface within Finland are positive, indicating that there were no areas larger than single localities in decline. The bi-linear saddle surface provides addition- al information. It is more depictive and shows the pattern of the 1970s (Fig. 3). The growth in the manufacturing employment diminishes when ap- proaching the north-eastern corner of the coun- try. In fact, the municipalities in the eastern parts of Lapland and Kainuu were not the most success- ful in gaining new jobs. They were too remote. FENNIA 180: 1–2 (2002) 223Spatial turns of manufacturing since 1970 The impacts of the decline of the manufacturing sector in the main industrial cities create a sad- dle structure that bends up in the northwest and southeast (Fig. 3). The deindustrialisation of the 1980s and early 1990s, as illustrated by municipalities on the maps in CD-Figure 1, resulted in sharp downward surfaces to the south, depicting job losses in in- dustrial Finland. Industrial cities in the eastern parts of the country declined as well. The surfac- es of 1980–1990 and 1990–1995 thus slope gen- tly to the east. The crisis period of the early 1990s (1990– 1995) produces a north–south slope that is a lit- tle smoother than the slope in the 1980s (1980– 1990), indicating that the manufacturing employ- ment degenerated towards the north as well dur- ing “the Great Depression II” (these equations and figures are not included). In the east–west dimen- sion, the eastern parts of the country suffered more in relative terms during “the Great Depres- sion II” than in the 1980s. This indicates that the western parts of the country were more flexible in adapting to new conditions. The Information Society dawned in the west and south. One planar equation (3) depicts the entire dein- dustrialisation period of 1980–1995: (3) z = – 5807 + 0.914y – 0.200x (1980–1995) Similarly, the bi-linear saddle surface reveals deindustrialisation in the south and southeast (Equation 4): (4) z = 60051 – 8.68y – 19.50x + 0.00281xy (1980–1995) Many industrial cities are located centrally in the country, which slopes the surface to decline towards the southeast (Equation 4; Fig. 3). On the other hand, the surface slopes up to the northeast. More clearly than the saddle surface of 1980– 1995 in Figure 3, the 1980s’ surface (not includ- ed in this article) demonstrates the impacts of the decline in industrial cities – Pori, Tampere, Lahti, and Kotka – in the formation of a U-shaped val- ley. This was the belt of deindustrialisation in the 1980s. The rise of the new economy in the late 1990s favoured southern and western localities as the seedbeds of growth (Equation 5). Western parts of Finland, such as Ostrobothnia, South Ostroboth- Fig. 3. Bi-linear saddle surface on the employment changes in the manufacturing sector in Finland 1970–1980, 1980– 1995, and 1995–1998. 224 FENNIA 180: 1–2 (2002)Markku Tykkyläinen nia, and North Ostrobothnia, grew (CD-Fig. 1). A net increase of industrial jobs in rural areas was 12,300 from 1995 through 1998, being 88 per- cent of the total net increase in the country (cal- culated from the same data file as CD-Fig. 1). The rural re-industrialisation was not necessarily a re- sult of the growth in the ICT industry, but rather a result of the rise in incomes and the growth in demand in other sectors. The growth in rural are- as has thus been mainly induced due to the long boom in the Finnish economy. The industrially- diversified economic areas in the west with many small and medium-sized companies have been responsive to these induced growth impacts. (5) z = – 74 + 0.066y – 0.103x (1995–1998) The trend surface of 1995–1998 is above the zero-level within Finland, indicating growth all over the country. Nevertheless, z-values are low- er than in the 1970s’ surface, but, on the other hand, the period 1995–1998 is short. The surface slopes upwards to the north at an angle similar to that of the 1970s, but in contrast, the surface slopes downwards to the east. The slide toward the east does not differ much from the situation in the 1980s. The bi-linear saddle surface is almost straight and visually is very similar to the planar surface (Equation 6; Fig. 3). (6) z = – 1535 + 0.279y + 0.325x – 0.0000624xy (1995–1998) Although the coefficient of x is positive, the negative xy-component in the equation forces the surface to slope downwards to the east. The growth in manufacturing employment thus de- clines slightly towards the east. The entire period from 1970 to 1998 is depict- ed by a planar surface that inclines to the east and climbs to the north (Equation 7). The employment change in manufacturing has been negative in the south and southeast of Finland. The zero-line runs from Central Ostrobothia to the northern parts of North Karelia, north of Kokkola to Lieksa. (7) z = – 5881 + 0.980y – 0.302x (1970–1998) The bi-linear saddle surface of the three dec- ades reveals the dominant effect of the decline caused by deindustrialisation but, as one can ex- pect, the surface is at a level higher than during the deindustrialisation period (Equation 8; Fig. 4): (8) z = 58515 – 8.40y – 19.16x + 0.00275xy (1970–1998) During the period 1970–1998, the urban mu- nicipalities lost a total of 90,000 manufacturing jobs. Rural municipalities gained a net increase of 26,000 jobs during the same period. If the change in the coverage of the industrial statistics is taken into account, the net loss of jobs was even higher. The strong decline in employment expli- cates the dominance of the deindustrialisation pattern over the entire period. Conclusions The late-Fordist boom of the 1970s, deindustrial- isation, and the emergence of the new economy brought about the geographical transformations of the Finnish manufacturing sector. There are com- petitive factors which explain the changes. For instance, in the 1970s, the boom resulted largely from the growth of the labour-intensive light in- dustry and the expansion of traditional manufac- turing. This spatial diffusion of manufacturing ap- pears to be related to the labour advantage of the peripheries, growth of new industrial activities, and regional policy. It was more profitable, with the help of state support, to expand production in the peripheries than in the industrial cores. There was strong political will to utilise the comparative advantage of the country geographically and, thus, to mobilise the production factors of Finland’s re- mote regions. This was part of the planning doc- trine of the time (e.g., Jatila 2001: 3–10). Deindustrialisation consisted of two types of job losses. First, many industrial cities, especially Helsinki, lost their comparative advantage in manufacturing employment to smaller cities, towns and rural municipalities in the late twenti- eth century. Nevertheless, Helsinki boomed with- out any expansion of the manufacturing sector. Many structural factors (skills, labour force, rents, congestion, etc.), as well as the economic policy, influenced this decline. The more cyclical dein- dustrialisation emerged when the global shift to Newly Industrialised Countries (NIC) took place and the old Fordist industrial infrastructure lost its competitiveness. The collapse of the Soviet trade FENNIA 180: 1–2 (2002) 225Spatial turns of manufacturing since 1970 delivered the final blow in Finland. Traditional production suffered and employment declined. Restructuring took place, and the state funded re- search and development to an increasing degree to develop new industrial activities. The growth of output and employment resumed in the late 1990s. The driving forces of growth were different from those in the 1960s and 1970s, however. A relatively narrow section of the econ- omy with specific locational requirements boost- ed the economy. The late 1990s’ geographical growth pattern of the manufacturing sector was clearly bifurcated into two phenomena: first, the R&D-driven ICT-industry which favoured certain urban or semi-urban localities and, second, the growth of the rural manufacturing sector which was a sort of spin-off from the long-lasted indus- trial growth in Finland. To sum up the main trends, the manufacturing growth of the peripheries in the early stage of the regional policy era changed gradually into a spa- tially more selective development of localities in advanced environments. Moreover, the 1970s’ growth of manufacturing employment in the east turned to growth in the west in the 1990s. Why it happened and whether this shift is a longer-last- ing phenomenon, deserves further research. REFERENCES Ali-Yrkkö J, L Paija, C Reilly & P Ylä-Anttila (2000). NOKIA – A big company in a small country. ETLA B 162. Helsinki. Barkley DL (1988). The decentralization of high-tech- nology manufacturing to nonmetropolitan areas. Growth and Change 19, 13–30. Bulletin of Statistics 2000: III. Statistics Finland, Hel- sinki. Bulletin of Statistics 2001: I. Statistics Finland, Hel- sinki. Ennakkotietoja teollisuudesta 2000 (2001). Teolli- suus 2001: 7. Statistics Finland, Helsinki. Glasmeier A (1988). Factors governing the develop- ment of high tech industry agglomerations: A tale of three cities. Regional Studies 22, 287–301. Haynes KE & ZB Machunda (1987). Spatial restruc- turing of manufacturing and employment growth in the rural Midwest: an analysis for Indiana. Eco- nomic Geography 63, 320–333. Hedegaard L (1999). The concentration of the struc- tural funds. North 10: 6, 7–10. Hjerppe R (1989). The Finnish economy 1960–1985. Government Printing Office, Helsinki. Jatila U (2001). Yritysten sijaintipaikan valinta pää- töksenteko-ongelmana. Unpublished Phil. Lic. thesis in economic geography. Helsinki School of Economics and Business Administration. Kansantalouden tilinpito 1975–1998 (2000). Kansan- talous 2000: 1. Statistics Finland, Helsinki. Fig. 4. Bi-linear saddle surface on the em- ployment changes in the manufacturing sec- tor in Finland, 1970–1998. 226 FENNIA 180: 1–2 (2002)Markku Tykkyläinen Kiander J & P Vartia (1998). Suuri lama. ETLA B 143. Helsinki. Korkiasaari J & I Söderling (1994). Muuttoliike. In Koskinen S, T Martelin, I-L Notkola & K Pitkänen (eds). Suomen väestö, 226–264. Gaudeamus, Helsinki. Moberg K (1899). Metallindustri och stenbrott. Atlas öfver Finland, Text 23, 1–30. Sällskapet för Fin- lands Geografi, Helsingfors. Municipalities 1998. Regional classifications hand- book (1998). Statistics Finland, Handbooks 28. North D (1998). Rural industrialization. In Ilbery B (ed). The geography of rural change, 161–188. Longman, Essex. Norton RD & J Rees (1979). The product cycle and the spatial decentralization of American manu- facturing. Regional Studies 13, 141–151. OECD Economic outlook 69: 1 (2001). OECD, Paris. Official statistics of Finland XVIII. Industrial statis- tics 1970–1990 (1973–1992). Statistics Finland, Helsinki. Osaamiskeskukset (2001). 13 August 2001. Palmén J (1899). Forsar. Atlas öfver Finland, Text 22, 1–30. Sällskapet för Finlands Geografi, Helsing- fors. Palmén K (1899). Industrin. Atlas öfver Finland, Text 24, 1–20. Sällskapet för Finlands Geografi, Hel- singfors. Palomäki M (1980). Industrial development and re- gional standard of living in Finland. In Walker DF (ed). Planning industrial development, 189–213. John Wiley & Sons, Chichester. Rautava J & J Hukkinen (1992). Russia’s economic reform and trade between Finland and Russia. Bank of Finland Bulletin 66: 4. Siirilä S, L Hautamäki, J Kuitunen & T Keski-Petäjä (1990). Regional well-being variations in Finland. Fennia 168, 179–200. STV 1982 = Statistical yearbook of Finland 1982 (1983). Statistics Finland, Helsinki. STV 1995 = Statistical yearbook of Finland 1995 (1995). Statistics Finland, Helsinki. STV 1997 = Statistical yearbook of Finland 1997 (1997). Statistics Finland, Helsinki. STV 2000 = Statistical yearbook of Finland 2000 (2000). Statistics Finland, Helsinki. Teollisuusaineisto (2001). Markku Tykkyläisen arkis- to, Kansio 97. Maakunta-arkisto, Joensuu. Tervo H (1983). Tuotantotoimintaan suunnatun alue- politiikan työllistävä vaikutus kehitysalueilla 1966–1980. Jyväskylän yliopisto, Taloustieteen laitos, Keski-Suomen taloudellinen tutkimuskes- kus, julkaisuja 57/1983. Tutkimus- ja kehittämistoiminta Suomessa 1998 (1999). Official statistics of Finland. Tiede, tek- nologia ja tutkimus 1999: 2. Statistics Finland, Helsinki. Tykkyläinen M (1987). Teollisuuden työvoimamuu- tokset 1980-luvun alkupuoliskolla Suomessa. Ter- ra 99, 49–56. Tykkyläinen M (1988). Periphery syndrome – a re- interpretation of regional development theory in a resource periphery. Fennia 166, 295–411. Tykkyläinen M (1996). The legacy of post-war set- tlement policy. Finnish Journal of Rural Research and Policy 3 (English Supplement), 85–93. Tykkyläinen M (1999). Metsästä elektroniikkateolli- suuteen. In Westerholm J & P Raento (eds). Suo- men kartasto, 96–99. Suomen Maantieteellinen Seura & WSOY, Helsinki. Yearbook of industrial statistics, Volume 1 1992 (1992). Official statistics of Finland XVIII. Manu- facturing 1992: 7. Statistics Finland, Helsinki. Yli-Jokipii P & A Koski (1995). The changing pattern of Finnish regional policies. Fennia 173, 53–67.