Shrinking Cities - Shrinking Economy?The Case of East Germany

Peter Franz

Shrinking Cities - Shrinking Economy? The Case of East Germany

1. Boom in Negative Scenarios for East German Cities
2. Growing and Shrinking Economy - Why Growth Indicators can be Ambiguous
3. Causal Links
    3.1 The Growth Rate as an Independent Variable
    3.2 Demographic Development as an Independent Variable
4. Empirical Findings on East German Cities and Urban Regions
    4.1 The Cities
          4.1.1 Growth indicators for per capita income
          4.1.2 Growth indicator gross value added
          4.1.3 Interim conclusion
    4.2 Urban Regions
5. Summary and Outlook

Notes
References

Abstract:
In the discussion on the future of shrinking cities, it is often assumed that a declining population means lagging economic growth. This article looks into the potential interplay between demographic factors and economic growth factors and indicators. Over the periods 1994-96 and 1998-2000, no concurrent demographic and economic decline was yet apparent in major East German cities. However, a pattern of growth without employment effects (jobless growth) was in evidence in many cities. Moreover, findings suggest that the economic development paths of cities are progressively diverging.

 

1. Boom in Negative Scenarios for East German Cities

In the fourteen years since German unification, East German cities initially made relatively uniform economic progress, but in recent years their development has diverged. Following the discussion on widespread, growing housing vacancies, a picture of East German cities as shrinking communities has become established, with far-reaching urban development, social, and economic consequences. Experts are increasingly claiming that demographic decline and economic dynamics are interlinked. Lang and Tenz draw the following conclusion: "From the urban development point of view, a shrinking city is characterised by two causal processes: a loss of population and faltering economic momentum. This demographic and economic decline triggers processes in all areas of urban development" (Lang/Tenz 2003, 130)

In the medium term, almost all East German cities are believed threatened by a downwards spiral: shrinking population, growing housing vacancies, urban decay, declining economic basis, diminishing attractiveness, growing outmigration.(1) Negative scenarios of this type can become self-reinforcing and self-fulfilling prophecies. But academic observers should subject such pessimism to thoroughgoing scrutiny. They must investigate the extent to which the posited contextual relationship is empirically grounded. There is no disputing that sections of the housing industry and municipal undertakings responsible for the infrastructure in cities with vacant housing are suffering losses (Franz 2001). But whether the regional economy as a whole is being negatively influenced is an open question that can only be answered empirically. Studies on the development of local government finances in 118 municipalities in Brandenburg and Saxony show that shrinking cities do not always experience a deterioration in income position (Seitz 2002). This article therefore looks at the question of whether a decline in the population of a city influences its economic performance from the point of view of the regional economy.

Chapter 2 examines individual factors that can be responsible for demographic and economic decline processes not running parallel. We go on to show that certain common growth indicators are themselves influenced by demographic change. The outcome is a typification of conceivable growth and decline constellations.

Chapter 3 presents various hypotheses about a causal link between demographic and economic decline. On the basis of statistical data, chapter 4 investigates what decline or growth constellations are to be found in major East German cities and urban regions. The article ends with a summary and a look at future prospects (chapter 5).

 

2. Growing and Shrinking Economy - Why Growth Indicators can be Ambiguous

In studying the effects of shrinkage, a bundle of complex interactions between economic and demographic factors have to be unravelled. This is rendered more difficult by the circumstance that important indicators for economic growth are influenced by both changes in economic performance and by changes in population or employment figures.

The thesis that cities with a shrinking population also suffer deteriorating economic performance implies a high positive correlation between demographic development and economic growth. The literature on macro-economic growth sees as established a high positive correlation between gross domestic product (GDP) growth and population growth. The link is so close that the population growth rates are often taken as a proxy indicator for economic growth.(2) A fall in population figures would accordingly indicate economic stagnation or even a decline in economic performance. However, it is not clear whether this long-term correlation can weaken in the shorter interim term or whether it ceases to operate in phases of economic upheaval, such as that experienced in East Germany after 1990.

There are two main grounds for assuming a looser correlation in the case of East Germany.

• First, owing to persistently high unemployment, a reduction in the labour supply due to a decline in population is hardly likely to limit growth, since the labour factor will continue to be available at an adequate level even if population figures continue to fall.
• Second, studies on the phenomenon of jobless growth indicate that economic and employment growth need not always show a high positive correlation in the shorter term.(3) For East German states, jobless growth could be particularly important, since various studies on productivity in Germany have noted a considerable and enduring productivity gap for firms in the region. Many East German enterprises are therefore likely to achieve substantial progress in productivity in the years to come. This could lead to gains in growth not necessarily accompanied by higher employment rates. This means that the productivity gap offers potential for growth, especially in East Germany, that could be achieved without a major impact on the level of employment - and thus relatively independently of supply changes on the labour market.

Before going on to look at causal links in chapter 3 - a statistically high correlation between two variables does not, of course, automatically mean they are causally linked - certain other aspects need to be discussed. In portraying regional economic growth, the standard measures are nominal gross value added per person employed and per capita income (Bode 2002, 363).(4) But taking persons employed or population as normalization denominators means that values are influenced by changed employment and population numbers in the region under study as soon as several points in time are examined and compared.(5) This can mean that per capita income in a region can increase at an above-average rate in spite of stagnating economic output because the number of inhabitants has fallen more strongly than in other regions. This is purely statistically determined growth. Another conceivable case is jobless growth, where growing output is not matched by an increase in employment. Both of these cases would mean that growth and demographic or employment development show only weak positive or even negative correlation (field I in overview 1). Weak or negative correlation also results from another conceivable constellation, namely where a region records a rise in population accompanied by below average economic growth since the additional jobs are in low-productivity sectors. This could occur in a region with a high proportion of jobs in tourism (field IV in overview 1). These constellations are summed with the "normal cases" of parallel change in population and growth in overview 1 (grey-shaded fields II and III in overview 1). While changes in economic growth indicators are considered relative to development at the national level, simple change rates are taken for changes in population and employment (Bode 2002, 369).

 

The subdivision in overview 1 shows how the individual regions under study distribute across fields I to IV. In chapter 4 this classification is made for selected East German cities. With the majority of cities being assigned to fields II and III, positive correlation between change in growth indicators and change in population or employment figures is to be expected. Vice versa, if most cities are assigned to fields I and IV, the two factors would correlate negatively.

 

3. Causal Links

Having considered various constellations which could lead to economic and demographic developments in a region correlating in different ways, causal links must now be examined. Since the two factors interact, each will be looked at in turn as an independent variable.

In concrete empirical cases the possibility of mutual influence must be taken into account. Bürkner points out that this uncertainty makes it easier for politicians to pick out the cluster of causes that suits them, and concludes that there is a trend among East German local politicians to thematize shrinkage and urban redevelopment while erecting "a wall of silence" around possible economic causes in order to hide their own impotence or to avoid negative effects for the image of the community concerned (Bürkner 2001, 46).

 

3.1 The Growth Rate as an Independent Variable

If it is true that the extent of regional economic growth influences population numbers in a region, it should be remembered that the population figure reflects natural population movement (births, deaths) and supraregional migration, i.e., the balance of inmigration and outmigration. As a rule, it is assumed that the growth situation is mainly responsible for the scale of migration. Prosperous regions attract people and households who take advantage of better employment and income opportunities. Vice versa, the number of departing people and households exceeds the number arriving in regions that are stagnating and suffering lower economic performance. The literature on spatial mobility refers to push and pull factors (Franz 1984, 58f.). From a micro-economic perspective these processes are termed "voting with your feet" (Tiebout 1956).

In the long-term, economic growth is ascribed influence not only on migration but also on reproductive behaviour. It is assumed that high (low) demographic growth in a region leads to an optimistic (pessimistic) view of the future with corresponding encouragement (discouragement) of reproduction.

If natural population movements and the migration balance point in the same direction, shifts in the migration balance are directly reflected in population figures: a birth deficit in conjunction with a negative migration balance is currently regarded as the "normal case" in East Germany. If the two factors change in opposing directions, the migration balance has only an attenuated impact on negative demographic development.

 

3.2 Demographic Development as an Independent Variable

In pursuit of these cognitive interests, this article concentrates on the economic effects of a decline in population. From an economic point of view, a falling population in a city or region has supply and demand-side effects. It reduces the labour supply available in the region as well as the overall demand of private households for consumer goods and services (overview 2).

 

It might, however, be the case that population losses in a city have a statistical impact but no economic effect. This is true of migration from cities to suburbs which involves a change in residential location but usually neither a change in job nor in consumer behaviour. Investigation within narrow administrative boundaries is inadequate in this regard. It needs to include the complete urban region (Barkjak et al. 2000, 51f.). This holds not only for residential suburbanisation but also for the suburbanisation of industry and services (Franz 2002, 128f.). The growth expected for East German cities after 1990 proved greater in the areas surrounding major cities than within city bounds (Barjak et al. 2000, 51).

These arguments for taking the urban field into account implies that different patterns of demographic decline influence growth factors in different ways. A fall in population caused primarily by suburbanisation appears to be less damaging to growth than a decline due mainly to the outmigration of younger carriers of high-quality human capital (overview 2). In East German cities birth deficits have caused the greater part of population decline in recent years.(6) A deficit of births affects the labour supply only with considerable delay; with regard to consumer demand, birth deficits lead to smaller households, which demand fewer consumer goods than larger households.

 

4. Empirical Findings on East German Cities and Urban Regions

The empirical testing of all the above hypotheses would exceed the scope of this article. The empirical findings to be presented are intended to provide a first well-founded answer to the question of the link between population and economic decline in East German cities with reference to the distinctions shown in overview 1.

The first step assigns the 26 East German independent cities (kreisfreie Städte) (excluding Berlin) to the four combinations of demographic and economic decline or demographic and economic growth listed in overview 1. This is done for the two periods 1994 to 1996 and 1998 to 2000 to determine whether individual cities have "shifted" from one category to another in the course of time. The two periods differ with regard to outmigration dynamics. Whereas migration from East Germany fell steadily after 1994 and almost came to a halt in 1996, population losses from outmigration rose again from 1998.

The second step includes urban regions in order to take account of the argument advanced in section 3.2 that an urban region perspective takes better account of economic interrelations between cities and their surroundings. Since it is not possible to define the urban regions of all East German independent cities in an identical way, only eleven urban regions have been selected, which reduces the general validity of the findings. The above hypotheses suggest that demographic decline at the urban region level would be less and growth and employment gains greater that in the core city alone.

The findings are discussed with the aid of three scatter diagrams for the periods 1994-1996 and 1998-2000. The space between the x and y axes is subdivided by a horizontal and a vertical line. In keeping with the differentiation shown in overview 1, the horizontal line marks the national average of change in per capita income or the gross value added per person employed for the periods concerned. To the left of the vertical line are cases with declining population or employment (fields I and III), to the right of the line instances of growing populations or employment in the relevant periods (fields II and IV).

 

4.1 The Cities

 

4.1.1.Growth indicators for per capita income

The first diagram in figure 1 shows that all East German cities lost residents between 1994 and 1996. In this period, most cities also experienced above-average economic growth (field I in overview 1).

 

The municipalities of Mecklenburg-West Pomerania concentrated in the top left corner show the greatest contrast between growth rate and population loss.(7) Only four cities, of which three in Thuringia, record below-average growth rates, thus belonging in field III (cf. overview 1). No city experienced "minus growth."(8) 

In the second period under study, 1998-2000, this relatively uniform picture presented by the preceding period becomes more differentiated (cf. lower diagram in figure 1). All four fields are now occupied, albeit by only one city in two cases. It is plausible that Weimar as a popular tourist destination should experience an expansion in labour-intensive sectors with relatively low productivity (field IV).(9) More cities than in the preceding period recorded below-average growth rates accompanied by a fall in population (field III). Only now are cities with "minus growth" recorded: Zwickau, Eisenach, and Plauen.(10) It is also interesting that the Saxon cities of Leipzig, Dresden, and Chemnitz, despite various moves to absorb surrounding municipalities in 1999, have not achieved any growth in population.(11) 

 

4.1.2 Growth indicator gross value added

In contrast to the per capita income indicator, the values of the "gross value added per person employed" indicator are more widely scattered across all fields I-IV already for the period 1994-1996 (cf. upper diagram in figure 2). The vast majority of municipalities record above-average growth in productivity, with the greatest leaps being achieved by Jena, Zwickau, and Chemnitz. In these cities, the effects of major industrial investment are clearly apparent. Common to the four cities Stralsund, Dresden, Dessau, and Neubrandenburg is above-average productivity growth and an increase in employment (field II). Only five municipalities (Weimar, Magdeburg, Erfurt, Hoyerswerda, und Görlitz) show below-average productivity development. Unlike for the per capita income indicator, there is no clustering of cities from a single state in one of the four fields.(12) 

Comparing the two periods, the most striking change is that in 1998-2000 more municipalities remained below the average figure for productivity growth (nine to six; cf. lower diagram in figure 2). At the same time, only slightly more cities (ten to eight) recorded growth in employment. They include the three largest Saxon cities and the four largest Thuringian municipalities. There is much to suggest that expanding employment in the service sector is responsible for the change over the preceding period, a sector with lower productivity than manufacturing. The shift by Weimar from the bottom left (upper diagram) to the top right field (lower diagram) supports the assumption that the service sector has grown particularly strongly there.(13) 

 

4.1.3 Interim conclusion

On the basis of the findings presented so far, the initial question as to whether cities experiencing demographic decline can at the same enter a phase of below-average or even negative economic growth can be answered as follows.

• In spite of shrinking populations, most East German cities recorded above average growth rates in the period 1994-1996. In the majority of cities, this growth was accompanied by a decline in the number of people in work. This pattern corresponds to the constellation of jobless growth. Statistically determined growth is however, not to be excluded precisely in the cities suffering the greatest drop in population and employment.
• In the period 1998-2000, there was an overall deterioration in growth momentum accompanied by persistent demographic decline in the cities under study. Few still experienced above-average productivity gains. At the same time, development paths tended to differentiate: the number of municipalities to see a revival in employment increased.  

  Figure 2: Changes in relative productivity and employment in 26 East German county boroughs (1994-1996, 1998-2000) GVA p.e.: gross value added per employeeSource: state statistical offices; own calculations.

• A minus sign before the correlation coefficients for all four diagrams means that the majority of cities more frequently occupy fields I and IV (cf. overview 1), where growth indicators and population or employment numbers often change in opposing directions. Parallel demographic and economic decline is the exception.

The following look at urban regions will show whether these findings are relativised when the urban field is taken into account.

 

4.2 Urban Regions

Figure 3 shows that falling population figures predominate not only in the city but also in the urban region. This means that, although the level of outmigration from cities into surrounding areas improves the migration balance of the urban field, the deficit of births in the city and in surrounding areas, as well as their negative migration balance with other regions cannot be compensated overall. Only the urban region of Weimar and - from 1998 on - that of Jena have managed to gain population.

The suspicion expressed in section 3.2 that per capita income developed more favourably in urban regions than the cities themselves is confirmed by the distribution of the eleven urban regions under study (cf. upper diagram in figure 3).(14) In 1998-2000, all four fields (I-IV) are occupied, which means that the urban regions, like the cities, had begun to differentiate more strongly as regards economic development. Like the cities (cf. figure 1), urban regions showed overall lower growth rates in the period 1998-2000, and some similarly "slid" over the line of average growth.(15) Findings relating to the growth indicator "gross value added per person employed" (not shown here) do not basically deviate.(16) 

   

5. Summary and Outlook

The question in the title of this article as to whether the shrinkage of cities is to be equated with a shrinkage of the economy can be answered with "No, almost not" for the period 1998-1996 and with "Yes for a large number of the cities under study but not for the majority" for the period 1998-2000. There are signs that, after a phase of relatively uniform development, East German cities increasingly embarked on diverging development paths in the second period (cf. Winkler 2004).

The findings invite further conclusions. In major East German cities, the prevailing constellation is economic growth coupled with a fall in population and employment. In this case the informative value of the two growth indicators needs to be assessed with particular caution, since without in-depth studies there is a risk of being misled by statistical artefacts.(17) 

This study shows the need for more research. In order to place the findings on a firmer empirical basis, it would be useful to include further smaller towns and cities and, above all, to increase the number of urban regions under study. But even with such supplementation, many other interesting questions will remain. In order to test whether a decline in population has a differential impact on sectors that concentrate on regional business and those that do business beyond the region, GDP and gross value added data would have to be differentiated by sector. This would, in turn, raise the question of the extent to which cities with different sectoral accents, e.g., service cities, industrial cities, administrative cities, or commercial cities, display different patterns of progress in productivity. In our discussion of the findings we suggested this might be the case for Weimar. In testing the various causal hypotheses presented in this article, it would also be possible, as German unification recedes into the past and longer time series become available, to undertake first causality tests on whether falls in population in East German cities are to be attributed to economic or demographic factors.

Greater scientific light on the links between economic and demographic developments could open new perspectives for local government development policy. Owing to intensive lobbying by the housing industry, the East German discussion on shrinking cities has been strongly reduced to the issue of eliminating vacant housing. But residents are likely to leave even the redeveloped or downsized city if they find no attractive economic opportunities there.

 

Notes

(1) For a description of such a downwards spiral see Bernt (2002, 41). Bürkner (2001, 53) distinguishes between "shrinkage chains" in stagnating and in prospering urban regions, implying that demographic decline can occur even in regions with above-average economic growth. (back)

(2) For a macroeconomic perspective on the link between economic and population development see Felderer (1990) and Birdsall et al. (2001). (back)

(3) In macroeconomics, the existence of "jobless growth," i.e. a weakening correlation between employment or unemployment and economic growth, has hitherto not been considered to have been established (cf. Döpke 2001). From a regional economic perspective it can be said that the East German regions with their smokestack industries have been characterised by a discrepancy between high industrial investment and low levels of employment since 1990 (cf. Barjak et al. 2000, 44). (back)

(4) A discussion of the substantive advantages and disadvantages of the indicators (cf. Franz 2003, 162 f.) is beyond the scope of this article. (back)

(5) "If production and employment are moving in the same direction, per capita income must necessarily change less than total income. In contrast, changes in total income will be exceeded if the denominator demographic and employment development is converse" (Bade/Niebuhr 1999, 135). (back)

(6) In Saxony-Anhalt, for instance, outmigration accounts for one third and the excess of deaths over births for two-thirds of demographic decline (Hardt et al. 2001, p. 68). (back)

(7) This means that these cities are most likely to find that growth in per capita income is only statistically determined. (back)

(8) The measure of association between growth indicator and demographic development is R = -0,21 (Bravais-Pearson correlation coefficient). For the correlation values presented here and below, only the sign is of interest. Owing to the small number of cases under study, statistically significant values are to be expected only beyond the threshold of R = +/-0,4. (back)

(9) Weimar can accordingly be expected to record below-average development in productivity. (back)

(10) In Eisenach, for which no data are available for the 1994-1996 period, a level effect is likely, since a great deal was invested there in the early 1990s in a new motor vehicle plant (Opel Eisenach). (back)

(11) The measure of correlation is R = -0,63 (back)

(12) The measure of correlation is R = -0,63 (back)

(13) The correlation value for the period 1998-2000 (R = -0,35) remains negative. (back)

(14) For the period 1994-1996 the correlation coefficient is R = -0.17. (back)

(15) R = 0,20. (back)

(16) The two correlation values of the indicators with employment development for 1994-1996 (R = -0.58) and 1998-2000 (R = -0.10) were negative. (back)

(17) One such statistically determined distortion is to be seen in the cities of the Ruhr District: "Gainful employment there has remained still farther below the national average than gross value added, so that per capita income, despite clear structural weaknesses, has risen more strongly than in the country as a whole" (Bade/Niebuhr 1999, 135). (back)

 

References

 Bade, Franz-Josef/Niebuhr, Annekatrin (1999), Zur Stabilität des räumlichen Strukturwandels, in: Jahrbuch für Regionalwissenschaft, H. 2, 131-156.

 Barjak, Franz/Franz, Peter/Heimpold, Gerhard/Rosenfeld, Martin (2000), Regionalanalyse Ostdeutschland: Die wirtschaftliche Situation der Länder, Kreise und kreisfreien Städte im Vergleich, in: Wirtschaft im Wandel, H. 2, 31-55. (back)

 Bernt, Matthias (2002), Risiken und Nebenwirkungen des "Stadtumbaus Ost". Leipzig (UFZ-Diskussionspapier 5/2002). (back)

 Birdsall, Nancy/Kelley, Allen/Sinding, Steven (2001), Population Matters: Demographic Change, Economic Growth, and Poverty in the Developing World, Oxford. (back)

 Bode, Eckhardt (2002), Aktive und passive Sanierung im Wachstumsprozess ostdeutscher Regionen, in: Die Weltwirtschaft, H. 4, 362-382. (back)

 Bürkner, Hans-Joachim (2001), Schrumpfung und Alltagskultur: Blinde Flecken im Stadtumbau-Diskurs, in: Keim, Karl-Dieter (ed.), Regenerierung schrumpfender Städte - zur Umbaudebatte in Ostdeutschland, Erkner, 41-67. (back)

 Döpke, Jörg (2001), The "Employment Intensity" of Growth in Europe, Kiel (Kiel Working Paper No. 1021). (back)

 Felderer, Bernhard (1990) (ed.), Bevölkerung und Wirtschaft, Berlin. (back)

 Franz, Peter (1984), Soziologie der räumlichen Mobilität, Frankfurt/M. und andere. (back)

 Franz, Peter (2001), Leerstände in ostdeutschen Städten: Keineswegs nur ein wohnungspolitisches Problem, in: Wirtschaft im Wandel, H. 2, S. 27-34. (back)

 Franz, Peter (2002), Suburbanisierung von Industrie und Dienstleistungen, in: Institut für Länderkunde (Hrsg.), Nationalatlas der Bundesrepublik Deutschland - Dörfer und Städte (Bd. 5), Heidelberg, 128-129. (back)

 Franz, Peter (2003), Wie schneiden die großen ostdeutschen Städte im gesamtdeutschen Vergleich ab? Daten zu ihrer wirtschaftlichen und finanziellen Lage, in: Raumforschung und Raumordnung, H. 3, 160-169. (back)

 Hardt, Gabriele/Kempe, Wolfram/Schneider, Hilmar (2001), Ost-West-Wanderung junger Menschen Ausdruck für allgemeinen Anstieg der Mobilität, in: Wirtschaft im Wandel, H. 4, 67-74. (back)

 Lang, Thilo/Tenz, Eric (2003), Von der schrumpfenden Stadt zur Lean City: Prozesse und Auswirkungen der Stadtschrumpfung in Ostdeutschland und deren Bewältigung, Dortmund. (back)

 Seitz, Helmut (2002), Kommunalfinanzen bei schnell schrumpfender Bevölkerung in Ostdeutschland: Eine politikorientierte deskriptive Analyse, Berlin/Frankfurt/O., Online-Manuskript: http://www.makro.euv-frankfurt-o.de. [Aktualisierter Link: http://www.tu-dresden.de/wwvwlemp/inhaber/publications/Seitz(2002)_KommunalfinanzenBeiSchnellSchrumpfenderBevoelkerungInOstdeutschland.pdf, 17.01.2011] (back)

 Tiebout, Charles (1956), A Pure Theory of Local Expenditures, in: Journal of Political Economy, 64. Jg., 416-424. (back)

 Winkler, Cordula (2004), Die Wirtschaftsstruktur der ostdeutschen Großstädte: Ökonomische Unterschiede nehmen zu, in: Wirtschaft im Wandel, H. 2, 46-53. (back)

[previous article | index | next article]

Weitere Verweise auf diesen Beitrag: 

Zeitschriftenartikel - Growth and Shrinkage in Germany - Trends, Perspectives and Challenges for Spatial Planning and Development.

Zeitschriftenartikel - Demographic Decline, Segregation, and Social Urban Renewal in Old Industrial Metropolitan Areas