Docstoc

Analyzing Socio-Economic Impacts of Large Investments by Spatial

Document Sample
Analyzing Socio-Economic Impacts of Large Investments by Spatial Powered By Docstoc
					Analyzing Socio-Economic Impacts of Large Investments by Spatial
Microsimulation

Urban Lindgren
Department of Social and Economic Geography, Umeå University, SE-901 87, Sweden;
email: urban.lindgren@geography.umu.se
Magnus Strömgren
Department of Social and Economic Geography, Umeå University, SE-901 87, Sweden;
email: magnus.stromgren@geography.umu.se
Einar Holm
Department of Social and Economic Geography, Umeå University, SE-901 87, Sweden;
email: einar.holm@geography.umu.se
Erling Häggström Lundevaller
Department of Social and Economic Geography, Umeå University, SE-901 87, Sweden;
email: erling.lundevaller@geography.umu.se (also affiliated with Department of Statistics, Umeå Uni-
versity)


ABSTRACT: In the near future, a nuclear waste repository will be located in either Östhammar or
Oskarshamn, two Swedish municipalities. This is a major investment that is likely to have socio-
economic implications at the local level for several decades. In order to analyze the indirect local effects
of such large investments, a spatial and dynamic microsimulation model (SVERIGE 3) has been con-
structed. The model simulates demographic events (e.g., fertility and migration) as well as education and
the labor market. In this study, the simulation model is utilized to evaluate a number of scenarios com-
prising various potential investments in Östhammar, one of which is the nuclear waste repository. As
part of the study, the direct local effect of the investments was estimated. When running the model, the
estimated direct local effects function as exogenous economic input to concerned labor market sectors.
The results of the simulations indicate that investments such as the nuclear waste repository will have
some economic and demographic effects. However, infrastructure projects that increase accessibility
seem to generate more profound and long-lasting effects at the local level. A municipality such as
Östhammar, located close the Stockholm metropolitan area, may be especially likely to benefit by such
infrastructure investments.




                                                     1
1. INTRODUCTION


In the near future, a nuclear waste repository will be located in either Östhammar or Oskarshamn, two
of Sweden‘s 290 municipalities. Establishing a nuclear waste repository is a major investment that will
involve a construction phase, but also an operation phase extending for several decades. Consequently,
such an investment is likely to have long-term socio-economic implications at the local level. The so-
called direct effects, such as local purchases of goods and services, can be predicted fairly accurately
based on the study of available planning material, investigations of the local economic structure, and
experiences of the local impacts of other large investments. However, more indirect, long-term effects
on the local economy and population structure—which may turn out to be just as significant—are
much more difficult to estimate.


Microsimulation is one possible approach for estimating such indirect local effects of an investment.
Micro-orientated modeling been used for the analysis of many different kinds of socio-economic phe-
nomena, such as population dynamics, labor markets, and welfare systems. However, many microsimu-
lation models are non-spatial in the sense that they cannot use geographical information for policy
analysis targeting small areas. Recently, efforts have been made to incorporate the spatial dimension in
microsimulation models (e.g., Hooimeijer, 1996; Veldhuisen et al., 2000). In the UK spatial microsimula-
tion models have been developed for analyzing local labor market problems and policies (Ballas and
Clarke, 2000), social and regional policy (Ballas and Clarke, 2001), and population dynamics (Ballas et
al., 2005). In this study, the estimation of various investments‘ direct effect is combined with a spatial
microsimulation approach, in order to also shed light on the more elusive indirect effects at the local
level. Specifically, a spatial and dynamic microsimulation model (SVERIGE 3) is utilized to evaluate a
number of scenarios comprising various potential investments in the municipality of Östhammar, one
of which is the investment in the nuclear waste repository.


The presentation of the study is structured in five sections. Following this short introduction, there is a
description of the studied investments and simulated scenarios, as well as the local setting they concern,
i.e., Östhammar municipality. In the next section, the microsimulation model is presented, with particu-
lar attention devoted to the way the simulation model handles the labor market and the studied scena-
rios. After the description of the model, simulation results are presented followed by a concluding
discussion.




                                                    2
2. SCENARIOS OF INVESTMENTS IN ÖSTHAMMAR


This study uses the microsimulation model SVERIGE 3 to evaluate the impacts of potential future
investments in the municipality of Östhammar. The investments, which are ―packaged‖ in scenarios,
include the investment in the nuclear waste repository and related facilities, but also certain other in-
vestments that are likely or at least possible to come about. The study concerns the time period 2000–
2060, the last year of the currently planned operation phase of the nuclear waste repository.


2.1 The investments and the scenarios
Since the beginning of the 1990s, Swedish Nuclear Waste and Fuel Management Co (Svensk
Kärnbränslehantering AB, also known as SKB) has carried out investigations in order to find a suitable
location for a nuclear waste repository. A number of potential sites has been considered and subse-
quently ruled out, either because of geological unsuitability or local political and public skepticism. To-
day, only two candidate municipalities remain, Östhammar and Oskarshamn. Both municipalities have
nuclear power plants, opened in 1980 and 1972, respectively. Today, there are three reactors in opera-
tion at each facility.


The nuclear waste repository is assumed to be completed by 2020 and be in operation until 2060. There are
other investments related to the nuclear waste repository as well, which will or may be located in
Östhammar. There will be an expansion of an already existing facility, SFR. SFR is a repository for low-
and intermediate-level radioactive waste, which has been in service in Östhammar since 1988. In addi-
tion, a “capsule factory” will be constructed, and may be located in Östhammar. The capsule factory will
produce the copper capsules that will be used to enclose the nuclear waste. Furthermore, SKB has
stated that part of its central management will relocate to the municipality where the repository will be
located.


Three other investments are part of the constructed scenarios. Considering the heavy demand for iron
ore and other metals on the world market, there are plans to reopen the Dannemora mine. The reopening
of the mine, which closed in the beginning of the 1990s, will entail substantial investments for purposes
of upgrading the standard of the mining process. An investment in increased local housing supply is also
part of the study. It is assumed that at least all planned waterfront housing in the municipality is con-
structed. Finally, an investment in upgrading the road between Östhammar and Uppsala (Road 288) to
expressway standard is considered. In practice, this means a much better road, with the speed limit in-
creased to the Swedish maximum of 110 km/h.




                                                    3
Figure 1 shows the timeline of the construction and operation phases of the studied investments. The
length of the construction phase ranges from 2 years (the Dannemora mine) to 25 years (the SKB-
related investments). Due to lack of reliable data, no construction phase has been incorporated in the
study concerning the relocation of the SKB management. During the later stage of the construction
process, operation of the nuclear waste repository and the capsule factory will commence. The mining
at Dannemora will start in 2009, once the mine has been modernized, and is assumed to continue dur-
ing the remainder of the study period.




                                    SFR


                 Nuclear waste repository
                                                                             Operation phase
                                                 Construction phase
                         Capsule factory
    Investment




                       SKB management


                        Dannemora mine


                                 Housing


                                    Road


                                        2000   2010     2020          2030       2040          2050   2060
                                                                      Year

Figure 1 Timeline of the construction and operation phases of the studied investments




2.1.1 Simulated scenarios
In total, nine different scenarios of what the future has in store for Östhammar were constructed and
subsequently simulated. In this study, the simulation results for seven of these scenarios are presented
(Table 1). Scenario 1, If nothing happens, is a ―null scenario‖; only the already decided expansion of SFR is
carried out. The outcomes of simulating this scenario is intended to function as a reference, to which
the results of the other scenario simulations can be compared. In Scenario 2, Dannemora mine is reo-
pened; this investment is present in all subsequent scenarios as well. Scenario 3, Nuclear waste repository,
also includes the nuclear waste repository and SKB management. Scenario 4, Repository and capsule factory,
adds the capsule factory. Scenario 5, All SKB investments and housing, also contains the investment in in-
creased local housing supply. Scenario 6, Road 288, comprises the expansion of SFR, the opening of
Dannemora mine, and the road investment. Scenario 7, Maximum local development, finally, is a scenario
where all the addressed investments are carried out.


                                                                 4
Table 1 The simulated scenarios
                                                            Investment
          Scenario
                                    1        2          3         4           5          6       7
1 If nothing happens                        SFR
                              Dannemora
2 SFR and mining                            SFR
                              mine
                              Dannemora         Nuclear waste             SKB
3   Nuclear waste repository                SFR
                              mine              repository                management
    Repository and            Dannemora         Nuclear waste   Capsule   SKB
4                                           SFR
    capsule factory           mine              repository      factory   management
    All SKB investments       Dannemora         Nuclear waste   Capsule   SKB
5                                           SFR                                      Housing
    and housing               mine              repository      factory   management
                              Dannemora
6   Road 288                                SFR                                                Road
                              mine
                              Dannemora           Nuclear waste Capsule SKB
7   Maximum local development               SFR                                    Housing Road
                              mine                repository    factory management




2.2 On Östhammar municipality
The setting of the experiments, Östhammar municipality (Figure 2), is a coastal municipality in the
Swedish county of Uppsala. Bordering Uppsala municipality, Östhammar has a large population in its
immediate surroundings. The urban locality of Uppsala, as well as the northern part of the Stockholm
metropolitan area, are within commuting distance. Still, as demographic information from Statistics
Sweden reveals, Östhammar is a municipality with some ―rural‖ characteristics. With a 2006 population
of 21,435 persons, the population density is 14.6 inhabitants/km2. This is considerably lower than the
Swedish average of about 22 inhabitants/km2. The degree of urbanization is also comparatively low, 67
% compared to the average figure of 84 %. Furthermore, the actual urban localities in the municipality
are quite small in terms of population size. Only one town, Östhammar, has a population exceeding
3,000 inhabitants. In contrast to many other municipalities with similar demographic characteristics, the
population has increased in recent decades; the net population increase 1970–2006 was about 3,000
inhabitants. In terms of economic structure, Östhammar is heavily reliant on manufacturing, as well as
energy production through the Forsmark nuclear power plant. A large share of the population com-
mutes to other municipalities in the local labor market region, primarily Uppsala.




                                                    5
                                 FORSMARK
                        (nuclear power plant)

                                                     Öregrund


                                         Norrskedika
                                                           Östhammar

                  Dannemora
                           Österbybruk
                                     Gimo                         Hargshamn
                 DANNEMORA
                 MINE

                                                ROAD 288


                                            Alunda

                                        Skoby
          10
                 Kilometers
                                                           Map design: Magnus Strömgren, 2007


Figure 2 Östhammar municipality




3. MODEL DESIGN


The model used in this study—SVERIGE 3—is inspired by earlier models of the same kind, i.e., spatial
and dynamic microsimulation models. In this section, the lineage of the SVERIGE 3 model is briefly
presented. Then, two major challenges in creating the model are addressed. They concern 1) modeling
of the labor market and 2) estimating and representing the direct local effect that the investments will
contribute to. Finally, the model design is presented by an outline of graphical user interface, but also
with a look ―under the hood.‖


3.1 From SVERIGE to SVERIGE 3
In the development of SVERIGE 3, previous dynamic and spatial microsimulation models have consti-
tuted sources of inspiration. One such precursor, SVERIGE, simulates the population development in
Sweden, addressing events such as migration, education, and family formation (Holm et al., 2002). The
SVERIGE model has been used to carry out impact studies on a broad range of socio-economic prob-
lems. For example, Alfredsson (2002) used the SVERIGE model to analyze the impact of individuals‘
changed patterns of consumption on energy requirements and CO2 emissions. Rephann and Holm

                                                     6
(2004) examined the economic-demographic effects of different immigration policies and were able to
demonstrate their impacts concerning, among other things, migration, labor force participation, and
earnings.


The successor to SVERIGE, called LISA (but for all intents and purposes, constituting SVERIGE 2) is
less spatially explicit: individuals are not allocated to 100 meter squares, but rather to local labor market
regions. However, the model is more sophisticated when it comes to representing the labor market and
individuals‘ different sources of income (work income as well as various a transfer payments). Holm et
al. (2004) used the LISA model to study the impact of changes in transfer payment rules. The results
indicate that economic incentives have a significant effect on the number of recipients of transfer pay-
ments, particularly concerning unemployment insurance and pensions due to early retirement.


There are a number of similarities between the SVERIGE 3 model on the one hand, and the
SVERIGE and LISA models on the other. The models are all primarily stochastic, with event probabil-
ities mostly a function of logistic regression equations applied to micro unit characteristics. Many mod-
ules in SVERIGE 3 also simulate events that were implemented in the previous models as well. The
data source for the starting populations and the estimation of regressions, etc. is also the same: the lon-
gitudinal database ASTRID, containing, among other things, the entire Swedish population 1990–2003
with a large number of associated variables and a high degree of spatial resolution. However,
SVERIGE 3 differs from its predecessors in some important respects. First, while SVERIGE allocates
individuals to 100 meter squares, and LISA assigns them to local labor market regions, SVERIGE 3
focuses on the municipal level. Second, SVERIGE 3 has a specific focus municipality—the one where
the simulated investments take place—that may be modeled in greater detail than its surroundings.


Since one main purpose with SVERIGE 3 is to study the local impact of large investments, considera-
ble attention has been devoted to coming up with a reasonable model representation of the labor mar-
ket. Furthermore, the direct local effects that the investments will result in have been estimated, and
subsequently structured in such a way that it can function as exogenous input to the model economy
and its associated labor market.


3.2 Modeling the labor market
A central part of SVERIGE 3 is to simulate how investments influence the local economy and labor
market. Therefore, the construction of an appropriate model labor market has been a key issue in the
development of the simulation model. The model labor market is supposed to handle the matching
between demand for and supply of labor in different branches. Naturally, the working of the actual
labor market is an important reference for the simulation model to make sense. In this context, a cen-

                                                     7
tral aspect is the actual definition of occupations that the individuals in the model can have and apply
for.


A classification that potentially could form the basis for the model labor market is the Swedish Stan-
dard Classification of Occupations (Standard för svensk yrkesklassificering, SSYK). However, a number of
factors contributed to a decision not to utilize this classification. First, there were concerns as to the
reliability of the data, partly because a number of different sources have been utilized to assign the
population their respective profession. It should also be noted that the number of persons in the vari-
ous categories are remarkably uneven, ranging from almost 650,000 individuals (―personal care and
related workers‖ to a mere 150 persons (―fashion and other models‖). The major problem, however,
was that the classification is available as a variable in the ASTRID database for only the years of 2002
and 2003. This made it difficult to estimate the transition between different kinds of jobs over a longer
period of time—an aspect of the actual labor market that is among the most important factors to grasp
for purposes of modeling it.


Instead another solution was ultimately devised. The Swedish Standard Industrial Classification (Svensk
näringsgrensindelning, SNI) designates workplaces and the persons working there in different branches
according to the main economic activity of the workplace. The SNI classification is based on, and large-
ly corresponds to, the European Union branch classification system NACE (Nomenclature statistique
des Activités économiques dans la Communauté Européenne). In the database, the SNI classification is
available for a much longer time period than is the SSYK classification. Based on this detailed classifica-
tion of over 700 branches, a system of 57 different categories of branches (―SNI 57‖) was designed.
The original SNI classification is very thorough in subdividing the primary (e.g., agriculture) and sec-
ondary sectors (i.e., industry), as well as wholesaling and retailing, but much less detailed when it comes
to defining large parts of the rest of the tertiary, service sector. Therefore, the SNI 57 categories
representing primary and secondary sector activities and wholesaling/retailing were generally con-
structed by aggregating a comparatively large number of SNI codes. For example, 62 different SNI
codes representing wholesaling of different kinds were recoded into one single category, ―wholesaling.‖


This subdivision of the SNI classification into SNI 57 was then combined with individuals‘ education
level, as defined by the Swedish Education Nomenclature (Svensk utbildningsnomenklatur, SUN). Here, a
binary classification was utilized, either ―high‖ education—representing university education of some
kind—or ―low,‖ at most a high school (Swedish gymnasium) diploma. In total, this scheme produces 57
× 2 = 114 occupations. In the empirical data, care personnel with low education level is the largest cat-
egory, followed by construction workers (low education), high school staff (high education), retail store
personnel (low education), and medical staff (high education). Since the SNI classification in fact desig-

                                                    8
nates branches rather than professions, this way of representing occupations is not entirely unproble-
matic. Still, when combined with education level, the recoded SNI branches was deemed an appropriate
way to represent the labor market in the simulation model.


3.3 Measuring the investments’ direct local effect
In order to use SVERIGE 3 to study the local impact of large investments, the direct local effect that
the investments will generate has to be estimated. Moreover, the effect has to be represented in such a
way that it can function as input to the model economy and its associated labor market. Of course, in-
vestments come in many shapes and forms. However, investments typically involve demand for goods
and services for purposes of construction. The extent to which the local trade and industry can deliver
goods and services that is in demand will depend on the local economic structure. In addition, many
investments also generate economic effects after construction is completed through the day-to-day
operation of the resulting facility. This is true for, for instance, the nuclear waste repository that is a
central project in this study.


A number of previous studies (Lindgren and Strömgren, 2005; Lindgren and Strömgren, 2006;
Lindgren and Strömgren, 2007) have addressed the direct local effect of building and operating the
nuclear waste repository and related investments in Östhammar. Lindgren and Strömgren (2007) esti-
mate that the construction of the repository in Östhammar will generate local purchases in the order of
668 million Swedish crowns (SEK), 17 % of the total worth of purchases. The operation of the reposi-
tory was assessed to represent a local effect of 2,475 million SEK. This estimation is based on the as-
sumption that 75 % of the operating cost enters the local economy one way or another through wages.
Concerning the mining operation, housing development, and road construction projects that are also
part of this study, additional research was conducted in order to estimate the direct local effects
(Lindgren and Strömgren, forthcoming). The estimated direct local effect of the studied investments
ranges from 77 million SEK (the SFR expansion) to over 3,000 million SEK (the nuclear waste reposi-
tory) (Table 2). Compared the investment in the nuclear waste repository, the housing and road invest-
ments are not that significant in terms of direct economic effect. It should be noted, however, that the
housing and road investments will be concentrated to a much smaller time period (Figure 1). Further-
more, as clarified further on, housing and road investments have additional effects that have been taken
into account in the model.




                                                    9
Table 2 Estimated direct local effect of the studied investments (million SEK)
        Investment           Construction Operation      Total
Dannemora mine                         15     1,959       1,974
SFR                                    77         0           77
Nuclear waste repository             668      2,415       3,083
Capsule factory                        35       346         381
SKB management                          0     1,421       1,421
Housing                              278          0         278
Road                                 198          0         198




The direct local effects of the day-to-day operation of the facilities invested in are comparatively easy to
handle in the model. The wage costs of the expected staff can continuously be allocated to occupation
categories that best fit the practice of the facilities. All in all, four SNI 57 codes are concerned: ―min-
ing‖ (Dannemora mine), ―waste management‖ (the nuclear waste repository), ―manufacturing, metal
goods‖ (the capsule factory), and ―research and development‖ (the SKB management). (Concerning the
SFR, housing, and road investments, there are only construction effects.) The share of highly educated
staff will naturally vary between the facilities. For instance, the relocation of the SKB management will
require the transfer or recruitment of a large number of personnel with high education, compared to,
for instance, the competence requirements of the capsule factory. We estimate that 75 % of the work-
force of the SKB management will have a ―high‖ (i.e., university) education level, whereas the situation
will be the opposite with the capsule factory.


The effects of the investments‘ construction phase are more complicated to express model-wise. For
some investments, such as the nuclear waste facility, there are available data that in some detail specify
what kinds of goods and services will be needed; in other cases, assessments were made based on expe-
riences from other similar investments. The estimations of demand were then related to local capacity
to deliver such goods and services, using information provided by a questionnaire that was sent to a
large number of companies in the municipality. Then, the expected local purchases was distributed to
the 57 defined branches, also taking into account the degree of highly educated personnel likely to be
involved.


Finally, a data set specifying the local direct effect of each scenario was put together. For each year, the
local economic input is allocated according to the expected distribution of concerned occupations. This
data set is the way the direct economic effects at the local level are translated into the simulation envi-
ronment. However, two of the scenario components, the housing and road investments, have addition-
al effects. New housing may constitute an incentive for in-migration, and improving the road will
increase accessibility. These effects, too, have to be expressed in ways that make sense to the simulation
                                                    10
model. Based on the type and location of the planned residential areas, the housing was grouped in four
categories, detached houses and apartments, either close to or further away from the waterfront. Water-
fronts are generally regarded as attractive locations. A geographical information system (GIS) analysis
also showed that waterfront land values in Östhammar generally are low, compared too much of the
rest of Stockholm County. Therefore, planned housing close to the sea was assumed to be completed
and sold according to plan, thereby facilitating extra in-migration (slightly more per housing unit in the
case of detached houses). The impact of the road investment on accessibility was also analyzed using
GIS. It was found that the investment reduces travel time between the urban localities Östhammar and
Uppsala from 60 to 45 minutes. Expressed in another way, this places Östhammar almost 30 kilometers
―closer‖ to the Stockholm metropolitan area. This is also how the increased accessibility of the road
investment has been implemented in the model.


3.4 Interface and structure of SVERIGE 3
SVERIGE 3 is a Windows application, coded using the programming language C# in the .NET envi-
ronment (Figure 3). The graphical user interface of the simulation model is visualized in Figure 4. To
run a simulation experiment, the user will have to open two data files, a model specification file and file
containing population data. The user can then configure the experiment through a number of tabs,
which partly corresponds to the functional structure of the simulation model. In addition to the overall
configuration of the experiment in one of the tabs, concerning aspects such as start and end year of the
simulation, focus municipality, and sample size, there are configuration options for just about all major
aspects of the simulation process. For instance, different parameters can be specified concerning the
workings of the fertility, immigration, and education modules.




                                                    11
Figure 3 Example of SVERIGE 3 program code (note: as the title bar indicates, the model was known
as ―Svesim2‖ in the of stage of development when this screenshot was taken)




Figure 4 The graphical user interface of SVERIGE 3




Placeholder for technical description in English

                                                 12
For performance reasons, SVERIGE 3 has the capability to simulate the focus municipality with a larg-
er sample size than its surroundings. It is also possible to limit the geographical scope of the simula-
tions, not simulating the entire country but rather all municipalities within a user-defined distance from
the focus municipality. During the process of debugging the model, the focus municipality was mod-
eled in full, while the rest of Sweden was represented by a rather small sample. As it turned out, this
resulted in severe and unrealistic matching problems on the model labor market. In certain sectors of
the labor market, there were hardly any available and qualified labor in the vicinity. The solution was to
switch to a larger sample, but more limited geographical scope. As explained later on, in the case of
Östhammar this brought about another problem, related to the municipality‘s geographical position in
the simulation environment.


4. SIMULATION RESULTS


The simulations carried out in this study starts the year 2000 and runs to 2060, the last year of the cur-
rently planned operation phase of the nuclear waste repository. The simulation model was used to eva-
luate each scenario three or (in most cases) four times. In total, 27 simulation runs were carried out.
Figure 5 illustrates Östhammar‘s basic economic and demographic development in the four separate
simulation runs of the reference scenario, ―If nothing happens.‖ The conditions for the experiments
are the same; the different results are due to the stochastic nature of the simulation model. As the fig-
ure illustrates, the model predicts that, in the long run, Östhammar will experience a downward trend
when it comes to employment and population size if no major investments take place.




                                                   13
            30,000


            25,000


            20,000      Population
  Persons




            15,000


            10,000
                        Employment
             5,000


                ,0
                 2000           2010   2020   2030        2040       2050        2060
                                              Year

Figure 5 Östhammar‘s population and employment development in four separate simulation runs of
the scenario ―If nothing happens‖




Compared to the reference scenario, all remaining scenarios have a positive economic and demographic
effect. In some cases, though, the effect is quite small—generally smaller or around the same size as the
margin of error in simulations of the reference scenario—and cannot reverse a long-term demographic
and economic downward trend. The scenarios ―SFR and mining,‖ ―Nuclear waste repository‖ (Figure
6), and ―Repository and capsule factory‖ clearly fall into this category. When adding the housing project
(scenario ―All SKB investments and housing‖), a more substantial effect can be seen, as Figure 7 illu-
strates.




                                                     14
             30,000
                         —– "Nuclear waste repository"
                         - - - "If nothing happens"
             25,000


             20,000      Population
  Persons




             15,000


             10,000
                         Employment demand

              5,000


                 ,0
                  2000           2010         2020        2030        2040   2050   2060
                                                          Year

Figure 6 Development of population and employment demand in Östhammar for scenarios ―Nuclear
waste repository‖ and ―If nothing happens‖ (average of four separate simulation runs)



             30,000


             25,000      Population


             20,000
                         —– "All SKB investments and housing"
   Persons




                         - - - "If nothing happens"
             15,000


             10,000
                         Employment demand
              5,000


                 ,0
                  2000           2010        2020         2030        2040   2050   2060
                                                         Year

Figure 7 Development of population and employment demand in Östhammar for scenarios ―All SKB
investments and housing‖ and ―If nothing happens‖ (average of four separate simulation runs)




The two scenarios with the largest economic and demographic effect are the ones containing improved
accessibility though infrastructure investments: ―Road 288‖ and ―maximum local development.‖ The
only difference between scenario ―Road 288‖ (Figure 8) and the scenario ―SFR and mining‖ is the road
investment, represented in the model by ―placing‖ Östhammar 27 kilometer closer to Uppsala. Follow-
ing the completion of the road in 2018, there is a large, long-term increase in population. In fact, the
population growth is so substantial that there are economic spin-off effects as well. There is a generally

                                                                 15
increased interaction in terms of migration and commuting. However, in-migration and out-commuting
is significantly larger than corresponding flows in the opposite direction.



           45,000
                       —– "Road 288"
           40,000      - - - "If nothing happens"
           35,000

           30,000
 Persons




           25,000

           20,000       Population

           15,000

           10,000
                        Employment demand
            5,000

               ,0
                2000           2010          2020   2030        2040   2050      2060
                                                    Year

Figure 8 Development of population and employment demand in Östhammar for scenarios ―Road
288‖ and ―If nothing happens‖ (average of four separate simulation runs)




In Table 3, the effects of the various combinations of investments are summarized, expressed as a year-
ly average 2007–2060. For each scenario, the table displays the direct economic effect, followed by indi-
rect effects in terms of employment demand, employment, and population. For the indirect effects, the
minimum, mean, and maximum difference to the reference scenario is shown. The substantial indirect
effects of the road investment are apparent. There is a notable discrepancy between economic demand
and actual employment in the scenarios ―SFR and mining,‖ ―Nuclear waste repository,‖ and ―Reposito-
ry and capsule factory,‖ presumably indicating a shortage of available and qualified labor.




                                                           16
Table 3 Average yearly economic and demographic effects of the scenarios
                                                 Economic effects                Demographic effect
                                      Employment demand
            Scenario                                              Employment      Population change
                                    Direct   Direct + indirect
                                            Min. Mean Max. Min. Mean Max. Min. Mean Max.
SFR and mining                           98    98 570 879 104 440 838              128 1,125 1,920
Nuclear waste repository                314 314 590 992            6 278 571        17    596 1,216
Repository and capsule factory          332 344 532 628           60 180 281       137    354     480
All SKB investments and housing         345 933 1,314 1,715 571 900 1,297 1,370 2,022 2,728
Road 288                                107 4,580 4,879 5,755 4,890 5,397 6,549 10,635 11,468 13,659
Maximum local development               355 5,035 5,769 6,262 5,091 5,905 6,489 11,524 12,952 13,576




5. DISCUSSION


When there are changes in the structure of the economy, there are inevitable local effects where the
events taking place. In addition, more elusive indirect effects may arise. Using a custom-made microsi-
mulation model directly related to the modeling tradition of SVERIGE, Lindgren (1999) investigated
the local labor market effects of a hypothetical closure of a paper mill. The simulation results showed
that there are substantial indirect labor market effects, i.e., there are many occupation categories not
related to the paper mill that would meet with increased unemployment in the case of the plant shutting
down. This study, focusing on industrial and other investments rather than plant closure, also demon-
strates indirect effects of structural local changes.


Compared to the reference scenario, all other investigated scenarios have beneficial indirect economic
and demographic effects at the local level. In some cases, though, the effect is quite small, generally
smaller or around the same size as the margin of error in simulations of the reference scenario. Moreo-
ver, the positive impact of the investments cannot reverse a long-term demographic and economic
downward trend in Östhammar. This may in fact be partly related to the municipality‘s geographical
position in the simulation environment. As mentioned earlier, the simulations are run with a large sam-
ple size to avoid excessive matching problems on the model labor market. However, for performance
reason this necessitates a more limited geographical scope. Thus, since the simulation runs don‘t model
the entire country, but rather only Östhammar and the municipalities in its vicinity, Östhammar is
forced to bear an excessive burden of being a peripheral area to the most expansive Swedish region—
the Stockholm metropolitan area. In the simulations, Östhammar may suffer from a ―Stockholm syn-
drome,‖ as it were. As computer technology continues to develop (and more efficient coding may help,
too), this trade-off dilemma is likely to be less of a problem.



                                                        17
The largest indirect effects are seen in the scenarios containing the road investment, an infrastructure
project that increases accessibility. In the long term, neither the SKB investments nor the reopening of
the mine can match the demographic and economic impacts of the road investment. Although there
are some indirect effects related to the construction phase of the road, far more important is the fact
that Östhammar gets ―closer‖ to the rest of the country, resulting in population growth and increased
economic activity. A municipality such as Östhammar, located close the Stockholm metropolitan area,
may be especially likely to benefit by such infrastructure investments.


REFERENCES


Alfredsson E (2002) Green consumption energy use and carbon dioxide emission, Umeå: SMC/Department of
Social and Economic Geography, Umeå University.

Ballas D and Clarke G P (2000) ‗GIS and microsimulation for local labour market analysis‘, Computers,
Environment and Urban Systems, 24, 305-330.


Ballas D and Clarke G P (2001) ‗Modelling the local impacts of national social policies: a spatial micro-
simulation approach‘, Environment and Planning C, 19, 587-606.


Ballas D, Clarke G P, Dorling D et al. (2005) ‗SimBritain: a spatial microsimulation approach to popula-
tion dynamics‘, Population, Space and Place, 11, 13-34.


Holm E, Holme K, Mäkilä K et al. (2002) The SVERIGE spatial microsimulation model: content, validation and
example applications, Umeå: SMC/Department of Social and Economic Geography, Umeå University.

Holm E, Lindgren U and Malmberg G (2004) Arbete och tillväxt i hela landet: betydelsen av arbetskraftsmobili-
sering, Östersund: ITPS.


Hooimeijer P (1996) ‗A life-course approach to urban dynamics: state of the art in and research design
for the Netherlands‘, in G Clarke (Ed.), Microsimulation for urban and regional policy analysis, London: Pion,
28-63.

Lindgren U (1999) ‗Simulating the long-term labour market effects of an industrial investment: a micro-
simulation approach‘, Erdkunde, 53, 150-162.




                                                      18
Lindgren U and Strömgren M (2005) ‗Geografiska effekter av en djupförvarsetablering‘, in SKB (Ed.),
Samhällsforskning 2005: Betydelsen för människorna, hembygden och regionen av ett slutförvar för använt kärnbränsle,
Stockholm: SKB, 42-64.


Lindgren U and Strömgren M (2006) ‗Slutförvaret och det lokala näringslivet: en undersökning om up-
phandlingsbehov och leverantörskapacitet i Östhammar och Oskarshamn‘, in SKB (Ed.),
Samhällsforskning 2006: Betydelsen för människorna, hembygden och regionen av ett slutförvar för använt kärnbränsle,
Stockholm: SKB, 48-65.


Lindgren U and Strömgren M (2007) Slutförvarets lokala effekter på befolkning och sysselsättning i Östhammar
och Oskarshamn, Stockholm: SKB.


Lindgren U and Strömgren M (forthcoming) Socioekonomiska effekter av ett slutförvar i Östhammar, Stock-
holm: SKB.


Rephann T J and Holm E (2004) ‗Economic-demographic effects of immigration: results from a dy-
namic spatial microsimulation model‘, International Regional Science Review, 27, 379-410.


Veldhuisen J, Kapoen L and Timmermans H (2000) ‗RAMBLAS: a regional planning model based on
the microsimulation of daily activity travel patterns‘, Environment and Planning A, 32, 427-443.




                                                         19

				
DOCUMENT INFO