Situated lifestyles: I. How lifestyles change along with the level of urbanization and what the greenhouse gas implications are—a study of Finland

Radical greenhouse gas (GHG) reductions in the near future are needed to mitigate the climate change to a level adaptable for both human settlements and the environment (IPCC 2007). Since the majority of all the global GHG emissions seem to be related to urban settlements (Satterthwaite 2008, Dodman 2009, Hoornweg et al 2011), spatial planning has been argued to have significant potential to shape more sustainable human settlements (Carter and Fowler 2008, Fields 2009, Antrobus 2011). There is an extensive body of literature demonstrating how higher density offers opportunities for more efficient energy use and lower GHG emissions from transport and housing (Norman et al 2006, VandeWeghe and Kennedy 2007, Jenks and Burgess 2000, Ewing and Cervero 2010, Glaeser and Kahn 2010, Fuller and Crawford 2011).

On the other hand, increased volume of economic activity is claimed as another effect of urbanization. Agglomeration economies have for a long time explained how denser agglomerations are able to create more wealth (Ciccone and Hall 1996, Broersma and Oosterhaven 2009). This generally leads to more consumption with associated environmental burdens. In addition, different forms of agglomeration open up certain possibilities of consumption and undermine others, which is likely to affect the lifestyles. In fact, some recent evidence would indicate that, even where higher urban density leads lower private vehicle usage and housing energy needs, the life in denser agglomerations may be more consumption intense up to the point where the overall emissions exceed those caused by residents of less dense areas (Sovacool and Brown 2010, Heinonen 2012, Lenzen et al 2004).

There thus seems to be a gap in our current understanding on the urban form–GHG emission relationship, namely how the urban form affects lifestyles and thus GHGs on a much wider scale than traditionally assumed. Looked at from the perspective of research focusing on energy requirements and environmental burdens related to different lifestyles, the tradition is long (e.g. Schipper et al 1989, Biesiot and Moll 1995, Biesiot and Noorman 1999, Wier et al 2001, Bin and Dowlatabadi 2005, Brännlund and Ghalwash 2008, Cayla et al 2011), but the variables concerned rarely include the urban form. Some of this literature, especially the carbon footprint assessments based on monetary expenditure data (e.g. Baiocchi et al 2010; Kerkhof et al 2009; Hertwich and Peters 2009; Hertwich 2011; Erickson et al 2012; Heinonen and Junnila 2011a) bring consumption closer to the spatial location, yet mostly without trying to explain how the form affects the consumption choices. Still, studies in this field have demonstrated, e.g., that a life cycle perspective and consumption based approach are actually mandatory to fully understand the emissions related to a certain society because of the outsourced emissions, that is, the emissions embodied in all the consumed goods that are produced elsewhere and imported (Peters and Hertwich 2008, Ramaswami et al 2008, Schulz 2010).

Monetary expenditure data alone may also give an insufficient picture of the actual lifestyles (Schipper et al 1989). Most consumption, be it services or products, involves phases of use in everyday life reflected in time allocation of a consumer. Hence, it has been claimed that time use data contribute to better understanding of consumption patterns and changes therein (Gershuny 1987, Minx and Baiocchi 2009). Further, together monetary expenditure and time use data can give us understanding of the actual lifestyles of certain consumers and thus a possibility to understand the urban form–lifestyle relationships.

Drawn from the above discussion, the purpose of this paper is to analyze the urban form–lifestyle relationships and the resulting GHG implications. We employ both monetary expenditure and time use data to portray lifestyles in different basic urban forms in Finland: metropolitan, urban, semi-urban and rural. The GHG implications are assessed with a life cycle assessment (LCA) method that takes into account the GHG emissions embedded in different goods and services. We analyze first the consumption patterns of the residents in selected urban forms and calculate the associated GHGs, and then proceed to time use data to understand in more detail how and where the money is consumed.

There are some studies that have addressed the issue of how the urban form might affect the environmental burdens (Høyer and Holden 2003, Holden 2004, Holden and Norland 2005), but the tradition is relatively thin. To our knowledge this letter (1) brings valuable actual data based evidence to the discussion, and (2) is the first concentrating on a climate change perspective.

We present in the letter that, while the direct structural factors, transportation and housing energy, seem to follow the patterns reported in numerous studies and decrease with higher density, the reductions may actually be rather small and can thus be easily overridden by sources of indirect GHG emissions. We also highlight that the indirect emissions actually seem to have strong structural determinants, often undermined in studies concerning sustainable urban forms. We introduce a concept of parallel consumption to explain how certain lifestyles lead to multiplication of consumption outside of the limits of time budget and the living environment, e.g. using a multitude of service spaces while possessing a home with similar possibilities. We also look for contextual factors, such as time spent in traffic, for describing actual fuel consumption better than the average mileage driven. The findings indicate that much deeper understanding on the urban form–lifestyle relationships is needed in order to create truly effective urban planning strategies from the climate change perspective.

The structure of the letter is as follows. In section 2 the concept of situated lifestyles is defined. Section 3 explains the data and the utilized methods. In section 4 we analyze concurrently the lifestyles based on monetary consumption patterns, time use and the resulting GHGs. Section 5 looks at the lifestyles through the concept of parallel consumption. Finally, in section 6 the results are discussed and some final conclusions are drawn.

Frequently, lifestyles have been seen as a product of the values of individuals (e.g. Bauman 2000, Firat and Dholakia 1998, Bin and Dowlatabadi 2005). Hence there is little need to consider the urban form as a determinant of individual behavior. There are other interpretations of lifestyles that take structures into account and argue that consumption is constrained by the surrounding structure. Baiocchi et al (2010) put emphasis on the physical location and residence of consumers. Yet, the variables they use are still to a large extent based on class, occupation and education. Schipper et al (1989) and Jalas (2002) go further to argue for paying attention to the activities and the locations of consumption. Spaargaren and Van Vliet (2000) has proposed a combination of the above, arguing that sociotechnical structures do most of the choosing for us, but that nevertheless consumers look for consistent 'style' across areas of consumption.

This study is premised on the claim that the form of a human settlement is a fundamental structural factor that underlies patterns of consumption. It affects housing types, commuting distances, availability of different goods and services, social contacts and emulation, and the alternatives for pastimes. Urban form thus is reflected in behavioral patterns, time allocation and purchasing decisions. We use the notion of situated lifestyles to capture this effect. We recognize other determinants of behavior, but argue that urban form affects time allocation as well as purchasing decisions deeply enough to create identifiable lifestyles even on a very highly aggregated level, and thus give an indication on how the urban form affects the GHG emissions.

3.1. Four types of urban form

3.2. Consumption and time use data

3.3. GHG implication assessment

Figures 1(A)–(C) present the overall results of the study in monetary purchases, GHGs and time use. Figure 1(A) show how the structure of consumption is surprisingly similar in all the samples, but the overall consumption levels increase from 12 200€ a⁻¹ in rural areas to 13 800€ a⁻¹ in semi-urban areas, 14 100€ a⁻¹ in cities and 17 600€ a⁻¹ in HMA. This indicates that the consumption patterns are actually quite different. All the consumers make the choices under the same time constraint, and thus time allocation needs to change to allow the overall higher consumption. Figure 1(C), depicting the daily allocation of time in the four types of area, shows that time use is in fact different. The important lifestyle related difference in time use patterns is the amount of leisure time spent outside of home. Leisure time spent outside of home varies from 8.7 h in HMA to 8.4 h in cities, 7.6 h in semi-urban areas and 6.6 h in rural areas. Thus, while the housing prices increase along with the level of urbanization (see section 3 also) and more money is spent on housing in the more urbanized areas, less time is spent at home and more actually allocated to taking advantage of the better service levels of the area. Figure 1(C) also shows that the higher income level in the more urbanized areas is mainly a result of higher salaries, as working time is very evenly distributed across the different urban forms.

Zoom In Zoom Out Reset image size

The increasing levels of monetary consumption and the data on how the time is allocated explain the GHG results of figure 1(B) that show increasing overall emissions towards the more urbanized areas, especially HMA. The overall carbon footprints, according to the assessment model, are 10 900 kg GHG a⁻¹ in HMA, 9600 kg a⁻¹ in cities, 9500 kg a⁻¹ in semi-urban areas, and 8900 kg a⁻¹ in rural areas. Thus, even though more services are consumed in the more urbanized areas, this does not solve the problem of increasing GHG emissions. The city type of living is actually not able to reduce even the emissions from housing and transport, as analyzed further in section 4.1, and thus the increased consumption of services only adds to the very similar large basis of emissions from the necessary consumption.

4.1. Transport related consumption patterns and GHGs

4.2. Consumption on housing and the resulting GHGs

4.3. Other consumption patterns and the GHG implications

An aspect of urban agglomerations is the possibility of sharing resources and potentially reducing the existing size of personally owned assets, especially space. Smaller living spaces in more dense areas can be interpreted as a tradeoff between own and shared spaces. Laundries, parks, restaurants, cafés and pubs as extensions of the home are obvious examples. At best this could extend to sharing of equipment, household utilities etc and decrease the need to own goods. With this logic, many services can also contribute to eco-efficiency by reducing material and energy consumption (Halme et al 2004).

However, commercial services outside of home do not always lead to reduced overall material and energy use. It may well be that the usage and resource needs of all the 'shared spaces' exceed the gain from reduced own living space. Restaurants serve their clientele only for short hours per day and the sites of cultural events may be even less frequently uses by their clients, which increase their GHG intensity per used service. Dense urban living may also create needs for summer cottages and second homes, particularly when such an arrangement is affordable, which further increase one's actual space needs. Summer cottages are also often equipped with all modern machinery to reduce the need to carry things along when visiting the cottage and to make the place more pleasant. Equipment in the summer cottage may also increase the need for heating even when nobody is present. All this increases the GHGs that we cause even when we choose to live in smaller apartments in city centers.

We call this phenomenon here 'parallel consumption' and define it as concurrent consumption of service spaces in different locations. Our homes are heated and many electric appliances are run while we use diverse service spaces that extend our living space outside of the actual home. The concept aims to capture the way that consumption of space can multiply and extend beyond the apparent limit of time budgets or the spatial constraint imposed by the home or even the surrounding settlement. The concept also helps to depict that our actual home can only give a limited picture of the space that we actually use for living and thus what counts for the GHG impacts that we cause.

The phenomenon of parallel consumption is strongly present with service use such as hotels, cafés and restaurants, as homes are heated and operated at the same time, and also equipped to offer similar services. In HMA 31 min d⁻¹ is spent in hotels and restaurants (table 4), in cities 24 min d⁻¹, but only 14 min d⁻¹ in semi-urban areas and 10 min d⁻¹ in rural areas. Correspondingly, 910€ a⁻¹ is spent for these services by the average resident of HMA (table 2), 620€ a⁻¹ in cities, 500€ a⁻¹ in semi-urban areas and 360€ a⁻¹ in rural areas. Concerning second homes the pattern is similar.

In the case of summer cottages and second homes, the parallel consumption phenomenon is continuous as two living spaces are concurrently in possession of those owning such premises. The consumption statistics show that in Finland 25% of the households in HMA possess a summer cottage, 22% in cities, 23% in semi-urban areas and 20% in rural areas, which depicts that also in this respect smaller living spaces are actually compensated by spaces elsewhere. Also, persons living in HMA and in cities have the highest time use in summer cottages with 44 min d⁻¹ compared to 37 min d⁻¹ in semi-urban areas and 22 min d⁻¹ in rural areas. As was depicted in section 4, this closes down the advantage in housing energy consumption of more dense living. Energy purchases for the summer cottages and second homes vary from approximately 60€ a⁻¹ by the HMA residents to 40€ a⁻¹ in cities and in semi-urban areas and 30€ a⁻¹ in rural areas. Possession of summer cottages may also relate to owning of private vehicles, since the vast majority of all the trips to summer cottages are made with private cars (Perrels and Kangas 2007). This hinders the full materialization of the potential of lower private driving needs in more dense agglomerations.

Movie theaters, swimming halls, spas, gyms, laundries etc have similar qualities, extending our living space to different locations. According to both monetary expenditure and time use data sets, the higher affluence level and higher service level in more urbanized areas lead to increased utilization of these services, followed by increased GHGs. When recognizing that many of the services where parallel consumption materializes depend on dedicated single-purpose space, parallel consumption highlights the 'hidden' burden of services and the logic of how service space gets included in the carbon footprints of individual consumers.

According to the purpose of the paper, we have demonstrated how there seem to be identifiable and from a GHG perspective significant urban form related lifestyle differences between the residents of different types of area even on a very aggregate level, making lifestyles actually situated instead of being mere personal projects. In a case study of Finland we found that even though the emissions related directly to the urban form, those from housing energy and private driving fuel combustion, decrease along an increase in the level of urbanization, when all the indirect emissions related to these categories were taken into account, in housing more emissions seem to be caused in the more urbanized areas and in transport the differences are significantly reduced, as shown by table 2. Furthermore, the more urbanized area types seem to associate with more GHG intensive lifestyles concerning emissions from other consumption activities. As indications of the consumption intensive urban lifestyles we pointed out for example that clothes and electronics are bought much more in the more urbanized areas, and that especially in all kinds of leisure service the consumption increases significantly along with the degree of urbanization. In addition, while in the less urbanized areas the daily life is much more home centered based on their time allocation, the smaller homes are at least equally equipped in cities, while more time is allocated to consuming services outside the home.

The overall carbon footprint of an average resident in HMA in Finland is 10,900 kg a⁻¹ according to the study, followed by city and semi-urban areas with 9600 kg a⁻¹ and 9500 kg a⁻¹ respectively and rural areas with 8900 kg a⁻¹. Consistently with the income hypothesis of agglomeration economies (e.g. Ciccone and Hall 1996), the first explanation is the higher affluence level in the more urbanized areas that seems to result in higher consumption. However, it seems that the urban type of living cannot very effectively reduce even the emissions related to the home or transport. Although a higher share of the overall purchases in more urban areas is allocated to services with lower GHG intensities, there is a very similar large basis of necessary consumption (housing, transportation and food) and emissions in each area. Thus the higher consumption volume only adds to the emissions instead of redirecting the whole consumption towards low GHG intensity activities. On the other hand, the overall GHG intensity of consumption decreases towards the more urbanized areas due to the increase in consumption of services from the lower end of GHG intensities of different commodity groups.

Based on the evidence on how we tend to extend our living space from the actual home to all kinds of service space while still equipping our homes to produce similar services, we brought into discussion the concept of parallel consumption. This phenomenon is not bound to any single urban form, but it would seem that the more urban lifestyles are more strongly based on this type of consumption, where reduced living spaces are actually compensated by service space use outside of the actual home. Increased use of summer cottages may indicate also that the hectic city lifestyle encourages possessing such spaces for more peaceful moments. Overall, the time use data showed that as much as 120 min d⁻¹ more of the time outside of workplace or school is spent at home in rural areas compared to the more urbanized areas of HMA and cities. As this difference is largely spent for consumption of services outside of the actual home, the concept of parallel consumption helps in explaining the finding that the more urban lifestyles seem to lead to the highest GHG loads. As a worst case scenario, the study also points out that living is still relatively home centered in Finland, and thus significant growth in the parallel consumption phenomenon is possible in the future. On the other hand, economic crises and the current trend of decreasing economic growth presumably constrain the scope for parallel consumption.

The study includes uncertainties that can be divided into three broad categories. First, both data sets are subject to errors and biases, especially regarding the less frequently occurring activities or purchased goods. We tried to avoid this type of problem by using primarily aggregated data, especially concerning less frequently purchased goods. Also, the 52 category aggregation of the ENVIMAT assessment model forced us to avoid the most detailed data. These problems relate to sample sizes as well, small samples increasing the uncertainty. In this study the samples remained rather large with over 500 observations in each sample. However, it is well known that people tend to underreport monetary amounts used in socially undesirable sectors such as alcohol (e.g. Kok et al 2006). In the employed Finnish data the situation is relatively good, as the consumption estimates from consumer expenditure survey cover 87% of the consumption data in the national accounts (70 285 and 80 439 m€). Further, according to the consumer survey user's handbook (Statistics Finland 2006), this is due to differences in the accounting principles and thus the difference cannot be considered very significant. Furthermore, as the data are formed by combining diary and interview data their precision is quite high; standard errors in all the main consumption categories with the exception of spending on education are less than four per cent (Statistics Finland 2006).

The second source of uncertainty is the IO LCA method itself and thus the ENVIMAT assessment model. While IO LCAs are argued to be the most suitable for wide system level assessments in the built environment (Crawford 2011), they do have some problems. The primary source of uncertainty arises from the aggregation error inherent to all IO models (e.g. Wiedmann 2009). In ENVIMAT there are only 52 sectors, meaning that relatively diverse production sectors and individual products are inevitably aggregated together and given one single emission factor. This leads to another problem in IO LCAs, namely proportionality and homogeneity assumptions. These mean the straightforward assumption of IO approach that the monetary purchases would be linearly proportional to the output (GHGs) (e.g. Treloar 1997). The significance of this problem can be diminished by concentrating on the average consumers, as in this study. With sufficient samples the average consumers should use average goods and the problem would not be significant. Notwithstanding, Girod and De Haan (2010) have provided evidence that the products purchased by the more affluent consumers are not average products and that the monetary differences actually reflect differences in quality instead of quantity. In this case it might be that the IO LCA overestimates the emissions of the more affluent consumers, which would decrease the differences between the areas in this study. However, this concerns mainly the daily consumption of goods and that of some durable goods, but a large part of the consumption activity targets to exactly the same goods and services. Overall, the GHG assessment related problems are also reduced, as the paper describes the lifestyle differences from two other data sets as well. Regarding the GHG results, Heinonen and Junnila (e.g. Heinonen 2012) have also tested earlier that the results of the ENVIMAT model are very similar to those of another IO model, EIO LCA (Carnegie Mellon University 2008), in the Finnish context.

Concerning the possible price and currency level asymmetries between the data and the model, the fit is excellent, since both ENVIMAT and the Household Budget Survey are based on the same CIOCOP classifications, and the 2006 Household Budget Survey that we utilize has been employed in building the ENVIMAT model. Further, in ENVIMAT the often problematic domestic technology assumption of the majority of IO LCAs is relaxed and the imports from the most important trade partners of Finland are taken into account (Seppälä et al 2011).

Another uncertainty perspective arises from the assumptions regarding energy. In this study we have assumed that the same average energy (from ENVIMAT) is used in all the areas. If the local energy production in a certain area were much more or less GHG intensive than the Finnish average, it would affect the carbon footprints significantly. From 23% (HMA) to 27% (rural areas) of the carbon footprint comes from just the housing energy, and the local energy profile affects the local services and locally produced goods as well. Now, in HMA the local energy production is actually very similar in GHGs to the Finnish average and the other three area types consist of locations from all around Finland. Thus the results of the study in general are not subject to this error, but within cities, semi-urban areas and rural areas there exist locations where the local energy production deviates significantly from the national average. Heinonen and Junnila demonstrate the impact comparing Helsinki to Porvoo (Heinonen and Junnila 2011b), which belongs to cities in the samples of this study, but actually has a very different energy production profile from the Finnish average.

Furthermore, district heat can be seen as a byproduct of electricity generation. The employed assessment model ENVIMAT follows the shared benefits method, meaning that primary energy use is allocated according to the need to produce a similar amount of heat or electricity separately. The sensitivity of the results to the method of dividing the emissions of combined heat and power (CHP) production is rather high, as the production in Finland is predominantly fossil-fuel based. Allocating more emissions to electricity would result in lower emissions for the city residents using primarily district heat for heating purposes.

Finally, the paper presents a case study of Finland, which sets limitations for the generalizability of the results. In other locations around the globe different factors may dominate the emissions, which could significantly change the outcome of a similar analysis. The phenomenon of parallel consumption is present in all types of urban setting, and the role of indirect emissions is certainly important in all the developed countries with high consumption volumes. However, the differences in housing energy use and private driving etc, dominant factors in the GHG emissions, may be much larger in other contexts and lead to different conclusions.

To conclude, as part I of a broader study, the purpose of the study was to describe the broad context and depict that the urban form impact is deep enough to show in carbon footprints even on a highly aggregated area type division. On a more detailed level there is actually a large variation of different kinds of urbanization within each of these area types and a diverse set of lifestyles. These present thresholds where a certain other impact is greater than that of the urban form, for example that of local energy production. Even in these, the urban form shapes the lifestyles, but understanding the connections in a more detailed level is imperative to materialize the potential of designing location specific GHG mitigation strategies. In part II we will focus on these variables that affect the GHGs when one step deeper from purely urban form is taken.