Portraying and Analysing Urban Shrinkage in Greek Cities—The Case of Larissa

The phenomenon of urban shrinkage is multidimensional, concerning regions, cities, towns, or metropolitan areas that are facing population reductions and simultaneously intense economic and social transformations. Greek cities are experiencing strong signs of urban shrinkage such as thousands of enterprises closing down, the emergence of a series of vacant commercial spaces, even whole buildings abandoned, as a result of the economic crisis, mainly from 2011 and onwards. This paper documents, analyses and investi-gates the changes brought by urban shrinkage in Larissa—a medium-sized Greek city. It highlights the characteristics of emerging new economic urban landscapes in Greek cities at a time of uncertainty; it also introduces a methodological research framework for urban shrinkage, which may be applied in other Greek cities.


Introduction
Urban shrinkage is a phenomenon that can encompass regions, cities, towns, or sections of metropolitan areas that are experiencing dramatic reductions in their economic and social base as well as their entire population (Gospodini, 2012), (Pallagst et al., 2013). In the case of Greek cities, the spatial impacts caused by the economic crisis of public debt (2009-) are particularly important since Greek cities had never in the past been experiencing urban shrinkage. New conditions of shrinkage produce spatial entities of obsoleteness, usually characterized by the term "non-productive space" (Newman & Kim, 2017), while simultaneously making obsolete the roles that residents once had in them (Gospodini, 2006). On the other hand, these abandoned structures can potentially be reclaimed, based on their probable use, as "micro-clusters" (Capdevila, 2014) promoting local circulation of information and knowledge, and social enhancement (Tranel & Handlin, 2006) or "natural nodes of a green city network" (Frazier & Bagchi-Sen, 2015). In the broadest sense, these abandoned structures, "land use alternation patterns" (Lee et al., 2018), may provide a particularly fertile ground to streamline the overall quality of the surrounding built environment.
For the purpose of this study, across the full spectrum of non-productive spaces, the concept is getting anchored in vacant commercial properties. Even though in Greece retail trade is the key pillar of the porosity of the urban tissue (Gospodini, 2012), the ground reality is stark. Commercial vacancies retain the stigma of economic crisis and remain unhelped by municipal and governmental apathy amid the broader neglect of small enterprises. Interestingly, these empty lots can reflect upon the municipal policy challenges on the way to a sustainable urban economic environment and promoting neighborhood live ability (Manika & Gospodini, 2016).
Even though conceptualizing urban shrinkage is not "uncharted territory", we realized the pressing need to dig deeper into the spatial impacts of commercial vacancies and provide actual numbers and real patterns. As a first step, we incorporate "firmography" (Maoh & Kanaroglou, 2013) principals and methods.
Hence, we adopt a conceptual reflection of crude rates, basic demographic equations, to the studying of firms' cycles of activation and obsolescence. In the second step, that is benefitted from the first, we shed light on a dynamically over time portraying of the spatial distribution of commercial vacancies, via spatial statistic methods. More specifically, kernel density was adapted to estimate and contributing to analyzing plot spatial patterns. In the last step, we endeavor to decipher changes in the place of residence of economically sound firms during the period considered. This mobility of economically sound firms within the urban fabric is the emblem of our concept put into practice.

Literature Review
As documented in recent history, the causes of urban shrinkage may vary but can be grouped as follows: • Suburbanization leading to degradation of inner-city areas (see Lever, 1993), , (Couch et al., 2005); • Economic Transformations in the industrial sector, as witnessed in European and US cities in the 80s (see (Gospodini, 2012), (Haase et al., 2012)); • Economic restructuring and its underlying potentials, as happened in East European countries in the 90s (see Pallagst et al., 2009); • Demographic reasons in developed regions, and processes for settling economic and political immigrants (Mykhnenko & Turok, 2008); • Climate change, environmental disasters or radical changes in the political and administrative system (see Großmann et al., 2013) • Potential Economic crises due to forces of globalization and other structural causes of the effects of various macro-process interactions at the local level (Gospodini, 2012), (Moss, 2008), (Pallagst et al., 2009).
In urban planning terms, non-productive spaces' impact reflecting on the shocks of building dereliction to the sense of local community and the vitality of neighborhoods. In inner-city areas of Greek cities, a chain of vacant shops, offices, and houses can be observed in the former lively commercial streets (Gospodini, 2012). Much has been written about vacant commercial properties and their study is intertwined with many different definitions. As an illustration, the most common definitions and approaches can be incorporated in "firmography" (Maoh & Kanaroglou, 2013), (Mejia-Dorantes & Martín-Ramos, 2013), "demography of the firms" (Giannetti & Velucchi, 2006), (Wissen, 2005) or "business demography" (Markowicz, 2014). Key pillars of the linkages between them are "Retail Location Theory" (Brown, 1993) or theories based on "the mobility of firms" (Mejia-Dorantes & Martín-Ramos, 2013), (Froment, 2015).
It is argued that a firm's decision to locate or relocate anchored from certain economic criteria to maximize profits and select an attractive or affordable area.
Studying locating and movement of firms, in parallel with the potential contributions of the resulting abandoned structures has a recent manifestation of urban shrinkage, can draw on for new ideas on analyzing this phenomenon.

Characteristics of Study Area
Many Greek cities can be described as shrinking since the 2011 financial crisis, affecting their economic and social well-being by the emergence of vacant commercial spaces (Manika, 2018). The city of Larissa is typical such a case, due to four significant characteristics: 1) The first concerns the morphological pattern of the city. More specifically the urban fabric of Larissa is characterized by the typical morphology (streets, squares, city blocks) of modern Greek cities, the result of post-war housing development patterns in Greece; 2) Its median population size for a permanent population equal to 145,000 according to the 2011 census, a city that represents the largest group of Greek cities except for Athens, Thessaloniki and Piraeus; 3) The lack of a specific economic identity like for instance tourism Greek ci-

Methodology and Data
In view of the fact that this paper has a quantitative and exploratory nature, we show analytically changes in economic activity of the study area, by sector and by year. This analysis is prompted and benefited from firmography approach.
Hence, as a first step, we apply crude mortality and birth rates. The crude mortality rate is used for the businesses that closed the year of examination and the crude birth rate for the businesses that opened the year of examination and given by the following equations based on Eurostat Statistics explained platforms The concept is that crude rates capture the voice and presence of the emerging economic transformation of the city. This paper endeavors to foster effective and fruitful regeneration policies for shrinking areas. Thus, we make efforts to overcome the leaky pipeline of studying economic sectors only as an entity. Hence, in some cases, the emphasis is shifting to distinguish specific sectors of commercial activity based on their contribution to the vibrancy of the urban landscape and their role in the economic identity of the city (Long & Huang, 2017), (Mehta, 2007).
The second step in our analysis is a spatial statistic approach based on kernel density. Kernel density allows the analysis of point events by utilizing their spatial concentration in a two-dimensional form (Zhao et al., 2017). The output area shows the intensity of events as a function of their density in the geographical area (Adolphson, 2010), (Srinurak et al., 2016). The method was applied based on existing businesses, as point data, for all sectors for the years 2008 to 2014. The raster datasets produced were sorted in such a way that the differentiation between the classes is noticeable in each case. The selection of the bandwidth was based on the amount of information and detail needed. Therefore, the spatial shrinkage of the economic activities of the study area and the reduction of their spatial intensity are observed.
The next step, for the accurately retracing of urban shrinkage, is to analyze the It was necessary to group the city streets into specific categories, based on 1) the current General Urban Planning of Larissa and 2) the commercial density of streets. The index used (Manika, 2018) applied per segment of each axis between nodes is expressed by the formula:

Number of businesses Street Commerciality Index
Length un 0 it *10 = (3) The data used comes from ( That is when the information at road-level is sufficient the data of the Ministry of Finance is used and when the point information is needed the Chamber of Commerce database is used since the Ministry could not provide such data. At the same time, data from the Ministry of Finance's database were used, where possible, to assess the validity of the Chamber of Commerce data before it was used. The database created the appropriate digital backgrounds (road axis and city blocks) and point databases. The procedure followed was that of geocoding, whereby the desired cartographic information was produced using information describing sites such as postal addresses and an address locator. Through this process, most of the businesses immediately acquired spatial reporting. Other businesses were manually matched with the help of location programs (such as Google Maps and Google Earth).

Surveying and Documenting Economic and Spatial Shrinkage
The city of Larissa in the period under study is facing a plethora of economic transformations. These transformations concern reductions in 1) the average annual taxable income of natural persons and 2) the total commercial activities of Larissa. Subsequently, based on the available data and the principles of business demography, 1) the city's economic identity and 2) its gradual transforma-

Economic Shrinkage
To better analyze changes in economic activity, by sector and by year, we use crude mortality and birth rates, based on the principles of business demography.
Therefore, the crude mortality rate is used for the businesses that closed the year of examination and the crude birth rate for the businesses that opened the year of examination.
For the purpose of this study, specific sectors of commercial activity were distinguished based on their contribution to the vibrancy of the urban landscape and their role in the economic identity of the city (Mehta, 2007), (Long and Huang, 2017). These sectors are presented in detail in the graphs below.
The graphs visualize the percentage change of each sector during the examined period. In particular, the red and green lines represent the crude mortal- growth and macroeconomic stability conditions are achieved where the economy has transformed and upgraded the manufacturing system of this sector (Kitson and Michie, 1997), (Pitelis and Antonakis, 2003).
Regarding the wholesale and retail trade, as shown in Graph 2, there was a relative increase in business in 2008 alone, with a downward trend in 2010 from a peak in 2010. The decline is mainly due to small businesses, which are also the largest proportion of Larissa's businesses, with significant impacts on the economic and social life of the city, as they stimulate the local community, bringing the smaller streets to life by restraining the social fabric and creating new employment opportunities.
The declining trend of the wholesaler and retail industry, throughout its business cumulatively, despite moving at high levels, seems to be gradually being Greece. It is an image that is found in many European and American states, with Europeans finding coffee in the various cafes an affordable part of their daily lives and Americans a luxury they can afford in times of crisis, a time in which they have restricted their needs in other categories of retail (Laniyonu, 2017), (Papachristos, Smith et al. 2011). The arts, entertainment and, recreation industry, following its first year of growth, shows a relative balance, with its down- In an attempt to analyze the course of the industry, its two categories must be considered separately. In the case of the private sector, this decline may be explained Graph 5. Education (Source: Greek Ministry of Finance (own processing)).
Similarly, in the case of the sector focusing on public works, its downward trend is primarily explained by the reduction of funding opportunities provided by the Community Support Framework for the regions of Greece and the resulting reduction in public works auctions. Another important factor, are the discounts of over 40% (on average) given to public works in a desperate survival effort of the companies and the opportunity is given to the large contractors during the period under consideration, who enjoy the advantage of access to banking financing due to their good creditworthiness to participate in large-scale projects. This possibility deprived the possibility of survival by the medium and small size contracting companies that dominate Larissa.
It is thus observed that the city's economic base is declining while at the same time facing severe transformations. The country's economic uncertainty during the study period could not leave the city's economic activity unaffected. In conclusion, it can be said that reductions in sectors such as construction and wholesale and retail are directly related to the disposable income and purchasing power of citizens and industries such as education, the arts, entertainment and recreation that tend to identify with a relative sense of luxury so they are the first hit. On the contrary, there is an increase in the complementary economic activities of citizens (insurers, brokers) and in private sector initiatives that are coming to fill the gaps in the public sector in areas such as health in times of economic downturn and in areas such as focusing on specific issues of consumer habits.

Spatial Shrinkage
The abandonment of a city's residences or businesses, whether it be the depar-  More specifically: 1) The area of increased intensity is gradually shrinking; 2) In the same context and area of significant intensity, it gradually loses a large proportion of its area and acquires a more perforated urban vacancy; 3) Of the significant intensity areas identified in 2008, most are lost over time.
Excluding cases of Agios Achilles and those found in the areas of Agios Athanassios and Philipoupoli which are maintained but with reduced intensity. Of course, during the examination period some areas of significant intensity were created as a result of the spatial reduction of areas of increased and significant intensity of existing businesses or as residual imprints of their larger extent.

Movements per Axis
Particularly important for the study of the spatial dimension of urban shrinkage is the examination of changes in the place of residence of economically sound businesses during the period considered. The data used were those of the Ministry of Finance and provide information on the place of business of the company at the start and the year of its establishment, its location at the time of its last movement, and the year of business, respectively.
To analyze the movements, it was necessary to group the city streets into seven categories, as shown in Map 3. In particular, categories K1, K2, K3 include the roads belonging to the respective K1, K2, K3 central reception areas of the current General Urban Planning of Larissa and Category A roads which are following the same G.U.P. (General Urban Plan) belong to areas of pure residence.
Category Π includes Larissa-Athens, Larissa-Thessaloniki ring roads, Trikala-Karditsa ring road, and Konstantinou Karamanlis Avenue which although part of the ring road due to the city's diffusion, has now lost its basic features.
The last section of the ring in the northeast of the city has been auctioned but has not yet begun construction. Finally, category E1 includes roads "exits" of the city and especially parts of these roads within the city limits, and similarly to E2 the sections that are continuations of parts of the previous category that extend outside the city limits.
As there are cases of roads crossing more than one of the former areas of G.U.P. Map 3 was used to allow them to be assigned to only one of the categories under consideration. The category that comprises the bulk of the concentration of the commercial activities of each street based on the previous map thus constituted the dominant characterization of that street. The data from the Larissa Chamber of Commerce was used for the map as they provide information on the exact business address at the beginning of the year. The index used (street commerciality index or commercial density of a street (Browning et al., 2010), as described in the methodology and data, and applied (Map 4) per segment of each axis between nodes is expressed by the formula:

Number of businesses Street Commerciality Index
Length u 0 nit 1 0 = * As for all businesses, it can be seen from Table 1  with Category A roads, shifts to more central and more commercial roads are likely to follow the fall in rental prices in the central districts.
2) The second major flow concerns the movement of businesses from category K1 to roads of the same category or roads of category A.
These movements can be accomplished in two ways. Whether they concern: a) Arc movements on Category A-roads adjacent to Category K1. That is, businesses that chose to move very shortly from their original location and intended for the same centralized streets but perhaps more favorable positions Finally, it can be seen in Graph 7 that most of the flows made during the study period involved businesses are mostly young, under 5 years old, to gain their market share in the city. More specifically, more than 35.5% of these businesses are up to 5 years old while 15% of them are up to 1 year old until their move.
In conclusion, therefore, businesses' flows during the study period can be grouped into two categories: a) Movement from Category A roads to roads of the same category or a lesser Graph 7. Age of moving businesses, Source: Greek Ministry of Finance (own processing). Moving businesses are mostly young, under 5 years old, in an effort to gain as strong as possible a good market position. Finally, the role of the price of renting business in a decision to move to a financially sound business is decisive, as it is one of the biggest overheads of the business and affects its financial prospects.

Results
The implication of the aforementioned triangle methodological framework shows that the economic base of the city of Larissa is shrinking while it is undergoing intense transformations. The economic uncertainty of the country in the time of crisis could not leave the economic activity of the city unaffected.
Reductions in sectors such as construction and wholesale and retail which are directly related to the disposable income and purchasing power of citizens are the first to be affected. Industries such as education, the arts, entertainment, and recreation tending to be identified with a relative sense of luxury also suffered from the crisis.
On the other hand, there is an increase in the complementary economic activities of citizens (insurers, brokers) and in private sector initiatives that are coming to fill public-sector gaps in sectors such as health during times of economic downturn, or in the focus-related sector of specific consumer habits.
The application of kernel density showed the geo-economic trends of the economic shrinkage in Larissa. Vacant commercial lots, as a function of their location, had a decisive influence on the spatial footprint of Larissa's economic base. In the period of crisis, this spatial footprint reduced its size, and its main form converted from solid and compact to perforated and fragmented. The resulting empty lots determinant the non-productive areas which behavior as contemporary urban shrapnels sprawled disproportionately within the fabric. Low commerciality areas in Larissa operated as "magnet" of urban shrinkage and tend to overcome the difficulty the threshold of viability in social and economic terms.
We can identify two main trends of the urban shrinkage in Larissa. On one hand, the urban shrinkage is identified through the creation of the spatial perforation form and the parallel intensity reduction of high-commerciality areas.
On the other hand, urban shrinkage is observed through the weakening or even elimination of low-commerciality areas. It was immediately obvious that the highest percentage of the declining amount of businesses concentrated in low-commerciality areas. These transformations can be interpreted not only by business write-offs but also by movements of economically sound businesses during the study period. We observed significant flows of movement especially of new businesses (less than 5 years) from less to more central axis or more privileged positions. The empirical part narrated here underlies the assumption that hotspots of spatial and economic shrinkage signs at the city level are firstly recognized in smaller, low commercial streets, perpendicular or parallel to the traditional high-streets and the ring road. These areas need to receive added scrutiny by local governments and seems to be the ideal locus for spatially located policies with benefit proliferating at social and economic scale. Hotspots of spatial and economic shrinkage signs in Larissa's traditional downtown area are limited because movements of economically sound businesses fill again the vacant lots, reducing their vacancy duration. It became obvious that a shift of public realm from central to low commerciality areas is needed.

Conclusion
Non-productive commercial lots, the catalyst for a negative domino effect, are a crucial problem anchored in the emerging new economic urban landscapes in Greek cities caused in the Crisis Period. The concepts of shrinking cities and business demography are at the center of several current urban planning debates. In light of these debates, vacant lots are analyzed given economic terms but have been plagued by little consensus on their contextualize relationship with urban morphology. Managerially, this study offers insight into a triangle methodological framework structured on surveying, portraying, and documenting economic shrinkage. Responds to this challenge, we analyze the economic transformation based on economic and spatial statistics methods. In parallel, we map and categorize movement per axis of economically sound businesses during the period considered. These "flows" of businesses reveal several salient points of the function of their location which are as manifold as the vitality and the cohesion of their surrounding area. A combination of these processes reveals the ideal locus for implementing adequate regeneration and renaissance policies. This framework can act as an urban observatory of a better and dynamic understanding of the evolution of the shrinkage phenomenon. It may addresses to (i) urban planners for supporting decision-making and implication of adequate regeneration strategies toward more efficient urban planning and (ii) investors, for the evaluation and management of selected areas depending on the trends of the shrinkage phenomenon. For a better in-depth study of the issue in the future, it is important to examine the economic and spatial dimensions of urban shrinkage combined with additional parameters. Population data, income, or real estate values can be considered as indicative parameters. Nevertheless, the study of urban shrinkage is important and has gained increased attention lately. This increasing trend was further supported by the COVID-19 outbreak crisis, the lockdown measures, and their expected impact on the economy of cities or certain sectors of economic activity.